CASE REPORT

In 1993, a 24 year old woman underwent a laparotomy because of bleeding in the liver. A hemi-hepatectomy was performed and revealed a tumour of 11 × 15 cm; in addition a 3 cm tumour was resected from the ileum. The tumours were diagnosed as GANT. In 1998, surgery was performed because of multiple GANT lesions in the pelvis. In May 2001, tumour cyst fluid was taken from a liver lesion and pleural fluid during an episode of pleurodesis. In 2002, the patient died because of bleeding in a lung metastasis.

MATERIAL AND METHODS

Sections (4 µm thick) were cut from formalin fixed, paraffin wax embedded sections and mounted on Superfrost Plus slides (Menzel, Germany), dried overnight, dewaxed, and dehydrated. Next, antigen retrieval was performed with 10mM citrate buffer (pH 6.0) at 100°C in a microwave oven, pretreatment was performed with either trypsin or pepsin in the Nexes IHC automated immunostainer (Ventana, Strassburg, France). Endogenous peroxidase activity was blocked with 0.03% hydrogen peroxide in methanol for 10 minutes. Slides were stained either by hand or in the automated immunostainer according to the manufacturer’s instructions (Ventana). Slides were washed with phosphate buffered saline between steps. Visualisation was performed with either diaminobenzidine or envision kit; subsequently, the slides were counterstained with haematoxylin, dehydrated, and mounted in DePeX mounting medium (BDH Laboratory Supplies, Poole, Dorset, UK). CD34, mouse monoclonal (ms-mono), CD117 (ms-mono), chromogranin (ms-mono), anti-NSE (ms-mono), anti-S-100 (rabbit polyclonal), K6, and K20 (ms-mono) were purchased from Dako (Ely, Cambridgeshire, UK). The antibody to vimentin (ms-mono) was a gift from Dr J Hilgers (Department of Gynaecology, VUMC, Amsterdam, the Netherlands).

For electron microscopy, fragments of the tumour were fixed in 2% glutaraldehyde, postfixed in osmium tetraoxide, dehydrated with ethanol, and embedded in Epon. Ultrathin sections were collected on Formavar coated copper grids. The sections were contrasted with uranyl acetate and lead citrate and examined in a JEOL 1200 EX electron microscope.

The VEGF and bFGF enzyme linked immunosorbent assay (ELISA) kits were from R&D Systems (Minneapolis, Minnesota, USA) and the endostatin ELISA was from Cytimmune (College Park, Maryland, USA).

The MMP-2 and MMP-9 activity assays were from Amersham Biosciences (Cardiff, UK). These assays separately measure active and total (active plus latent) activity. Protein concentrations of MMP-2 and MMP-9 were measured with a standard quantitative ELISA, as described previously.⁸ The urokinase PA (uPA) protein concentration was measured by means of the immunoassay kit: u-PA EIA HS (Taurus, Leiden, the Netherlands). The activity of uPA and plasmin was determined as described previously.⁹ The tissue PA (tPA) antigen was determined using a specific tPA ELISA (Biopool Int, Umea, Sweden).

RESULTS

Histopathology revealed a mesenchymal proliferation, negative for keratins, and with focal positivity for c-kit and CD34. Vimentin, S-100, NSE, and chromogranin were positive. Eleven mitoses/2 mm² were found. The vascularisation of the tumour tissue was exuberant. VEGF was expressed abundantly in this tumour and large cysts with VEGF positivity were present (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Immunohistochemical staining for vascular endothelial growth factor showing pseudocyst formation with positively (brown) stained cyst fluid, and positive tumour cells surrounding the cyst (original magnification, ×10).

Electron microscopic examination revealed the focal accumulation of desmosomes in the tumour cells. Dense core granules were seen in the cytoplasm of the tumour cells and microtubules were present focally. Intercellular skeinoid fibres were absent but collagen fibres, approximately 20 nm thick, were seen.

Tables 1 and 2 show data from the patient’s tumour fluid, pleural fluid, and blood. There were high concentrations of both VEGF and endostatin in all three sample types. The production and activity of MMP-2, MMP-9, and uPA was reduced in tumour fluid. The concentration of tPA and the activity of plasmin were higher than those seen in tumour fluid from nine other patients with sarcoma.

DISCUSSION

GANTs are associated with bleeding, haemorrhagic cysts, and a high vascularity,^1,2 as was the case in our patient. Large intratumoral cysts have been described as a frequent characteristic of soft tissue sarcomas (STS) in general, and these cysts often contain high amounts of VEGF,¹⁰ which is a permeability factor also. The concentration of VEGF, but not of bFGF, was extremely high in tumour fluid, pleural fluid, and in serum from our patient.

A remarkable finding was the increased concentration of endostatin in tumour fluid, pleural fluid, and plasma of our patient. This is an unusual finding because the concentration of endostatin in the tumour fluid from nine other STS was not different from the normal plasma value, and the concentrations of endostatin in pleural fluid and ascites were also not increased in a series of 92 samples from a broad spectrum of cancer types.¹¹ This finding suggests that this GANT tumour generated (an) enzyme(s) capable of degrading CXVIII to endostatin. Increased serum concentrations of endostatin in patients with STS, associated with an increased risk of tumour recurrence after resection, have been reported previously.¹² The role of endostatin in the biology of these tumours warrants further study.

“A remarkable finding was the increased concentration of endostatin in tumour fluid, pleural fluid, and plasma of our patient”

Mechanistic studies have not yet firmly established a unique receptor or target for endostatin. Binding to cell surface integrins (α₅β₁), blocking the VEGF–KDR/Flk-1 interaction, and binding the catalytic domain of MMP-2 have been suggested.^13–15 In the tumour fluid of our patient, the activity of MMP and uPA was reduced, but the tPA protein concentration and plasmin activity were greatly increased, which is in accordance with the mechanism of action suggested by Reijerkerk et al.¹⁶ These workers suggested that VEGF induces hyperpermeability of vessels, resulting in the local influx of plasma proteins and the generation of a fibrin containing “provisional matrix”, which facilitates the migration of endothelial cells. However, excessive degradation of this matrix by increased plasmin activity may have largely contributed to a high turnover of blood vessels and the bleeding tendency in this tumour.

In the case reported here, increased production of VEGF was accompanied by the generation of endostatin. Because VEGF initiates angiogenesis, resulting in the generation of immature vessels, and endostatin inhibits angiogenesis, ultimately resulting in apoptosis of immature vessels, this combination might have had a negative effect on the integrity of the endothelial lining. Tumour associated neovascularisation has been described as an active, but abnormal and inefficient process, because of the absence of a balanced angiogenesis programme. In this GANT the presence of both VEGF and endostatin could have led to dysfunctional angiogenesis. It remains to be determined whether other GANTs have this angiogenic profile and whether this can explain the relatively mild progression of this disease and bleeding as a complication.

This is the first description of a tumour producing both the angiogenesis stimulator VEGF and the angiogenesis inhibitor endostatin. The investigation of tumour fluid, which can be seen as the in vivo “conditioned medium” of a tumour, was of great help in understanding the dysfunctional vascularisation of this tumour.