Canine Pemphigus Foliaceus with Concurrent Immune-Mediated Thrombocytopenia

Introduction

Canine pemphigus foliaceus (PF) is the most common autoimmune pustular skin disease in dogs characterized by pustules, crusting, erosions, and alopecia of the head, face, nasal planum, pinnae, and footpads.^1–3 When generalized, skin lesions spread over the entire body, and symptoms include fever, lethargy, and limb edema. Histopathologically, acantholytic keratinocytes are found within intraepidermal-to-subcorneal neutrophilic and eosinophilic pustules. Direct immunofluorescence (IF) has shown that antikeratinocyte autoantibodies are deposited in vivo in the subcorneal layers, especially the stratum spinosum and granulosum.^1–9 Indirect IF using canine PF sera revealed the presence of circulating antikeratinocyte immunoglobulin (Ig) G in vitro.¹⁰ Recently, several different epidermal staining patterns were identified by indirect IF using healthy canine footpad epithelium as a substrate.¹¹ Comparisons of IF staining patterns of desmosomal and nondesmosomal adhesion molecules of keratinocytes with those of canine PF sera showed that desmocolin-1 was the major autoantigen of canine PF.^11,12 Although circulating antikeratinocyte autoantibodies are considered to cause acantholysis only on the epidermis, canine PF reportedly occurs either concurrently with or following hypothyroidism, leishmaniasis, and systemic lupus erythematosus (SLE).^3,13–15 Dysregulated antibody production based on polyclonal B cell activation, cross-reacting antigens, and the innocent bystander reaction due to adsorption of circulating immune complexes onto keratinocytes are thought to cause these autoimmune disorders.^3,13–15

Thrombocytopenia in dogs is often immune mediated.¹⁶ The pathogenesis of canine immune-mediated thrombocytopenia (IMT) involves platelet surface-associated (PSA) Ig binding, which leads to the increased removal of antibody-coated platelets by macrophages through phagocytosis in the reticuloendothelial system.¹⁶ Canine IMT is difficult to diagnose because it occurs not only primarily without obvious underlying causes, but also secondarily to various systemic diseases including infectious, neoplasias, adverse reactions to drugs (e.g., sulfonamide, cephalosporin, cyclophosphamide, etc.), and systemic autoimmune diseases such as SLE.^16–20 Typically, IMT is diagnosed by excluding other causes of thrombocytopenia, such as disseminated intravascular coagulation and deficient platelet production by bone marrow.^16–20 Kristensen et al. (1994) established a direct assay for measuring PSA-Ig using flow cytometry, which can be used to diagnose canine IMT.^16,17,19 Tsuchiya et al. (2010) recently improved the accuracy of the assay using established artificial positive-control platelets.²¹ The authors of this study routinely use this assay when canine IMT is suspected in their hospital.

The purpose of this article is to describe a rare association of PF with concurrent IMT. Diagnosis was made based on histopathological findings, immunohistochemical detection of IgG deposition around keratinocytes, and flow cytometric detection of PSA-Ig.

Case Report

A 3 yr, 10 mo old male wirehaired fox terrier was presented to his primary care veterinarian with pruritus and generalized dermatitis. The owner noticed that the dog's skin lesions had worsen 6 days before presentation. At the time of examination, the dog was pyrexic (39.7°C) and the skin lesions were bilateral, symmetrical, and manifested as pustules, crusting, and erosions. The pustules initially appeared on the ventral trunk, in the ear canal, and on the pinnae. Multifocal crusting and erosion then progressed over the entire body to cover the nasal planum, top of the head, muzzle, dorsal and ventral trunk, feet, and paw pads over the next 18 days (Figures 1A, B, and C). Prior to referral, commercially-available rapid drug sensitivity testing (Monoris Inc., Tokyo, Japan) was performed without bacterial identification from swabs collected from underneath a crust on the ventral trunk. Although the crusting dermatitis did not respond to a systemic antibiotic (cephalexin^a, 25 mg/kg per os [PO] q 12 h) selected by the drug sensitivity test, the fever was reduced with a single dose of diclofenac sodium^b (1.25 mg PO). Because leukocytosis, anemia, and thrombocytopenia were diagnosed after the administration of cephalexin and diclofenac sodium, the dog was referred to the study authors' hospital for further examination, diagnosis, and treatment.

