Cutaneous lymphangiosarcoma in an African green monkey from St. Kitts, and a literature review

Abstract

A 22-y-old, intact male African green monkey (Chlorocebus sabaeus) with a pedunculated cutaneous ulcerated spheroid mass on the left flank was autopsied. Histologically, the mass was infiltrative, with neoplastic cells forming empty vascular channels lined by a monolayer of polygonal neoplastic cells, many with eosinophilic intracytoplasmic granules, supported by abundant fibrous stroma. Immunohistochemically, neoplastic cells had strong immunolabeling with vimentin, moderate immunolabeling with LYVE1 and PROX1 antibodies, and no cytokeratin or CD31 immunolabeling. Ultrastructurally, a basal lamina was absent, nuclei had a 1:4 heterochromatin:euchromatin ratio, with a mild-to-moderate increase in thickness and asymmetrically distributed nuclear fibrous lamina, enlarged compacted nucleolus, round mitochondria, conspicuous rough endoplasmic reticulum, and membrane-bound electron-dense granules. Intercellular desmosomes were present in clusters of neoplastic cells. Lymphangiosarcoma has not been previously reported in non-human primates, to our knowledge. Our case highlights the importance of integrating histopathology, immunohistochemistry, and ultrastructural analysis to accurately diagnose rare vascular tumors, such as lymphangiosarcoma.

A 22-y-old, intact male African green monkey (Chlorocebus sabaeus), housed in a research facility, had an ulcerated skin mass localized in the left flank area and, with deteriorating general condition, was euthanized and submitted for autopsy. The mass was first noticed ~4 mo before euthanasia when it was movable, with no apparent involvement of the underlying musculature.

Grossly, the 55 × 42 × 40-mm, firm mass was pedunculated, round, white, pink, and red, and ulcerated ( Fig. 1A ). On cross-section, it was white with several round-to-irregular cystic spaces containing clear yellow fluid, or red areas with blood (hemorrhage and necrosis). No other gross lesions were observed during autopsy.

Figure 1.

Gross and histopathology of a cutaneous lymphangiosarcoma in an African green monkey. A. A pedunculated and ulcerated mass on the left cranioventral abdomen. Inset: on cross-section, the mass was white-tan with cystic and hemorrhagic areas. B. Histologically, empty vascular channels are surrounded by abundant stroma. Inset: many neoplastic endothelial cells lining the channels contain intracytoplasmic eosinophilic granules. H&E. C. A single layer of plump neoplastic endothelial cells lines the vascular channels, with prominent (hobnail) luminal projections and intracytoplasmic eosinophilic granules. Inset: an atypical mitotic figure. Masson trichrome.

Multiple sections of the cutaneous mass and internal organs were fixed in 10% neutral-buffered formalin and processed routinely; 4-µm sections were stained with H&E. Immunohistochemistry (IHC) was performed ( Table 1 ) using the following antibodies: CD31, vimentin, lymphatic vessel endothelial hyaluronan receptor 1 (LYVE1), Prospero-related homeobox gene-1 (PROX1), and pancytokeratin. All antibodies were optimized for use in non-human primates for our project. Staining was performed using an automated IHC staining system (Bond Max; Leica), in accordance with the manufacturer’s instructions. A negative control, omitting the primary antibody, and a positive control (canine tissue), demonstrating the normal distribution of the antigen, were included in each batch of slides.

Table 1.

Antibodies, methodology, and results of immunohistochemical characterization of a cutaneous lymphangiosarcoma in an African green monkey.

Antibody	Target	Antibody source/clone	Dilution	Type	Incubation time	Antigen retrieval	Detection system	Immunolabeling
CD31	Endothelial cells	Leica/ JC70A	RTU	Mouse	30 min	ER2	Bond Polymer Refine Red	–
Vimentin	Mesenchymal cells	Leica/V9	RTU	Mouse	30 min	ER2	Bond Polymer Refine Red	+
LYVE1	Lymphatic endothelial cells	Abcam/Polyclonal	1:100	Rabbit	30 min	ER2	Bond Polymer Refine Red	+
PROX1	Lymphatic endothelial cells	AngioBio/ Polyclonal	1:600	Rabbit	30 min	ER2	Bond Polymer DAB	+
Cytokeratin (pan)	Epithelial cells	Leica/AE1/AE3	RTU	Mouse	30 min	ER2	Bond Polymer Refine Red	–

– = lack of immunolabeling in neoplastic cells; + = immunolabeling present in neoplastic cells; DAB = 3,3′-diaminobenzidine; ER2 = epitope retrieval solution 2; RTU = ready to use.

