Kaposi sarcoma (KS), a previously rare tumor lesion, has due to its association with the expanding HIV pandemic 1,2 become a global clinical problem. Cases of the epidemic, AIDS-associated KS (AKS) present with more rapidly disseminating lesions compared with the sporadic/classic, endemic African (EKS) and iatrogenic (transplant) forms. 3–5 All these clinically different KS forms are associated with a novel herpesvirus (HHV-8 or Kaposi sarcoma herpesvirus [KSHV]), which expresses a latency-associated nuclear antigen (LANA) in infected cells 6–11 including cells of body cavity–based lymphomas/pleural effusion lymphoma of AIDS patients. 12 Previous studies have indicated that circulating monocytes-macrophages and lymphocytes of KS patients can also be HHV-8 infected and may represent an important reservoir for viral transmission. 13–16 The different KS forms (AKS and EKS) have the same histopathologic appearance and are characterized by angiogenesis with inflammatory cell infiltrates, accompanied by the appearance of so-called spindle cells (SCs), considered to be the KS tumor cells, which typically express CD34 (hematopoietic stem cell and endothelial marker). 9,17–20 The dermal KS develops from an initial small patch/plaque into a locally growing, eventually ulcerating, nodular, tumorlike lesion. It is still a controversy whether KS should be regarded as a true monoclonal neoplastic proliferation of SC or a predominantly hyperplastic, reactive process, 21,22 in view of the absence of monoclonal, genomic KS markers 23 and that primary SC are diploid with relatively low proliferation rate. 27 Furthermore KS usually presents and disseminates as multicentric new lesions rather than metastasis of the primary tumor. 21,24–30
Furthermore, we and others have demonstrated a stage-and Bcl-2-related, progressive reduction in apoptotic cells as a possible pathogenic factor in KS development. 21,25,27,29,30 The present study analyzed the possible relationship of HHV-8-infected (LANA+) and noninfected (LANA−) tumor SCs (CD34+), infiltrating CD3+, CD68+, CD20+, and CD45+ leukocytes as well as proliferating (Ki67+) cells during AKS and EKS development using a 4-color immunofluorescence technique.
MATERIAL AND METHODS
All KS tissues (n = 29) studied were diagnostic surgical biopsies classified as nodular AKS (n = 11), patch/plaque AKS (n = 7), and nodular EKS (n = 11) fixed in buffered formalin (10%) and embedded in paraffin at the Department of Pathology, Muhimbili University College of Health Science, Dar-es-Salaam, Tanzania. Tonsil biopsy material from a case with tonsillitis was used as performance control of immunostainings.
Deparaffinized and rehydrated 4-μm sections were microwave heated 6 minutes in 0.1 M citrate buffer (pH 6.0) for antigen retrieval and quenched for endogenous peroxidase activity by treatment with 0.3% hydrogen peroxidase. 27 All antibody incubations of the specimens were done at room temperature unless stated otherwise and the immunostained sections were mounted with Vectashield (Vector Laboratories, Inc., Burlingame, CA) 31 (Table 1). Antibodies were diluted in Tris Buffer Saline (TBS) containing 1% bovine serum albumin and 0.05% sodium azide.
Sections were blocked for nonspecific antibody binding with normal horse or goat serum (30 minutes), before incubation with the antibodies (Table 1), according to 2 different protocols: 1) After incubation with a primary antibody (37°C, 1 hour), binding was detected by application (40 minutes) of biotinylated horse anti-mouse, goat anti-rabbit, or goat anti-rat, respectively, followed by incubation with peroxidase-conjugated avidin-biotin complex (ABC) and 3.3 diaminobenzidine. 27 4′-6-Diamidino-2-phenylindole-Hcl (DAPI) (5 μM DAPI in 500 m M trisodium citrate) was used for DNA staining. 32 2) For triple antibody immunostaining, serum blocked sections (see above) were incubated (37°C, 1 hour) with primary mouse antibody washed and labeled with anti-mouse fluorescein isothiocyanate (FITC) (40 minutes). Subsequently, the sections were incubated (37°C, 1 hour) with the rabbit (second) antibody followed by biotinylated anti-rabbit (40 minutes) and by Avidin-Cy5 (Amersham, Buckinghamshire, UK) (30 minutes), and finally with rat (third) antibody (37°C, 1 hour) followed by anti-rat Cy3 (40 minutes) labeling and DAPI. Between all incubations the sections were thoroughly rinsed in TBS.