View Figure

• Download as Powerpoint
• Download Figure

Twelve days after the initial presentation at the primary care hospital (day 1), the dog had lost 10% of its body weight (new body weight was 9 kg). Body temperature, pulse, and respiratory rates were 37.9°C, 120 beats/min, and 25 breaths/min, respectively. A physical examination revealed lameness; swelling of all limbs in the synovial joints of the carpus, stifle, and hock; conjunctivitis; and peripheral lymphadenopathies on the superficial cervical and popliteal lymph nodes. A grade 2/6 systolic heart murmur, loudest over the left thorax, was audible during cardiac auscultation. The skin showed generalized erythema, pustules, crusting, and erosions. Those lesions were distributed symmetrically on the dog's face (including the muzzle, nasal planum from the top of the muzzle to the front of the glabella, the periocular region, ear canal and pinnae), ventral and dorsal trunk, feet, and paw pads. The oral mucosa and mucocutaneous junctions, such as the eyelids and anus, were free of lesions. Impression smears were obtained from the erosive skin lesions under the crusting. Cytology findings showed eosinophilic, pyogranulomatous inflammation comprising numerous degenerate and nondegenerate neutrophils, many eosinophils and macrophages, and a few small lymphocytes. A few oval and intensely basophilic acantholytic keratinocytes with a centrally placed nucleus were individually surrounded by abundant neutrophils. Infectious organisms such as bacteria and Malassezia spp. were unremarkable in smears. Swabs collected from pustules in the axillary area were examined for bacterial and fungal cultures at Hoken Kagaku Laboratory (Kanagawa, Japan) using sheep blood agar and Sabouraud agar, respectively. Sarcoptic mange and Demodex canis were not detected in skin scrapings or plucked hairs.

Considering the diagnosis of hematologic abnormalities by the primary care veterinarian and based on the physical exam findings conducted at the authors' hospital, immunologic tests, diagnostic imaging, arthrocentesis, fine-needle aspiration of lymph nodes, skin biopsy, and a histopathological examination were performed.

Table 1 shows the hematologic and serum biochemical findings. Complete blood cell counts showed a slightly elevated white blood cell count (1.76 × 10⁹/L; reference range, 0.6–1.7 10⁹/L), a low red blood cell count (4.3 × 10¹²/L; reference range, 5.5–8.5 10¹²/L), a packed cell volume of 0.29 (reference range, 0.37–0.55), and a remarkably low platelet count (PLT; 70 × 10⁹/L; reference range, 200–500 10⁹/L). The ratio and cell counts of reticulocytes were 0.0032 proportion of red blood cells and 13.76 × 10⁹/L, respectively (the range of reticulocyte counts in regenerative response to mild anemia (a packed cell volume of 0.26–0.38) are 100–150 × 10⁹/L). A slightly low serum iron concentration (16.11 μmol/L; reference range, 16.83–21.84 μmol/L) and low unsaturated iron binding capacity (6.80 μmol/L, reference range; 22.73–60.86 μmol/L) indicated that chronic inflammation caused the mild nonregenerative anemia. Blood coagulation tests showed an elevated plasma fibrinogen (13.35 μmol/L; reference range, 2.59–9.88 μmol/L), a normal prothrombin time (8.1 s; reference range, 6.8–8.6 s), a slightly prolonged activated partial prothrombin time (28.1 s; reference range, 13.1–26.9 s), and a normal fibrinogen-fibrin degradation product (<2.5 mg/L; reference range, 0–2.5 mg/L). The blood coagulation test results excluded disseminated intravascular coagulation. The serum biochemical analysis showed moderately low total protein (47 g/L; reference range, 51–77 g/L) and albumin (18 g/L; reference range, 25–40 g/L) with high alkaline phosphatase (68.17 μkat/L; reference range, 0.6–4.33 μkat/L), and bilirubin (16.93 μmol/L; reference range, <5.13 μmol/L). C-reactive protein (CRP) was measured using a laser nephelometric immunoassay^c to evaluate the degree of systemic inflammation. Examination of globulin protein fraction by serum protein electrophoresis was performed because hypoalbuminemia was observed. CRP was obviously elevated at 114.29 nmol/L (reference, <9.52 nmol/L) and serum protein electrophoresis revealed elevated α₂-globulin level.²²

