Cutaneous Schwannomas in 22 Horses

S. Schoniger1, B. A. Valentine2,3, C. J. Fernandez2,4, andB. A. Summers2,5

AbstractSchwannomas are uncommonly recognized in horses. This study describes cutaneous schwannomas in 22 horses aged 8 to25 years: 12 male, 7 female, and 3 of unknown sex. The horses had solitary cutaneous masses: 9 on the head, 3 on the neck,and the others on the shoulder, hip, thorax, abdomen, rump, extremities, or tail. The location of 1 tumor was unknown. Thedermal tumors were well demarcated and expansile. Twelve had a multinodular pattern, whereas 10 formed a single nodule.Antoni A areas were observed in all tumors, and 10 tumors contained Antoni B areas. In Antoni A areas, the densely packedspindle-shaped neoplastic cells were arranged in short fascicles with nuclear palisading. In the hypocellular Antoni B areas, neo-plastic cells were separated by abundant myxomatous stroma. Tumors commonly had hyalinization of stroma and vessel walls andancient change. Cellular vacuolation was observed in 18 tumors. In all 22 cases, neoplastic cells were immunopositive for S100protein. Expression of laminin and glial fibrillary acidic protein was observed in all 6 tumors evaluated by immunohistochemistryfor these markers. One tumor was examined ultrastructurally: Neoplastic cells had branched cytoplasmic processes and weresurrounded by an external lamina. Follow-up information was available 8 months to 10 years postexcision for 9 horses, for whichsurgical excision of the tumor was curative. The equine cutaneous schwannomas in this study had microscopic features like thoseof human schwannoma and had benign clinical behavior. Correct classification of equine cutaneous schwannoma will facilitateaccurate prognosis and appropriate treatment.

Keywordsbenign peripheral nerve sheath tumor, electron microscopy, equine neoplasia, horse, immunohistochemistry, schwannoma, skin

In human medicine, benign peripheral nerve sheath tumors(PNSTs) are a diverse group of neoplasms that includes 4 maintypes: schwannoma (synonym: neurilemmoma), neurofibroma,perineurioma, and ganglioneuroma. The 4 types of benignPNST have characteristic diagnostic features, and their distinc-tion is important for prognosis and treatment. In particular, theidentification of patients with multiple neurofibromas may leadto the diagnosis of von Recklinghausen disease (neurofibroma-tosis I), which has implications for family members.3,41 In theveterinary literature, however, the terms benign PNST, schwan-noma, and neurofibroma are often used interchangeably,7,15,31

whereas ganglioneuroma is classified as a separate entity.15,16

Although rare cases of perineurioma (originally designated asbenign hypertrophic neuropathy) have been reported in domes-tic animals,2,9,18 these tumors are not included in most text-books of veterinary pathology,15,19 or they are mentionedonly briefly.16 The generic diagnosis of ‘‘benign PNST’’instead of schwannoma and neurofibroma has been recom-mended in the veterinary literature because the criteria to dis-tinguish between these two are not well established foranimals.7,15 In contrast, we believe that benign PNSTs can besubclassified in veterinary medicine; indeed, variants of neuro-fibroma have been recently described in several species.30

In domestic animals, benign PNSTs may develop in theskin, in peripheral or cranial nerves, and, rarely, in other tissues

or organs. Cutaneous benign PNSTs are rarely reported in dogs,cats, horses, and cattle.5,12,30 Cutaneous schwannomas can beobserved as spontaneous tumors in mice;23 a plexiform cuta-neous schwannoma has been reported in a pig.38 In horses, theeyelids are considered a predilection site for cutaneousPNSTs.5,22 In addition to forming in the skin,5,22,36 equinebenign PNSTs have been recognized in the intestinal wall.14,21

1Department of Pathology and Infectious Disease, Royal Veterinary College,Hatfield, Hertfordshire, UK2Department of Biomedical Sciences, College of Veterinary Medicine, CornellUniversity, Ithaca, New York3Current address: College of Veterinary Medicine, Oregon State University,Corvallis, Oregon4Current address: Animal and Plant Health Laboratories Division, Agri-Foodand Veterinary Authority, Animal and Plant Health Center, Republic of Singa-pore5Current address: Department of Pathology and Infectious Diseases, RoyalVeterinary College, Hatfield, Hertfordshire, UK

Corresponding Author:Sandra Schoniger, Department of Pathology and Infectious Disease, RoyalVeterinary College, Hawkshead Lane, North Mymms, Hatfield, Herts AL97TA, UKEmail: sschoeniger@rvc.ac.uk

Veterinary Pathology48(2) 433-442ª The American College ofVeterinary Pathologists 2011Reprints and permission:sagepub.com/journalsPermissions.navDOI: 10.1177/0300985810377072http://vet.sagepub.com

433 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

The diagnosis of human schwannoma and neurofibroma isusually based on the identification of characteristic histologicfeatures with immunohistochemistry for cellular markersand/or the detection of ultrastructural features.3,25,28,33,41