Incubation with TBS instead of respective primary antibody was used as a negative control for the detection system. Performance of the respective antibodies was evaluated on sections of human tonsil as well as a body cavity–based lymphoma cell line (BCBL-1). 12 To evaluate possible cross-reactions between different primary antibodies and the respective secondary antibodies, adjacent sections were stained with either all 3 different primary antibodies, but only one secondary antibody and the corresponding detection system at a time; or only one primary antibody, but all 3 secondary antibodies and the detection system at a time. The triple immunostaining observations were confirmed by ABC peroxidase immunostaining with corresponding antibodies in parallel sections.
A fluorescence microscope (Olympus BX60, Tokyo, Japan) equipped with a digital camera (Sony DKC-5000, Tokyo, Japan) and filter cubes was used to document bright field as well as specific FITC, Cy3, Cy5, and DAPI fluorescence images (Fig. 1), which were edited and overlaid by Adobe Photoshop 6.0. Cells were counted and evaluated by visual scoring of color micrographs in eight adjacent fields (256 × 190 μm each) of characteristic lesions and the relative frequency of cells positive for respective marker (combination) was calculated.
Percentage, median, range, and t test P value were analyzed by Microsoft Excel (Redmond, WA).
CD34 and LANA Expression
The CD34-expressing cells were the most frequent in early (patch/plaque) (43.2%) as well as in nodular AKS (79.1%) and EKS lesions (74.0%) and increased significantly during KS evolution to late stages (Table 2). Most of these CD34+ cells had the morphology of SCs but also of vessellining endothelial cells (VLECs). LANA was expressed as a distinct granular, nuclear reaction in most SCs of nodular AKS and EKS (46.2 and 38.8%, respectively) but by significantly fewer at early stages (22.1%) (Table 2). LANA was also expressed on some VLECs of both AKS and EKS (Fig. 1A). The cell reactivity for LANA correlated with the frequencies for CD34+ SCs as evident from the doubly labeled (CD34+LANA+) cells (Fig. 1B and C). However, the double-positive (CD34+LANA+) cells (Table 2, subgroups “b” and “e”) represented only a fraction of the total number of CD34+ cells (Table 2) in nodular AKS (41.8% of 79.1%) and EKS (32.6% of 74%) as well as in early AKS (15.1% of 43.2%). Thus a significant number of CD34+ cells in late (36.8–38.7%) and also in early lesions (22.0%) were LANA− (Table 2, subgroups “c” and “f”) (Fig. 1C). In contrast, relatively few LANA+ cells were CD34− (Fig. 1C); these cells usually were more frequent in early (6.5%) as compared with late (3.4–4.1%) (Table 2, subgroups “a” and “d”) stages. However, in general no significant difference in the total frequency of LANA+ cells was observed between AKS and EKS cases at comparable stages (P > 0.05) (Table 2).
Immunostaining for the proliferation marker Ki67 showed significant higher frequencies in early AKS stages (11.5%) as compared with late (4.5%) (P < 0.05), but there was a considerable variation between individual cases (Table 2). No significant difference in cell proliferation was observed between nodular AKS and EKS, although nodular EKS tended to have a higher proliferation rate (median 10.2%) as compared with nodular AKS (median 4.5%) (Table 2). Cells belonging to the noncycling Ki67−/LANA+/CD34+ SC population (Table 2, subgroup “b”) showed a clear increase during evolution from patch/plaque (median 13.5%) to nodular (median 40.3%), as well as the Ki67−/LANA−/CD34+ SC population (Table 2, subgroup “c”). Most proliferating cells belong to the LANA+/CD34+ SC (Table 2, subgroup “e”) population, but Ki67+/LANA+/CD34− non-SC cells (Table 2, subgroup “d”) were also found particularly in patch/plaque lesions. Thus both LANA−/CD34+ SC and LANA−/CD34− cells are represented in the proliferating cell pool. However, the proliferation rate among infected SCs (Table 2, subgroup “e”) was significantly higher than noninfected SCs (Table 2, subgroup “f”) (P < 0.05). LANA+/CD34+ cycling cells (Table 2, subgroup “e”) appeared to be more frequent in EKS (median 4.4%) as compared with AKS (median 1.5%), although it is not statistically significant (P > 0.05). A relatively small population of LANA+/CD34− cells either Ki67+ or Ki67− (Table 2, subgroups “a” and “d”) was observed in all KS lesions and appeared to be more frequent in the early patch/plaque stages (median 6.5%) as compared with the nodular stage (median 4.1%).
