Histology and Immunohistochemistry of Seven Ferret Vaccination-site Fibrosarcomas

Vet Pathol 4:288 293 (23) Histology and Immunohistochemistry of Seven Ferret Vaccination-site Fibrosarcomas J. S. MUNDAY, N.L.STEDMAN, AND L. J. RICHEY Athens Diagnostic Laboratory and Department of Veterinary Pathology, College of Veterinary Medicine, University of Georgia, Athens, GA Abstract. The anatomical location, histology, and immunohistochemistry of 1 ferret dermal and subcutaneous fibrosarcomas were examined. Seven of the 1 tumors were from locations used for vaccination. All fibrosarcomas contained spindle-shaped cells surrounded by variable quantities of connective tissue stroma. However, vaccination-site fibrosarcomas (VSFs) subjectively contained a higher degree of cellular pleomorphism. Multinucleated cells were present in three of seven VSFs but not in any of the nonvaccination-site fibrosarcomas (NVSFs). Large histiocytic cells, interpreted as macrophages, containing intracytoplasmic basophilic granular material were observed in two VSFs but not in any of the NVSFs. Five VSFs contained peripheral lymphoplasmacytic aggregates. Immunohistochemically, three VSFs stained with anti smooth muscle actin antibodies and one stained with antibodies against desmin. No expression of muscle cytoskeletal filaments was observed in any NVSF. Filaments interpreted as actin were visible in both the VSFs examined ultrastructurally. One of the VSFs examined ultrastructurally contained intracytoplasmic crystalline material. The preferential development of subcutaneous fibrosarcomas in vaccination sites suggests that, as in cats, vaccination may promote local sarcoma development in ferrets. Additionally, some of the histologic, immunohistochemical, and ultrastructural features of these tumors are similar to those reported for feline vaccine-associated sarcomas. To the authors knowledge, vaccination has not previously been reported to be oncogenic in any species other than cats. Key words: Adjuvant; ferrets; fibrosarcoma; oncogenesis; rabies; ultrastructure; vaccination. In cats, a relationship between vaccination and sarcoma development was first suggested in 1991. 6 An increased tumor incidence was observed during the following 5 years, and the feline vaccine-associated sarcoma subsequently became a well-defined clinical entity. 12 Factors suggested to have contributed to the increased sarcoma development in the early 199s include increased use of killed-virus adjuvanted rabies vaccines, the development and use of feline leukemia virus vaccines, and increased subcutaneous vaccine administration. 1,12,14 The precise mechanisms by which vaccination promotes sarcoma development in cats is unknown, but it is considered likely that chronic inflammation is important. 2,1 Histologic, 1,2 ultrastructural, 11 and immunohistochemical 5 features of feline vaccine-associated sarcomas have been reported. When compared with nonvaccine-associated sarcomas, vaccine-associated sarcomas are significantly more frequent in the subcutis. 2 They also are more likely to contain areas of necrosis, 2 peripheral lymphocyte aggregates, 2,11 cellular pleomorphism, 2 and blue-gray material within macrophages. 11 Using electron microscopy, this intrahistiocytic crystalline material was identified as aluminum based. 11 A partial myofibroblast differentiation of neoplastic cells has been demonstrated in some vaccine-associated sarcomas using immunohistochemistry and electron microscopy. 5,11 Cats, dogs, and ferrets routinely receive killed-virus rabies vaccinations. However, to the authors knowledge, vaccine-associated oncogenesis has only been reported in cats. There are an estimated six million ferrets in the United States, making them the third-most numerous interactive pet. 16 Ferrets are recommended to be vaccinated against rabies at 3 months of age and then annually. 17 Ferrets also are recommended to receive modified live, nonferret cell culture origin canine distemper vaccine. 