Schwannomas are well-demarcated expansile tumors entirelycomposed of neoplastic Schwann cells. The spindle-shaped toovoid neoplastic cells form Antoni A and Antoni B areas.Antoni A areas are composed of short fascicles of denselypacked neoplastic cells that may have nuclear palisading andVerocay bodies, the formation of which may be incom-plete.3,10,26,28,33,41,43 In the hypocellular Antoni B areas, theneoplastic cells are embedded in abundant myxoid stroma.Hyalinization of stroma and vessel walls is common. As a man-ifestation of degenerative change, some tumors contain scat-tered neoplastic cells with enlarged bizarre nuclei (referred toas ancient change) or areas of tumor sclerosis.28 Immunohisto-chemistry for S100 protein, a calcium-binding proteinexpressed by Schwann cells, labels most (if not all) neoplasticcells in benign schwannomas.33,41 Moreover, neoplasticSchwann cells often produce glial fibrillary acidic protein(GFAP) to a variable extent20,27 and are surrounded by anexternal lamina containing laminin and collagen IV.28,44 Onultrastructural examination, neoplastic Schwann cells are iden-tified by their complete and often-reduplicated external laminaand complex interdigitating cytoplasmic processes.3,28,41 Incontrast, neurofibromas are well to poorly demarcated tumorsthat contain a mixture of Schwann cells, perineurial cells, andfibroblasts. Their histologic hallmark is the close association ofthe elongated cells with ropey collagen fibers. Immunohisto-chemistry for S100 protein labels the Schwann cells but notfibroblasts or perineurial cells. Ultrastructural examinationconfirms the presence of Schwann cells, perineurial cells, andfibroblasts.3,28,41

Human neurofibroma and schwannoma are sporadic ordevelop in association with the inherited diseases neurofibro-matosis 1 and 2, respectively. They are subclassified into sev-eral variants based on their growth pattern and microscopicfeatures. Recognized growth patterns of schwannoma are unin-odular and multinodular (plexiform). Microscopic variants ofschwannoma are classical (synonym: conventional), cellular,pigmented (synonym: melanotic), lipoblastic, ancient, collage-nous, epithelioid, and glandular.3,10,24,28,41 To our knowledge,this comprehensive study is the first of the histologic featuresand clinical behavior of equine cutaneous schwannoma. Thepurpose of this investigation was to facilitate accurate identifi-cation of equine schwannomas and distinguish them from otherspindle cell tumors of the skin.

Materials and Methods

Clinical Cases

All cases were excisional biopsy specimens that were sub-mitted in formalin over a 29-year period (1979–2008). CaseNos. 1–21 were submitted to the Department of BiomedicalSciences, College of Veterinary Medicine, Cornell University.

Case No. 22 was submitted to the Diagnostic Laboratory at theRoyal Veterinary College.

The signalment, tumor location, and duration were obtainedfrom the clinical history. The tumor size was obtained from thehistory or by measuring the histologic section(s). The latter wasused only if the tumor sizewas not reported and neoplastic tissuewas contained within the margins of the examined section(s).In cases with longitudinal and transverse sections of the mass,the length, width, and thickness of the neoplasmweremeasured.In cases with only 1 section, the width and the thickness of themass was measured.

In March 2009, we asked the following question of theveterinary surgeons who submitted the biopsy samples: Wasremoval of the tumor curative or was recurrence and/or metas-tasis observed? In 9 cases, the requested follow-up informationwas received (Table 1).

Histopathology and Immunohistochemistry

The formalin-fixed biopsies were processed routinely,embedded in paraffin, sectioned, and stained with hematoxylinand eosin (HE). All tumors were evaluated immunohistochemi-cally for S100 protein (Dako, Carpinteria, CA; no pretreat-ment); 6 tumors (Nos. 5, 7–10, 22) were evaluated forlaminin (Dako; steaming in citrate buffer) and 6 (Nos. 5–10) forGFAP (Novocastra, Newcastle Upon Tyne, UK; steaming incitrate buffer). As chromogen, 3,30-diaminobenzidine tetrahy-drochloride was used. In negative controls, the primary anti-body was replaced by nonimmune serum. Appropriatepositive controls were used.

Electron Microscopy

Electron microscopy was performed on tumor No. 3. Formalin-fixed tissue was further fixed in 2.5% glutaraldehyde in 0.1Msodium cacodylate, washed 3 times in cacodylate buffer, andthen postfixed in 2% osmium tetroxide. After further washing,the tissue was dehydrated in graded ethanol solutions, incu-bated in 100% acetone, and embedded in epon araldite plastic.From 1-mm-thick toluidine blue–stained sections, ultrathin sec-tions of selected areas were contrasted with lead citrate and ura-nyl acetate and examined in a Philips 300 transmission electronmicroscope.

Results

Signalment, Tumor Location, Size, and Duration

Table 1 cites the signalment and clinical history for each of the22 horses. The horses were 8 to 25 years of age, with a meanand median age of 16 years. The age was not provided in5 cases. Twelve horses were geldings, 7 were female, and in3 cases, the sex was unknown. There were 6 Quarter Horses,4 Thoroughbreds, 3 of unspecified breeding, 2 Arabians, and1 Appaloosa, Clydesdale, Irish Draught, Paint, Thoroughbredmix, Warmblood, and pony. Nine tumors were on the head,3 on the neck, 2 on the thorax, 2 on the shoulder, and 1 on the

434 Veterinary Pathology 48(2)

434 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

Table

1.SignalmentandNeo

plasticFeatures

in22

Horses

WithCutaneo

usSchw

anno

maa

CaseNo.

Breed

Sex

Age,

Years

Tum

or

Location

Duration

Size

(cm)

MicroscopicFeatures

bRecurrenceor

Metastasis

1NR

NR

NR

Tail

NR

0.9!

0.7c

MN;A

ntoni

A;intraneural

NA

2Tho

roughbred

NR

NR

Ear

*1yr

1.5!

1.0c

UN;A

ntoni

A;"

"vacuolarchange

NA

3Appaloosa

NM

16Lateraln

eck

*1yr

1.5!

