Primary cutaneous undifferentiated round cell tumor with concurrent polymyositis in a dog

Abstract

A cutaneous poorly differentiated round cell tumor with concurrent, non-suppurative, polymyositis was diagnosed in a hovawart dog. Histochemical staining, immunohistochemistry, and transmission electron microscopy findings suggested that the tumors cells were of myeloid, or possibly natural killer cell origin. The possibility that the concurrent polymyositis may represent a pre-neoplastic or paraneoplastic process is discussed.

A 4-year-old, male hovawart dog with a 3-month history of progressive weakness was referred to the Veterinary Teaching Hospital (VTH) of the Faculty of Veterinary Medicine of Turin (Italy). The dog was regularly vaccinated, and trained weekly as a working dog. There was no history of previous systemic disease or trauma. Two weeks prior to referral, empirical treatment with amoxicillin-clavulanate (Synulox; Pfizer, Rome, Italy), 20 mg/kg body weight (BW), BID, PO for 1 wk, for an ulcerated cutaneous lesion on the left thoracic wall was initiated by the referring veterinarian; the dog had no response to the treatment.

Case description

Upon presentation 2 skin lesions were noted. An approximately 4-cm in diameter, slightly raised, firm, ulcerated, nodular focus was located on the left lateral thoracic wall caudal to the scapula (Figure 1). The second similar mass, approximately 1 cm in diameter, was present on the left abdominal wall. Neurological examination revealed normal mentation and severe exercise intolerance. Proprioception was normal, postural reactions were decreased. Cranial nerve reflexes were intact but spinal nerve reflexes were decreased. There was generalized, moderate muscle atrophy; range of movement of the jaw and mastication were normal. The dog did not show pain on palpation or limb manipulation. Joints were neither swollen nor painful. Clinical signs were most consistent with generalized peripheral nervous system disease or possibly chronic, generalized polymyopathy.

Figure 1.

Ulcerated cutaneous lesion caudal to the left scapula.

The results of a complete blood (cell) count (CBC), blood chemistry tests, urinalysis, protein electrophoresis, and ACTH stimulation test were within reference intervals, except for an elevated creatine kinase (CK) [4942 U/L; reference interval (RI): 40 to 254 U/L]. Thyroid profile (tT4 and TSH) revealed a normal TSH and a decreased tT4 (14.2 nmol/L; RI: 16.7 to 18.0 nmol/L); euthyroid sick syndrome was suspected. Antinuclear antibody and acetylcholine receptor antibody tests were negative. Serological tests for Borrelia spp., Toxoplasma spp., Neospora spp., Dirofilaria immitis, Ehrlichia canis, and Leishmania infantum were negative. A polymerase chain reaction (PCR) test of blood samples for Babesia spp. was negative. Abdominal ultrasound and thoracic radiographs were unremarkable. Arthocentesis was not performed because the joints were neither swollen nor painful. Electromyography (EMG) and nerve conduction velocity (NCV) were normal.

Biopsies from the 2 cutaneous nodules, triceps brachial and cranial tibial muscles and the peroneal nerve were collected under general anesthesia. The muscle biopsy specimens were snap-frozen in isopentane pre-cooled in liquid nitrogen. Serial sections (8 μm thick) were stained by a standard panel of histological and histochemical stains and reactions, including haematoxylin and eosin (H&E), periodic acid-Schiff (PAS), oil-red O (ORO), Gomori trichrome, ATPase with pH 4.3 and pH 9.8 preincubation, reduced nicotinamide adenine dinucleotide dehydrogenase-tetrazolium reductase (NADH-TR), succinate dehydrogenase (SDH), cytochrome oxidase (COX), and non-specific esterase. Peroneal nerve biopsy specimens were post-fixed in osmium tetroxide, and dehydrated in serial alcohol solutions and propylene oxide prior to embedding in epoxy resin. Transverse sections (1 μm) were stained with toluidine blue for light microscopy.