Considering the possible phenotype of the observed LANA+/CD34− non-SC cells, triple antibody labeling for LANA, CD68, and CD3 was performed on the various KS lesions (Table 3). A median of 7.3% CD68+ cells was observed in nodular AKS, 4.9% in nodular EKS, and 8.8% in patch/plaque AKS. Similarly, a median of 5.2% CD3+ cells was found in nodular AKS, 3.1% in nodular EKS, and 3.2% in patch/plaque AKS. Very few (0.5%) CD68+ macrophages and no CD3+ cells were found to be positive for LANA (Table 3, Fig. 1D). Double immunostaining of the KS sections with LANA and CD20 or CD45 showed that no CD45+ or CD20+ cells were LANA+, but up to 10% of the cells were CD45+, and 2% were CD20+. Most CD20+ leukocytes were located to the borders of the tumor area, around vascular endothelium in both early and late KS lesions.
Our findings that HHV-8-infected LANA+ SCs increase from early patch to late KS nodular lesions of AKS are consistent with previous observation of increased cell survival due to expression of Bcl-2 and apoptosis inhibitory viral gene. 25,27,29,30 The observed increase in SCs appears thus to reflect the accumulation of HHV-8-infected cells in late lesions. Correspondingly, the lower frequency of LANA+ SCs in early lesions seems to correlate with an observed higher apoptotic activity. 25,27,29,30 These and other findings are apparently most compatible with a bimodal model of KS tumor development from an early (patch/plaque) reactive lesion to a late (nodular) progressively “clonal” tumor cell proliferation. 17,18,21,22,24,26 The origin and clonality of the SCs in KS are controversial 21,23 although endothelial-derived cells have been proposed from finding of infected VELCs, as seen in this study and previous studies. 33 However, a relatively constant restriction in viral infected SC (LANA−/CD34+) was also noted during tumor evolution. This apparent heterogeneity in viral permissiveness of SCs seems less compatible with a clonal transfer of virus and may alternatively indicate that a fraction of noninfected CD34+ SCs could be continuously recruited from progenitor cells and locally triggered to develop permissiveness to HHV-8 infection. Our findings of comparable LANA expression in EKS and AKS are consistent with the notion of basically similar pathogenic mechanisms for these clinically distinct lesions. 22,27,28 Interestingly, we also found in all KS forms and stages a minor population of LANA+ non-SC (CD34−) cells usually more frequent in early as compared with late lesions, and these cells did not express a leukocytic phenotype and may represent CD34− progenitor cells. However, the precise phenotype of these cells is yet to be ascertained.
Immunostaining for cycling cells (Ki67+) showed, as previously observed, 22,27,28 comparable rates for AKS and EKS, but significantly higher in early KS stages as compared with late. Due to a considerable individual variation in both early- and late-stage biopsies, no clear significant difference in proliferation rate was found between AKS and EKS, although nodular EKS appeared to have a tendency for higher proliferation as compared with nodular AKS. Interestingly, a statistically significant difference in proliferation was found between LANA+ and LANA− SCs, which suggests that LANA+ SCs could have a proliferative advantage. 34 The observed proliferating CD34− cells were either LANA+ or LANA− and were more frequent in the early stages. Thus, as previously described, 27 cell proliferation was found in different cell populations as SCs, endothelial cells, and probably other cells in all KS lesions. Considering that AKS is clinically more aggressive and more often becomes disseminated than EKS, the observed data of comparable proliferation rates in AKS and EKS appear paradoxical. These findings could therefore indicate that clinical aggressiveness in AKS may reflect a higher rate of SC progenitor recruitment than in EKS lesions, as suggested by a nonproliferative increase in HHV-8-infected cells during KS (AKS) evolution to nodular stage.