1.0c

MN;A

ntoni

A"

B;"

"vacuolarchange

None

in8yr

d

4Arabian

F16

Lateralh

ead

Severalm

os

1.5!

1.5!

1.5e

MN;A

ntoni

A"

B;"

"vacuolarchange

None

in10

yrd

5Po

nyNM

18Throatlatch

*1yr

2.0!

2.0!

2.0c

MN;A

ntoni

A"

B;"

"vacuolarchange

None

in5yr

f

6Tho

roughbredmix

NM

11Abo

veeye

NR

1.5!

1.0!

0.5c

UN;A

ntoni

A;"

vacuolarchange

None

in5yr

f

7Quarter

Horse

F10

Tho

rax

Severalyr

4.0!

2.0!

1.5e

MN;A

ntoni

A"

BNone

d

8Quarter

Horse

NM

NR

Forehead

6–10

mos

1.0!

1.0!

1.0e

MN;A

ntoni

A;"

""

vacuolarchange

NA

9Quarter

Horse

NM

25Face

4mos

0.5!

0.5!

0.5c

UN;A

ntoni

A;"

"vacuolarchange

NA

10Tho

roughbred

NM

18Tho

rax

NR

1.0!

0.5!

0.5c

UN;A

ntoni

A"

B;"

"vacuolarchange

NA

11NR

NM

18Head

NR

1.5!

1.5!

1.5c

UN;A

ntoni

A;"

vacuolarchange

NA

12Tho

roughbred

NM

12Neck

NR

1.2!

1!

0.5c

UN;A

ntoni

A;"

"vacuolarchange

None

in1yr

d

13Quarter

Horse

NM

Aged

Shoulder

NR

4.0!

4.0!

4.0e

MN;A

ntoni

A"

B;"

"vacuolarchange

NA

14Quarter

Horse

F16

Metacarpus

NR

1.0!

1.0!

1.0e

UN;A

ntoni

ANA

15Arabian

NM

12NA

NR

2.0!

1.5c

MN;A

ntoni

A"

B;"

vacuolarchange

NA

16NR

NR

NR

Face

NR

1.0!

1.0!

0.5c

UN;A

ntoni

A;"

"vacuolarchange

NA

17PaintHorse

F16

Face

NR

1.0!

1.0!

1.0c

UN;A

ntoni

A"

B;"

vacuolarchange

NA

18Clydesdale

F17

Rum

pNR

3.0!

2.0!

1.0e

MN;A

ntoni

A;"

vacuolarchange

None

in6yr

g

19Tho

roughbred

NM

25Abd

omen

>1yr

>1.7!

1.2c

,hMN;A

ntoni

A"

B;"

vacuolarchange

NA

20Quarter

Horse

NM

16Face

3yr

>3.0!

1.5c

,hMN;A

ntoni

A;"

""

vacuolarchange

NA

21W

armblood

F8

Shoulder

NR

1.5!

1.0!

1.0c

UN;A

ntoni

A;"

"vacuolarchange

None

in4yr

g

22IrishDraught

F11

Hip

NR

0.5!

0.5!

0.5c

MN;A

ntoni

A"

BNone

in8mosg

aNA,n

otavailable;NR,n

otrecorded;F,fem

ale;NM,n

euteredmale

bMN,m

ultino

dular;UN,u

nino

dular;",m

ild;"

",m

oderate;"

"",m

arked.

cMeasurements

obtainedfrom

thetissue

section(s);eReported

size;h

Size

ofexam

ined

portionoftumor.

dUntildeath;

fUntilloss

ofcontact;

gAliveat

timeofsurvey

(March

2009).

435 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

abdomen, rump, hip, metacarpus, tail base, and unspecifiedlocation. All but 4 tumors were less than 2 cm in maximaldimension. Tumors were described as soft (n, 1) to firm (n,2) and fleshy (n, 1) and not painful (n, 2) to extremely painful(n, 1). For 9 horses, the duration of the neoplastic growth wasreportedly from a few months to 1–3 years. The tumors of caseNos. 3, 5, 9, 14, and 20 had a reported recent increase in size.

Microscopic Features

Table 1 lists the growth patterns and other microscopic featuresof the 22 schwannomas. In 10 cases, the tumor was uninodular(Fig. 1). In the other 12 cases, multiple distinct nodules weresurrounded and separated by fibrous tissue (Fig. 2). Growthwithin a nerve was detected in one (No. 1) of the multinodulartumors. Thirteen tumors were dome shaped; all formed well-demarcated and expansile dermal masses surrounded by afibrous capsule or compressed fibrous tissue. The neoplasticcells were spindle shaped to ovoid, had indistinct cell borders,single nuclei, and moderate amounts of eosinophilic cyto-plasm. The elongate to ovoid nuclei contained stippled chroma-tin and one to several inconspicuous nucleoli. Cellularpleomorphism, anisocytosis, and anisokaryosis were generallymild, but scattered neoplastic cells had enlarged or lobulatednuclei (Fig. 3, inset) consistent with ancient change, a degen-erative change well recognized in human schwannoma28 andrecently described in equine cutaneous melanocytic nevi.29 Themitotic index (counted in Antoni A areas) ranged from 0 to 6mitotic figures (average, 2) per 10 high-power fields (400!).The mitotic activity in 2 tumors that had doubled in size (No.5) and had recent growth (No. 20) was 4 and 3 mitotic figuresper high-power field (10!), respectively, just slightly abovethe average. All tumors contained Antoni A areas; 10 tumorshad smaller Antoni B areas. Within the Antoni A areas, theelongated neoplastic cells were arranged in short fasciclesseparated by small to moderate amounts of stroma (Fig. 3).Nuclear palisading was observed in all tumors to a variableextent (Fig. 3). Tumors contained arrangements of neoplasticSchwann cells in Verocay bodies and rosettes (Figs. 4–7).Novel features included a pattern resembling half of a Verocaybody with a single nuclear palisade and eosinophilic Schwanncell processes that formed the margin of a tumor nodule(Fig. 8).