TABLE 1 Complete Blood Cell Count and Serum Biochemical Profile Results at the Time of Initial Presentation

View Fullscreen

To determine the cause of the hypoalbuminemia, diagnostic imaging was performed, focusing on possible renal and hepatic abnormalities. Thoracic and abdominal radiography revealed no apparent abnormalities. An abdominal ultrasound showed hepatic venous congestion and mild hyperechogenicity of the liver parenchyma and renal cortex. The dog underwent a cardiac work-up to determine the cause of its heart murmur and hepatic venous congestion. Echocardiography revealed mild pulmonary hypertension on the basis of increased tricuspid regurgitation velocity (3.41 m/s, reference range <3.0 m/s; estimated pressure gradient across the tricuspid valve, 46 mm Hg). The dog's systolic blood pressure was 120 mmHg, which was considered normal for dogs. Urinalysis revealed a specific gravity of 1.045, a protein content of 1.8 g/L, and a protein/creatinine ratio of 1.5, indicated mild proteinurea (reference ranges, 1.020–1.050, <0.5 g/L, and <0.5, respectively). The findings of proteinurea and a hyperechoic renal cortex indicated protein-losing nephropathy. The leakage of antithrombin III from the kidneys is considered a possible cause of pulmonary thromboembolism, which leads to pulmonary hypertension; however, further work-up was not performed.

To exclude lymphoma, a fine-needle aspirate of the lymph nodes was obtained, which showed mild-to-moderate reactive lymphoid hyperplasia with mild neutrophilic infiltration. the authors performed arthrocentesis because idiopathic polyarthritis was suspected on the basis of the history of fever, physical findings of lameness, and increased serum CRP level. A gross examination of joint fluid showed turbidity and decreased viscosity. A cytologic examination identified low-to-moderate cellularity and mild inflammation consisting of large mononuclear cells (94%), small mononuclear cells (5%), and neutrophils (1%). Those findings were consistent with polyarthropathy associated with chronic inflammation.

At that point, a mild anemia, polyarthropathy (possibly due to chronic inflammation), thrombocytopenia, hypoalbuminemia (probably due to protein-losing nephropathy), and pulmonary hypertension were identified. Based on those findings, the authors' differential diagnoses included PF, pemphigus erythematous, panepidermal pustular pemphigus, cutaneous lymphoma, erythema multiforme, cutaneous vasculitis, IMT, and SLE. An additional skin biopsy and immunologic tests were conducted to reach a definitive diagnosis.

A commercial laboratory (Monoris Inc, Tokyo, Japan) performed a direct Coombs' test and measured serum antinuclear antibodies and rheumatoid factor (RF) to determine the presence of SLE. The direct Coombs' test (37°C and 4°C) and antinuclear antibodies were negative, but RF was positive (titer was 320; reference value; 0). Nine days after the initial presentation, flow cytometry for PSA-IgM, -IgG, and -C3 was performed as described by Tsuchiya et al (2010) for the diagnosis of canine IMT.²¹ The gates for positive events were set based on healthy control samples at 0–5%. Values >10% were judged positive. The results showed apparently positive anti-canine IgM (89.3%) and anti-canine IgG (47.1%) antibodies, but negative anti-canine C3 (7%) antibody compared with the healthy control (Figure 2).

View Figure

• Download as Powerpoint
• Download Figure

Skin biopsies obtained using a 4 mm biopsy punch from the crusting skin lesions on the dog's head, neck, and dorsal trunk were examined by routine histopathology. Those findings revealed severe crusting of the superficial epidermis and large intraepidermal to subcorneal pustules (Figure 3A) that mainly comprised neutrophils and eosinophils, and some included acantholytic keratinocytes (Figure 3B). The pustules involved the hair follicles and extended into the follicular infundibula. Neutrophilic and eosinophilic inflammation was seen in the pustules. Primarily neutrophils and then mononuclear cells infiltrated the superficial-to-middle dermis. Direct IF using anti-canine IgG antibodies^d (1:800) identified IgG deposition throughout the superficial-to-middle epidermis layer, particularly in the stratum spinosum (Figure 3C). Canine PF and IMT were diagnosed based on those histopathological and immunologic findings.