In sections of the triceps brachii, myofibers were frequently small, atrophied, or lost and were separated by diffuse and severe endomysial infiltrates of mononuclear inflammatory cells composed mainly of macrophages and lymphocytes. Moderate endomysial fibrosis was evident. Similar, multifocal interstitial inflammatory infiltrates accompanied by moderate endomysial fibrosis and mild perimysial fibrosis were noted in sections of the cranial tibial muscle (Figure 2). In both muscles, inflammatory infiltrates were associated with invasion and phagocytosis of hypereosinophilic myofibers which lacked normal striations and often had slightly floccular to granular sarcoplasm. Foci of myofiber regeneration were present in sections of the cranial tibial muscle. No significant abnormalities were present in muscle sections stained with PAS and ORO. Oxidative stain reactions and fiber type distribution appeared normal. Protozoal organisms were not detected. Intramuscular nerve branches within muscle sections and sections of the peroneal nerve were microscopically unremarkable. These findings were consistent with chronic, non-suppurative, polymyositis.

Figure 2.

Section of the cranial tibial muscle. There are endomysial inflammatory infiltrates composed of macrophages, lymphocytes and rare neutrophils, associated with phagocytosis of non-necrotic fibers and myonecrosis (H&E, Bar = 100 μm).

Histological examination of incisional biopsies from the 2 cutaneous nodules revealed a monomorphous population of atypical round cells with round to oval to indented and pleomorphic nuclei, one variably evident nucleolus, variably distinct cell borders, and an abundant eosinophilic homogenous cytoplasm (Figure 3). Severe anisocytosis, anisokaryosis, atypical mitoses (mitotic index of > 30 per 10 HPF) and necrotic areas were also observed. The monomorphic infiltrates were locally invasive, expanding and effacing the dermis, and extending into the subcutis. The overlying epidermis and hair follicles were also involved. Poor hyaline stroma separating tumors cells was observed. A provisional diagnosis of an undifferentiated round cell tumor was made. Toluidine blue and lysozyme stains were negative. Anti-CD3, anti-CD79a, and anti-CD117 stains performed on paraffin-embedded tissue sections were also negative.

Figure 3.

Haired skin, deep dermis: multifocal aggregates of round neoplastic cells invading vascular walls and associated with microhemorrhages and necrosis (H&E, Bar = 50 μm).

Given the complexity of the case, all histological sections of the skin and new cytological specimens were submitted for a second opinion and for additional immunocytochemical stains to provide a precise diagnosis. All primary antibodies that were applied are listed in Table 1. Neoplastic cells were CD18, and CD11a intensely positive, while only 30% of the cells were CD11b positive. Tumor cells did not exhibit positive staining with the following antibody markers, CD3, CD4, CD8-alpha, CD20, CD21, CD79a, CD1c, CD1a, CD117, CD11c. At this time, differential diagnosis for these undifferentiated, cutaneous round cell tumors was early myeloid, immature monocytic/macrophage or possibly natural killer cell origin. After the diagnosis of chronic, non-suppurative polymyositis, and while awaiting the results of the additional immunocytochemical stains, the dog was treated with a tapering course of oral prednisone (Deltacortene; Bruno Farmaceutici Spa, Rome, Italy), beginning at 1 mg/kg q12h over a 7-week period. Concurrently, because of the ulcerated cutaneous nodule, oral amoxicillin-clavulanate (Synulox), 20 mg/kg BW, q12h, was given for 4 wk.

Table 1.

Antigen targets for nodules located in internal organs and on the skin

Antigen	Clone	Type	Dilution	Expression	Source
CD3-ɛ	CD3-12	Rat monoclonal	1:20	negative	Seroteca
CD20	Polyclonal	Rabbit polyclonal	1:400	negative	Neomarkersb
CD79a	M7051	Mouse monoclonal	1:20	negative	DAKOc
CD1a	CA9.AG5	Mouse monoclonal	1:20	negative	L.A.B.L.d
CD1c	CA13.9H11	Mouse monoclonal	1:20	negative	L.A.B.L.d
CD4	CA13.1E4	Mouse monoclonal	1:20	negative	L.A.B.L.d
CD8-alpha	CA9.JD3	Mouse monoclonal	1:20	negative	L.A.B.L.d
CD18	CA1.4E9	Mouse monoclonal	1:20	100% intensely positive	L.A.B.L.d
CD21	CA2.1D6	Mouse monoclonal	1:20	negative	L.A.B.L.d
CD117	A4502	Rabbit polyclonal	1:500	negative	DAKOc
CD11a	CA16.1B11	Mouse monoclonal	1:20	100% positive	L.A.B.L.d
CD11b	CA16.3E10	Mouse monoclonal	1:20	30% positive	L.A.B.L.d
CD11c	CA11.6A1	Mouse monoclonal	1:20	negative	L.A.B.L.d

^aSerotec, Kidlington, Oxford, UK.