Most tumors had vacuolar change, mainly in Antoni A areas(Figs. 9, 10). Eighteen tumors contained a variable number ofcells with a single large clear vacuole surrounded by a narrowrim of cytoplasm and an eccentric nucleus. In 16 of thesetumors, a few cells had multiple smaller clear, distinct cyto-plasmic vacuoles and a single central or eccentric nucleus (Fig.9). Vacuolated cells constituted various percentages of cellswithin the tumors: < 5% (n, 6), 10 to 40% (n, 10), and> 60% (n, 2). A few vacuolated cells contained enlarged nucleisuggestive of ancient change (Fig. 9). Antoni B areas, whichoccupied approximately 5 to 40% of the tumor mass, wereeither intermingled with Antoni A areas (Fig. 2) or separatedfrom them by fibrous septa (Fig. 11). Within the hypocellular

Antoni B areas, the neoplastic cells were embedded in abun-dant myxomatous stroma (Fig. 11). In 2 tumors, a transitionwas observed between highly vacuolated Antoni A and AntoniB areas (Fig. 12), suggesting that Antoni B areas could developsecondary to cellular degeneration in Antoni A areas, as hasbeen proposed.32

A variable amount of stromal and capillary vessel hyaliniza-tion was observed in 21 tumors (Fig. 10) and gave the tumors acharacteristic appearance that facilitated their recognition.Some tumors contained small hemorrhages (n, 10) and a lowto moderate number of hemosiderin-laden macrophages (n,11). Tumor necrosis (n, 4), ectatic vessels (n, 3), microcysticchange (n, 2), ulceration (n, 1), and lymphocytic infiltration(n, 2) were uncommon findings.

Immunohistochemistry and Electron MicroscopicExamination

Neoplastic cells, including those with cytoplasmic vacuoles,uniformly expressed S100 protein in all 22 cases. The immu-noreactivity was cytoplasmic and, in some cells, also intranuc-lear (Fig. 13). All 6 tested tumors expressed GFAP; moderateimmunoreactivity was observed in the neoplastic cells: 5%(No. 7), 30% (No. 6), and 75 to 90% (Nos. 5, 8–10). Some ofthe vacuolated cells were GFAP immunopositive (Fig. 14).Neoplastic cells were surrounded by laminin-positive materialin all 6 tumors examined. Some tumor cells contained intracy-toplasmic laminin-positive granular material (Fig. 15). Theintensity and extent of laminin expression varied. Of the2 tumors composed of Antoni A tissue only, the entire tumorwas strongly immunolabeled in one case (No. 9); in the othercase (No. 8), 40% of the tumor had weak laminin expression.The other 4 tumors—composed of Antoni A and B tissue—hadweak to moderate laminin expression in Antoni A areas to avariable extent: 100% (No. 22), 50% (No. 5), 35% (No. 7), and10% (No. 10). In 1 tumor (No. 22), moderate laminin expres-sion was also observed in the Antoni B area.

On ultrastructural examination of 1 tumor, neoplastic cellswere identified as Schwann cells by their complete and focallyredundant external lamina and cytoplasmic processes thatbranched and interdigitated (Fig. 16).

Biologic Behavior

The surgical excision of the tumor appeared to have been cura-tive for all horses for which follow-up information was avail-able (n, 9; Table 1). No recurrence or metastasis wasdetected between 8 months and 10 years after surgical excision,although several horses (Nos. 3, 4, 7, 12) died for reasons unre-lated to the schwannoma.

Discussion

In these 22 equine cutaneous schwannomas, the tumors werewell-demarcated uninodular or multinodular dermal masseswith Antoni A areas (and, in some tumors, Antoni B areas),

436 Veterinary Pathology 48(2)

436 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

Figure 1. Equine cutaneous schwannoma, No. 10. This tumor has a uninodular growth pattern and is largely composed of Antoni A tissue withmultiple smaller less cellular Antoni B areas (asterisks). HE. Figure 2. Equine cutaneous schwannoma, No. 17. This tumor has a multinodulargrowth pattern. The distinct tumor nodules are separated by fibrous connective tissue. Some tumor nodules contain pale Antoni B areas(asterisks) in addition to Antoni A areas. HE. Figure 3. Equine cutaneous schwannoma, No. 19. In Antoni A areas, the spindle-shapedneoplastic Schwann cells form short fascicles separated by a mild amount of stroma. Nuclei of adjacent cells may be arranged in parallel rows(nuclear palisading). Most neoplastic cells have mild pleomorphism. HE. Inset: Equine cutaneous schwannoma, No. 7. A neoplastic cell contains

Schoniger et al 437

437 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

nuclear palisading, and hyalinization of stroma and vesselwalls. The neoplastic cells expressed S100 protein, laminin,and GFAP. We have observed in domestic animals variantsof schwannoma and neurofibroma30 similar to those recog-nized in human beings and therefore propose that the humanclassification can be cautiously applied to the diagnosis ofthese tumors in domestic animals.

