Glycoprotein K8.1A of Kaposi's Sarcoma-Associated Herpesvirus Is a Critical B Cell Tropism Determinant Independent of Its Heparan Sulfate Binding Activity

Elucidating the molecular mechanisms by which KSHV infects B lymphocytes is critical for understanding how the virus establishes lifelong persistence in infected people, in whom it can cause life-threatening B cell lymphoproliferative disease. Here, we show that K8.1A, a KSHV-encoded glycoprotein on the surfaces of the virus particles, is critical for infection of B cells. This finding stands in marked contrast to previous studies with non-B lymphoid cell types, for which K8.1A is known to be dispensable. We also show that the required function of K8.1A in B cell infection does not involve its binding to cell surface heparan sulfate, the only known biochemical activity of the glycoprotein. The discovery of this critical role of K8.1A in KSHV B cell tropism opens promising new avenues to unravel the complex mechanisms underlying infection and disease caused by this viral human pathogen.

ABSTRACT

B lymphocytes are the major cellular reservoir in individuals infected with Kaposi’s sarcoma-associated herpesvirus (KSHV), and the virus is etiologically linked to two B cell lymphoproliferative disorders. We previously described the MC116 human B cell line as a KSHV-susceptible model to overcome the paradoxical refractoriness of B cell lines to experimental KSHV infection. Here, using monoclonal antibody inhibition and a deletion mutant virus, we demonstrate that the KSHV virion glycoprotein K8.1A is critical for infection of MC116, as well as tonsillar B cells; in contrast, we confirm previous reports on the dispensability of the glycoprotein for infection of primary endothelial cells and other commonly studied non-B cell targets. Surprisingly, we found that the role of K8.1A in B cell infection is independent of its only known biochemical activity of binding to surface heparan sulfate, suggesting the possible involvement of an additional molecular interaction(s). Our finding that K8.1A is a critical determinant for KSHV B cell tropism parallels the importance of proteins encoded by positionally homologous genes for the cell tropism of other gammaherpesviruses.

IMPORTANCE Elucidating the molecular mechanisms by which KSHV infects B lymphocytes is critical for understanding how the virus establishes lifelong persistence in infected people, in whom it can cause life-threatening B cell lymphoproliferative disease. Here, we show that K8.1A, a KSHV-encoded glycoprotein on the surfaces of the virus particles, is critical for infection of B cells. This finding stands in marked contrast to previous studies with non-B lymphoid cell types, for which K8.1A is known to be dispensable. We also show that the required function of K8.1A in B cell infection does not involve its binding to cell surface heparan sulfate, the only known biochemical activity of the glycoprotein. The discovery of this critical role of K8.1A in KSHV B cell tropism opens promising new avenues to unravel the complex mechanisms underlying infection and disease caused by this viral human pathogen.

INTRODUCTION

Kaposi’s sarcoma-associated herpesvirus (KSHV) (also known as human herpesvirus 8) is a lymphotropic gammaherpesvirus that is the causative agent of not only the endothelial neoplasm for which it is named, but also the two B cell lymphoproliferative disorders multicentric Castleman’s disease (MCD) and primary effusion lymphoma (PEL) (1–4). As for other human and nonhuman gammaherpesviruses (5–7), B lymphocytes are the major KSHV reservoir during lifelong viral persistence in the infected host and are a likely origin of reactivation, shedding, and dissemination of the virus (8, 9). Thus, the mechanism(s) of B cell infection is of critical importance for understanding the natural biology of KSHV, as well as the pathogenic mechanisms of KSHV-mediated diseases.

The investigation of KSHV infection of B cells has been greatly hampered by limitations of experimental systems, most notably the refractoriness of established B cell lines to infection by cell-free KSHV (10–14). We previously reported a human B lymphoma cell line, MC116, as unusually susceptible to KSHV cell-free infection (15), thus offering a potential model in which to study molecular mechanisms governing the early steps in KSHV infection of B cells. As a major focus, we sought to identify critical determinants among the KSHV-encoded virion glycoproteins associated with B cell tropism. Here, we describe a critical tropism function for a specific KSHV glycoprotein previously deemed to be dispensable for infection of non-B cell types.