^bNeomarkers, Fremont, California, USA.

^cDAKO, Glostrup, Germany.

^dDr. Moore’s Leukocyte Antigen Biology Laboratory, Davis, California, USA.

Seven weeks after the initial evaluation, although clinical signs of weakness had markedly improved, the dog was presented to the VTH for decreased appetite, weight loss, and development of multiple lesions on the skin, penis mucosa (Figure 4), and oral cavity. The dog was rehospitalized and scheduled for a complete work-up, but died suddenly the following night. A bone marrow aspirate and core biopsy were performed immediately after death. The marrow exhibited normal cellularity and a normal myeloid:erythroid ratio with no indications of neoplasia. At necropsy, multifocal whitish nodules 1- to 3-cm were present in the dermis of the preputial haired skin and in internal organs including liver, kidney, oral cavity, and lungs. The remaining viscera and the central nervous system (CNS) were unremarkable. Microscopic evaluation of each of the pale nodules revealed neoplastic infiltrates similar to that described in the previously examined skin lesions. Immunohistochemical and immunocytochemical staining of sections from these visceral and skin tumors was also similar to that described in the previously examined skin masses.

Figure 4.

Ulcerated lesions of the preputial skin and glans penis.

Samples of skin nodules were fixed in glutaraldehyde, postfixed in osmium tetraoxide, and embedded in epoxy resin for transmission electron microscopy (TEM). Ultrastructural evaluation revealed round to polygonal neoplastic cells with round nuclei, fibrillo-granular nucleoli, moderate amount of cytoplasm containing mitochondria, rough endoplastic reticulum, free ribosomes, and variable numbers of round to oval electron-dense granules (0.25 to 0.3 μm in diameter) (Figure 5). Clinical history, histopathology, immunocytochemistry, immunohistochemical analysis, and TEM supported a diagnosis of a primary cutaneous undifferentiated leukocyte neoplasia of putative myeloid or, less likely, natural killer cell origin. The polymyositis that developed concurrently with the tumor was hypothesized to have a paraneoplastic origin.

Figure 5.

Electron micrograph of a skin biopsy: note the electron-dense granules (arrow) and rough endoplasmic reticulum (arrowhead) within the cytoplasm of a neoplastic cells (Uranyl acetate and lead citrate, Bar = 0.3 μm).

Discussion

The current case report describes a cutaneous undifferentiated round cell tumor of leukocyte origin in a dog with concurrent polymyositis. Within neoplastic infiltrates, cytoplasmic granules were not noted with toluidine blue staining and tumor cells did not exhibit CD117 staining, thus a diagnosis of disseminated mast cell tumor was thought unlikely. In addition, lack of staining with CD3, CD20, CD79a, CD4, CD8-alpha, CD21, CD1a, CD1c, CD11c and lysozyme was not supportive of a lymphoid or well-differentiated macrophage/dendritic cell origin. Tumor cells stained strongly with CD18 and CD11a, consistent with leukocyte-derived cells. Approximately 30% of the tumor cells also stained positively with CD11b, which is a marker expressed by monocytes/macrophages, neutrophils, and NK cells (1). Specific markers for canine NK cells are not available. Based on these findings, tumors cells were thought to most likely be derived from early myeloid or immature monocyte/macrophage cell lines. To our knowledge, the prominent expression of CD18 in tumor cells in this case is not typical of NK-related cells, thus although the latter was considered, it was thought a less likely possibility.

Blastic transformation of neoplastic cells or aberrant loss of receptor expression could explain the lack of staining with some markers, such as lysozyme, a highly specific myeloid marker, as observed in human medicine (2,3). Development of cutaneous lesions in human patients with myeloproliferative disorders has been reported and can be due to the course of the disease itself or induced by some medication (4,5). Rarely, skin involvement may be the first manifestation of acute leukemia, in which the bone marrow biopsy shows a precursor lesion (myeloproliferative or myelodysplastic syndrome) or normal findings (6). The diagnosis of these disorders may be challenging. In addition to extensive immunocyto- and immunohistochemical analysis, often additional diagnostic testing and correlation with cytogenetic information is required (7). Based on these considerations, the development of skin lesions as the first manifestation of a myeloproliferative disorder was considered but could not be confirmed in this dog, even with a normal bone marrow and CBC.