Because the veterinary literature lacks established criteria forthe diagnosis of benign PNSTs, schwannoma, and neurofibroma,their application varies from one pathologist to another. Recentpractice in veterinary pathology has been to classify tumors withfeatures of schwannoma or neurofibroma as benign PNSTs.15

One textbook of equine dermatology uses the term schwannomafor all types of benign PNSTs of horses and describes 3 histo-pathologic patterns of schwannoma—namely, neurofibroma,neurilemmoma, and plexiform.31 Thus, the comparison of thisseries of schwannomas with published cases is problematic.

The expression of S100 protein in schwannomas of dogs andcats is reportedly variable,16 whereas in our experience, benignschwannomas of horses (this study) and cats (unpublished) areuniformly positive for S100 protein. Certain Schwann cellpopulations are GFAP positive,11 and a variable GFAP expres-sion has been observed in most schwannomas.20,27 GFAPexpression by neoplastic Schwann cells can be helpful for thedifferentiation of benign PNSTs from other spindle celltumors.7,27 Schwann cells synthesize external lamina compo-nents and are surrounded by a complete external lamina.25,41

Thus, schwannomas are commonly immunopositive for lami-nin and collagen IV, which are components of an externallamina.7,44 The variation in the extent and intensity of the lami-nin immunoreactivity in the 6 tumors examined in the currentstudy was likely caused by differences of antigen preservationin the formalin-fixed paraffin-embedded tissues because theinternal positive controls (the external lamina of vessels andSchwann cells in peripheral nerves) also had similar variationsin immunoreactivity.

Benign andmalignant PNSTs are said to be rare in horses.31,39

In a survey of equine cutaneous neoplasia (in total, 536 skintumors), only 1.1% were schwannomas.39 In agreement with theliterature,we rarely encounter equine PNSTs among our diagnos-tic cases, although it is our experience that equine schwannoma ismore common than equine neurofibroma.30

The equine cutaneous schwannomas identified in this studyoccurred as small solitary benign masses in 22 mature horses(mean and median age, 16 years) of a variety of breeds—espe-cially, Quarter Horses and Thoroughbreds. A breed predilection

for equine PNSTs has not been reported, and the observed pre-dominance of Quarter Horses and Thoroughbreds might reflectan increased prevalence of schwannoma in these breeds or ahigher submission from these 2 breeds compared to otherbreeds. PNSTs aremore frequently observed inmale horses thanin female horses.37 In this study, males (all gelded) were slightlymore common than females (12 versus 7), but 3 tumors occurredin horses of unknown sex, and the overall case numbers aresmall. PNSTs are more commonly diagnosed in middle-agedto older horses,5,35 and all equine schwannomas of this studywere observed in this age group.

The 22 schwannomas had been present for weeks to monthsto years (3 years) and had remained quiescent or, in some cases,had local growth. Microscopic examination revealed uninodu-lar or multinodular proliferation of spindle cells in the dermis,arranged in Antoni A (and, in some cases, Antoni B) patterns.In some schwannomas, a transition could be observed betweenAntoni A areas containing numerous vacuolated cells andAntoni B areas. This supports the theory that Antoni B areasdevelop secondary to degenerative changes.32 Many of theequine tumors contained arrangements of neoplastic Schwanncells similar to conventional Verocay bodies and to others that,in humans, have been designated incomplete or abortive.10,26,43

Some schwannomas contained rosettes resembling those iden-tified in human schwannoma.28

Stromal and vascular hyalinization and ancient change,which are common features of human schwannoma,28,41 werecommonly observed in equine schwannomas. Some of theequine tumors had a multinodular growth pattern, which canbe observed in some variants of human schwannoma.10,28,41,43

In human medicine, the modifier plexiform is used for neu-rofibromas with tumor growth in adjacent nerves, which resultsin marked multinodular enlargement.41 The term plexiform isalso applied to schwannomas and a variety of other tumorscomposed of multiple adjacent but separate nodules. Inhumans, the differentiation of plexiform schwannoma fromplexiform neurofibroma has prognostic implication becauseplexiform neurofibromas are not solitary tumors but occur inpatients with neurofibromatosis 1 (von Recklinghausen dis-ease).41 A solitary cutaneous plexiform schwannoma has beenreported in a pig,38 and a plexiform growth pattern of equinecutaneous schwannoma has been recognized.31 Only 1 of theequine tumors in the current study grew within preexistingnerves, but in 12 tumors, the growth pattern was multinodularand in a human classification would probably be deemedplexiform.

Figure 3 (continued) an enlarged nucleus consistent with a degenerative alteration called ancient change. HE. Figure 4. Equine cutaneousschwannoma, No. 19. This tumor nodule is mainly composed of Antoni A tissue and a small amount of Antoni B tissue. The Antoni A areacontains neoplastic Schwann cells arranged as rosettes (asterisks). HE. Figure 5. Equine cutaneous schwannoma, No. 8. Neoplastic Schwanncells form a rosette (asterisk); centrally located Schwann cell processes are surrounded by Schwann cell nuclei. HE. Figure 6. Equinecutaneous schwannoma, No. 20. This Verocay body (asterisk) has a central meandering core of eosinophilic cell processes, bordered at bothsites by Schwann cell nuclei. HE. Figure 7. Equine cutaneous schwannoma, No. 20. Verocay body (asterisk) with a central core of eosinophiliccell processes, bordered at both sites by Schwann cell nuclei. HE. Figure 8. Equine cutaneous schwannoma, No. 19. Two tumor nodules areseparated by stroma (asterisk). A single nuclear palisade with densely aggregated eosinophilic Schwann cell processes resembling half of aVerocay body is at the margin of one tumor nodule (arrowheads).