As also previously found, 15,16 a minor population of cells in the examined KS lesions had a myelomonocytic immunophenotype. Previously some circulating monocytes and macrophages have been shown to be HHV-8 infected and suggested to function as a route of transmission. 13,14 However, in this study only a very small number of macrophages and no T cells or B cells appeared to be infected (LANA+) and would therefore less likely account for HHV-8 transmission within the KS lesions. Also, few VLECs were found to be infected and they are less likely to represent a source of infection. The mechanism of primary HHV-8 infection of SCs is therefore still unclear.
In summary, our observations indicate that HHV-8 in AKS and EKS biopsies is associated predominantly with SCs, but also with other cell types. A relative large fraction of SCs did not express LANA, suggesting heterogeneity among SCs for virus permissiveness and possibly a continuous recruitment of noninfected progenitor SCs to the lesions, particularly in aggressive AKS. A minor population of LANA+ cells is CD34− and probably not SCs and needs to be further characterized. Cell proliferation was higher in early as compared with the late KS stages, which is consistent with the notion that KS tumors develop from a predominantly reactive early lesion. Few macrophages are LANA+ but not lymphocytes, suggesting that these cells are a less important vector for transmission of HHV-8 to the KS lesions.
Drs. L. Lema, J. Luande, and B. Kalyanyama of the Muhimbili University College, Dar-Es-Salaam, Tanzania, provided the biopsies. The technical assistance of Mariane Ekman, Angelika Magogo, and Vera Nelson is highly appreciated.
1. Amir H, Kaaya EE, Manji KP, et al. Kaposi's sarcoma before and during a human immunodeficiency virus epidemic in Tanzanian children. Pediatr Infect Dis J
2. Ensoli B, Cafaro A. HIV-1 and Kaposi's sarcoma. Eur J Cancer Prev
3. Boshoff C, Chang Y. Kaposi's sarcoma-associated herpesvirus: a new DNA tumor virus. Annu Rev Med
4. Lessan-Pezeshki M, Einollahi B, Khatami MR, et al. Kidney transplantation and Kaposi's sarcoma: review of 2050 recipients. Transplant Proc
5. Schulz TF. Epidemiology of Kaposi's sarcoma-associated herpesvirus/human herpesvirus 8. Adv Cancer Res
6. Chang Y, Cesarman E, Pessin MS, et al. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi's sarcoma. Science
7. Dupin N, Fisher C, Kellam P, et al. Distribution of human herpesvirus-8 latently infected cells in Kaposi's sarcoma, multicentric Castleman's disease, and primary effusion lymphoma. Proc Natl Acad Sci U S A
8. Hall KT, Giles MS, Goodwin DJ, et al. Characterization of the herpesvirus saimiri ORF73 gene product. J Gen Virol
9. Kanitakis J, Narvaez D, Claudy A. Expression of the CD34 antigen distinguishes Kaposi's sarcoma from pseudo-Kaposi's sarcoma (acroangio-dermatitis). Br J Dermatol
10. Olsen SJ, Sarid R, Chang Y, et al. Evaluation of the latency-associated nuclear antigen (ORF73) of Kaposi's sarcoma-associated herpesvirus by peptide mapping and bacterially expressed recombinant western blot assay. J Infect Dis
11. Renne R, Barry C, Dittmer D, et al. Modulation of cellular and viral gene expression by the latency-associated nuclear antigen of Kaposi's sarcoma-associated herpesvirus. J Virol
12. Cesarman E, Chang Y, Moore PS, et al. Kaposi's sarcoma-associated herpesvirus-like DNA sequences in AIDS-related body-cavity-based lymphomas. N Engl J Med