17 This vaccine is initially given at 9 and 11 weeks of age and is then repeated every 2 3 years. 17 Ferrets are routinely vaccinated in the subcutaneous tissues of the interscapular region (S. Hernandez-Divers, personal communication). Although a ferret vaccination-site fibrosarcoma (VSF) was reported in a letter to the Journal of the American Veterinary Medical Association, 15 no histologic description of the tumor or evidence supporting an association between vaccination and tumorigenesis was included. The aim of the present study was to determine whether an increased proportion of ferret subcutaneous sarcomas develop at sites used for vaccination. The histology, im- 288

Vet Pathol 4:3, 23 Ferret Vaccination-site Fibrosarcomas 289 Table 1. Location on the body, age of the ferret, vaccination history, histology, and immunohistochemistry of 1 ferret fibrosarcomas.* Ferret No. 1 2 3 4 5 6 7 8 9 1 Site Interscapular Interscapular Dorsal neck Interscapular Dorsal thorax Dorsal neck Dorsal neck Base of tail Carpus Ventral abdomen Age (years) Vaccine Inflammation 4 9 6.5 4 1 2 3.5 1 Rabies Rabies 2 2 Giant Cells Necrosis SMA (%) Des (%) * SMA percentage of neoplastic cells that demonstrated cytoplasmic staining with antibodies against smooth muscle actin; Des percentage of neoplastic cells that demonstrated cytoplasmic staining with antibodies against desmin; CDV canine distemper virus; no records available. Inflammation, giant cells, and necrosis are graded on a scale of (none) to (high). Vaccination sites. Vaccinations received in the last 12 months. 2 5 3 2 1 munohistochemistry, and ultrastructure of the ferret VSFs were then compared with feline vaccine-associated sarcomas. Materials and Methods A retrospective survey of pathologic accessions to the University of Georgia between 1996 and 22 revealed 18 sarcomas that were excised from the skin and subcutaneous tissues of ferrets. Of these, 1 were histologically diagnosed as fibrosarcomas, whereas eight were diagnosed as leiomyosarcomas. Immunohistochemical staining revealed that greater than 9% of the neoplastic cells in all leiomyosarcomas expressed both smooth muscle actin and desmin. Cutaneous leiomyosarcomas have not been associated with vaccination in cats, and a recent review of 12 cutaneous leiomyosarcomas in ferrets did not reveal a predilection for vaccination sites. 13 Therefore, cutaneous leiomyosarcomas in ferrets were considered unlikely to be related to vaccination and were excluded from the study. No subcutaneous rhabdomyosarcomas, malignant fibrous histiocytomas, myxosarcomas, or chondrosarcomas had been diagnosed in ferrets during the last 6 years. All 1 fibrosarcomas had been submitted by veterinarians from clinical practices in Georgia. Information regarding the location of the tumor on the body, the vaccination history, and the age of the ferret was derived both from the original submission form and by contacting the submitting veterinarian. Histologic examination of the neoplasms was performed using 4- to 6- m-thick hematoxylin and eosin stained sections of formalin-fixed, paraffin-embedded tissue. Immunohistochemistry was performed using formalin-fixed, paraffin-embedded 4- to 6- m-thick sections. Sections were heated for 3 minutes in a 6 C oven, deparaffinized, and stained using a TechMate 5 Autostainer (BioTek Solutions Inc., Santa Barbara, CA). Antibodies against vimentin (Biogenex, San Ramon, CA), smooth muscle actin (Dako Corporation, Carpinteria, CA), and desmin (Biogenex) were used. The avidin biotin peroxidase complex system (Vector Laboratories, Burlingame, CA) was used to observe all immunohistochemical reactions. Neoplasms from two cases (ferret Nos. 2 and 3) were examined ultrastructurally. Representative specimens of paraffin-embedded tissue were deparaffinized in xylene, rehydrated, and osmicated for 1 hour. Sections of these were stained with toluidine blue and examined using light microscopy to identify areas for ultrastructural