For ultrastructural analysis, neoplastic tissues were placed in modified Karnovsky solution and washed with buffered 0.166 M sodium cacodylate (Electron Microscopy Sciences [EMS]), followed by a secondary post-fixation step using 1% osmium tetroxide (EMS). Tissue blocks of ~1.0 mm³ were dehydrated through a graded ethanol series and subsequently embedded in Embed resin (EMS). Embedded samples were trimmed and sectioned (UC6 ultramicrotome; Leica). Ultrathin sections (60–90 nm) were mounted on 100-mesh copper grids (EMS), stained with 5% uranyl acetate for 20 min, followed by Sato lead citrate for 6 min. Specimens were examined with a transmission electron microscope (TEM; 1200 EX II, JEOL). Digital images were captured with a camera and integrated software (Veleta 2k × 2k camera, iTEM software; Olympus).

Histologically, a poorly demarcated, unencapsulated, infiltrative neoplasm expanded the dermis and elevated the ulcerated epidermis. The neoplastic cells formed empty channels lined by a single layer of neoplastic cells supported by abundant fibrous stroma ( Fig. 1B ). The neoplastic cells were polygonal with prominent cell borders, often with a plump and prominent luminal projection (hobnail appearance) and a moderate amount of basophilic cytoplasm, which, for many cells, contained variable numbers of 1–3-um eosinophilic granules ( Fig. 1C ). The nuclei were oval, centrally to paracentrally located, with finely stippled chromatin and one prominent magenta nucleolus. Mitotic figures were 9 per 2.37 mm² (10 consecutive hpf, 40× objective, 10× ocular FN22), including some atypical mitotic figures. Multifocally, moderate numbers of bi- and multinucleate neoplastic cells were present.

The epithelial surface covering the neoplasm was ulcerated and replaced by extravasated RBCs, viable and degenerate neutrophils, fibrin, bacterial aggregates, and proteinaceous material. Our morphologic diagnosis was a malignant vascular neoplasm. We ruled out hemangioma and lymphangioma based on the larger cells with cellular and nuclear criteria of malignancy and mitotic activity. The differential diagnoses at this stage included hemangiosarcoma (HSA) and lymphangiosarcoma (LSA).

On IHC, the neoplastic cells had strong diffuse intracytoplasmic immunolabeling for vimentin ( Fig. 2A ) and mild-to-moderate granular membranous staining to LYVE1 ( Fig. 2B ) and PROX1 ( Fig. 2C ). The neoplastic cells did not immunolabel with CD31 or cytokeratin.

Figure 2.

Immunohistochemical and ultrastructural features of a cutaneous lymphangiosarcoma in an African green monkey. A. Diffuse cytoplasmic immunolabeling of neoplastic endothelial cells lining vascular channels. Vimentin immunohistochemistry (IHC). B. Mild granular membranous immunolabeling of neoplastic cells. LYVE1 IHC. C. Moderate granular membranous immunolabeling of neoplastic cells. PROX1 IHC. D. Neoplastic cells surrounded by an edematous collagenous matrix. This neoplastic cell is binucleate, with an undulated nuclear envelope and marginal and central heterochromatin. The nucleus on the left has a compact nucleolus (n). Note: anchoring junction (black arrows); collagen (Cg); collagen fibrils (white arrow); distended rough endoplasmic reticulum (black arrowheads); extracellular matrix (Mx); swollen mitochondria (white arrowhead). Bar = 5.0 µm. E. Two neoplastic cells with anchoring intercellular junctions (black arrows). The cell on the right contains cytoplasmic granules (G) lined by a single membrane and a homogeneous gray to electron-dense matrix. Note: distended rER (asterisks); mitochondria (m); nucleus (N). Bar = 2.0 µm.

