Diagnosis and treatment of Merkel cell carcinoma in a puppy: A comparative oncology approach

Abstract

A 5-month-old male golden retriever dog was presented with a solitary, rapidly growing cutaneous mass on the left thoracic wall. The medical history was unremarkable and no other clinical abnormalities were identified. After wide surgical resection, histopathological evaluation revealed a malignant tumor of small, round, atypic basophilic neoplastic cells positive for neuron-specific enolase and pancytokeratin, providing a diagnosis of Merkel cell carcinoma. With clean histological margins and absence of local or distant metastasis, no recurrence was reported during a minimum follow-up period of 20 mo. The dog developed an oral nodule 12 mo after surgery. Histopathological evaluation revealed a papilloma, and PCR was positive for Papillomavirus sp. Although DNA extraction and PCR for papillomavirus detection were carried out on the Merkel cell carcinoma tumor, no papillomavirus was detected. Merkel cell carcinomas can easily be confused with other cutaneous neoplasms in young dogs. Despite being malignant, they are frequently associated with a good prognosis after complete surgical excision.

Key clinical message:

i) Merkel cell carcinoma should be included in the differential diagnosis of a cutaneous mass in a puppy. ii) Wide excision provided an excellent long-term outcome in this clinical case.

Merkel cells are believed to have a role in sensory mechanoreceptors located in various tissues (1). In humans, Merkel cell carcinoma (MCC) is uncommon and carries a poor prognosis (2–4). Comparatively, canine MCC also represents a rare neoplastic disease but is associated with better outcomes (5–8). In both humans and dogs, excision with wide margins is recommended (8–10). Confirmation of the MCC diagnosis can be achieved using immunostaining for various epitopes, such as pancytokeratin, chromogranin A, synaptophysin, or neuron-specific enolase (NSE) (6–8,11). In humans, the implication of Merkel cell polyomavirus in MCC has been clearly established (12). Merkel cell polyomavirus has not been identified in either canine or feline MCC tissues (13). A feline papillomavirus (Felis catus papillomavirus type 2) was recently associated with feline MCC (14). However, to date in dogs, no clinical or genomic association has been made between the virus and canine MCC.

Malignant skin tumors are rarely reported in dogs < 1 y of age and represent 1.5% of all skin tumors in this group. Mast cell tumors and soft-tissue sarcomas are most commonly reported and skin carcinomas represent < 1% of cases (15). Moreover, MCC has been reported in adult humans and dogs (5–7,9,11) but it has never been reported in dogs < 5 y of age and was only reported in 5 humans < 20 y of age among 27 105 cases of MCC in the United States (2).

In the literature, dogs affected by MCC are always older than 5 y. The dog in this case report was very young (5 mo). In the current case, MCC was confirmed using clinical and immunohistological features and treated solely with surgical removal.

CASE DESCRIPTION

A 5-month-old male golden retriever dog was referred because of a cutaneous mass on the left thoracic wall that had progressed over the course of 2 wk. The medical history was unremarkable and no other clinical abnormalities were reported.

A non-painful, firm, non-ulcerated, erythematous, cutaneous mass 2.5 cm in diameter was located on the left ventral thoracic wall, ~5 cm caudal to the axillary region. The rest of the clinical examination was unremarkable.

Peripheral blood count, electrolytes, and serum biochemical parameters were within physiological ranges. Fine-needle aspirations of the mass were done, but cytological evaluation was not diagnostic because of low cellularity. Surgical resection was discussed and approved by the owner. The dog was anesthetized (fentanyl: 2 μg/kg, IV; propofol: 4 mg/kg, IV; midazolam: 0.2 mg/kg, IV) and maintained with isoflurane in oxygen and fentanyl constant-rate infusion (2 μg/kg per hour, IV). A skin incision was made with 1.5-centimeter lateral margins and a deep fascial plane was obtained underlying the left latissimus dorsi muscle. After excision, the mass was sent in 10% formaldehyde solution for histopathological evaluation. Latissimus dorsi fascia and subcutaneous tissue (2-0 USP, poliglecaprone) and skin (3-0 USP, nylon) were closed using continuous patterns. Postoperatively, morphine was given subcutaneously (0.2 mg/kg, q4h). The dog was discharged from the hospital on the following day. An antiinflammatory drug was prescribed (meloxicam: 0.1 mg/kg, PO, q24h for 5 d).