13. Blasig C, Zietz C, Haar B, et al. Monocytes in Kaposi's sarcoma lesions are productively infected by human herpesvirus 8. J Virol
14. Monini P, Colombini S, Sturzl M, et al. Reactivation and persistence of human herpesvirus-8 infection in B cells and monocytes by Th-1 cytokines increased in Kaposi's sarcoma. Blood
15. Uccini S, Ruco LP, Monardo F, et al. Co-expression of endothelial cell and macrophage antigens in Kaposi's sarcoma cells. J Pathol
16. Uccini S, Sirianni MC, Vincenzi L, et al. Kaposi's sarcoma cells express the macrophage-associated antigen mannose receptor and develop in peripheral blood cultures of Kaposi's sarcoma patients. Am J Pathol
17. Henry M, Uthman A, Geusau A, et al. Infection of circulating CD34+ cells by HHV-8 in patients with Kaposi's sarcoma. J Invest Dermatol
18. Nickoloff BJ. The human progenitor cell antigen (CD34) is localized on endothelial cells, dermal dendritic cells, and perifollicular cells in formalin-fixed normal skin, and on proliferating endothelial cells and stromal spindle-shaped cells in Kaposi's sarcoma. Arch Dermatol
19. Russell Jones R, Orchard G, Zelger B, et al. Immunostaining for CD31 and CD34 in Kaposi sarcoma. J Clin Pathol
20. Traweek ST, Kandalaft PL, Mehta P, et al. The human hematopoietic progenitor cell antigen (CD34) in vascular neoplasia. Am J Clin Pathol
21. Roth WK, Brandstetter H, Sturzl M. Cellular and molecular features of HIV-associated Kaposi's sarcoma. AIDS
22. Kaaya EE, Parravicini C, Ordonez C, et al. Heterogeneity of spindle cells in Kaposi's sarcoma: comparison of cells in lesions and in culture. J Acquir Immune Defic Syndr Hum Retrovirol
23. Delabesse E, Oksenhendler E, Lebbe C, et al. Molecular analysis of clonality in Kaposi's sarcoma. J Clin Pathol
24. Ensoli B, Sturzl M. Kaposi's sarcoma: a result of the interplay among inflammatory cytokines, angiogenic factors and viral agents. Cytokine Growth Factor Rev
25. Hayward GS. Human herpesvirus 8 latent-state gene expression and apoptosis in Kaposi's sarcoma lesions. J Natl Cancer Inst
26. Judde JG, Lacoste V, Briere J, et al. Monoclonality or oligoclonality of human herpesvirus 8 terminal repeat sequences in Kaposi's sarcoma and other diseases. J Natl Cancer Inst
27. Kaaya E, Castanos-Velez E, Heiden T, et al. Proliferation and apoptosis in the evolution of endemic and acquired immunodeficiency syndrome-related Kaposi's sarcoma. Med Oncol
28. Kaaya EE, Parravicini C, Sundelin B, et al. Spindle cell ploidy and proliferation in endemic and epidemic African Kaposi's sarcoma. Eur J Cancer
29. Simonart T, Degraef C, Noel JC, et al. Overexpression of Bcl-2 in Kaposi's sarcoma-derived cells. J Invest Dermatol
30. Sturzl M, Hohenadl C, Zietz C, et al. Expression of K13/v-FLIP gene of human herpesvirus 8 and apoptosis in Kaposi's sarcoma spindle cells. J Natl Cancer Inst
31. Florijn RJ, Slats J, Tanke HJ, et al. Analysis of antifading reagents for fluorescence microscopy. Cytometry
32. Sanna PP, Jirikowski GF, Lewandowski GA, et al. Applications of DAPI cytochemistry to neurobiology. Biotech Histochem
33. Sturzl M, Blasig C, Schreier A, et al. Expression of HHV-8 latency-associated T0.7 RNA in spindle cells and endothelial cells of AIDS-associated, classical and African Kaposi's sarcoma. Int J Cancer
34. Friborg J Jr, Kong W, Hottiger MO, et al. p53 inhibition by the LANA protein of KSHV protects against cell death. Nature