Electron microscopy revealed that neoplastic cells were oval to polygonal, present individually or in clusters. The nuclear:cytoplasmic ratio was 1:1–1:1.5 with some binucleate neoplastic cells ( Fig. 2D ). The nuclei were oval, crenated with invaginations. The heterochromatin:euchromatin ratio was 1:4. The heterochromatin was in a granular central and marginal distribution. Mild-to-moderate increases were found in the thickness of nuclear fibrous lamina, which was distributed asymmetrically. Enlarged nucleoli were compacted with occasional segregation. Cytoplasmic organelles included round mitochondria, significant amounts of rough endoplasmic reticulum with distended cisterns, lysosomes, and membrane-bound electron-dense granules (presumably lysosomes and phagolysosomes). Intercellular desmosomes were present when neoplastic cells were in cluster of cells ( Fig. 2E ). Weibel–Palade bodies and pinocytosis vesicles were absent. The basal lamina was absent in neoplastic cells. Continuity of the neoplastic endothelial cells was not present in the few vascular-like structures examined. However, continuity of endothelium was challenging to evaluate because most of the tumor cells were individualized or in small clusters. The walls of small vessels within the stroma were composed of pericytes, basal lamina, and endothelial cells with pinocytotic vesicles, and the vessels contained erythrocytes within their lumens.

The diagnosis of cutaneous LSA was supported by combining findings from gross and histologic features consistent with a malignant lymphatic endothelial neoplasm; IHC labeling for vimentin, LYVE1, and PROX1, markers characteristic of lymphatic endothelium; and ultrastructural evidence from electron microscopy, indicating a lymphatic rather than blood vascular origin, including the absence of a continuous basal lamina, lack of Weibel–Palade bodies (organelles typical of blood vascular endothelial cells), and desmosomes rather than tight junctions. These features, taken together, support lymphatic endothelial, rather than blood endothelial, phenotype. We retrieved no cases of LSA in a search of Google, PubMed, Web of Science, and Scopus, using the search terms “lymphangiosarcoma”, “primate”, and “monkey”, suggesting that this condition has not been reported previously in non-human primates.

LSA is a rare malignant neoplasm that originates from lymphatic endothelial cells. LSA has been rarely reported in dogs and cats, most frequently arising on the ventrum, with additional isolated cases documented in horses and a cow.^5,8,13,15 The occurrence of LSA in non-human primates is unknown, given that vascular tumors are scarcely reported in non-human primates. One case series reported a higher incidence of LSA in male dogs. All published cases in horses have involved mainly females; however, the true sex predilection remains unclear.^6,15,19

Vascular tumors arising from either blood or lymphatic vessels are called angiosarcomas. ⁹ Ultraviolet radiation exposure has been proposed as a potential contributing factor in the pathogenesis of angiosarcoma in humans and may similarly influence tumor development in animals. ⁷ The affected monkey resided in the tropical environment of St. Kitts, where high levels of year-round sunlight could conceivably serve as an initiating or promoting factor in tumorigenesis.