The viral K8.1 gene gives rise by alternative RNA splicing to two open reading frames (ORFs) encoding the highly immunogenic glycoproteins K8.1A and K8.1B (16). K8.1A has long been recognized as a prominent component of the KSHV virion envelope (17–20). The early demonstrations of its binding activity to cell surface heparan sulfate (HS) (21, 22) engendered the notion that the glycoprotein contributes to the virion-target cell attachment steps associated with efficient entry/infection (reviewed in reference 23), analogous to the HS-dependent interactions mediating entry of diverse herpesviruses (24, 25). Perhaps surprisingly, K8.1A was found to be dispensable for KSHV entry/infection based on findings with endothelial cells (21) and the 293 cell line (26), a highly susceptible KSHV target commonly used for infectivity studies (10, 27–31). In the present report, we identify a critical role for K8.1A in B cell infection while confirming the previous findings of its dispensability for entry into non-B cell types. In addition, we show that the K8.1A requirement for KSHV B cell tropism is independent of the only known activity of the glycoprotein, i.e., binding to HS.

RESULTS

Critical role of the KSHV K8.1A glycoprotein in infection of the MC116 B cell line.

As one approach to test the involvement of K8.1A in MC116 cell infection, we examined the effects on KSHV infection of previously reported anti-K8.1A monoclonal antibodies (MAbs) (18). We employed rKSHV.219, a recombinant virus containing the enhanced green fluorescent protein (eGFP) gene linked to a strong constitutive cellular promoter (30). We had previously demonstrated the suitability of the eGFP reporter as an indicator of MC116 cell acute infection by virtue of its correlation with KSHV latent gene expression (15). For comparison, Vero (African green monkey kidney) cells and 293F (human embryonic kidney) cells were included during the initial analysis. 293 cells have previously been shown not to require K8.1A during initial infection (26). In order to expose virion K8.1A to antibody, cell-free virus was pretreated for 2 h at 4°C with a range of concentrations (0 to 10 μg/ml) of the appropriate MAb and then added to the above-mentioned cell types. After 3 days in culture at 37°C, cells were harvested and infection was quantitated by flow cytometric measurement of eGFP-positive cells.

The data in Fig. 1 indicate that two anti-K8.1 MAbs (4C3 and 4A4, both mouse IgG1) potently neutralized MC116 cell infection, whereas a mouse IgG1 isotype control had minimal effect (Fig. 1, top). This potent activity against the MC116 B cell line was consistently observed (i.e., in 10 experiments, high concentrations of 4C3 produced 83 to 99% inhibition of MC116 cell infection). In marked contrast, the anti-K8.1 MAbs over the same dose range displayed negligible activity against Vero and 293F cells (Fig. 1, middle and bottom), cell lines commonly used as KSHV targets.

FIG 1.

Effects of anti-K8.1 MAbs on KSHV infection of different cell types. Dilutions of rKSHV.219 stock (titer, 1.5 × 10⁶ IU/ml) were preincubated for 2 h at 4°C with the indicated murine anti-K8.1A MAbs (4C3, 4A4, or isotype control [ISO]) at the indicated concentrations. In duplicate wells of a 24-well plate, the virus-MAb mixtures (0.2 ml) were added to the indicated cell types (MC116, Vero, or 293F, each plated at 10⁵ cells/well in 0.5 ml). The volumes of virus stock per well corresponded to 10 μl for MC116 cells, 2 μl for Vero cells, and 1 μl for 293F cells. The cultures were incubated at 37°C for 3 days. Infection was assessed by flow cytometry and is depicted as percentages of cells infected (eGFP positive). In each graph, the data indicate the means of duplicate samples ± standard errors of the mean (SEM); significance refers to comparisons of the 4C3 and 4A4 MAbs to the isolype control at 10 μg/ml. The results shown are representative of experiments repeated >10 times for MC116 cells and at least 3 times for Vero and 293F cells. NS, not significant; ****, P < 0.0001.

Role of K8.1A in KSHV infection of human primary cell types relevant to viral pathogenesis.

We next examined the importance of the K8.1A glycoprotein for infection of human primary cell types most relevant to the KSHV-associated pathologies, namely, B lymphocytes for primary effusion lymphoma and multicentric Castleman’s disease and endothelial cells for Kaposi’s sarcoma (2, 32). We performed inhibition studies in primary B cells from explanted tonsil tissue. Tonsillar cultures have been actively studied for B cell infection by KSHV (14, 33–35). As shown in Fig. 2A, the 4C3 and 4A4 MAbs showed strong neutralization in tonsillar B cells from two different donors (∼60 to 80%) over the concentration range in which MC116 cells, tested in parallel (Fig. 2B), were fully neutralized. Thus, the K8.1A involvement observed for the MC116 B cell line also holds for a large fraction of infection of tonsil-derived B cells.

FIG 2.

