The CD8 and CD4 T-Cell Response against Kaposi���s Sarcoma-Associated Herpesvirus Is Skewed Towards Early and Late Lytic Antigens Rebecca C. Robey1,2, Dimitrios Lagos2, Fiona Gratrix2, Stephen Henderson2, Nick C. Matthews1, Richard J. Vart2, Mark Bower1, Chris Boshoff2, Frances M. Gotch1* 1 Department of Immunology, Imperial College London, London, United Kingdom, 2 UCL Cancer Institute, University College London, London, United Kingdom Abstract Kaposi���s sarcoma-associated herpesvirus (KSHV) is causally related to Kaposi���s sarcoma (KS), the most common malignancy in untreated individuals with HIV/AIDS. The adaptive T-cell immune response against KSHV has not been fully characterized. To achieve a better understanding of the antigenic repertoire of the CD8 and CD4 T-cell responses against KSHV, we constructed a library of lentiviral expression vectors each coding for one of 31 individual KSHV open reading frames (ORFs). We used these to transduce monocyte-derived dendritic cells (moDCs) isolated from 14 KSHV-seropositive (12 HIV-positive) and 7 KSHV-seronegative (4 HIV-positive) individuals. moDCs were transduced with up to 3 KSHV ORFs simultaneously (ORFs grouped according to their expression during the viral life cycle). Transduced moDCs naturally process the KSHV genes and present the resulting antigens in the context of MHC class I and II. Transduced moDCs were cultured with purified autologous T cells and the CD8 and CD4 T-cell proliferative responses to each KSHV ORF (or group) was assessed using a CFSE dye-based assay. Two pools of early lytic KSHV genes ([ORF8/ORF49/ORF61] and [ORF59/ORF65/K4.1]) were frequently-recognized targets of both CD8 and CD4 T cells from KSHV seropositive individuals. One pool of late lytic KSHV genes ([ORF28/ORF36/ORF37]) was a frequently-recognized CD8 target and another pool of late genes ([ORF33/K1/K8.1]) was a frequently-recognized CD4 target. We report that both the CD8 and CD4 T-cell responses against KSHV are skewed towards genes expressed in the early and late phases of the viral lytic cycle, and identify some previously unknown targets of these responses. This knowledge will be important to future immunological investigations into KSHV and may eventually lead to the development of better immunotherapies for KSHV-related diseases. Citation: Robey RC, Lagos D, Gratrix F, Henderson S, Matthews NC, et al. (2009) The CD8 and CD4 T-Cell Response against Kaposi���s Sarcoma-Associated Herpesvirus Is Skewed Towards Early and Late Lytic Antigens. PLoS ONE 4(6): e5890. doi:10.1371/journal.pone.0005890 Editor: Douglas F. Nixon, University of California San Francisco, United States of America Received April 24, 2009 Accepted April 30, 2009 Published June 17, 2009 Copyright: �� 2009 Robey et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by grant G0800168 from the Medical Research Council (http://www.mrc.ac.uk) and grant JRC 08/09 SG 006 from the Westminster Medical School Research Trust and the Chelsea and Westminster Health Charity (http://www.chelwestcharity.org.uk). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: f.gotch@imperial.ac.uk Introduction Kaposi���s sarcoma-associated herpesvirus (KSHV also known as human herpesvirus 8 [HHV-8]) is the etiological agent of Kaposi���s sarcoma (KS), the most frequently-arising malignancy in untreated individuals with HIV/AIDS [1] and consequently one of the most common cancers in Sub-Saharan Africa [2]. KSHV is also involved in the pathogenesis of at least two lymphoproliferative disorders, primary effusion lymphoma (PEL) [3] and multicentric Castleman���s disease (MCD) [4]. In immunocompetent individuals KSHV can establish life-long, asymptomatic infection. However, when immune control declines (for example, during AIDS) KSHV-related tumors may develop. KS is over 100 times more common in HIV-infected individuals than in immunocompetent individuals [1]. Moreover, spontaneous tumor regression is seen in individuals with KS when immuno- suppression is reversed through highly-active antiretroviral therapy (HAART) [5], and this has been shown to correlate with a quantitative increase in KSHV-specific CD8 T-cell responses [6��� 8]. Likewise, KSHV-specific CD8 responses have been found to be of higher frequency and with greater diversity in their antigenic repertoire in asymptomatic carriers of KSHV compared to individuals with