View Figure

• Download as Powerpoint
• Download Figure

The dog was hospitalized for 7 days and received fluid therapy with acetate Ringer's solution (3 mL/kg/hr) supplemented with low molecular weight heparin (dalteparin sodium^e, 75 U/kg/day) and monoammonium glycyrrhizinate^f (1 mg/kg intravenous injection [IV] q 12 hr). Cefazolin sodium hydrate^g (25 mg/kg IV q 8 hr) and enrofloxacin^h (5 mg/kg BW q 24 hr) were administered to prevent bacterial growth, and ursodeoxycholic acidⁱ (10 mg/kg PO q 12 hr) was administered as liver function support. Hand-mixed topical corticosteroid (0.12% betamethasone valerate^j and 0.3% heparinoid^k) and 0.5% vitamin A oil^l was applied to reduce the inflammation and as a moisturizer. After cutaneous lymphoma had been excluded by frozen-section histopathology by day 3 from the first presentation at the authors' institution, cyclosporine^m (5.5 mg/kg PO q 24 hr) was started. A bacterial culture and sensitivity tests were performed on day 7. Escherichia coli and Staphylococcus spp. with methicillin resistance (i.e., penicillin, oxacillin, cephalexin, cefazolin, imipenem, lincomycin, enrofloxacin, minocycline, gentamicin, and sulfamethoxazole-trimethoprim) were isolated on the basis of the Clinical Laboratory Standard Institute guidelines.²³ Based on the culture, the previously prescribed antibiotics were substituted with fosfomycin calcium hydrateⁿ (25 mg/kg PO q 12 hr). The results of fungal cultures were negative on day 17.

Follow-up examination on day 22 after initial therapy revealed a 50% improvement in the skin lesions. The physical findings, PLT (359 × 10⁹/L), and CRP (46.67 nmol/L) were also improved. Because of persistent erythema and pruritus on the skin lesion and hyperthermia (39.8°C), azathioprine^o (2 mg/kg PO q 24 hr) was coadministered with cyclosporine. On day 36, the red blood cell count and packed cell volume had recovered (5.14 × 10¹²/L and 0.37, respectively) and CRP had decreased to 11.43 nmol/L. Cyclosporine was discontinued immediately. On day 64, 90% of the skin lesions were improved and were completely resolved by day 120 (Figure 4). The dog's physical condition and complete blood cell count had also improved. The azathioprine dosage was gradually tapered to 0.8 mg/kg PO q 48 hr. On day 309 from the time of referral, PSA-IgM, -IgG, and -C3 were re-evaluated by flow cytometry. The results were negative for IgM (7%), IgG (5.1%), and C3 (3.6%). The skin symptoms were controlled by azathioprine monotherapy until day 785.

View Figure

• Download as Powerpoint
• Download Figure

Discussion

To the author's knowledge, this is the first report to describe a dog definitively diagnosed with concurrent PF and IMT. Some forms of immune-mediated skin disease have been considered as possible causes of IMT and an epidemiological survey by Grindem et al. (1991) found that IMT in pemphigus complex is very rare.^16,18 The following forms of histological types of pemphigus are recognized in dogs and cats: PF, pemphigus erythematosus, panepidermal pustular pemphigus, pemphigus vulgaris, and paraneoplastic pemphigus.^1–9 The most common form of pemphigus complex is PF, of which symptoms are commonly less severe than in other forms, such as pemphigus vulgaris.^1–4 Although Grindem et al. (1991) did not specify the forms of pemphigus, recent retrospective studies have not reported those diseases to occur simultaneously.^7,8,19,20 Therefore, the presence of PF and IMT seems to be rare.

The present findings of PSA-Ig analysis reflect immunological signs of canine PF in addition to known physiological and hematological abnormalities. In canine PF, symptoms are generally localized to the skin. Although systemic symptoms, including anorexia, depression, fever, and weight loss are rare, they become obvious when erosive skin lesions progress over the entire body.³ Various hematologic abnormalities, such as moderate to severe leukocytosis and neutrophilia, mild nonregenerative anemia, thrombocytopenia, mildly to moderately elevated α₂-, β₂-_, and γ-globulins, and mild hypoalbuminemia have also been reported.^7,8,14,15 Those features are usually associated with the severity of skin symptoms; however, the results of routine blood tests, diagnostic imaging, and urinalysis have received relatively little focus because they are not considered to support a specific diagnosis.³ Focusing on thrombocytopenia among systemic symptoms, IMT was diagnosed on the basis of PSA-Ig levels measured using flow cytometry. Therefore, if dogs with PF have obvious systemic signs, systemic work-ups and immunological tests should be conducted to identify complicating factors and associated diseases.

