Mucous membrane pemphigoid in dogs: a retrospective study of 16 new cases

Introduction

Mucous membrane pemphigoid (MMP), also known as cicatricial pemphigoid, is a chronic autoimmune subepidermal blistering disease (AISBD) recognized in humans, dogs and cats.1-4 In all of these species, it is characterized by a variable combination of vesicles, ulcerations and scarring affecting predominantly mucosae and mucocutaneous junctions. Although exact epidemiological data on canine MMP are unavailable, this disease is considered to be the most common AISBD in this species, with German shepherd dogs appearing to be an over represented breed.2

In humans and animals, MMP is associated with the presence of circulating autoantibodies that target variable antigens of the basement membrane zone (BMZ); these include collagen XVII (BP180), BPAG1e (BP230), integrin α6/β4, laminin-332 and, more rarely, collagen VII.1-4 The demonstration of skin fixed-antibodies and/or activated complement and/or the presence of circulating anti-BMZ antibodies is proposed as one of the diagnostic criteria for MMP in humans.4 Likewise, it has been suggested that, if available, the demonstration of anti-BMZ antibodies and/or complement should be part of the diagnostic criteria for the diagnosis of MMP in dogs.2 Nonetheless, due to autoantigen overlap among the AISBDs, the demonstration of an elevated titre of anti-BMZ antibodies would not help distinguish MMP from other AISBDs.5-7 Furthermore, in humans, especially when appropriate biopsy cannot be obtained, the presence of anti-BMZ antibodies is used predominantly to eliminate other blistering diseases that do not affect the BMZ directly; these include, for example, lichen planus, pemphigus vulgaris, erythema multiforme/Stevens Johnson syndrome and angina bullosa haemorrhagica.6

In humans, the treatment of MMP depends on the “risk group” into which patients are classified. This classification is based on the severity, extent and location of lesions (i.e. “low-risk” group: oral involvement with or without skin lesions; “high-risk” group: eye, genital, respiratory tract and upper gastrointestinal tract lesions).4 Because of the better prognosis in low-risk patients, the treatment is usually more conservative and utilises drugs such as topical glucocorticoids with or without oral low-dose glucocorticoids, tetracycline/niacinamide or dapsone.8, 9 In contrast, patients in the high-risk group are treated with more potent drugs from the onset to avoid, or reduce, the irreversible scarring and strictures associated with this disease. In this group, the use of systemic drugs is mandatory and a combination of adjunctive immunosuppressive medications such as high-dose glucocorticoids, azathioprine, cyclophosphamide and/or intravenous immunoglobulin (IVIG) infusions are often used.8, 9 Unfortunately, because of the chronic and progressive nature of the disease and its tendency to relapse, MMP is challenging to manage in people, and the reported treatment success rates can vary greatly.10

Because of the rarity of MMP in dogs, only limited information about the treatment and outcome of this disease is available. To the best of the authors’ knowledge, there is only a single publication discussing the treatment outcome of 11 dogs with MMP.2 More information about canine MMP will assist to better understand this syndrome and improve our ability to successfully manage this chronic, frequently relapsing disease.

The objectives of this study were as follows: (i) to report information regarding the signalment, clinical presentation, histological and immunological features of additional cases of canine MMP; and (ii) to assess the therapeutic interventions and treatment outcomes in dogs with MMP.

Material and Methods

Results

Clinical summary

Skin lesions were symmetrical in 14 of 16 dogs (88%) and they consisted of erosions and ulcers (16 of 16 dogs; 100% – an inclusion criterion), crusting (nine of 16; 56%), erythema (five of 16; 31%), vesicles/bullae (five of 16; 31%) (Figure 1a), scarring (three of 16; 19%) and hypopigmentation (two of 16; 13%).

The body regions first reported to be affected by lesions were oral/perioral (12 of 16; 75%), ocular/periocular (six of 16; 38%), nasal (four of 16; 25%) and genital (two of 16; 13%) regions and/or concave pinnae (two of 16; 13%). At the time of presentation to the veterinarian, all dogs (100%) exhibited lesions on mucosae or mucocutaneous junctions assuming the nasal planum to be a modified mucosa. Fifteen dogs exhibited lesions on two or more mucosae or mucocutaneous junctions (median: 3, mean: 3). Haired skin involvement was less common (six of 16; 38%). The distribution of lesions at the time of diagnosis is summarized in Table 1 and representative lesions are shown in Figure 1.

