Cutaneous lupus erythematosus in dogs: a comprehensive review

Direct immunofluorescence

Direct immunofluorescence revealed the presence of IgG at the basement membrane zone in 7/14 (50%) dogs with VCLE [22]. The deposition of IgG around blood vessels was observed in 13/14 dogs (93%). Finally, cytoplasmic basal keratinocyte IgG was detected in 6/14 subjects (43%); the deposition of activated complement was not seen [22].

Indirect immunofluorescence

Indirect immunofluorescence did not reveal anti-basement membrane circulating IgG autoantibodies in the serum of five dogs with VCLE [22]. Similarly circulating antinuclear IgG autoantibodies were not detected in the serum of any of 11 dogs with VCLE using human Hep2 cells as a substrate [22].

Immunoblotting and ELISA

Using Hep2 cell extracts, immunoblotting permitted the detection of autoantibodies against soluble nuclear antigens in 9/11 tested sera (82%) [22]. When an ELISA was performed with purified human soluble nuclear antigens, the serum from 8/11 dogs with VCLE (73%) was found to have IgG autoantibodies that bound to these antigens. Antibodies were found to target Ro/SSA (45% of dogs), La/SSB (45%), Sm/RNP (45%), Scl70 (36%), Jo-1 (36%) and Sm-SnRNP (18%) [22]. Overall, and as seen in humans with SCLE, most dogs with VCLE (6/11; 55%) were found to have IgG antibodies that targeted Ro/SSA and/or La/SSB antigens [22].

Direct immunofluorescence

In one study [34], direct immunofluorescence testing performed on paraffin-embedded sections revealed the presence of in situ deposition of IgG, IgM, IgA and C3 in the epidermal basement membrane of 100%, 47%, 11% and 5% of GSHPs, respectively. The multifocal or continuous fine deposition of IgG was recorded in 61%, 35% and 77% of skin biopsy sections, respectively. Interestingly, the follicular basement membrane deposition of IgG was found in 41% of tested biopsies.

Indirect immunofluorescence

Indirect immunofluorescence testing on sections of normal canine haired and salt-split-skin revealed the existence of circulating anti-follicular IgG antibodies in the serum of 57% of tested GSHPs with ECLE [34]. In addition, anti-sebaceous gland IgG antibodies were also detected in these dogs. Circulating anti-epidermal basement membrane antibodies were not observed, however. In three studies, antinuclear antibody serology usually remained below positive thresholds in GSHPs with ECLE [32, 34, 38].

Immunohistochemistry

Immunohistochemical staining confirmed the predominance of CD3-bearing T lymphocytes in the lower epidermis, superficial dermis, in the infundibulum of hair follicles and around sweat glands [34]. These CD3-positive T lymphocytes infiltrated sebaceous glands and their associated ducts in samples collected from two dogs.

Treatment and outcome

The review of published reports has yielded inconsistent information on the treatment and outcome of this disease. The early descriptions of ECLE suggested some benefit of dietary changes, supplementation with fatty acids, anti-seborrheic shampoos, antibiotics and/or oral retinoids [31, 32] The most recent case series [34, 38] reported the limited efficacy of immune-modulating drugs prescribed either as single or combination therapy (e.g. tetracycline-niacinamide combinations, doxycycline, oral glucocorticoids, azathioprine, ciclosporin, leflunomide, or hydroxychloroquine).

Hydroxychloroquine, an first-line antimalarial drug used in human CCLE, appeared to slow down the clinical progression in some dogs with ECLE; in contrast, high-dose ciclosporin reportedly was not able to halt lesion worsening [38]. As the response to immunomodulators is heterogeneous in human CCLE variants [40], the use of high-dose oral glucocorticoids and adjunctive immunosuppressive regimens need to be investigated on an individual patient basis [34, 38, 39].

Taking into account all GSHPs with ECLE for which a long-term outcome has been reported [31, 32, 34, 38, 39], over half of dogs are eventually euthanized for their lack of disease response to therapy. This makes this CLE variant the most challenging to treat among all those of canine CCLE.