Formalin-fixed, paraffin-embedded tumor samples were cut into 3.0-micrometer-thick sections and stained with hematoxylin-eosin-saffron (HES). Macroscopically, the tumor measured 2.3 cm. No skin or muscle invasion was noted, but extensive intratumoral necrosis and hemorrhage were present. Microscopically, dense sheets of small, round, basophilic cells were observed with scant cytoplasm, high nuclear-cytoplasmic ratio, oval nuclei with finely dispersed chromatin, and inconsistent small nucleoli. The dermis and subcutaneous tissue were extensively and focally expanded by a neoplastic proliferation predominantly arranged in broad cellular sheets, with occasional delicate trabecular formations. The first diagnostic hypothesis, after evaluation by 2 Board-certified veterinary pathologists, was a malignant neuroendocrine tumor. Differential pathological diagnoses included large cell lymphoma, poorly differentiated mast cell tumor, histiocytoma, undifferentiated soft-tissue sarcoma, and rhabdomyosarcoma. The lateral clean margins were 10, 11, 15, and 13 mm, respectively, and the deep margins were 11, 7, 15, and 13 mm, with a deep muscular layer (Figure 1).

FIGURE 1.

Histology of Merkel cell carcinoma in a dog. A — Subgross view of the subcutaneous tumor. Absence of skin or muscle invasion. Extensive intratumoral necrosis and hemorrhage. Original magnification: 10×; scale bar = 1 cm. B — At higher magnification, dense sheets of small, round, basophilic cells with scant cytoplasm, high nuclear-cytoplasmic ratio, oval nuclei with finely dispersed chromatin, and inconsistent small nucleoli. Original magnification: 400×; scale bar = 50 μm. Stain: hematoxylin-eosin-saffron (HES) (A and B).

Immunohistochemistry (IHC) was done using 3-micrometer-thick serial tumor sections and antibodies for vimentin, CD3, chromogranin A, Iba-1, and synaptophysin, using an automated slide stainer (Autostainer 360; Labvision Epredia, Brada, Netherlands). For desmin, pancytokeratin, and NSE, a Benchmark XT Ventana (Roche Diagnostics, Indianapolis, Indiana, USA) was used as the IHC platform. Positive controls were internal, whereas negative controls were created by replacing the primary antibody with normal mouse or rabbit serum (Table 1). Immunohistochemistry showed a very intense cytoplasmic signal for NSE in 100% of neoplastic cells, a marked cytoplasmic signal for vimentin in 80% of neoplastic cells, and a moderate to marked signal for pancytokeratin (AE1–AE3) in 25% of cells (Figure 2). All other markers were negative for the signal. The definitive diagnosis was a high-grade MCC, though a solitary skin metastasis of a neuroendocrine carcinoma could not be totally excluded.

TABLE 1.

Antibodies used for immunochemistry and their corresponding expressions.

Antibody antigen	Source	Clone	Isotype	Dilution
AE1–AE3 (pancytokeratin)	(1)	M3515	IgG1, mouse	1:50
CD3	(2)	CD3-12	IgG, rat	1:100
Chromogranin A	(1)	DAK-A3	IgG, mouse	1:50
Desmin	(1)	D33	Monoclonal, mouse	1:30
Iba-1	(3)	EPR16589	IgG, rabbit	1:500
Neuron-specific enolase	(1)	BBS/NC/VI-H14	Monoclonal, mouse	1:200
Synaptophysin	(4)	Polyclonal	Rabbit	1:200
Vimentin	(1)	V9	IgG1, mouse	1:50

Sources: (1) Agilent Dako, Santa Clara, California, USA; (2) Bio-Rad, Hercules, California, USA; (3) Abcam, Cambridge, UK; (4) Cell Marque, Rocklin, California, USA.

FIGURE 2.

Immunohistochemistry of Merkel cell carcinoma in a dog. A — Cytokeratin AE1–AE3: strong cytoplasmic expression by 25% of tumor cells. Inset: positive internal control, cutaneous sebaceous gland with strong cytoplasmic AE1–AE3 expression. B — Vimentin: moderate cytoplasmic expression by most neoplastic cells. Inset: positive internal control, vimentin-positive peritumoral fibroblasts. C — Desmin: total lack of desmin expression by neoplastic cells. Inset: positive internal control, skeletal muscle. D — Neuron-specific enolase: systematic and strong cytoplasmic expression by tumor cells. Inset: negative control, with replacement of primary antibody with irrelevant mouse immunoglobulins. Original magnification: 400×; scale bars = 50 μm (A to D).

