HHV-8/KSHV association with tumor cells during development of Kaposi sarcoma

Abstract

Kaposi's sarcoma (KS) is a highly and abnormally vascularized tumor-like lesion which usually presents as a cutaneous disease and progresses to extracutaneous, systemic spread usually in the gastrointestinal (GI) tract, lungs, oral cavity and lymph nodes. It is the most common neoplasm of AIDS patient and included in the staging of HIV disease progression. The Kaposi's sarcoma-associated herpesvirus or human herpesvirus-8 (KSHV/HHV8), a gamma-herpes virus was discovered in AIDS related KS tumor samples and subsequently in all clinicoepidemiological forms of KS as Classic KS (CKS), Iatrogenic KS (IKS), Endemic KS (EKS) and AIDS-associated KS (AKS) but also in some rare lymphoid disorders such as body cavity-based lymphoma (BCBL) and multicentric Castleman's disease (MCD). It is considered to be the main pathogenic factor of KS but apparently not sufficient to cause KS and other factors like immunosuppression, production of angiogenic cytokines and HIV-tat in AIDS evidently also play an important role. However the consistent demonstration of HHV-8 by polymerase chain reaction (PCR) and immunohistochemistry in tumor lesions, blood cells, serum and saliva of KS patients validates the pathogenic role of HHV-8 in KS oncogenesis.

Early KS lesions develop from patch/plaque to late, nodular tumors. The histology is characterized by an early infiltration of mononuclear inflammatory cells, formation of atypical small blood vessels and slits (angiogenesis) and the appearance of an increasing number of tumor spindle cells (SC). Various studies favour an endothelial origin (CD34+) of the KS SC but whether of vascular (VEC) or lymphatic endothelial cell (LEC) origin has not been settled. The HHV-8 latency-associated nuclear antigen type 1 (LANA-1) protein is the most frequently expressed viral antigen in infected cells. Whether KS represents a predominantly monoclonal neoplastic cell proliferation or a hyperplastic, reactive polyclonal process is still controversial.

In the present KS histopathological studies (paper I&II) by triple antibody immunofluorescence we observed that: (a) the frequency of LANA+/CD34+ cells increased from early patch to late KS nodular lesions; (b) 15-25% of the CD34+ SC were LANA- in both early and late KS, suggesting a continuous recruitment of non-infected endothelial cell into the KS lesion; (c) LANA+/CD34-cells were more frequent in early as compared to late KS and most of them expressed LEC markers such as LYVE-1 and D2-40, suggesting that the resident LECs represent an early target of primary HHV-8 infection; (d) LANA+/CD34-/LYVE-1+ cells decreased from early (18.7%) to late (2.9%) KS suggesting a phenotype switch from LEC to VEC.

We also established and optimized a semi-quantitative PCR assay for HHV-8 detection in formalinfixed paraffinembedded KS biopsies (paper II) and two different protocols for DNA extraction were compared namely the Chelex 100 and Qia-gene kit method. Our results indicate a better performance for Chelex-extracted DNA from paraffin embedded archival biopsies. In late, nodular stage of both AKS and EKS the virus load per unit tissue actin (HHV-8/actin) is higher than in early stages (patch/plaque), which corroborates our findings from double immunostaining for LANA and CD34 of the same cases. With quantitative real time PCR on DNA from oral and cutaneous AKS and EKS biopsies (paper III), we have not found significant higher HHV-8 load per unit Rnase-P in AKS compared to EKS in accordance with the immunohistochemistry findings and consistent with the notion of basically similar pathogenic mechanisms for these clinically distinct lesions. HHV-8 load per unit Rnase-P was significantly higher in the late than early tumor stages is in accordance with the immunohistochemistry findings in AKS as well as EKS of an increase in the frequency of infected spindle cells at late stages, and not an increased production of virus by the tumor SC. HHV-8 load per unit Rnase-P in oral was higher than in cutaneous AKS concordant with the finding of more LANA+ cells and LANA+ granules per nucleus of SC in oral compared to cutaneous AKS. This may reflect increased effects of LANA with more SC proliferation and less apoptosis in oral compared to cutaneous lesions. The high viral load in oral KS lesions is consistent with the known high risk of HHV-8 horizontal transmission.

Our findings also substantiate other studies that PCR analysis of HHV-8 is a simple and sensitive assay for differential diagnosis of all KS forms in biopsies processed for conventional histopathology.