Historical perspective

In the mid 1990’s, two German shepherd dogs (one in France and one in Québec, Canada) were described as having a genital-predominant DLE [41, 42]. In 1998, we proposed the disease name of MCLE for dogs with perimucosal ulcerative lesions and microscopic characteristics of CLE (Olivry T: British Veterinary Dermatology Study Group, York, 1998). Additional cases with identical phenotypes were later published with the diagnoses of MCLE [43], DLE [44], or, more recently, perianal/perivulvar LE [45]. Finally, we reported a large series of 21 additional dogs with MCLE was reported in 2015 [9] and a single case report from Chile was later published in 2017 [46].

Incidence and prevalence

There are no available data to estimate the incidence of prevalence of MCLE in dogs.

Signalment

Collating the signalment of all published cases of canine MCLE yielded pertinent information. Of the 36 dogs [9, 41–46], there were 17 German shepherd dogs and their crosses (47%); adding the two Belgian shepherds [43] leads to about half of the dogs with MCLE belonging to breeds related to German shepherds. Altogether, females appear nearly twice over-represented with a female-to-male ratio of 1.8; there was an equal representation of intact and neutered individuals. Interestingly, this female-to-male ratio increases to 3.8 if we only collate data from German/Belgian shepherds and their crosses. In all, the age of onset of skin lesions of MCLE varied between 3 to 13 years (median and means: 6 years). Most dogs for which this information was available (17/28; 61%) began exhibiting noticeable mucocutaneous lesions in mid-adulthood (i.e. between 4 and 8 years of age).

Odds ratios for breed, sex or age predispositions for the development of MCLE cannot be estimated, as dogs come from multiple continents (North and South America, Japan, Europe), and a reference population therefore is not available.

Clinical signs

The owners of dogs with MCLE often report perimucosal ulcerative skin lesions with vocalization suggesting pain why defecating or urinating.

At the time of presentation to the veterinarian, lesions have been reported to occur most commonly on or around the anus (24/36; 67%) (Fig. 7a) or on the genitalia or perigenital region (17/36, 47%) (Fig. 7b,c) [9, 41–46]. Similar lesions can also be seen, but less commonly, abutting the lips, but they usually do not cross into the mucosa itself (10 dogs; 28%) (Fig. 7e,f). More rarely, lesions have been noted around the eyes (6 dogs; 17%) (Fig. 7d) and nasal planum (4 dogs; 11%); oral lesion are rarest (3 dogs; 9%) [9, 41–46]. In the largest case series, most dogs had two or more areas affected, and the lesions were usually symmetrically distributed [9].

The characteristic lesions of MCLE are erosions and ulcers (Fig. 7a-f), but the latter do not tend to heal with scarring [9, 41–46], an important difference with the lesions of facial and generalized DLE. Crusts are present when lesions extend into haired skin. Hyperpigmentation can be seen often around ulcerative lesions or at the site of previous ones, thus leaving a figurate or reticulated pattern [9, 41–46]. Pruritus is normally absent or mild, but pain is described when defecating and urinating or when touching the lesions; systemic signs have not been reported [9, 41–46].

The most relevant clinical differential diagnoses of MCLE are mucocutaneous pyoderma (MCP), MMP and EM variants.