Table 1. Lesion distribution at the time of diagnosis in 16 dogs with mucous membrane pemphigoid (MMP)

Affected areas		Number of dogs	Percentage
Mucosae/mucocutaneous junctions	Oral cavity	11	69%
	Gingiva	10 of 11	91%
	Hard/soft palate	9 of 11	82%
	Tongue	4 of 11	36%
	Labial/perilabial	9	56%
	Nasal planum/perinasal	9	56%
	Eyelids	8	50%
	Genitalia	6	38%
	Anus/perianal	4	25%
Haired skin	Concave pinnae	5	31%
	Pressure points	2	13%
	Periungual	1	6%

• Note: For this disease and based on previous experience, the nasal planum was considered to be a modified mucosa; the affected pressure points in both cases were elbows.

Nondermatological complaints reported in six of 13 dogs (46%) were associated with the oral cavity and consisted of malodour (two dogs), pain when eating (two dogs) and excessive salivation/drooling (two dogs). A loss-of-function of affected organs secondary to chronic scarring (e.g. blindness, genital strictures, breathing difficulties) was not reported in any of the dogs.

Histopathology

For 11 dogs, skin biopsy material was available for further review, with 2–20 histological sections of mucocutaneous junctions, mucosae and/or haired skin per dog. Vesicles were formed by subepidermal clefts (Figure 2a), the majority of which were ruptured and bordered an ulcer. Vesicles ranged from very small to large, occasionally spanning the width of the biopsy, and/or involved follicular infundibula. Most vesicles were empty; however, a rupture of the majority of vesicles limited evaluation of vesicle contents for inflammatory cells, haemorrhage and fibrin. Subepidermal microvacuoles along the BMZ were uncommon and mild (two of 11; 18%) or, less often, marked (one of 11; 9%). A superficial dermal or submucosal band of fibrosis (seven of 11; 64%) was common below vesicles, or below intact epithelium near vesicles (Figure 2b), and ranged from mild to marked. Fibrosis was reorganizing or sometimes had numerous small blood vessels and shared features with a thin band of granulation tissue. Biopsies with and without fibrosis were seen in the same case.

Dermal and submucosal inflammation varied greatly in the extent and type of inflammatory cells present. Perivascular inflammation was occasionally mild and often moderate to marked. Marked inflammation was typically associated with a mucosa, chronic ulceration and/or secondary bacterial infection characterized by neutrophilic crusts, luminal neutrophilic folliculitis and bacterial colonization. Mucosal and mucocutaneous inflammation sometimes formed a band-like infiltrate of lymphocytes and plasma cells just below the epithelium (a lichenoid infiltrate). In at least one biopsy from each dog, mild and occasionally moderate amounts of dermal neutrophils (eight of 11; 73%) and eosinophils (six of 11; 55%) below vesicles or intact epithelium were present. Lymphocytes and plasma cells were common and ranged from mild to marked, when present. However, inflammation did not appear to localize to (target) the BMZ in most cases (nine of 11; 82%). Rowing of individual neutrophils along the BMZ was uncommon (one of 11; 9%) as was rowing of histiocytes along the BMZ (two of 11; 18%), but when present these changes were moderate to marked. Lymphocytic exocytosis was common (nine of 11; 82%) and was mostly mild but sometimes moderate. A few lymphocytes infiltrated infundibula and external root sheath epithelium of hair follicles in one dog.

Regular epidermal or mucosal hyperplasia was common (11 of 11; 100%) and was most often moderate. Although apoptosis of basal, and less often suprabasal, keratinocytes was common (nine of 11; 82%), apoptosis usually only involved rare individual keratinocytes, and only in a few cases was apoptosis numerous enough to be warrant a mild grade. Vacuolation of basal keratinocytes was not a feature. Pigmentary incontinence was absent or ranged from mild to marked, but many sections lacked epidermal pigment to disperse to the dermis.