438 Veterinary Pathology 48(2)

438 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

Figure 9. Equine cutaneous schwannoma, No. 21. This tumor contains numerous vacuolated cells with either a single large clearintracytoplasmic vacuole and an eccentric nucleus (thick arrow) or multiple smaller vacuoles and a central to eccentric nucleus (thin arrow). Amultivacuolated cell contains an enlarged nucleus (thin arrow). HE. Figure 10. Equine cutaneous schwannoma, No. 18. There is hyalinization ofvessel walls (arrowheads) and stroma (asterisk). HE. Inset: Equine cutaneous schwannoma, No. 5. Vascular hyalinization is depicted at highermagnification. HE. Figure 11. Equine cutaneous schwannoma, No. 3. Antoni A (A) and Antoni B (B) areas are separated by fibrous tissue. InAntoni B areas, neoplastic cells are in abundant myxomatous stroma. HE. Figure 12. Equine cutaneous schwannoma, No. 5. This Antoni B areacontains numerous vacuolated cells (arrows). HE. Figure 13. Equine cutaneous schwannoma, No. 22. Neoplastic cells uniformly express S100

Schoniger et al 439

439 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

Many of these 22 schwannomas contained variable numbersof vacuolated cells, some of which had enlarged and occasion-ally lobulated nuclei resembling the ancient change of the moreconventionally spindle-shaped neoplastic Schwann cells. Somevacuolated cells expressed S100, GFAP, and laminin, corrobor-ating their identity as neoplastic cells. The cause for the cellularvacuolation is unknown, but based on the histologic appear-ance, lipid accumulation has to be considered. Cellular vacuo-lation due to adipocyte differentiation is observed in lipoblasticschwannoma, a rare human variant.24 Adipocyte differentiationis also a rare finding in human neurofibroma24 and has beenreported in human meningioma.1

Differentiation of equine cutaneous schwannoma from sar-coid, the most common equine dermal tumor,5 should not bedifficult, because sarcoids lack Antoni A and B patterns, Ver-ocay bodies, hyalinization, and cellular vacuolation. Perhapssome variants of intradermal melanocytic tumors with spindlecell morphology and limited pigmentation (which would alsobe S100 protein positive) would present a greater diagnosticchallenge, although (1) melanophages will often be present andneither Verocay bodies nor hyalinization has been reported as asalient feature of these tumors34 and (2) neoplastic melanocytesare immunonegative for GFAP.17 Equine cutaneous schwanno-mas with vacuolar change would need to be distinguished fromspindle cell lipomas should these tumors be recognized inhorses. So far, spindle cell lipomas have been described onlyin human beings4,40,45 and dogs.6 Due to the presence of

multivacuolated cells that resemble lipoblasts, as well as thescattered enlarged nuclei, the equine cutaneous schwannomaswould need to be distinguished from spindle cell liposar-coma,42 although this tumor has not been reported in domesticanimals to date. Another differential is that of well-differentiated lipoma-like liposarcoma, which contains lipo-blasts and has been described in dogs.6 Mature adipocytes andwell-differentiated neoplastic adipocytes are S100 immunopo-sitive and are surrounded by a laminin-positive pericellularmembrane.8,45 In lipomas and liposarcomas with spindle cells,the spindle cell population is immunonegative for S100 andlaminin, or it shows only a mild multifocal labeling for oneor both antigens.4,8,42,45 Lipomas and liposarcomas are GFAPimmunonegative6 and lack Antoni A and B areas, Verocaybodies, and prominent vascular hyalinization.

Case outcome was known for 9 horses, none of which hadany evidence of recurrence or metastasis 8 months to 10 yearsafter surgical excision. Therefore, complete surgical excision islikely to be curative for equine cutaneous schwannomas. Therecognition of schwannoma and neurofibroma as separate enti-ties of benign PNSTs in veterinary medicine will allow moreaccurate diagnosis. It is likely that the diagnosis of schwan-noma and neurofibroma and their variants in the horse andother domestic animals has implications for prognosis and ther-apy similar to those for the sporadic forms of these tumors inhumans. For example, diffuse neurofibromas are infiltrative41

and therefore often difficult to excise; in fact, removing plexi-form neurofibromas may involve significant nerve sacrifice.13

The complete excision of schwannoma and solitary neurofi-broma, however, is usually curative.13 Finally, although malig-nant transformation occurs in some neurofibromas in humanpatients with neurofibromatosis type I, malignant transforma-tion of sporadic schwannoma (or neurofibroma) is rare, andwe are not aware of documented progression of schwannomato malignancy in horses or other animals.

Acknowledgements

We thank Professor K. C. Smith (Royal Veterinary College) for help-ful comments on the article, Ms J. Cramer (Cornell University), MsM.E. Beissenherz (University of Missouri), and the histopathologylaboratory at the Royal Veterinary College for immunohistochemistry.We are grateful to Drs D. T. Lamb, J. Kolb, R. Austin, L. F. Karcher,D. Pantano, A. Dwyer, K. T. Kay, C. E. Juul-Nielsen, and their associ-ates and the Liphook Equine Clinic for providing the follow-upinformation.

Declaration of Conflicting Interests

The authors declared that they had no conflicts of interest with respectto their authorship or the publication of this article.

Figure 16. Equine cutaneous schwannoma,No. 3. Neoplastic Schwanncells have an external lamina and branched, interdigitated cytoplasmicprocesses. Inset: The richly folded cytoplasmic processes are surroundedby a complete external lamina (arrows). Electron microscopy.