examination. Areas of tumor containing marked cellular pleomorphism, immunohistochemically identified areas of smooth muscle actin expression, and intracellular granular material were selected for examination. Ultrathin sections were then cut using an ultramicrotome with glass knives and stained with uranyl acetate and lead citrate. Results Seven of the 1 ferret fibrosarcomas developed in locations routinely used for vaccination (interscapular, dorsal aspect of the neck, and dorsal aspect of the thorax) (Table 1). The three neoplasms that developed at sites not used for vaccination were located at the base of the tail, on the carpus, and on the ventral aspect of the abdomen. Vaccination records were available for 6 of the 1 ferrets that developed fibrosarcomas; five of these had VSFs, whereas one had a nonvaccination-site fibrosarcoma (NVSF). Three ferrets with VSFs had been vaccinated by the same veterinarian who had submitted

29 Munday, Stedman, and Richey Vet Pathol 4:3, 23 the tumor, and it could be confirmed that the site of previous vaccination and tumor development were similar in these cases. Three ferrets with VSFs and one with an NVSF had received both rabies and canine distemper vaccinations within the last 12 months. Two ferrets, both with VSFs, had been vaccinated only against rabies within the previous 12 months. However, complete vaccination histories of neither ferret were available, and it is possible that they had previously received canine distemper vaccinations. No vaccination records were available for 4 of the 1 ferrets, two with VSFs and two with NVSFs. None of the ferrets was reported to have had other injections in this area. However, because of the incomplete histories of most of these cases, other injections cannot be definitively excluded. Histologically, all VSFs were moderately well demarcated and nodular (Fig. 1). All seven were situated in the subcutis and ranged in size from 1 to 25 mm in diameter. Skeletal muscle fibers of the panniculus muscle were seen ventral to all seven tumors. Three VSFs were not in contact with the panniculus muscle, two abutted the muscle causing localized stretching and distortion, and two infiltrated the panniculus muscle (Fig. 2). One NVSF was situated within the superficial dermis. This tumor was poorly demarcated and sessile. Although the other two NVSFs were within the subcutis, panniculus muscle was not observed in sections of either NVSF. All ferret fibrosarcomas consisted of neoplastic cells arranged in elongated interwoven bundles supported by a moderate fibrovascular stroma. Compared with NVSFs, cells in VSFs appeared less organized, with most tumors containing areas of shorter, more tightly packed bundles along with areas that contained loose whorls. Neoplastic cells in five VSFs showed marked cellular pleomorphism. Larger cells (up to 3 m) in these neoplasms often contained large, round, central nuclei with prominent, often multiple, nucleoli and large quantities of eosinophilic cytoplasm, whereas smaller cells (typically around 1 m) more often contained homogenous, elongated, central nuclei and small quantities of faintly eosinophilic cytoplasm. Large oval to round histiocytic cells containing intracytoplasmic basophilic granular material were observed in two VSFs (ferret Nos. 2 and 3) (Fig. 3) and zero of three NVSFs. The cells were interpreted as macrophages and were more frequent toward the periphery of the tumor. All 1 fibrosarcomas had at least one mitotic figure/4 field with up to four mitotic figures/4 field present in some tumors. Multinucleated cells were present in three of seven VSFs but in none of the NVSFs. The multinucleated cells were generally scattered throughout the neoplasm and ranged in number from 1 to 1 per 1 4 fields. Necrosis was observed in all VSFs and two of three NVSFs. Small foci of necrosis were present throughout five VSFs and one NVSF. However, two VSFs and one NVSF contained more marked necrosis with the