Detection of tissue-bound and circulating anti-BMZ autoantibodies and complement

The detection of tissue-bound IgG, IgM, IgA antibodies and C3 complement using direct IF was performed in 13 dogs. Twelve of these 13 samples (92%) exhibited a linear deposition of IgG along the BMZ. The only dog with negative direct IF (Case 15) tested positive for circulating anti-BMZ IgG. Basement membrane-bound IgA, IgM (Figure 3a) and C3 were detected in one (8%), four (31%) and two (15%) cases, respectively.

Sera from 11 dogs were available for detection of circulating anti-BMZ IgG, IgA and IgE antibodies. A positive deposition of circulating IgG antibodies in the form of a linear fluorescence deposit at the BMZ of the salt-split buccal mucosa was detected in eight of 11 (73%) sera. The binding of these circulating IgG autoantibodies was nearly always restricted to the epidermal side of the salt-split mucosa (seven of eight; 88%) (Figure 3b), or it was observed on both epidermal and dermal sides (one of eight; 13%). One of the three dogs without detectable circulating anti-BMZ IgG had detectable tissue-bound anti-BMZ IgG (Case 1). Tissues from the other two dogs with negative indirect IF were not available for direct IF staining (cases 3 and 4). Circulating anti-BMZ IgA and IgE antibodies were not detected in any of the 11 dogs.

Discussion

Although MMP is the most common AISBD in dogs, the overall rarity of this condition limits our knowledge about this entity, especially regarding the treatment and prognosis.1, 2, 7 Therefore, the goal of our retrospective study was to provide additional information about the clinical, histological and immunological aspects of this syndrome, with an emphasis on treatment and outcome. This study adds 16 new cases of canine MMP to those previously published.1, 2

The analysis of the signalment of these 16 dogs with MMP is consistent with the previously reported finding that the German shepherd dog is a frequently affected breed.2 A possible breed predisposition for MMP is predictable, as genetics have been linked to the susceptibility of development of many autoimmune diseases in people, including MMP.6 Unfortunately, the relative risk for development of MMP in German shepherd dogs could not be calculated because a control population was not available for such comparison.

The female-to-male ratio in this study was one, which is in contrast to the previously reported data where males were more often affected than females.2 The reason for this disparity is unknown, although the relatively small sample size compared to that of the previous meta-analysis could have prevented the detection of a subtle difference in the ratio.2 Interestingly, in people with MMP, women appear to be affected more often than men by a factor of 1.5 to 2.14

Although the median age of the disease onset reported in this study was similar to that of the previous meta-analysis, most dogs in this study developed the initial signs of MMP in their mid-adulthood (i.e. aged 4–7 years).2 This latter finding was in contrast to the data reported in the previous review in which dogs generally developed the disease during their late adulthood (older than 7 years).2 It is possible that increased awareness of this disease over the past decade has led to an earlier detection of clinical lesions and confirmation of the diagnosis. In humans, MMP is a disease of elderly people with a mean age of onset between 60 and 80 years.6, 14

Most dogs in our study presented initially with oral and/or perioral lesions, an observation similar to that reported in the previous study.2 Over the course of the disease, this region remained the most commonly affected site, although most dogs developed lesions affecting other mucosae or mucocutaneous junctions. Indeed, over 90% of dogs exhibited lesions affecting two or more mucosae or mucocutaneous junctions, a feature also reported in previous cases of canine MMP.2 In a similar way to dogs, oral cavity lesions are present in almost all human patients (85%) with MMP.6 Ninety one percent of dogs with oral involvement had lesions on the gingiva, which early in the course of disease could lead to the misdiagnosis of MMP as another ulcerative gingival disease.15 People affected with MMP often develop blisters on the nasal mucosa, larynx, pharynx or oesophagus. Although none of the dogs in this study exhibited signs suggesting such involvement, an endoscopic examination was not performed in any of the dogs.

