Kaposi's sarcoma

Coupling herpesvirus to angiogenesis

Article metrics

Since the discovery of Kaposi's sarcoma-associated herpesvirus (KSHV) only three years ago1, and its detection in every Kaposi's sarcoma biopsy (Fig. 1), there have been many hypotheses to explain how this virus contributes to the abundant vasculature and spindle-cell proliferation that are associated with the disease. Spindle cells are considered to be the ‘tumour’ cells, and they contain (and secrete) large amounts of a powerful angiogenic agent called vascular endothelial growth factor (VEGF)2. Angiogenesis — the sprouting of new blood vessels from pre-existing ones — is essential for tumour progression and, on page 86 of this issue, Bais et al.3 report that the G-protein-coupled receptor (GPCR) encoded by KSHV induces an ‘angiogenic switch’ in transformed NIH3T3 cells.

Figure 1: Kaposi's sarcoma biopsy.

a, Numerous vascular spaces and extravasated red blood cells are surrounded by spindle (‘tumour’) cells. b, Higher magnification of spindle cells, showing abundant magenta staining for the powerful angiogenic protein, vascular endothelial growth factor. (Reproduced, with permission, from ref. 2.)

Like experimentally mutated cellular GPCRs, KSHV-GPCR is fully active in the absence of chemokine ligands4. Bais et al.3 show that KSHV-GPCR can, indeed, act as a viral oncogene to transform NIH3T3 cells, and that this transformation is accompanied by secretion of VEGF. Spindle cells in Kaposi's sarcoma belong to the endothelial lineage, but they often express antigens characteristic of endothelial, macrophage and smooth-muscle cells. This indicates that they either represent pluripotent mesenchymal precursors, or are endothelial cells undergoing aberrant differentiation. KSHV DNA is present in most spindle cells, and also in endothelial cells and monocytes in Kaposi's sarcoma lesions.

The KSHV-GPCR gene is not the only viral oncogene to induce tumour formation and angiogenesis. Cells transformed by v-Ha-ras and v-raf (the constitutively active retroviral forms of cellular ras and raf) express increased VEGF5. But these genes induce a different signalling pathway from KSHV-GPCR. Bais et al. have found that GPCR triggers the JNK/SAPK and p38MAPK pathways, although they have not shown whether these cascades are involved in the induction of VEGF. So as many as three major signalling pathways that link membrane receptors with the nucleus in mammalian cells could, potentially, be involved in tumour progression by switching on angiogenesis. The viral interferon regulatory factor is another KSHV-encoded protein that can transform NIH3T3 cells and induce tumour formation in nude mice6.

The KSHV-encoded GPCR is part of a potential ‘oncogenic cluster’ of proteins within the viral genome. It is transcribed with OX2 (S. Talbot, unpublished data), which is a protein involved in intercellular signalling. This transcript is in the opposite direction — and from a different promoter — to three viral genes (Fig. 2), which encode cyclin D, vFLIP (viral FLICE inhibitory protein), and an immunogenic latent nuclear antigen, LNA-1. The vFLIP protein contains a death-effector domain, and it is homologous to the cellular FLIP protein7. Like cellular FLIP, vFLIP might block one of the main pathways by which immune mechanisms cause cell death — induction of apoptosis by the tumour-necrosis factor family of receptors. LNA-1 resembles the Epstein-Barr virus nuclear antigens (EBNAs). It may transactivate other cellular or viral genes to promote cellular growth or, like EBNA-1, it could maintain the viral DNA in an extrachromosomal circular (episomal) form.

Figure 2: ‘Oncogenic cluster’ in the genome of Kaposi's sarcoma-associated herpesvirus (KSHV).

The G-protein-coupled receptor (GPCR) is encoded by a viral gene that is transcribed with a homologue of OX2. Four other viral genes are transcribed in the opposite direction, and from different promoters. Bais et al.3 have shown that GPCR acts as a viral oncogene, and that cellular transformation is accompanied by secretion of vascular endothelial growth factor. K12 is a latent membrane protein (LMP) that may act like one of the latent membrane proteins of Epstein-Barr virus.

Herpesviral proteins can be classified as being either ‘latent’ or ‘lytic’. Latent proteins are expressed during non-replicative infection, whereas lytic proteins are expressed during viral replication, and they are associated with cell death. In KSHV, the expression pattern of the pirated genes (see box) is more murky — many genes are expressed during both the lytic and latent phases, although they are upregulated during lytic infection8. The pattern of viral gene expression also seems to differ in haematopoietic compared with endothelial cells. For example, KSHV-encoded interleukin-6 is expressed in latently infected B cells, but not in spindle cells8.

KSHV-GPCR is expressed in latent B cells, and upregulated with lytic induction. Both types of KSHV infection are present in Kaposi's sarcoma9, and both probably contribute to the complex pathology of the disease. Most spindle cells in Kaposi's sarcoma are latently infected, although many might undergo lytic infection at some point. Lytic infection, with production of viral particles, might be necessary to drive the formation of lesions. It is also possible that ‘abortive lytic’ infection (meaning that most early lytic genes are expressed) contributes to the formation of tumours in Kaposi's sarcoma.

The complex histology and expression pattern of KSHV proteins in Kaposi's sarcoma indicate that the model of tumorigenesis might not be like any other virus-induced malignancy. The irony of current KSHV research is that, despite having all this ammunition to trigger a malignancy, nobody has yet shown that KSHV can transform any cell type in vitro. Moreover, only a fraction of immunocompetent infected patients will actually develop a tumour associated with this virus.


  1. 1

    Chang, Y.et al. Science 266, 1865–1869 (1994).

  2. 2

    Cornali, E.et al. Am. J. Pathol. 149, 1851–1869 (1996).

  3. 3

    Bais, C.et al. Nature 391, 86–89 (1998).

  4. 4

    Arvanitakis, L., Geras-Raaka, E., Varma, A., Gershengorn, M. C. & Cesarman, E. Nature 385, 347–349 (1997).

  5. 5

    Grugel, S., Finkenzeller, G., Weindel, K., Barleon, B. & Marme, D. J. Biol. Chem. 270, 25915–25919 (1995).

  6. 6

    Gao, S.-J.et al. Oncogene 15, 1979–1985 (1997).

  7. 7

    Irmler, M.et al. Nature 388, 190–195 (1997).

  8. 8

    Moore, P. S., Boshoff, C., Weiss, R. A. & Chang, Y. Science 274, 1739–1744 (1996).

  9. 9

    Orenstein, J. M.et al. AIDS 11, 735–745 (1997).

  10. 10

    Neipel, F., Albrecht, J.-C. & Fleckenstein, B. J. Virol. 71, 4187–4192 (1997).

  11. 11

    Swanton, C.et al. Nature 390, 184–187 (1997).

  12. 12

    Boshoff, C.et al. Science 278, 290–293 (1997).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.