Figure 13 (continued) protein, whereas the adjacent stroma is immunonegative (asterisk). Inset: Equine cutaneous schwannoma, No. 22.S100 protein immunoreactivity of a vacuolated neoplastic cell. 3,30-Diaminobenzidine tetrahydrochloride is used as chromogen. Figure 14.Equine cutaneous schwannoma, No. 5. Neoplastic cells are glial fibrillary acidic protein immunopositive. 3,30 0-Diaminobenzidinetetrahydrochloride is used as chromogen. Inset: Glial fibrillary acidic protein immunostaining of a vacuolated cell. Figure 15. Equinecutaneous schwannoma, No. 9. Neoplastic cells are surrounded by a laminin positive external lamina. Inset: Equine cutaneous schwannoma, No.22. A few neoplastic cells express intracytoplasmic laminin. 3,300-Diaminobenzidine tetrahydrochloride is used as chromogen.

440 Veterinary Pathology 48(2)

440 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

Financial Disclosure/Funding

The authors declared that they received no financial support for theirresearch and/or authorship of this article.

References

1. Burger PC, Scheithauer BW: Tumors of the central nervous sys-

tem In: AFIP Atlas of Tumor Pathology, ed. Silverberg SG, 4th

series, fascicle 7. Armed Forces Institute of Pathology, Washing-

ton, DC, 2007.

2. Cummings JF, de Lahunta A: Hypertrophic neuropathy in a dog.

Acta Neuropathol 29:325–336, 1974.

3. Fletcher CDM: Peripheral neuroectodermal tumors. In: Diagnos-

tic Histopathogy of Tumors, ed. Fletcher CDM, 3rd ed., vol. 2, pp.

1733–1761. Churchill Livingstone Elsevier, Edinburgh, Scotland,

2007.

4. Fletcher CDM: Soft tissue tumors. In: Diagnostic Histopathology

of Tumors, ed. Fletcher CDM, 3rd ed., vol. 2, pp. 1527–1592.

Churchill Livingstone Elsevier, Edinburgh, Scotland, 2007.

5. Goldschmidt MH, Hendrick MJ: Tumors of the skin and soft tis-

sues. In: Tumors in Domestic Animals, ed. Meuten DJ, 4th ed.,

pp. 45–117. Iowa State Press, Ames, IA, 2002.

6. Gross TL, Ihrke PJ, Walder EJ, Affolter VK: Lipocytic tumors.

In: Skin Diseases of the Dog and Cat, 2nd ed., pp. 766–777.

Blackwell Publishing, Oxford, UK, 2005.

7. Gross TL, Ihrke PJ, Walder EJ, Affolter VK: Neural and peri-

neural tumors. In: Skin Diseases of the Dog and Cat, 2nd ed.,

pp. 786–796. Blackwell Publishing, Oxford, UK, 2005.

8. Haraida S, Nerlich AG, Wiest I, Schleicher E, Lohrs U: Distribu-

tion of basement membrane components in normal adipose tissue

and in benign and malignant tumors of lipomatous origin. Mod

Pathol 9:137–144, 1996.

9. Higgins RJ, Dickinson PJ, Jimenez DF, Bollen AW, Lecouteur

RA: Canine intraneural perineurioma. Vet Pathol 43:50–54, 2006.

10. Iwashita T, Enjoji M: Plexiform neurilemmoma: a clinicopatholo-

gical and immunohistochemical analysis of 23 tumours from 20

patients. Virchows Arch A Pathol Anat Histopathol 411:305–

309, 1987.

11. Jessen KR, Thorpe R, Mirsky R: Molecular identity, distribution

and heterogeneity of glial fibrillary acidic protein: an immunoblot-

ting and immunohistochemical study of Schwann cells, satellite

cells, enteric glia and astrocytes. J Neurocytol 13:187–200, 1984.

12. Kameyama M, Ishikawa Y, Shibahara T, Kadota K: Melanotic

neurofibroma in a steer. J Vet Med Sci 62:125–128, 2000.

13. Kim DH, Murovic JA, Tiel RL, Moes G, Kline DG: A series of

397 peripheral neural nerve sheath tumors: 30-year experience

at Louisiana State University Health Science Center. J Neurosurg

102:246–255, 2005.

14. Kirchhof N, Scheidemann W, Baumgartner W: Multiple periph-

eral nerve sheath tumors in the small intestine of a horse. Vet

Pathol 33:727–730, 1996.

15. Koestner A, Bilzer T, Fatzer R, Schulman FY, Summers BA, Van

Winkle TJ: Tumors of the peripheral nervous system. In: Histolo-

gical Classification of Tumors of the Nervous System in Domestic

Animals, 2nd series, vol. 5, pp. 36–38. Armed Forces Institute of

Pathology, Washington, DC, 1999.

16. Koestner A, Higgins RJ: Tumors of the nervous system. In:

Tumors in Domestic Animals, ed. Meuten DJ, pp. 697–738. Iowa

State Press, Ames, IA, 2002.

17. Levidou G, Korkolopoulou P, Papetta A, Patsouris E, Agapitos E:

Leptomeningeal melanoma of unknown primary site: two cases

with an atypical presentation of acute meningitis. Clin Neuro-

pathol 26:299–305, 2007.

18. Matiasek K, Kostlin R, Schmahl W, Hecht S: Plexiformer Ner-

venwurzeltumor vom Type eines Perineurioms bei einem Hund.

Wien Tierarztl Mschr 88:192–200, 2001.