The diagnosis of LSA is challenging, as it requires differentiation from other vascular neoplasms, particularly hemangioma and HSA. ¹¹ HSAs and hemangiomas typically form blood-filled vascular channels. ³ LSAs are also reported to have a discontinuous or absent basal lamina and discontinuous endothelial linings; HSAs have both a basal lamina and continuous endothelial coverage. ¹⁷ Immunohistochemically, factor VIII–related antigen and CD31 are typically expressed in HSAs and variably in LSAs, which may explain the lack of CD31 immunolabeling in our case. ⁴ Definitive confirmation of lymphatic endothelial origin relies on the use of endothelial cell–specific markers, such as LYVE1 and PROX1.^4,8 LYVE1 is a cell-surface receptor for hyaluronan, predominantly expressed in lymphatic capillaries—where it demonstrates diffuse cytoplasmic staining of lymphatic endothelium—and absent in blood vessels.^3,4 PROX1, a nuclear transcription factor essential for lymphatic endothelial differentiation, is expressed exclusively in lymphatic endothelial cells at all developmental stages, and has distinct nuclear staining.^3,4,10 These markers are increasingly used in veterinary medicine to assist in the diagnosis and classification of vascular tumors. In veterinary diagnostic pathology, particularly in canine cases, these markers are used to differentiate LSA from HSA, especially in cases with positivity of pan-endothelial markers, such as CD31 or factor VIII–related antigen. ⁴ In human medicine, podoplanin (D2-40) immunohistochemically labels lymphatic vessels, normal mesothelium, and tumor cells in various neoplasms, such as malignant mesothelioma and seminoma ¹⁸ ; in veterinary species, podoplanin is expressed in various tumor types.^2,16

Given that defined diagnostic criteria are not established in veterinary pathology, LSA is diagnosed by integration of histology, ultrastructural features (when available), immunophenotype, and exclusion of HSA ( Table 2 ), based on the following: 1) irregular vascular channels devoid of erythrocytes, lined by attenuated endothelial cells with residual abluminal basal lamina; 2) collagenous anchoring filaments in direct contact with the endothelial plasmalemma; 3) small intercellular gaps between adjacent endothelial processes with rare tight junctions or overlapping processes; 4) absence of endothelial fenestrations; and 5) lack of surrounding pericytes.^11,12,14 Of note, nuclear crenation seen in our case on electron microscopy is likely an artefactual modification of a round or oval structure (nuclear fibrous lamina).

Table 2.

Key diagnostic features of lymphangiosarcoma in animals.

Technique	Feature	Description
Histology	1. Irregular, anastomosing vascular channels	• Typically empty or nearly devoid of erythrocytes • Poorly circumscribed and infiltrative
TEM	2. Attenuated endothelial lining	• Flattened, spindle-to-elongated endothelial cells • Often discontinuous or poorly cohesive
	3. Residual or incomplete basal lamina	• Often fragmented or abluminal • Less continuous than blood vascular endothelium
	4. Anchoring filaments	• Fine collagenous filaments from endothelial plasmalemma to stroma • Classic ultrastructural hallmark of lymphatic endothelium
	5. Intercellular junctional pattern	• Small Intercellular gaps • Few tight junctions • Minimal overlapping endothelial processes
	6. Absence of endothelial fenestrations	• Helps distinguish from some blood vascular endothelia
	7. Lack of pericytes or smooth muscle support	• Reflects lymphatic rather than blood vascular differentiation
IHC	8. Immunohistochemistry	• PROX1 (most specific lymphatic marker) • Podoplanin (D2-40) (variable in animals) • LYVE1 (useful but not entirely specific) • CD31/CD34 positivity shared with hemangiosarcoma

TEM = transmission electron microscopy.

An interesting feature in our case was the granularity seen in the cytoplasm of many neoplastic cells. The granules (presumably lysosomes and phagolysosomes) are histologically and ultrastructurally similar to those described in a case of cutaneous epithelioid HSA in a dog with granular cell differentiation. ¹ Thus, our case appears to confirm that granular cell differentiation in a neoplastic cell is a phenotypical appearance seen in multiple cell types as a metabolic defect rather than being specific or indicative of a cell type. ¹

Prognosis of LSA is guarded (poor to extremely poor in dogs) given its rapid, infiltrative nature and spread along the deep muscular fascia; most reported cases have been fatal.^3,19 However, insufficient studies in veterinary medicine currently limit the identification of histologically relevant prognostic markers for LSAs. Widespread metastases of LSA have been reported in dogs, cats, and a horse^3,8,14,15; however, we found no evidence of metastasis in our case.

Our case brings awareness of cutaneous LSAs in non-human primates, which should be considered as a differential diagnosis for skin masses and as a possible background lesion for African Green Monkeys in a research setting. A combination of histology, IHC, and electron microscopy is necessary to reach a final definitive diagnosis.