Involvement of internal organs with a normal marrow noted at necropsy in this case strongly suggests primary cutaneous neoplasia with widespread metastasis. In humans, NK cell primary skin tumors are aggressive neoplasms characterized by CD56 expression in the absence of lineage-specific markers of T-cells, B-cells, or myelomonocytic cells, and are associated with a poor prognosis (8,9). Additionally, mixed phenotype myeloid/natural killer (NK) cell precursor acute leukemia with multiple subcutaneous nodules has also been reported in humans (10). The rapid progression of disease in this dog is similar to that described in primary NK cell cutaneous disorders in humans.

In order to identify the specific cellular lineage, TEM was selected as an additional diagnostic tool. Cytoplasmic granules common to both myeloid and NK cells were noted. Ultrastructurally, NK cells are generally characterized by coarse marginated heterochromatin and inconspicuous intracytoplasmic organelles (11). The abundant organelle content noted in the tumor cells further supported the likelihood that the tumor cells were derived from myeloid cells. The presenting complaint of generalized weakness in this dog was the most pronounced clinical finding which was attributed to generalized, chronic, non-suppurative polymyositis. Polymyositis in humans is considered a rare disease (12). In dogs the incidence is not known, but the results of one study suggest that it is relatively common (13). Clinical findings in dogs with generalized inflammatory myopathies may include some or all of the following: muscle weakness, muscle atrophy, myalgia, increased serum muscle enzymes (especially CK), and abnormal electromyographic activity. In chronic cases, such as noted in this case, microscopic lesions consist predominantly of endomysial and perimysial, perivascular to interstitial lymphocytic and histiocytic infiltrates with variable degrees of myofiber degeneration, atrophy and regeneration, and endomysial fibrosis. Lesions are typically multifocal (14). Differential diagnoses for inflammatory myopathies include infectious, immune-mediated, preneoplastic, paraneoplastic, toxic or drug-induced, and nutritional (13,15). Based on negative serological titers, PCR results, CBC, blood chemistry, urinalysis, serum protein electrophoresis, and absence of organisms in muscle biopsies, an infectious etiology was considered unlikely. Furthermore, since neoplastic cells were not identified within skeletal muscles initially or at postmortem examination, polymyositis could not be explained by direct neoplastic infiltration, as recently hypothesized (16). Although the definitive cause of inflammatory myopathies often remains unknown, most cases are thought to have an immune-mediated pathogenesis. Interestingly, the dog in this case exhibited clinical signs of polymyositis approximately 3 mo before developing cutaneous lesions. Whether this was a simple coincidence of timing or there was association between the 2 conditions is not known.

Recent evidence suggests an association between malignancy and inflammatory myopathies in humans (17,18). Ovarian, lung, pancreatic, breast, and stomach cancer have been commonly associated with myopathies (19,20). In dogs, few reports have also linked polymyositis with lymphoma (13,16,21), thymoma (22), myeloid leukemia, bronchogenic carcinoma, tonsillary carcinoma (23,24), anaplastic round cell tumor and plasmacytoma (13). In humans, acute and chronic myeloid leukemia associated with polymyositis has been rarely reported (25,26), but no report described primary nodular skin tumors. An altered immune surveillance, in which antigens targeted for an immune response are expressed at high levels in the inciting tumor and muscle tissue, has been proposed in the pathogenesis of this relationship (17–21). In this case, proof of an association between neoplasia and polymyositis was not demonstrated; therefore, a paraneoplastic pathogenesis of inflammatory polymyopathy remains speculative at best, and the concurrent presence of these 2 diseases in the same dog could have been a simple coincidence.

In conclusion, this report described a malignant undifferentiated cutaneous round cell neoplasm with widely disseminated visceral and mucosal tumors which we speculated to be of early myeloid, immature monocytic/macrophage or possibly NK cell origin based on immunohistochemical, histochemical and TEM findings. In addition, concurrent polymyositis which may represent a pre- or paraneoplastic condition is reported. CVJ