KS [9,10]. Longitudinal studies of two individuals with KS found a correlation between reduced levels of KSHV- specific CD8 T cells and recurrence of active KS [10,11]. Together, these findings infer that KS oncogenesis is associated with loss of T cell-mediated control of KSHV-infected cells. T-cell responses have been detected against several lytic and latent KSHV proteins [8���10,12���19]. Some of these responses have been shown to be functionally cytotoxic in vitro [15,18] and there is evidence that they exert evolutionary pressure on the virus in vivo [17]. A few KSHV-specific T-cell epitopes have been identified [8���10,12���14,16,17,20] but these are almost exclusively CD8 epitopes and they elicit weak responses compared to epitopes from other viruses such as HIV-1 and EBV [12,13], indicating that there may be immunodominant epitopes yet to be determined. There has been a limited number of investigations into the CD4 T-cell response against KSHV: one group reported the identifi- cation of two CD4 T-cell epitopes in K12 and K15 in one individual with AIDS-KS [9]. Thus, neither the breadth of the antigenic repertoire of the KSHV-specific T-cell immune response, nor its immunodominant targets, are fully understood. PLoS ONE | www.plosone.org 1 June 2009 | Volume 4 | Issue 6 | e5890

In previous studies, a necessary limiting factor has been the size and complexity of the KSHV genome. Each study has been confined to analysis of a handful of genes, selected according to their homology with immunogenic genes from other c-herpesvi- ruses [10] their expression profile [14,16] or evidence of sequence variation arising from immunological pressure [17]. Epitope identification has been performed using overlapping peptides for smaller genes [8,9,12,17] or predictive algorithms for peptide HLA-binding affinity for larger genes [9,13,14,16,20]. This study further investigates the immunogenic profile of KSHV using monocyte-derived dendritic cells (moDCs) lentivi- rally-transduced to express 31 different KSHV open reading frames (ORFs) to perform a large-scale screen for immunogenicity. Lentiviral vectors efficiently deliver foreign genetic material into non-dividing cells such as moDCs [21] and integrate into the cellular genome resulting in sustained transgene expression [21��� 23]. Using overlapping peptides for this number of gene products would be impractical, and even the use of epitope prediction software (Immune Epitope Database www.immuneepitope.org. uk) yields over 1000 potential nine-mer epitopes (IC50 value less than 5000 nM) from these 31 genes for HLA-A*0201 alone (unpublished data). Lentiviral delivery of an antigenic gene into moDCs allows the moDCs to present the naturally-occurring optimal CD8 and CD4 epitopes, thereby avoiding the limitations associated with the use of pre-determined peptides. No prior knowledge of the optimal peptide or HLA-restriction is required [24,25]. Lentiviral-transduced moDCs have been used to stimulate both primary and recall antigen-specific T-cell responses in vitro [23,26���29]. Such moDCs efficiently process and present both MHC class I- and II-restricted epitopes and thereby prime both CD8+ and CD4+ T cell responses [30,31]. One group has reported the use of lentiviral-transduced moDCs expressing a melanoma-specific antigenic protein to isolate an anti-melanoma CD8+ T cell that recognized a previously unknown peptide [32]. Using this system, we report that both the CD8 and CD4 T cell responses against KSHV are directed predominantly towards genes expressed in the early and late lytic phases of the viral life cycle, and we have identified novel immunogenic targets for future investigations into host immune control of KSHV infection. Results moDCs can be lentivirally transduced to stably express GFP or KSHV ORFs The green fluorescence protein (GFP)-encoding lentivirus pCSGW (the vector from which our pSIN vector was derived) was used to demonstrate that moDCs can be successfully transduced using our lentiviral vector and to investigate the kinetics of transgene expression and the optimal multiplicity of infection (MOI) for our vector. Transduction with pCSGW resulted in GFP expression by moDCs (Figure 1a). A time-course experiment revealed that GFP transgene expression increased steadily over several days (data not shown), as has been reported by other groups [21,23]. A