Histopathology

In the largest case series, and per inclusion criteria, skin biopsies contained a cell-rich lymphocytic interface dermatitis with basal keratinocyte damage (i.e. basal cell apoptosis, loss and/or hydropic degeneration) [9] (Fig. 8a-c). This pattern was often patchy, or in limited areas, sometimes only being observed at close proximity to an ulcer margin. Interface dermatitis commonly extended to the infundibula of hair follicles (Fig. 8d), while inferior segments of hair follicles are sometimes also involved (Fig. 8e). Basement membrane thickening was found to be multifocal, patchy to diffuse (Fig. 8c). Pigmentary incontinence varied from mild to marked. Plasma cells were present in all cases (Fig. 8b,c), mixed with lymphocytes and were often numerous in subepidermal, perivascular, periadnexal and in dermal areas below erosions and ulcers. Erosions and ulcers were common but granulation tissue was limited and fibrosis (scarring) was not seen. Occasional suprabasal keratinocyte apoptosis was noted in half of the cases, but suprabasal lymphocytic satellitosis, when present, was always mild. Nonetheless, superficial keratinocyte cell death can lead to confusion with EM  and morphologically related conditions. Not surprisingly, for a perimucosal ulcerative disease, lesions of concurrent bacterial infection were common, including neutrophilic crusting, pustules, perifolliculitis and folliculitis, as well as presence of bacteria in surface exudates. Such infection will complicate the histological diagnosis and the successful treatment of pyoderma is warranted prior to biopsy.

Immunopathology

In dogs in whom this information was reported, direct IF almost always revealed a positive IgG lupus band test (LBT) [9, 44]. Positive LBTs were sometimes also uncovered for IgA, IgM and C3. Positive ANA titers were rarely found, however.

Treatment and outcome

The skin lesions of canine MCLE appear to respond best to immunosuppressive dosages of oral glucocorticoids [9, 41–46]. The complete remission of signs is generally obtained within one month of treatment induction [9]. A combination of a tetracycline antibiotic, with or without niacinamide, appears beneficial either alone or as adjunctive combination in some dogs [9, 41, 45]. In most patients, the tapering of oral glucocorticoids leads to the prompt relapse of skin lesions, which will undergo remission once the dosage is re-escalated again. The usefulness of adding additional immunosuppressive drugs (e.g. azathioprine, ciclosporin, mycophenolate mofetil etc.) to permit the reduction of oral glucocorticoid doses needs further investigations.

Historical perspective

Among the several variants of human chronic CLE (e.g. discoid LE [DLE], verrucous (hyperkeratotic) LE, chilblain LE, lupus tumidus and lupus profundus), DLE represents the most common form: it is divided into a localized variant where skin lesions are confined to the head and neck, and a generalized form, in which skin lesions also occur below the neck [47].

In 1979, Griffin and colleagues reported clinical, histopathological and immunological characteristics of two dogs with localized facial lesions that were diagnosed as being affected with the canine counterpart of human DLE [1]. In these two dogs, the nasal-predominant dermatitis was associated with microscopic focal interface dermatitis, basement membrane thickening and a superficial lymphocytic and plasmacytic dermatitis. Since then, there were three large case series describing dogs with nasal skin-predominant DLE lesions [2–4], two of them including some of the same cases [2, 4]. The then-proposed terminology resulted in the widespread acceptance of “canine DLE” being equated mainly to facial localized lesions. In the 2010s, we began reporting dogs with a more widespread phenotype that resembled that of the generalized variant of human DLE [48–50]; this was followed with the publication of a case series of ten dogs with generalized DLE (GDLE) [10], this article encompassing the three cases already published by the NC State Dermatology group [48–50].

Signalment

The four largest series of dogs with the “classic” localized facial-predominant DLE (FDLE) allows the analysis of a cohort of 104 dogs [3, 4, 45, 51]. Among these cases, there were 32 German shepherd dogs and their crosses (31%). The age of onset of FDLE skin lesions varied between 1 and 12 years of age (median: 7 years); while the female-to-male ratio was 0.7, there was an equal representation of intact and neutered individuals.

A retrospective study recently evaluated the historical and outcome information in ten dogs with GDLE [10]. Amongst these dogs, there were two Chinese crested dogs and two Labrador retrievers; there was one each of the following pure breeds: miniature pinscher, Leonberger, Shih-Tzu and toy poodle. The age of onset of GDLE skin lesions varied between 5 and 12 years of age (median 9 years). The female-to-male ratio was 1.0 and all dogs were castrated. Interestingly—and surprisingly—German shepherd dogs, a breed predisposed to develop several forms of LE, such as SLE, localized FDLE and MCLE, did not seem affected by GDLE. This discrepancy may be explained by the German shepherd breed not being predisposed to this disease, by the small size of the reported cohort or by a possible clinical misdiagnosis of GDLE as one of the” idiopathic lichenoid dermatoses” as they were diagnosed in the 1980’s solely based on the histopathological identification of a “lichenoid tissue reaction” in dogs [52].

