Published online before print April 2, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Laboratory of Cellular Oncology, Center for Cancer Research, National Cancer Institute, NIH, Bethesda, Maryland, USA
1 Correspondence: Laboratory of Cellular Oncology, Center for Cancer Research, National Cancer Institute, Building 37, Room 4124, Bethesda, MD 20892, USA. E-mail: tosatog{at}mail.nih.gov
ABSTRACT
Kaposis sarcoma (KS)-associated herpesvirus is associated with the proliferative/malignant disorders KS, primary effusion lymphoma (PEL), and multicentric Castlemans disease (MCD) in patients with AIDS. In spite of recent advances in the treatment of KS, PEL and MCD represent therapeutic challenges. Recent advances in dissecting the pathogenesis of these diseases have indicated that the viral cytokine IL-6 and the cellular cytokines/growth factors IL-10, IL-6, stromal cell-derived factor 1, and vascular endothelial growth factor are important contributors to the growth, survival, and spread of PEL and MCD and are therefore potential targets for drug development.
Key Words: primary effusion lymphoma multicentric Castlemans disease angiogenesis cell migration
INTRODUCTION
The introduction of highly active antiretroviral treatment (HAART) has significantly extended the survival of patients with AIDS, but as more patients with AIDS succumb to cancer rather than from infectious complications, there is a need for innovative approaches in the treatment of AIDS-associated malignancies.
Primary effusion lymphoma (PEL) and multicentric Castlemans disease (MCD) are distinct B cell lineage malignancies associated with Kaposis sarcoma-associated herpesvirus (KSHV) infection in patients with AIDS. KSHV (also named human herpesvirus 8) is a
-herpesvirus discovered in 1994 in association with KS tissues [1
]. It has structural similarity to EBV and herpesvirus saimiri, and besides its association with KS, PEL and MCD are detected in a small proportion of normal individuals [2
]. PEL is a large-cell immunoblastic lymphoma, which usually presents as an effusion malignancy in the body cavities [3
4
5
]; MCD is a devastating lymphoproliferative disorder, which characteristically presents with severe general symptoms and lymphadenopathy [6
, 7
]. Typically, PEL is latently infected with KSHV [1
] and is often coinfected with EBV, which contributes to KSHV persistence in PEL [8
, 9
]. Instead, most MCD lesions are EBV-negative, and there is evidence of KSHV reactivation and replication, particularly during clinical exacerbations [10
, 11
]. KSHV plays a pathogenetic role in PEL and MCD through expression of the latency-associated nuclear antigen (LANA), viral G-protein-coupled receptor, viral (v)FLICE-inhibitory protein, viral cyclin, and other gene products that variously perturb cell division and survival [12
].
PEL cells are of B cell lineage but lack the phenotypic markers of mature B cells while expressing CD138 (syndecan-1) and multiple myeloma 1/IFN regulatory factor 4, which are typical of plasma cells [13
,14
]. Unlike PEL, KSHV-infected cells in MCD express a high level of IgM
and CD20 but do not express CD138 [15
, 16
]. PEL cells have been successfully propagated in vitro and represent a unique source of KSHV-infected cells, whereas no KSHV-infected cell line has yet been derived from MCD tissues.
Despite treatment with high-dose chemotherapy and other drugs, PEL and MCD usually display aggressive clinical courses, with median survival times of approximately 6 months from diagnosis for PEL and 12 months for MCD [7 , 17 , 18 ]. In this review, we will summarize the current understanding of the role played by the viral cytokine IL-6 and the cellular chemokine receptor CXCR4 in the pathogenesis of the KSHV-associated malignancies PEL and MCD, with a particular focus on how this information could be translated into novel therapies for these lethal diseases.
STRUCTURAL AND FUNCTIONAL CHARACTERISTICS OF vIL-6
vIL-6, a cytokine product of KSHV that is encoded at open reading frame K2, displays
25% amino acid identity to cellular IL-6 [19
, 20
]. It was originally identified in KSHV-infected PEL cells, where it is constitutively expressed to a limited degree and to a much greater extent when viral replication is induced [19
].