The histopathological and immunochemical results were available 6 wk after surgery. In consideration of the diagnosis, a computed tomographic examination of the thorax revealed a soft-tissue attenuating band along the scar and a mild enlargement of the left axillary lymph node (5 mm). No thoracic lesions were detected. In addition, an abdominal ultrasound was done at that time and was unremarkable. Six punch biopsies were obtained from the subcicatricial tissue, and a histopathological examination identified granulomatous and fibrous inflammatory scar tissue associated with suture material. Aspiration cytology of the left axillary lymph node was also done and no metastatic cells were identified. Considering the complete surgical resection, the frequent benign behavior of canine MCC, and financial/logistical concerns, surveillance only was elected, with follow-up visits every 2 mo for 6 mo.

On follow-up evaluation at 5 mo after surgery, the tumefaction had disappeared. Superficial ultrasound assessment of the scar and left axillary lymph node confirmed the resolution of subcutaneous swelling and lymphadenopathy. Ultrasound-guided fine-needle aspiration of the left axillary lymph node was done and the cytological aspect was normal. Nine months after surgery, general clinical and ultrasound evaluations of the abdominal cavity, surgical site, and left axillary region were unremarkable. At 17 mo of age (i.e., 12 mo after initial presentation), the dog was presented with a 3-millimeter gingival nodule located on the labial surface, ventral to the left mandibular canine (304). Similar nodules were also observed in 2 of the dog’s littermates at the age of 10 mo. The nodule was resected marginally. Histopathological evaluation of the gingival nodule revealed epithelial proliferation compatible with a papilloma.

The lesion was further tested using PCR (commercial laboratory: Laboklin GmbH & Co. KG, accreditation to standard DIN ISO 17025:2018, in-house validated PCR protocol) for the presence of papillomavirus, confirming a positive result for papillomavirus DNA. After this result was obtained, papillomavirus was sought in the excised MCC. For this purpose, DNA was extracted from the paraffin-embedded MCC tissue after a 24-hour lysis step at 56°C in the presence of proteinase K, and extraction of nucleic acids was completed according to an internal procedure on silica membrane (NucleoSpin; Macherey-Nagel, Dueren, Germany). To assess efficacy of the extraction step, a DNA control was added to the sample and detected using real-time PCR. Two PCR tests were conducted to detect papillomavirus DNA from the MCC; one was to detect canine oral papillomavirus as well as feline papillomavirus and closely related papillomavirus (“Canine oral papillomavirus assay”), whereas the other was to detect a broad range of feline and canine papillomaviruses as well as many types of human papillomavirus (Papillomavirus sp.) (16). Papillomavirus detection within the MCC was negative in both PCR tests.

At 20 mo after surgery, the dog’s owner was contacted by phone and reported the dog was in very good general health. No lesion compatible with local recurrence was reported.

DISCUSSION

This case report presents the first description of MCC in a puppy. The tumor was classified as stage IIa high-grade according to the staging classification scheme of MCC for humans (17). Wide surgical resection was completed and resulted in clean margins and long-term control of the disease (follow-up of 20 mo) without recurrence.

Although juvenile neoplasms are well documented in human children (18), veterinary reports are scarce (15). In children, only 1 to 2% of cutaneous neoplasms are malignant, with skin sarcomas most frequently reported. In contrast, cutaneous carcinoma is very rarely reported (19). Various cutaneous neoplasms have been reported in young dogs, but the incidence of skin carcinoma is very low (15). In the present case, the skin mass presented multiple clinically suspicious criteria of malignancy — namely, rapid growth, firm consistency, and adhesion to the surrounding tissues. However, cutaneous histiocytoma, which is the most frequently encountered cutaneous neoplasm in dogs younger than 2 y, often exhibits rapid growth despite its benign nature (20,21). In the present case, a histiocytoma was primarily suspected, but the malignant appearance on histology prompted Iba-1 immunochemistry, which excluded a histiocytic origin (22). Moreover, desmin and CD3 were negative, which ruled out a rhabdomyosarcoma or T-cell cutaneous lymphoma, respectively (23,24).