A diagnosis of canine IMT is difficult because it is usually only concluded by excluding other potential causes of thrombocytopenia.^16–20 Dogs with IMT develop an increased risk of spontaneous hemorrhage when PLT fall below 10–30 × 10⁹/L, whereas O'Marra et al. (2011) reported that signs of bleeding were absent in 19% of 73 dogs showing platelet counts <50 × 10⁹/L.^16–20,24 Those asymptomatic dogs with moderate thrombocytopenia are not thoroughly evaluated in routine clinical practice because an invasive bone marrow examination is needed to exclude nonimmune-mediated thrombocytopenia. The authors of the current study diagnosed IMT in a dog by measuring PSA-Ig using flow cytometry at the time of the initial presentation at their hospital. Furthermore, the positive PSA-IgM and -IgG findings became negative when the thrombocytopenia was improved by immunosuppressive therapy with azathioprine. The study authors thus considered that the PSA-Ig level correlated with disease status and was useful for the diagnosis and evaluation of treatment in dogs with IMT.

Vaughan et al. (2010) reported that complicated cases of canine PF concurrent with allergic skin disease and other systemic diseases are significantly associated with the presence of eosinophil infiltration in intraepidermal pustules.⁸ In the present case, eosinophil infiltration was detected in the pustules, and the dog presented with pruritus at onset and a history of recurrent pyoderma. Concurrent IMT was also diagnosed. Those finding are consistent with those of Vaughan et al. (2010).⁸ One case, positive for eosinophil infiltration, was provisionally diagnosed with drug hypersensitivity against ampicillin, amoxicillin, enrofloxacin, and metronidazole.⁸ Because PF and IMT, as observed in the present case, are clinical symptoms in drug hypersensitivity, it was possible that those symptoms persisted at the initial presentation to the authors' hospital.^2,6,8,16 If canine PF is accompanied with eosinophilic infiltration, it may be necessary to examine underlying diseases.

Skin lesions had been evident in the case described herein since the first examination at the primary care veterinary hospital. However, the PLT (194 × 10⁹/L; reference range, 200–500 × 10⁹/L) at that time were essentially normal and gradually decreased thereafter. The possible causes of secondary IMT include infectious diseases, neoplasia, adverse reactions to drugs, and systemic autoimmune diseases such as SLE. In the current case, PF and adverse drug reaction were considered as the possible etiologies. IMT may also arise in association with circulating antikeratinocyte autoantibodies.³ Secondary IMT develops following SLE in some dogs, suggesting that immune complexes bound to platelets by complement-mediated immune adherence or nonspecific interactions cause IMT concurrent with SLE.¹⁶ The current case had IgG antibodies on the epidermis, had IgM and IgG antibodies on the platelet-cell surface, and was RF positive. Those findings indicate that in this case, PF may have been the cause of hematologic immune-mediated diseases such as IMT. The other possible cause of IMT was an adverse drug reaction. Prior to referral, IMT occurred after administration of cephalexin, which is a known cause of secondary IMT. In the present case, cefazolin sodium hydrate, which belongs to the same drug class as cephalexin, had been administered IV for 3 days before administration of cyclosporine. Because the symptoms did not worsen after cefazolin administration without any immunosuppressant drug, the possibility of an adverse drug reaction was considered low. Epidemiological studies are warranted to determine the prevalence and association of IMT with PF in dogs.

Conclusion

Canine PF is autoimmune pustular disease with symptoms that are generally localized to the skin. Although a low incidence of thrombocytopenia associated with pemphigus complex is reported, the study authors confirmed the diagnosis of concurrent PF and IMT based on histology findings, direct IF, and the presence of PSA-Ig. The dermatologic and hematologic signs were treated with immunosuppressive therapy using cyclosporine and azathioprine, and the dog has survived for >2 yr since the day of first presentation at the authors' animal hospital. Positive findings for PSA-IgM and IgG became negative when thrombocytopenia was improved after immunosuppressive therapy. Those findings indicate that IMT can develop with canine PF; thus, systemic evaluation should include precise hematologic assessment.