The IL-6 family of cytokines exerts its biological activities through a receptor complex composed by a nonsignaling receptor, which is cytokine-specific, and by the signal-transducing chain gp130 [21 ]. Use of the gp130 common transducer by all members of the IL-6 cytokine family explains the overlapping biological activities of these cytokines [21 ]. Cellular IL-6 binds to the specific receptor subunit, the IL-6R, and the complex of IL-6/IL-6R associates with gp130, promoting its homodimerization and signaling. Extensive structural studies have revealed that human (hu)IL-6 forms an antiparallel 4-helix bundle and have mapped the contact points of huIL-6 with IL-6R and gp130 [22 , 23 ].
vIL-6 differs from other IL-6 family members in that it can directly bind to gp130 and promote intracellular STAT/MAPK signaling without a requirement for IL-6R or another nonsignaling subunit [24
, 25
]. By ELISA, we found that vIL-6 can bind to gp130 directly, and by plasmon surface resonance, we measured that vIL-6 binding to gp130 displays an affinity that is
1000-fold lower than that of the huIL-6/IL-6R complex [24
] (Fig. 1
). Certain studies, including our own [24
, 26
], have provided evidence that IL-6R plays no role in vIL-6 signaling, but others have shown that vIL-6 can signal through gp130, with or without IL-6R [27
]. Studies in vitro have demonstrated that vIL-6 displays similar biological activities to those of huIL-6, including its ability to promote the proliferation of the murine hybridoma B9 and human myeloma INA-6 cell lines and to induce IgM secretion from the SKW6.4 human B cell line, but is
1000-fold less potent than huIL-6. In this regard, recent studies have provided evidence that vIL-6 is largely an intracellular cytokine, which is inefficiently secreted but can be biologically active intracellularly [28
].
![]() View larger version (24K): [in a new window] |
Figure 1. Properties of vIL-6 binding to gp130. (A) Binding of huIL-6 (50 mg/ml) to immobilized, soluble gp130, with or without soluble IL-6R (sIL-6R), analyzed by plasmon resonance (BIAcore 2000 system). (B) Binding of recombinant (Escherichia coli-derived), purified maltose-binding protein (MBP)-vIL-6 to gp130. (C) Schematic representation of the vIL-6/gp130 complex based on crystallographic analyses. gp130 (red and yellow) binds to vIL-6 helices in a tetrameric complex. Site I denotes an epitope recognized by a vIL-6-neutralizing mAb that we have generated.
|
-globulinemia. These results indicated that vIL-6 is a multifunctional cytokine that promotes tumor angiogenesis and tumor growth, at least in part by inducing VEGF expression, and induces plasmacytosis in lymphoid organs. Recent experiments have confirmed that vIL-6 can promote VEGF expression and additionally described that vIL-6 induces the VEGFRs-1 and -2 and angiopoietin-2 [30
]. Cellular IL-6 has similar biological activities, a result that is consistent with vIL-6 and cellular IL-6 signaling through the common gp130 signal transducer [21
, 31
, 32
].
![]() View larger version (120K): [in a new window] |
Figure 2. Tumorigenicity of NIH3T3 cells overexpressing vIL-6 in immunodeficient mice. (A) Representative tumors in nude mice 4 weeks after they were injected s.c. with vIL-6-NIH3T3 cells. Control NIH3T3 and vIL-6-NIH3T3 cells were injected at 0.5 x 106 cells/mouse (BALBc nude). (B) Increased vascularization in vIL-6-NIH3T3 tumors. Representative image from tumor tissue stained with H&E (original magnification, x100). (C, E, and G) VEGF visualized by immunohistochemical staining in tumor tissue (original, x40), lymph node (original, x200), and spleen (original, x100), respectively. (D, F, and H) Control immunostaining of tumor tissue, lymph node, and spleen, respectively.
|
Immunohistochemical analysis of MCD lesions revealed that KSHV-infected cells, which we detected by their expression of the viral nuclear antigen LANA, also expressed vIL-6 [33 ]. Indeed, most cells that were positive for LANA were also positive for vIL-6 (Fig. 3 ). In addition, we found that most sera from AIDS patients with MCD contained measurable levels of vIL-6 [34 ].
![]() View larger version (29K): [in a new window] |
Figure 3. vIL-6 detection in AIDS-associated MCD tissue and blood. Immunohistochemical detection of KSHV LANA (A) and vIL-6 (B) in the mantle zone of a lymph node affected with MCD (original magnification, x200). The characteristic, nuclear-speckled staining of LANA and the cytoplasmic staining of vIL-6 are shown in the insets (original, x1000). (C) Serum IL-6 and vIL-6 levels in the circulation of a patient with AIDS-MCD during an active phase of the disease and subsequent clinical remission. Antiretroviral therapy was with stavudine (d4T), lamivudine (3TC), and nelfinavir (NFV); prednisolone was added at the outset of general symptoms and tapered; the antiviral drug foscarnet was added as the clinical symptoms subsided.
|
The PEL cell lines BC-1, BCBL-1, and BCP-1 secrete vIL-6 (1680–19,200 pg/ml/48 h) when propagated in culture under standard culture conditions (1x106 cell/ml), and vIL-6 levels markedly increase after culture with the phorbol ester 12-O-tetradecanoylphorbol 13-acetate (2630–118,100 pg/ml/48 h). In addition, most PEL body-cavity effusions contain high levels of vIL-6 (1390–66,630 pg/ml). They additionally contain variable amounts of huIL-6 (957–37,494 pg/ml), IL-10 (<8–1,342,318 pg/ml), and VEGF (1133–11,417 pg/ml) [37 ].
