Angiogenesis in cancer of unknown primary: clinicopathological study of CD34, VEGF and TSP-1

Cancer of unknown primary (CUP) is a unique clinical entity that accounts for an approximately 3% of human cancers[1]. Patients with CUP present with metastases for which the site of origin cannot be identified at initial workup. Early dissemination, unpredictability of metastatic pattern and aggressiveness constitute fundamental characteristics of these tumors. Although the clinical characteristics of CUP have been established, little is known about the underlying biology of these tumors [2, 3].

Angiogenesis, the formation of new vessels, is essential for tumor growth and the development of metastases. It evolves though a complex multifactor process that involves interaction of pro-angiogenic and anti-angiogenic signals from tumor, endothelial and stromal cells. The angiogenic activity is reflected in the development of novel microvessels in tumor tissue that is quantified by the intratumoral microvessel density (MVD). Among several molecules implicated, Vascular Endothelial Growth Factor (VEGF) and Thrombospondin-1 (TSP-1) appear to be most relevant. Much evidence indicates that VEGF is a key activator of angiogenesis[4, 5] and TSP-1 a primary endogenous inhibitor of angiogenesis[6] Up to now, no useful prognostic factors have been established other than the classic pathologic and laboratory ones and immunohistochemical detection of various factors did not add prognostic value in CUP.[7, 8] Moreover, investigation of the expression of crucial angiogenesis factors that can be therapeutically targeted is today of great interest for the oncologists who deal with CUP clinical research.[9]

We were prompted to investigate angiogenesis in unknown primary cancer in an attempt to enrich our understanding of the biology of these tumors. We studied by immunohistochemistry the tissue expression of VEGF and TSP-1 in CUP and correlated with MVD and clinicopathological parameters. In a recently published study vascular endothelial growth factor, and CD34, factors were not found to be of prognostic value in adenocarcinoma of unknown primary.[8]

A total of 81 patients diagnosed with CUP and treated in three University Medical Oncology Settings (Ioannina, Patras and AHEPA, Thessaloniki, Greece) between January 1997 and December 2002 were selected on the basis of availability of archival tumor tissues and accessibility to medical notes. Pathology diagnosis was reviewed by two pathologists blinded to written pathology report and representative paraffin blocks were selected for immunohistochemistry.

Demographics of studied cases are depicted in Tables 1 and 2. Overall survival of patients included in this study is shown in Figure 1. (Median survival 10 months).

The investigation of the biological profile of CUP and the understanding of molecular pathways underlining these tumors has been limited. We have worked on these issues and found several oncoproteins overexpressed, but failed to establish any clinically relevant correlations[10, 11]. We now investigated neo-angiogenesis by assessing MVD and the tissue expression of two representative molecules involved in angiogenesis; the major stimulator of angiogenesis VEGF (A), and the intrinsic angiogenic inhibitor TSP-1. Overall, we demonstrated that a high angiogenetic activity occurs in CUP tumors, which is higher in unfavorable when compared with more favorable subsets.

VEGF is known to play a key role and MVD is considered to reflect the final result of the tumor angiogenesis cascade. In the present study, all cases were found to be VEGF-positive and in the majority VEGF was overexpressed. Tumor VEGF and MVD were strongly correlated that is in line with findings in solid tumors[12, 13]. We failed to demonstrate any significant correlations of angiogenic activity with regard to clinical outcome, but this was obviously due to universal expression of both CD34 and VEGF in our cases.

We also demonstrated that in our series TSP-1 was overexpressed in 50% and was absent or weak in approximately half of the cases. TSP-1 correlated inversely with microvessel counts. The role of TSP-1 in epithelial tumor growth and metastases remains controversial. In vitro studies suggest that TSP-1 may promote tumor cell adhesion and invasion by up-regulating urokinase plasminogen activator and its receptor[14] but in clinical studies overexpression has been associated with a lower MVD score and a better clinical outcome in several carcinomas[15]. Moreover, other studies suggest that TSP-1 inhibits tumor progression and may serve as an indicator of less aggressive potential and of favorable prognosis in solid tumors.[16] We consider that low TSP-1 in our material reflects a suppression of anti-angiogenic mechanism of TSP-1 that possibly contributes to the aggressiveness of these tumors.

CUP patients have in general a brief life expectancy with a median survival approximately of 3–9 months.[17, 18] It must be emphasized that CUP diagnosis applies to a heterogeneous group of patients who are usually grouped together in biological and therapeutic studies to obtain statistically meaningful results. However several patients fare better and enjoy longer survival and within this more favorable prognostic subgroup, unique subsets, such us young patients with midline tumors and women with peritoneal carcinomatosis or isolated axillary adenocarcinoma, have a distinct clinical biology compared to others also classified as unknown primary cancer. [19‐23]. In our study MVD score were found low in the group of more favorable tumors compared to unfavorable, but neither MVD nor VEGF or TSP-1 were associated with known prognostic factors.[24] Similarly, Hillen et al, in a small study, evaluated MVD as a prognostic factor for patients with liver metastases of unknown primary and found that MVD score correlated with marginally shorter survival.[25]

In conclusion, we found that angiogenesis is very active and VEGF expression is universal in cancer of unknown primary, which supports the clinical investigation of VEGF targeted therapy in this clinical setting.[9] To identify additional druggable molecular targets in cancer of unknown primary we need to advance our knowledge on the biology of these tumors and validate novel molecular therapeutics.