The nasal planum appeared to be the second most common area to be affected (nine of 16; 56%); it was also the third most common area in which the initial lesions were reported to develop (four of 16; 25%). Dogs with this lesion distribution could be misdiagnosed clinically as having discoid lupus erythematosus (DLE). However, facial DLE rarely affects the oral cavity and this alone distinguishes MMP from facial DLE.16-18

Our study did not reveal any apparent correlation between the location of lesions, their severity or the number of affected sites and disease prognosis. In contrast, people affected with ocular, laryngeal, genital or oesophageal MMP tend to have poorer prognosis due to the possibility of a functional impairment of these tissues (e.g. blindness in ocular MMP, breathing difficulties in nasal MMP, sexual dysfunction in genital MMP).6 The location of lesions, their severity and the speed of the progression are important factors dictating the nature of the treatment approach in human MMP.4, 9 For example, on one hand, people with severe ocular, genital, oesophageal and/or laryngeal MMP usually are treated with systemic immunosuppressive drugs such as glucocorticoids, cyclophosphamide, azathioprine, mycophenolate mofetil or, in refractory cases, with intravenous immunoglobulins and/or biologics such as chimeric anti-CD20 antibody. On the other hand, the initial treatment proposed to people with mild disease involves topical glucocorticoids, dapsone or tetracycline antibiotics with niacinamide.4, 9

In this study, CR was obtained in 60% of dogs that received tetracycline antibiotics and niacinamide alone or in combination with another immunosuppressive drug. Likewise, five of 11 dogs (46%) included in the previous review of canine MMP had had a partial or complete resolution of clinical signs with a combination of tetracycline and niacinamide;2 combining these data suggests that this combination alone or with additional immunosuppressive drugs should be considered as a first-line treatment for canine MMP. Although the exact pathogenesis of MMP remains unknown, several lines of evidence suggest that antibody-mediated, complement-dependent as well as complement-independent pathways play a role in the dermoepidermal separation occurring in this disease.6, 14 The first pathway is initiated by the binding of autoantibodies to the BMZ antigen(s), which leads to activation of complement, mast cell degranulation, recruitment of polymorphonuclears and BMZ injury by released proteases.6 In addition, there is increasing evidence that complement-independent processes involving keratinocytes, proteases and/or endocytosis with subsequent degradation of the components of the BMZ could be involved in blister formation.19, 20 Similar mechanisms leading to complement-independent blister formation have been investigated in human bullous pemphigoid (BP).21, 22 The inhibitory effect of tetracycline antibiotics and niacinamide on the activity of proteases, along with their wide range of antiinflammatory properties (e.g. inhibition of leukocyte chemotaxis and proinflammatory cytokines), is a logical hypothesis for the observed benefit of these drugs in canine and human patients suffering with pemphigoid diseases, although the mode of action in dogs remains to be elucidated.23-25

Our data indicated that oral glucocorticoid monotherapy was the most frequent treatment regimen that failed to induce CR. Indeed, most dogs reported in this study and in the previous review received a combination therapy including two or more immunosuppressants.2 Although CR was achieved in more than 90% of dogs in this study, a high rate of a disease relapse was observed. The relapsing nature of this disease should be considered when discussing the prognosis and a long-term treatment plan with the owner.

Microscopic findings in this study were similar to those reported previously for canine MMP.1, 2 Subepidermal clefts, an inclusion criterion, were empty or ruptured more often than filled with inflammatory cells. Ulcers were common. Dermal inflammation often contained eosinophils, which is also described in canine epidermolysis bullosa acquisita (EBA) and BP,11, 26 but their numbers were usually low in the MMP. Although neutrophils and eosinophils were common in the dermis, inflammation targeted the BMZ in only few cases, where it overlapped with the inflammatory patterns described for canine EBA.11 Much of the moderate-to-marked inflammation in the dermis and submucosa was considered secondary and it was likely attributable to bacterial infection, ulcers and/or a chronic generic response of mucosal or mucocutaneous tissue to injury; this was often seen as a lymphoplasmacytic (lichenoid) band pattern. These nonspecific inflammatory features could potentially complicate a histological diagnosis.