We examined the effects of vIL-6 on PEL cell growth. Using neutralizing antibodies to vIL-6, IL-6, and IL-10, we determined that vIL-6 and to a greater extent huIL-10 serve as autocrine growth factors for the PEL cell lines BCBL-1 and BC-1 but found no evidence that huIL-6 plays such role [38 ] (Fig. 4 ). The prominent role of IL-10 in sustaining PEL autocrine growth was confirmed by blocking experiments using a soluble form of the recombinant IL-10R [38 ]. Other investigators, however, have reported that huIL-6 can also stimulate the growth of selected PEL cells under certain culture conditions and have suggested that the in vitro, antiproliferative effects that the drug rapamycin exerts on PEL cells are largely attributable to its inhibition of autocrine growth factor activity by IL-6 and IL-10 [39 , 40 ]. VEGF, secreted at biologically active concentrations by PEL cell lines, has no direct, biological activity on the PEL cells, which do not express the principal VEGF signaling receptor VEGFR2 and variably express VEGFR1. Thus, PEL cells generally secrete IL-10, IL-6, vIL-6, and VEGF; IL-10 and vIL-6 exert autocrine growth factor activity in PEL (Fig. 5 ).
![]() View larger version (23K): [in a new window] |
Figure 4. Effects of vIL-6 and IL-10 neutralization on PEL cell proliferation. BCBL-1 PEL cells were cultured for 3 days, with or without 25% autologous conditioned medium alone or with anti-vIL-6 IgG, control IgG, mAb to IL-10, or a combination of anti-vIL-6 and anti-IL-10 antibodies. Cell proliferation was measured by radioactive thymidine uptake.
|
![]() View larger version (29K): [in a new window] |
Figure 5. Schematic representation of autocrine growth factor activity in PEL. KSHV-encoded vIL-6, cellular IL-10, IL-6, and VEGF are constitutively expressed and secreted in the culture supernatants of PEL cells. Cellular IL-10 and vIL-6 bind to the IL-10R and gp130, respectively, activate STAT3, and promote PEL cell growth.
|
Noteworthy, treatment with anti-human VEGF antibodies did not eradicate PEL from the mice injected i.p. with BCBL-1 cells, as a number of mice had evidence of PEL cell dissemination to the blood and distant sites [41 ]. This result is consistent with the experiments in vitro showing that PEL cells do not express VEGF receptors and fail to respond to VEGF [41 ] and provides evidence that the beneficial effects of VEGF neutralization derive from blocking host vascular hyperpermeability.
Recently, we have analyzed the pattern of PEL dissemination in a new mouse model of PEL (unpublished results). Immunodeficient NOD/SCID mice inoculated i.p. with BC-1 cells (20x106 cells/mouse) usually develop a progressive, lymphomatous ascite, which is visible on inspection 8–10 days after injection and is documented by recovery of PEL cells from the peritoneal fluid (Fig. 6 ). At autopsy, 3–4 weeks after PEL cell injection, 50–75% of the animals display a visible tumor mass growing from the parietal and/or visceral mesothelial layer of the peritoneal cavity, which is histologically consistent with PEL (Fig. 7A ). This is accompanied by evidence of thickening of the diapham, which is infiltrated with PEL (Fig. 7B) , PEL infiltration of the abdominal muscles (Fig. 7C) , and colonization to distant sites (Fig. 6) . We reasoned that PEL transition from a localized effusion malignancy to a disseminated lymphoma compromising essential body functions is critical to disease progression and patient demise.
![]() View larger version (33K): [in a new window] |
Figure 6. Schematic representation of a PEL mouse model. Immunodeficient NOD/SCID mice are injected i.p. with PEL cells (BC-1 cells, 20x106/mouse) on Day 1. On Days 8–10, the mice develop a lymphomatous ascite; in the example shown, the peritoneal cavity is closed. On Days 15–18, 50–75% of the mice develop visible tumors attached to the peritoneal mesothelium. From Day 29 on, most mice have evidence of PEL dissemination outside the peritoneal cavity. In the example shown, PEL cells are identified microscopically under the renal capsule and in the blood.