formation of a central cavity that was lined by necrotic cell debris. Mild inflammation was seen, associated with necrotic cells in all tumors that contained necrosis. However, five VSFs additionally contained small to moderate numbers of lymphocytes and plasma cells that were not associated with areas of necrosis. These cells were diffusely scattered throughout the neoplasms and also were present as peripheral nodular lymphocytic aggregates (Fig. 4). The peripheral lymphocytic aggregates were numerous and large (3 cells in diameter) in two tumors but less frequent and small (1 cells in diameter) in the other three. Cells in all 1 fibrosarcomas had cytoplasmic staining with antibodies against vimentin. Three VSFs demonstrated cytoplasmic staining with anti smooth muscle actin antibodies (Fig. 5). This staining was variable in intensity, and between 2% and 5% of cells expressed the antigen. The smooth muscle actin positive cells were within foci randomly scattered throughout two VSFs, whereas antigen expression in one VSF (ferret No. 6) was predominantly confined to peripheral cells that partly encapsulated the tumor. Approximately 1% of the peripheral cells in ferret No. 6 also stained with antibodies against desmin (Fig. 6). Because of their peripheral location, these cells may be reactive rather than neoplastic. Expression of muscle cytoskeletal filaments was not observed in any of the NVSFs. Ultrastructurally, both examined VSFs were composed of a pleomorphic population of spindle cells that have oval to convoluted nuclei, euchromatin, and prominent nucleoli. Bi- and trinucleate cells with similar nuclei also were observed in both tumors. In both VSFs, occasional spindle cells contained bundles of subplasmalemmal, elongate, thin filaments that were consistent with actin. Rare intracytoplasmic aggregates of dense, needlelike, crystalline material were observed in the VSF from ferret No. 3. Because of the artifacts formed during formalin fixation, the type of cell that contained this material could not be definitively determined. Ferrets that had VSFs ranged in age from 1 to 9 years with a median age of 4 years. One ferret that had an NVSF was 1 year old. The ages of the other two ferrets with NVSFs had not been recorded. Follow-up information was available for three ferrets with VSFs and one with an NVSF. One ferret that had a VSF was euthanatized because of tumor recurrence 3 months after excision (ferret No. 4). In two ferrets, no VSF recurrence had been observed at 8 (ferret No. 1) and 12 (ferret No. 6) months after surgical

Vet Pathol 4:3, 23 Ferret Vaccination-site Fibrosarcomas 291 Fig. 1. Subcutaneous fibrosarcoma; ferret No. 1. Sarcoma with a central area of necrosis (*) and moderately well defined fibrous capsule (arrows). HE. Bar 3 m. Fig. 2. Subcutaneous fibrosarcoma; ferret No. 3. A pleomorphic population of spindle-shaped cells infiltrate between skeletal muscle fibers of the panniculus muscle (*). The cells have a high mitotic rate, and small numbers of lymphocytes are seen surrounding the tumor (lower left). HE. Bar 2 m. Fig. 3. Subcutaneous fibrosarcoma; ferret No. 3. Neoplastic cells are accompanied by round cells that have large quantities of basophilic granular cytoplasm (arrows). HE. Bar 2 m. Fig. 4. Subcutaneous fibrosarcoma; ferret No. 2. The periphery of the sarcoma contains a nodular infiltration of lymphocytes. HE. Bar 1 m. Fig. 5. Subcutaneous fibrosarcoma; ferret No. 2. Cytoplasm of approximately 5% of the neoplastic cells stains positively with antibodies against smooth muscle actin. Avidin biotin peroxidase complex system and Gill s hematoxylin. Bar 2 m. Fig. 6. Subcutaneous fibrosarcoma; ferret No. 6. Small numbers of peripheral cells stain positively with antibodies against desmin. These cells partly encapsulate the tumor, and definitive distinction between a reactive and a neoplastic cells population is not possible. Avidin biotin peroxidase complex system and Gill s hematoxylin. Bar 2 m.