19. Maxie MG, Youssef S: Nervous system. In: Jubb, Kennedy, and

Palmer’s Pathology of Domestic Animals, ed. Maxie MG, 5th

ed., vol. 1, pp. 281–457. Saunders Elsevier, St Louis, MO, 2007.

20. Nielsen AB, Jansen ECL, Leifsson PS, Jensen HE: Immunoreac-

tivity of bovine schwannomas. J Comp Path 137:224–230, 2007.

21. Pascoe PJ: Colic in a mare caused by a colonic neurofibroma. Can

Vet J 23:24–27, 1982.

22. Pascoe RR, Summers PM: Clinical survey of tumours and

tumour-like lesions in horses in south east Queensland. Equine

Vet J 13:235–239, 1981.

23. Peckham JC, Heider K: Skin and subcutis. In: Pathology of the

mouse, ed. Maronpot RR, pp. 555–612. Cache River Press, St

Louis, MO, 1999.

24. Plaza JA, Wakely PE Jr, Suster S: Lipoblastic nerve sheath

tumors: report of a distinctive variant of neural soft tissue neo-

plasm with adipocytic differentiation. Am J Surg Pathol

30:337–344, 2006.

25. Requena L, Sangueza OP: Benign neoplasms with neural differ-

entiation: a review. Am J Dermatopathol 17:75–96, 1995.

26. Santos LD, Currie BG, Killingsworth MC: Case report: plexiform

schwannoma of the vulva. Pathology 33:526–531, 2001.

27. Sarlomo-Rikala M, Tsujimura T, Lendahl U, Miettinen M: Pat-

terns of nestin and other intermediate filament expression distin-

guish between gastrointestinal stromal tumors, leiomyomas and

schwannomas. APMIS 110:499–507, 2002.

28. Scheithauer BW, Woodruff JM, Erlandson RA: Tumors of the

peripheral nervous system. In: Atlas of Tumor Pathology, ed.

Rosai J, 3rd series, fascicle 24, pp. 1–421. Armed Forces Institute

of Pathology, Washington, DC, 1997.

29. Schoniger S, Summers BA: Equine skin tumours in 20 horses

resembling three variants of human melanocytic naevi. Vet Der-

matol 20:165–173, 2009.

30. Schoniger S, Summers BA: Localized, plexiform, diffuse and

other variants of neurofibroma in twelve dogs, two horses and a

chicken. Vet Pathol 46:904–915, 2009.

31. Scott DW, Miller WH: Neoplastic and non-neoplastic tumors. In:

Equine Dermatology, pp. 698–795. Saunders, St Louis, MO,

2003.

32. Sian CS, Ryan SF: The ultrastructure of neurilemoma with

emphasis on Antoni B tissue. Hum Pathol 12:145–160, 1981.

33. Skovronsky DM, Oberholtzer JC: Pathologic classification of per-

ipheral nerve tumors. Neurosurg Clin N Am 15:157–166, 2004.

Schoniger et al 441

441 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from

34. Smith SH, Goldschmidt MH, McManus PM: A comparative

review of melanocytic neoplasms. Vet Pathol 39:651–678, 2002.

35. Stoica G, Tasca SI, Kim H-T: Point mutation of neu oncogene in

animal peripheral nerve sheath tumors. Vet Pathol 38:679–688,

2001.

36. Sturgeon BP, Milne EM, Smith KC: Benign peripheral nerve

sheath tumor of the perianal region in a young pony. J Vet Diagn

Invest 20:93–96, 2008.

37. Sundberg JP, Burnstein T, Page EH, KirkhamWW, Robinson FR:

Neoplasms of Equidae. J Am Vet Med Assoc 170:150–152, 1977.

38. Tanimoto T, Ohtsuki Y: Cutaneous plexiform schwannoma in a

pig. J Comp Pathol 109:231–240, 1993.

39. Valentine BA: Survey of equine cutaneous neoplasia in the

Pacific Northwest. J Vet Diagn Invest 18:123–126, 2006.

40. Weiss SW, Goldblum JR: Benign lipomatous tumors. In: Enzin-

ger and Weiss’s Soft Tissue Tumors, 5th ed., pp. 429–476. Mosby

Elsevier, St Louis, MO, 2008.

41. Weiss SW, Goldblum JR: Benign tumors of peripheral nerves. In:

Enzinger and Weiss’s Soft Tissue Tumors, 5th ed., pp. 825–901.

Mosby Elsevier, St Louis, MO, 2008.

42. Weiss SW, Goldblum JR: Liposarcoma. In: Enzinger and Weiss’s

Soft Tissue Tumors, 5th ed., pp. 477–516. Mosby Elsevier, St

Louis, MO, 2008.

43. Woodruff JM, Marshall ML, Godwin TA, Funkhouser JW,

Thompson NJ, Erlandson RA: Plexiform (multinodular) schwan-

noma: a tumor simulating the plexiform neurofibroma. Am J Surg

Pathol 7:691–697, 1983.

44. Zhang HY, Yang GH, Chen HJ, Wei B, Ke Q, Guo H, Ye L, Bu H,

Yang K, Zhang YH: Clinicopathological, immunohistochemical,

and ultrastructural study of 13 cases of melanotic schwannoma.

Chin Med J (Engl) 118:1451–1461, 2005.

45. Zelger BWH, Zelger BG, Plorer A, Steiner H, Fritsch PO: Dermal

spindle cell lipoma: plexiform and nodular variants. Histopatho-

logy 27:533–540, 1995.

442 Veterinary Pathology 48(2)

442 by Dr Miquel VILAFRANCA COMPTE on April 26, 2011vet.sagepub.comDownloaded from