Incidence and prevalence

At this time, there is no usable information to determine the frequency of occurrence of FDLE and GDLE in dogs.

Clinical signs

The classic skin lesions of human DLE usually consist of early erythematous and variably scaly macules or papules that slowly evolve into a coin-shaped (i.e. discoid), plaques with adherent scales, follicular plugging (i.e. comedones) and peripheral hyperpigmentation presumed to occur secondarily to inflammation [47]. These discoid plaques can coalesce and develop central scarring and depigmentation [47]. Atypical presentations of GDLE have been reported in patients of differing ethnic groups; the morphological appearance of lesions in these patients varies from hyperpigmented macules to hyperkeratotic, hyperpigmented plaques with an erythematous border [53].

The early skin lesions in canine FDLE consist of erythema, depigmentation and scaling that progress into erosions and ulcerations with atrophy and loss of the architecture of the nasal planum (Fig. 9a-f); crusting may be present if the epithelial integrity is damaged [3, 4]. Skin lesions usually affect the nasal planum (Fig. 9a-f) and might even involve the nares (Fig. 9c,d,f); several dogs exhibit additional skin lesions on the dorso-proximal muzzle (Fig. 9a,b), lips, periorbital skin and pinnae [3, 4]. Squamous cell carcinoma was reported to develop from chronic DLE nasal lesions in dogs [54], as in humans [55]. Pruritus has been reported to be variable in dogs with FDLE [3, 4].

Clinicians should remember that cutaneous (epitheliotropic) T-cell lymphomas can have localized lesions that affect the nose and could mimic those of FDLE. Other differential diagnosis for depigmentation and inflammation on the nasal planum are MCP and the uveodermatological syndrome, which resembles the Vogt-Koyanagi-Harada syndrome of humans. One should keep in mind that the "so-called MCP" is a poorly described disease that, if it were to even exist as a primary disease, is likely to occur secondarily to other diseases such as FDLE, MMP and MCLE and other nasal-targeting auto-immune and immune-mediated diseases.

Dogs with GDLE present with generalized or multifocal, annular (discoid) to polycyclic plaques with dyspigmentation, an erythematous margin, adherent scaling, follicular plugging and central alopecia; these predominate on the neck, dorsum and lateral thorax (Fig. 10a,f) [10]. In many of these dogs, the plaques evolved into ulcerations healing with a central atrophic or hypertrophic scar and dyspigmentation (depigmentation and hyperpigmentation) (Fig. 10a,f). Four of ten of the reported dogs (40%) had mucocutaneous regions involved with plaques usually appearing on or around the genitalia. An unusual pattern of reticulated (net-like) hyperpigmentation was visible on the ventral abdomen and lateral thorax in two of these cases, a feature also seen in other CCLE variants such as MCLE [9]. In the largest series of cases, systemic signs were not reported; pruritus and pain at the site of lesions were observed in four (40%) and three of ten dogs (30%), respectively [10]. There are only two canine skin diseases that could closely mimic GDLE: generalized (and often vaccine-induced) ischemic dermatopathies and the very rare hyperkeratotic EM (a.k.a. “old dog” EM).

Laboratory evaluation

In humans affected with the generalized variant of GDLE, a positive ANA titer is frequently found, and it represents a risk factor for development of SLE within five years after the initial diagnosis of skin lesions [56]. So far, out of the 104 dogs with classic FDLE included in the four largest series of cases, there were no reports of progression to SLE [3, 4, 45, 51]. Seven dogs with GDLE had a low positive ANA serum titer, but a progression with acquisition of additional criteria for SLE was not seen in any dog within the median follow up of 2.5 years (ranging 0.5 to 6 years) in the published series [10]. To our knowledge, the progression of a DLE variant to “clinical” SLE has been reported only in one dog [57].