292 Munday, Stedman, and Richey Vet Pathol 4:3, 23 excision. Histologically, all three of these VSFs had appeared completely excised. One ferret that developed an NVSF (ferret No. 1) was euthanatized 6 months after incomplete surgical excision because of tumor recurrence. Discussion Seven of 1 dermal and subcutaneous ferret fibrosarcomas submitted to the University of Georgia during the last 6 years were from areas routinely used for injection. The preferential development of fibrosarcomas in these areas suggests that vaccine administration may predispose ferrets to sarcoma development. Although the injection of other medications cannot be excluded, observations in felines suggest that adjuvanted vaccines are most likely to be involved in tumor development. Histologic features of subcutaneous sarcomas that are suggestive of an association with vaccination have been described in cats. 1,2 These features were compared with ferret VSFs and NVSFs. Sixty-nine percent of feline vaccine-associated sarcomas contain lymphocytic inflammation, predominantly in peripheral areas. 2 Peripheral lymphocytic inflammation was observed in five of seven ferret VSFs. Intracellular basophilic material, thought to be adjuvant, was present in 22% of feline vaccine-associated sarcomas. 1 A similar intracellular basophilic granular material was present in two of seven ferret VSFs. Neither peripheral lymphoid aggregate nor intracellular basophilic granular material was observed in NVSFs. Subjectively, cells in VSFs appeared larger, rounder, and more pleomorphic than cells in NVSFs. Greater cellular pleomorphism has been described as a feature of feline vaccine-associated sarcomas. 2 Although the presence of lymphocytic inflammation and basophilic material provides evidence for vaccine association in ferrets, comparison of other features of feline vaccine-associated sarcomas did not provide additional supportive evidence. In cats, a subcutaneous location has been reported to be more common in vaccine-associated sarcomas than in nonvaccine-associated sarcomas. 2 In ferrets, all seven VSFs were within the subcutis, however, two thirds of NVSFs also had a subcutaneous location. Cavitating tumor necrosis has been described as a feature of feline vaccine-associated sarcomas. 1 Necrosis with the formation of a central cavity was observed in two of seven VSFs but also was present in one of three NVSF. Three VSFs but no NVSFs contained multinucleated cells. Although 5% of feline vaccine-associated sarcomas have been reported to contain multinucleate cells, 1 multinucleate cells were not reported to be significantly more frequent in feline vaccine-associated sarcomas when compared with nonvaccine-associated sarcomas. 2 Similarities between the immunohistochemical features of feline vaccine-associated sarcomas and ferret VSFs were observed. As has been reported in 64% of feline vaccine-associated sarcomas, 1 three of seven ferret VSFs contained neoplastic cells that stained positively with antibodies against smooth muscle actin. Desmin expression has been reported to be less common in feline vaccine-associated sarcomas 5 and was observed in only one of seven ferret VSFs. The peripheral distribution of muscle intermediate filaments that was observed in ferret No. 6 has previously been reported in feline vaccine-associated sarcomas. 1 The expression of smooth muscle actin in some tumors is suggestive of, as in feline vaccine-associated sarcomas, 11 partial myofibroblast differentiation. The ultrastructural features of the ferret VSFs provide supportive evidence of a role of vaccination in tumorigenesis. The cytoplasmic crystalline material observed in one tumor was morphologically similar to the cytoplasmic aluminum inclusions present in feline vaccine-associated sarcomas. 11 These aluminum inclusions are hypothesized to originate from vaccine adjuvant. 7 Although the chemical composition of the material was not identified, its presence suggests that a foreign material may have contributed to the development of this tumor. Actin filaments were reported in 16 of 2 feline vaccine-associated sarcomas. 11 Filaments that were interpreted as actin were observed in both ferret VSFs. In cats, additional evidence supporting a causal relationship between vaccination and sarcoma development was derived from epidemiologic studies. 