six-day period was selected as a suitable length of time for culture of transduced moDCs, representing a balance between obtaining good transgene expression and optimal moDC viability. After six days, a transduction efficiency of 12.262.5% (mean6s.d.) GFP-positive immature moDCs was observed (n=3) (Figure 1b). Interestingly, a downregulation of GFP transgene expression was observed after maturation of moDCs (7.561.7% GFP-positive cells). A titration experiment was performed to assess the optimal multiplicity of infection (MOI) for the vector (data not shown) and a target MOI of between 3 and 8 for each transduction was selected. With our GFP construct this achieved good transgene expression (between 11 and 15% GFP-positive immature moDCs, data not shown), with no notable improvement if the MOI was increased above 8. Furthermore, whilst there is a consensus that lentiviral transduction of moDCs at MOIs of less than 10 does not affect moDC viability, immunophenotype or antigen- presenting function [23,30,33,34], the evidence regarding transduc- tion with higher MOIs is less clear [35]. We used RT-PCR to ensure that all KSHV ORFs were expressed by moDCs after lentiviral transduction (Figure 2a) and quantitative PCR to titre all KSHV-gene-encoding lentivirus preparations (Figure 2b). The volume of each lentivirus used in all subsequent experiments was then adjusted to achieve a uniform MOI of between 3.4 and 7.2 lentiviral copies per cell for all preparations (median 4.5 interquartile range 3.9 to 5.4 mean 4.75). This range was selected based on the results from experiments with our GFP construct discussed above. As the KSHV lentiviral library consists of 31 KSHV ORFs we decided initially to perform our immunogenic screen with moDCs transduced with up to three different KSHV ORFs simultaneous- ly, in order to make the experiments more manageable and to make the best use of clinical samples. KSHV ORFs were grouped according to their expression profile to determine whether latent, immediate-early, early or late lytic gene products elicit the strongest T-cell responses. The classification of KSHV ORFs was based on their expression in PEL cells [36,37]. We performed a multiple transduction experiment and used RT-PCR to demonstrate that moDCs can be transduced to express one, two or three KSHV ORFs simultaneously (Figure 1c). Lentiviral transduction does not affect moDC���s antigen- presenting surface phenotype or moDC maturation We used GFP-transduced moDCs to examine MHC-I and CD80 surface expression by non-transduced and lentiviral- transduced moDCs. There was no difference between the MHC-I or CD80 mean fluorescence intensities (MFIs) of non- transduced or transduced immature moDCs, and these markers were equally upregulated by non-transduced and transduced moDCs after exposure to maturation stimuli (Figure 1d and 1e). This indicates that lentiviral transduction does not affect the antigen-presenting surface phenotype of moDCs or moDC maturation. T-cell proliferation responses to moDCs transduced to express KSHV genes moDCs were isolated from 14 KSHV seropositive (12 HIV+) individuals and 7 KSHV-seronegative (4 HIV+) individuals. Characteristics of all study participants are shown in Table 1. moDCs were transduced to express up to three KSHV ORFs (grouped according to their expression profile) and then cultured with autologous CFSE-stained T cells. KSHV ORFs known to affect MHC-I expression (K3 and K5 [38] and K9 and ORF 71 [39]) were used to singly transduce moDCs, since these genes��� function may affect T-cell priming by moDCs thus skewing the results. After six days, T cells were harvested and flow cytometry was used to assess the CD8-positive cytotoxic lymphocyte (CTL) response and the CD8-negative (CD4) helper T-cell response to each KSHV ORF or pool, as measured by the proportion of CFSE-low proliferating cells (Figure 3a). An example of CD8 and CD4 responses by one HIV-positive, KSHV-seropositive individ- ual (P7) are shown in Figures 3b and 3c, respectively. We used strict criteria to designate positive and borderline positive responses (see Materials and Methods) as we wished to ensure that no false positives were recorded as a result of the slight variation we observed in individuals��� background response to T-Cell Response against KSHV PLoS ONE | www.plosone.org 2 June 2009 | Volume 4 | Issue 6 | e5890