|
![]() View larger version (49K): [in a new window] |
Figure 7. Microscopic evidence of PEL dissemination in a murine model of the disease. Immunodeficient NOD/SCID mice were injected i.p. with PEL cells (BC-1 cells, 20x106/mouse) and were killed on Days 21–30. PEL infiltration of the peritoneal mesothelium (A), the diaphram (B), and the abdominal muscles (C) is visualized by microscopic morphology. Tissue sections were stained with H&E; original magnification, x20.
|
TARGETING vIL-6 AND CELLULAR CYTOKINES FOR THE TREATMENT OF MCD AND PEL
We have developed a series of mAb against vIL-6 [24 ]. Six of these were characterized in detail. All of these antibodies recognized specifically vIL-6 but not cellular IL-6 in Western blotting and ELISA. In functional studies, we determined that four of the monoclonal did neutralize vIL-6 activity measured in the B9 proliferation assay and the SKW6.4 Ig secretion assay, but they did not neutralize cellular IL-6 activity, also measured in these assays. One of these antibodies (v6m31.2.4) had the highest neutralizing capacity for vIL-6 [24 ]. Mapping studies using recombinant vIL-6 mutants that we produced indicated that all of the neutralizing antibodies recognize a 13-residue epitope in vIL-6, which includes Asp81 to Cys93 [24 ]. Structural analyses of vIL-6 binding to gp130 have indicated that this vIL-6 region is not found in the binding interface, suggesting that the neutralizing antibodies may alter the conformation of vIL-6 at the critical binding interface with gp130 [50 ]. Based on the recent, successful application of monoclonal antibodies (mAb) to the treatment of various human diseases, we believe that vIL-6 neutralization by monoclonal could be tested for the treatment of KSHV diseases, particularly MCD, where the cytokine appears to contribute significantly to disease morbidity.
The viral cytokine vIL-6 and the cellular cytokines/growth factors IL-10, IL-6, and VEGF all induce phosphorylation of STAT3 for intracellular signaling [51
52
53
]. PEL cells consistently express a constitutively active STAT3, which is attributable, at least in part, to the autocrine growth factor activities of vIL-6 and IL-10. To target activated STAT3 in PEL cells, we used a dominant-negative form of STAT3, which contains a mutated DNA-binding motif and acts as a competitive inhibitor of wild-type STAT3 [54
]. When PEL cells were transfected with this construct, significant apoptosis rapidly ensued. At 3 days after transduction,
90% of the PEL cells were dead, and control cells that do not express active STAT3 displayed minimal cell death. Similar PEL cell apoptosis was induced by tyrphostin AG490, a JAK2 inhibitor that was shown to inhibit STAT3 phosphorylation. Levels of AG490-induced apoptosis in PEL cells were markedly reduced by the broad-spectrum caspase inhibitor Z-Val-Ala-Asp-fluoromethylketone, providing evidence for caspase-dependent cell death. Interestingly, we found that inhibition of STAT3 activity by AG490 was associated with decreased levels of the antiapoptotic protein survivin, which inhibits the activity of caspase proteases [55
]. By contrast, levels of the other antiapoptotic proteins Mcl-1, Bcl-2, and Bcl-XL were unaffected in PEL treated with AG490. Furthermore, the overexpression of survivin in PEL cells rendered them resistant to apoptosis by AG490. Together, these results provide evidence that STAT3 is a desirable, therapeutic target for PEL [54
]. There is much interest in the development of STAT3 inhibitors [56
, 57
]. Our results indicate that such drugs could be extremely useful in the treatment of PEL.
CONCLUDING REMARKS
The use of HAART aimed at controlling HIV replication and promoting immune reconstitution has been successful at prolonging the survival of patients with AIDS. Treatment of KSHV-induced malignancies is now a challenge, as more patients survive longer. Intense investigations of KSHV-induced malignant transformation and the pathogenesis of KSHV malignancies provided important clues for the design of rational therapies. Viral proteins, such as vIL-6, and cellular cytokines and their receptors have emerged as promising targets for treatment as they contribute to the pathogenesis of PEL and MCD.
Received November 18, 2007; revised February 11, 2008; accepted February 11, 2008.
REFERENCES
-receptor/gp130 complex Science 300,2101-2104
4 integrin clustering by immobilized chemokines stimulates leukocyte tethering and rolling on endothelial vascular cell adhesion molecule 1 under flow conditions J. Exp. Med. 192,495-506This article has been cited by other articles:
![]() |
Z. He, Y. Liu, D. Liang, Z. Wang, E. S. Robertson, and K. Lan Cellular Corepressor TLE2 Inhibits Replication-and-Transcription- Activator-Mediated Transactivation and Lytic Reactivation of Kaposi's Sarcoma-Associated Herpesvirus J. Virol., February 15, 2010; 84(4): 2047 - 2062. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||