8 In the present study, five of the animals that developed VSFs had received a rabies vaccine within the previous 12 months. However, vaccination records were not available for the remaining two ferrets that developed VSFs. Three ferrets also had received vaccinations against canine distemper within a year of developing a VSF. Nonadjuvanted live virus vaccines have not been shown to cause sarcoma development in cats 12 and, if ferret fibrosarcomas are associated with vaccination, the rabies vaccine is considered more likely to be implicated than the distemper vaccine. Cats with vaccine-associated sarcomas have been reported to develop tumors at a younger age than do cats with nonvaccine-associated sarcomas. 2 Because the age of only one of the three ferrets that developed NVSFs was known, it is not possible to compare the ages of ferrets with VSFs and NVSFs in the present study. The involvement, if any, of retroviruses in the development of feline vaccine-associated sarcomas is not yet clear. 3,9 To the authors knowledge, no retrovirus has been identified in ferrets. It is possible to interpret this as evidence against retroviral involvement in the

Vet Pathol 4:3, 23 Ferret Vaccination-site Fibrosarcomas 293 development of ferret VSFs. However, a retroviral infection was strongly suspected to have caused an outbreak of lymphoma in a ferret colony. 4 Therefore, it is difficult to definitively exclude retroviral infection as a contributing factor in the development of these tumors in ferrets. In conclusion, the high proportion of ferret subcutaneous fibrosarcomas that developed in vaccination sites suggests that vaccination may predispose ferrets to local tumor development. Additionally, feline vaccine-associated sarcomas and ferret VSFs share some histologic, immunohistochemical, and ultrastructural characteristics. To the authors knowledge, this is the first evidence that vaccination may cause neoplasia in a nonfeline species. Acknowledgements We thank Louri Caldwell, Melissa Gay, Michelle Graham, Julie Joiner, Elaine Barrentine, W. L. Steffens, and Mary Ard for their expert technical assistance. References 1 Couto SS, Griffey SM, Duarte PC, Madewell BR: Feline vaccine-associated fibrosarcoma: morphologic distinctions. Vet Pathol 39:33 41, 22 2 Doddy FD, Glickman LT, Glickman NW, Janovitz EB: Feline fibrosarcomas at vaccination sites and non-vaccination sites. J Comp Pathol 114:165 174, 1996 3 Ellis JA, Jackson ML, Bartsch RC, McGill LG, Martin KM, Trask BR, Haines DM: Use of immunohistochemistry and polymerase chain reaction for detection of oncornaviruses in formalin-fixed, paraffin-embedded fibrosarcomas from cats. J Am Vet Med Assoc 29:767 771, 1996 4 Erdman SE, Kanki PJ, Moore FM, Brown SA, Kawasaki TA, Mikule KW, Travers KU, Badylak SF, Fox JG: Clusters of lymphoma in ferrets. Cancer Invest 14:225 23, 1996 5 Hendrick MJ, Brooks JJ: Postvaccinal sarcomas in the cat: histology and immunohistochemistry. Vet Pathol 31: 126 129, 1994 6 Hendrick MJ, Goldschmidt MH: Do injection site reactions induce fibrosarcomas in cats? J Am Vet Med Assoc 199:968, 1991 7 Hendrick MJ, Goldschmidt MH, Shofer FS, Wang YY, Somlyo AP: Postvaccinal sarcomas in the cat: epidemiology and electron probe microanalytical identification of aluminum. Cancer Res 52:5391 5394, 1992 8 Kass PH, Barnes WG Jr, Spangler WL, Chomel BB, Culbertson MR: Epidemiologic evidence for a causal relation between vaccination and fibrosarcoma tumorigenesis in cats. J Am Vet Med Assoc 23:396 45, 1993 9 Kidney BA, Ellis JA, Haines DM, Jackson ML: Comparison of endogenous feline leukemia virus RNA content in feline vaccine and nonvaccine site-associated sarcomas. Am J Vet Res 62:199 1994, 21 1 Macy DW, Hendrick MJ: The potential role of inflammation in the development of postvaccinal sarcomas in cats. Vet Clin North Am Small Anim Pract 26:13 19, 1996 11 Madewell BR, Griffey SM, McEntee MC, Leppert VJ, Munn RJ: Feline vaccine-associated fibrosarcoma: an ultrastructural study of 2 tumors (1996 1999). Vet Pathol 38:196 22, 21 12 McNiel EA: Vaccine-associated sarcomas in cats: a unique cancer model. Clin Orthop 382:21 27, 21 13 Mikaelian I, Garner MM: Solitary dermal leiomyosarcomas in 12 ferrets. J Vet Diagn Invest 14:262 265, 22 14 Morrison WB, Starr RM: Vaccine-associated feline sarcomas. J Am Vet Med Assoc 218:697 72, 21 15 Murray J: Vaccine injection-site sarcoma in a ferret. J Am Vet Med Assoc 213:955, 1998 16 Rupprecht CE, Gilbert J, Pitts R, Marshall KR, Koprowski H: Evaluation of an inactivated rabies virus vaccine in domestic ferrets. J Am Vet Med Assoc 196:1614 1616, 199 17 Ryland LM, Bernard SL, Gorham JR: A clinical guide to the pet ferret. Comp Contin Educ Pract Vet 5:25 32, 1983 Request reprints from John S. Munday, Athens Diagnostic Laboratory, College of Veterinary Medicine, University of Georgia, Athens, GA 362 (USA). E-mail jmunday@vet.uga.edu.