In humans, MCC is an uncommon type of skin cancer that is predominantly seen in those older than 70 y. Viral infection and radiation exposure are 2 leading factors implicated in its etiology (4). Particularly, infection by Merkel cell polyomavirus has been reported in up to 80% of human MCC cases (25). In cats, Felis catus papillomavirus has been reported in 85 to 95% of MCC cases with viral genomic integration (14,26). In dogs, the etiology of MCC is unknown (5–9,11). In cats, 80% of MCC cases also present with other proliferative cutaneous lesions that are generally other types of epithelial neoplasms (13). Similarly, in the present case, the dog subsequently exhibited an oral papilloma with positive papillomavirus detection, though papillomavirus was not detected in the MCC. However, as PCR was done after DNA extraction from paraffin tissue several months after formaldehyde fixation, the test’s sensitivity may have been reduced (27). In addition, the cuts made in the tumor could have been unrepresentative of an area with viremia, contributing to the negative result. Papillomavirus carrier status in dogs can be associated with immunodeficiency, which is also occasionally reported in human MCC (28). Perhaps the dog in the present case had some form of immunodeficiency, promoting MCC development.

In pediatric and young adult human patients, MCC is exceptionally rare. Among 27 105 incident cases of MCC, only 30 individuals were younger than 30 y and 5 were younger than 20 y (2). The average age at diagnosis of dogs with canine MCC is 10 y and the youngest reported was 5 y old (6,7). At the time of diagnosis, the puppy in the present case was 5 mo old, but the mass had been diagnosed by the primary veterinarian at 4 mo. A rapid increase in size is commonly reported in both humans and dogs, and was also observed in our case, corresponding to a high mitotic rate confirmed microscopically (4,7). In human patients, suspicion of MCC is confirmed via biopsy and immunochemistry (4). In the present case, the mass was excised after inconclusive fine-needle aspiration without initial excisional biopsy because histiocytoma was the first hypothesis and the objective of the surgery was complete removal of the lesion. Histopathological analysis of the excised MCC revealed the following common features reported in canine MCC: 10-micrometer, oval to polygonal neoplastic cells; fibrous connective and vascular tissue; and high mitotic index (7).

On H&E evaluation, MCC can easily be confused with other tumors, particularly lymphoma and histiocytoma (4,7). Therefore, immunochemistry is necessary for a definitive diagnosis in both humans and dogs. In humans, IHC positive staining for NSE and cytokeratin 20 and negative staining for S100 protein, leukocyte common antigen, vimentin, and HMB45 are enough to confirm the diagnosis (4). In dogs, only 3 studies reported on immunochemical characterization for MCC. Pancytokeratin (AE1–AE3)-positive labelling was observed in all 3 studies and in the present case (6,9,11). In the present case, NSE provided the most intense and consistent signal, as reported (11). However, chromogranin A and synaptophysin, which were commonly reported in previous observations, did not signal any neoplastic cells in our sample (6,9,11). Moreover, in our case, vimentin revealed an intense signal, which could suggest epithelial-mesenchymal transition of the neoplastic cells. This could also explain the low prevalence of pancytokeratin-positive cells in our case, contrary to other studies (6,9,11).

To differentiate the primary origin from a metastatic skin lesion of a visceral neuroendocrine carcinoma, a full staging procedure is necessary. For example, in a previous study, multiple cutaneous nodules were probably confused with cutaneous metastases of a visceral primary neuroendocrine neoplasm (6). In the present case, this scenario was unlikely, as the mass was solitary and the tumor was staging negative for the presence of any other neoplastic lesion.

In humans and cats, MCC carries a poor prognosis because of frequent lymph node and distant metastasis (3,4). Most metastases are observed in regional lymph nodes and the liver, lungs, subcutaneous tissue, and bones (29). Surgery is the primary treatment option for human MCC, with wide surgical resection and positive lymph node excision (10). Adjuvant radiotherapy, chemotherapy, and, more recently, immunotherapy are associated with variable responses (10). To date in dogs, MCC seems to exhibit less aggressive behavior (7,9,11). Nevertheless, 2 studies reported metastases with fatal evolution in the skin, lymph nodes, liver, and adrenal glands, despite chemotherapy (5,6). In the present case, staging was negative, and wide excision resulted in a disease-free status at 20 mo without any adjunctive therapy, similar to previous studies (7–9). In humans, juvenile forms of MCC have poorer prognoses, with more frequent metastases in those younger than 30 y (2). The more aggressive course of MCC observed in younger human patients was not observed in this young dog during 20 mo of follow-up.

This is the first description of an MCC in a puppy. The outcome was favorable in this case, with very good general health outcomes, no lesion compatible with local recurrence reported, and survival of at least 20 mo with surgical treatment. No viral association was confirmed in the present case, unlike in human and feline MCC. Further studies are required to evaluate the involvement of canine papillomavirus in canine MCC and the potential of canine MCC as a model of non-virus-induced MCC.