Histopathology

The histology of DLE in dogs is similar to that of humans and is characterized by a lichenoid cell-rich, lymphocytic interface dermatitis reaction pattern with basal keratinocyte vacuolar degeneration, apoptosis, loss of basal cells and basement membrane thickening [1, 10].

In canine FDLE, the interface reaction (vacuolar degeneration, apoptosis and loss of basal cells) is often subtle or mild in biopsy samples (Fig. 11a-c) [1, 10]. Only small areas might exhibit an active interface reaction and these lesions are easily missed, as nasal planum biopsies tend to be few and small. Interface changes can involve the follicular infundibula (Fig. 11d), when lesions extend off of the nasal planum; however, folliculitis has not been specifically investigated in canine FDLE. Pigmentary incontinence occurs secondarily to the interface reaction (Fig. 11a,b) but it is not specific to this type of injury and it can be found, persistent, in the nasal planum of dogs without concurrent nasal dermatitis [58, 59] Thickening of the basement membrane zone is patchy or multifocal but is not specific, as it occurs with other chronic inflammatory disorders of the nasal planum, such as leishmaniosis, where geographically relevant [60]. Superficial dermal fibrosis can be absent or range from mild-to-marked. Secondary bacterial colonization is common in FDLE and often complicates the histological diagnosis. These issues are compounded by the fact that, historically, the diagnosis of nasal-predominant “canine DLE” was given to dogs when microscopic examination of nasal planum skin biopsy specimens revealed a superficial dermal “band-like” pattern of inflammation rich in lymphocytes and plasma cells (a so-called “lichenoid infiltrate”), without any emphasis on the presence of an interface reaction. In fact, it is now believed that such “lichenoid” lymphocyte and plasma cell rich inflammation is a nonspecific inflammatory pattern seen in and near mucosae or related tissues (oral cavity, nasal planum, eyelids, genitalia, etc.). In a retrospective histological study of nasal dermatitis in dogs, a cell-rich lichenoid infiltrate was common, but only a small subset of subjects with nasal lesions exhibited the interface dermatitis associated with CLE [61].

In canine GDLE, in contrast to FDLE, the interface reaction is usually well developed, when an adequate number of biopsies are examined from the active margins of lesions (Fig. 12a,b) [10]. The epidermis may be atrophic or mildly hyperplastic (Fig. 12a,b) as a consequence of regional variation in severity of the interface reaction. Pigmentary incontinence can be pronounced, especially at the margins of lesions, where the interface reaction extends into zones of secondary hyperpigmentation induced by chronic inflammation (Fig. 12a-d). In chronic lesions, dermal fibrosis occasionally displaces the cell-rich inflammatory infiltrate from the superficial dermis (Fig. 12c,d). Cell-poor zones of lesion occasionally occur but often individual lymphocytes can be found in the basal layer of the epidermis in good numbers, with satellitosis of apoptotic basal keratinocytes. In GDLE, superficial epidermal apoptosis occurs, with or without lymphocytic satellitosis, which can erroneously suggest the diagnosis of erythema multiforme or morphologically related conditions. However, the collection of multiple biopsies reveals apoptosis to be most prominent in the basal epidermal layer in cases of GDLE.

In the recent case series of canine GDLE [10], alopecia occurred in nearly all patients; lymphocytic interface folliculitis involved the infundibulum and extended into the isthmus. A lymphocytic mural folliculitis was also common, but it was usually milder and involved the infundibulum, isthmus and inferior hair follicle segments, typically sparing the bulbs. This mural pattern mirrors that of human DLE, where it is also called a panfollicular pattern; it is usually minimally severe, but such a pattern is insufficiently described [62]. Sebaceous gland atrophy occurred in GDLE cases, where it was mostly mild and partial in biopsies but sometimes it was complete [10]. It should be noted that diagnostic biopsies typically focus on epidermal changes at the margins of skin lesions where hair follicle and sebaceous gland changes might not be fully developed.