Occasional basal keratinocyte apoptosis and pigmentary incontinence were observed and were considered nonspecific, as both can be seen with other subepidermal blistering diseases11 and with injured hyperplastic epidermis or mucosa generically. Additionally, the occasional keratinocyte apoptosis in AISBDs could also be the result of anoikis, a form of apoptosis triggered by a loss of the cell attachment to the appropriate matrix.27 Basal cell apoptosis in combination with a generic mucosal lichenoid inflammatory pattern might lead to a misdiagnosis of mucocutaneous lupus erythematosus (MCLE), which shares clinical features with MMP including a propensity to form mucocutaneous ulcers and a breed predisposition for the German shepherd dog. Mucocutaneous lupus erythematosus, however, rarely affects the oral cavity.28 It is possible that some dogs develop both conditions concurrently, because we have observed one case with good histological evidence of both MCLE and MMP.28 Suprabasal apoptosis of keratinocytes might lead to a histological misdiagnosis of the erythema multiforme (EM) group of diseases, but usually EM presents with more suprabasal apoptosis than that observed in this study.29 It is recommended that clinicians collect multiple biopsies to increase their probability of obtaining an accurate diagnosis. Fibrosis is a recognized clinical feature of MMP in humans and dogs.1, 4 In this study, superficial dermal or submucosal fibrosis was seen more commonly histologically than in previous reports of canine MMP or EBA.1, 11 However, the degree of fibrosis varied from absent to marked in biopsies from the same dog and, by itself, fibrosis is not expected to differentiate canine MMP from other subepidermal clefting diseases. Furthermore, chronic secondary bacterial infection might induce fibrosis in some cases nonspecifically, complicating identification of any primary fibrosing features of canine MMP.4

The presence of a continuous deposit of anti-BMZ antibodies and/or C3 complement in the skin of affected people is a diagnostic criterion in human MMP.4 A similar criterion has also been suggested for canine MMP in the past, but, because direct IF is not commercially available, this criterion is not usually and readily fulfilled in veterinary medicine.2 In our study, 13 dogs were available for detection of tissue-bound and/or circulating anti-BMZ antibodies; anti-BMZ antibodies, most frequently IgG, were detected in all dogs using direct and/or indirect IF. Sera from 11 dogs were available for indirect IF using salt-split buccal mucosa, which distinguishes between an autoimmune response directed against antigens localized to the epidermal (collagen XVII, BP230, integrins) or the dermal (laminins, collagen IV, collagen VII) side of the salt-split tissue.30, 31 Canine MMP has been shown to be immunologically heterogeneous, with antibodies targeting various BMZ antigens such as collagen XVII, BP230 or laminin-3322 and, therefore, an antibody deposit can be detected at either the epidermal or the dermal side of the salt-split mucosa. Consequently, although the detection of anti-BMZ antibodies is helpful in confirming the autoimmune nature of the syndrome and narrowing down the differential diagnoses, it is in no means specific for MMP, because other AISBD will often exhibit similar positive results and patterns.11, 26, 32 This opinion is echoed in an editorial that emphasised the importance of the clinician in the diagnostic process.7 Finally, anti-BMZ antibodies and/or complement are not always detected by IF in people and dogs affected with MMP.2, 33, 34 In this study, three of 11 dogs (27%) had undetectable levels of circulating anti-BMZ autoantibodies, a percentage similar to that previously reported in canine MMP.2 Although the exact explanation of the negative indirect IF in the three dogs remains unknown, it is conceivable that these three dogs might have been treated with immunosuppressive medications before blood collection, which could have reduced anti-BMZ autoantibodies below the detectable level. The degradation of autoantibodies over time is an alternative hypothesis.

In conclusion, canine MMP is a chronic, slowly progressive, mucosal/mucocutaneous-dominant AISBD frequently diagnosed in middle-aged German shepherd dogs with histological and immunological findings similar to those reported in a previous canine MMP meta-analysis.2 Treatment with tetracycline antibiotic and niacinamide, often in combination with other immunosuppressive drug(s), was found to be beneficial in more than half of these cases. Unfortunately, the rarity of this syndrome, the low number of dogs with reported treatment outcomes in the literature and the retrospective nature of the reports limits our ability to make definitive conclusions on the best therapeutic approach for canine MMP.