Immunopathology

A linear deposition of IgG and IgM at the dermo-epidermal basement membrane zone (i.e. a positive LBT) of lesional skin was found in 90% of dogs with GDLE, and this proportion is similar to what is seen in human DLE lesions [10]. Interestingly, the most commonly detected immunoreactant deposited in one series of dogs with classic FDLE was C3 (90–100%), while IgG and IgM were revealed in 40–70% of cases, respectively [4]. In contrast, in the second case series, a positive LBT showed immunoglobulins (all classes together) and activated complement (C3) in 85–90% of 22 cases [3]. These variable results between canine localized and generalized DLE could be related to differences in tissue fixation techniques (frozen vs. formalin), antigen retrieval methods and/or immunofluorescence staining protocols that were performed 30 years apart. To investigate the value of performing DIF in canine CLE diagnostic work-up, further studies regarding the sensitivity and specificity of a positive LBT for the diagnosis of CLE variants are warranted.

Treatment and outcome

Besides the obvious need for photoprotection (sun avoidance), the 2017 update of the Cochrane systematic review of interventions for human DLE reported evidence for the benefit of a potent topical glucocorticoid and the oral drugs hydroxychloroquine and acitretin (a retinoid) [63] Furthermore, there was insufficient evidence for the efficacy of other interventions, such as topical calcineurin inhibitors (e.g. tacrolimus), [63].

Since 1992, antibiotics of the tetracycline family, with or without concurrent niacinamide (a.k.a. nicotinamide), have been suggested to be helpful for the treatment of canine immune-mediated skin diseases including FDLE. An initial report by White and colleagues showed that 14/20 (70%) dogs with FDLE had a good-to-excellent response using a tetracycline-niacinamide combination [64]; a recent retrospective study revealed a similar positive response rate in dogs with FDLE [45]. While tetracycline-niacinamide therapy is considered to be safe, tetracycline is no longer available commercially in many countries. Although tetracycline and doxycycline were shown to be relatively similar in their effectiveness to treat the so-called canine lupoid onychodystrophy, a poorly-understood onychitis [65], therapeutic equipotency data for other canine auto-immune and immune-mediated diseases, such as DLE are unavailable; additional studies are necessary to confirm the effectiveness of substituting doxycycline or minocycline for the tetracycline used beforehand to treat dogs with CLE.

Topical tacrolimus ointment has been used successfully for the topical treatment of canine FDLE. At first, Griffies and colleagues evaluated the use of 0.1% tacrolimus ointment applied topically to the lesional (facial) skin of ten dogs with DLE, most of these dogs receiving topical tacrolimus as an adjunctive therapy to oral glucocorticoids [66]. There was a positive response in eight dogs (80%), three of them having had an excellent improvement in skin lesions [66]. Recently, Messinger and colleagues conducted a randomized, double-blinded, placebo-controlled crossover study to evaluate the efficacy of a lower concentration of tacrolimus ointment (0.03%) in 19 dogs with FDLE [51]. Tacrolimus ointment, applied twice daily as monotherapy for up to 10 weeks, appeared safe and effective. A noticeable clinical improvement was seen in 13/18 (72%) of the dogs, whereas only three dogs receiving the placebo had lesions that improved. To summarize, limited outcome data suggest that topical tacrolimus ointment and/or a niacinamide-cyclin combination therapy should be considered as potentially effective therapeutic options for canine FDLE.

Skin lesions of canine GDLE appear to respond to a wide range of treatments but half of the patients experienced relapses upon the tapering of drug dosages. In a recent report [10], a remarkable improvement or a complete remission in GDLE skin lesions followed treatment with oral ciclosporin (mean 4.8 mg/kg once daily) along with a short course of glucocorticoids at treatment onset. Furthermore, oral hydroxychloroquine, in conjunction with topical 0.1% tacrolimus ointment application, helped induce and maintain remission of skin lesions in two dogs with GDLE [10].