Role of K8.1A in KSHV infection of primary B cells from human tonsil tissue. Dilutions of rKSHV.219 stock (titer, 0.6 × 10⁶ IU/ml) were preincubated for 2 h at 4°C with the 4C3, 4A4, or isotype (ISO) MAb at the indicated concentrations. In duplicate wells of a 24-well plate, the virus-MAb mixtures (0.2 ml) were added to 0.5 ml containing tonsillar aggregate lymphocyte cultures from donors 1 and 2 (10⁶ cells/well) (A) or MC116 cells (10⁵ cells/well) (B). The volume of rKSHV.219 stock per well corresponded to 10 μl for both the tonsillar aggregate lymphocyte cultures and the MC116 cells. The cultures were incubated at 37°C for 3 days and analyzed for infection as for Fig. 1. For the aggregate lymphocyte cultures, B cell infection was assessed by quantitation of CD20⁺ eGFP-positive cells. The data represent the means of duplicate samples ± SEM; significance refers to comparisons of the 4C3 and 4A4 MAbs to the isotype control at 10 μg/ml. The results shown are representative of experiments repeated up to 5 times for tonsillar B cells and >10 times for MC116 cells. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

To confirm previous reports arguing against a role for K8.1A in initial infection of human microvascular endothelial cells (HMVEC) (21), we performed antibody inhibition analysis in HMVEC and human umbilical vein endothelial cells (HUVEC). The 4C3 MAb had minimal effect against both primary endothelial cell types (Fig. 3A), while again strongly blocking MC116 cell infection run in parallel (Fig. 3B).

FIG 3.

Role of K8.1A in KSHV infection of primary endothelial cells. Dilutions of rKSHV.219 stock (titer, 3.5 × 10⁶ IU/ml) were preincubated for 2 h at 4°C with MAb 4C3 at the indicated concentrations. In duplicate wells of a 24-well plate, the virus-MAb mixtures (0.2 ml) were then added to endothelial cells (A) or MC116 cells (0.5 ml; 10⁵ cells/well) (B) and analyzed at 3 days postexposure. The volumes of rKSHV.219 per well corresponded to 10 μl for MC116 cells, 2 μl for HUVEC, and 1 μl for HMVEC. Infection was assessed by quantitation of eGFP-positive cells. The data are presented as percentages of cells infected (eGFP⁺) within the total population of cells. In each graph, the data indicate the means of duplicate samples ± SEM; in panel B, significance refers to infection at 10 μg/ml compared to no antibody. The results shown are representative of experiments repeated up to 5 times each for HUVEC and HMVEC and >10 times for MC116 cells. NS, not significant; **, P < 0.01.

These results confirm that K8.1A is not required for infection of endothelial cells.

Cell-type-dependent effects observed with a K8.1 deletion mutant virus.

As a complementary approach to test the selective requirement for K8.1A in KSHV infection of B cells but not endothelial cells, we analyzed two KSHV recombinant viruses generated from the corresponding genomes cloned into a bacterial artificial chromosome (BAC), i.e., wild-type (WT) KSHV BAC36 (29) and the deletion mutant BAC36ΔK8.1, in which the majority of the K8.1 open reading frame has been replaced with a kanamycin resistance marker (26). Western blot analysis of infected 293F cells (using the 4C3 MAb, as previously reported [18]) verified that the mutant displayed selective loss of the K8.1A glycoprotein, but not the gH glycoprotein (Fig. 4A). Infectivity assays revealed markedly different effects of the deletion on a B cell versus a non-B cell target, based on analysis of eGFP expression by fluorescence microscopy (Fig. 4B) and flow cytometry (Fig. 4C). In the experiment shown in Fig. 4C, the mutant virus displayed infectivity (85%) comparable to that of the wild-type virus in 293F cells, consistent with the reported dispensability of K8.1 for infection of this epithelial cell line (26). The results were markedly different for the MC116 B cell line, for which the mutant virus produced only 5% infectivity compared to the wild type.

FIG 4.

Effects of the K8.1 deletion mutation on KSHV infection of different cell types. The wild-type and K8.1 deletion mutant viruses were analyzed for specific glycoprotein expression (by Western blotting) and infectivity on different target cells (by fluorescence microscopy and flow cytometry). For infectivity assays, 293F and MC116 cells were challenged with identical inputs of either BAC36 (WT) or BAC36ΔK8.1 (deletion mutant). For each virus, the input dose was chosen to yield approximately 5% infection of 293F cells, based on previous titrations with 293F cells; the volumes corresponded to 60 μl for BAC36 and 30 μl for BAC36ΔK8.1. The cultures were maintained at 37°C for 3 days and analyzed for infection based on eGFP expression. (A) Western blot analysis. Cell lysates from sodium butyrate-induced 293F cells containing KSHV BAC36 (WT) or BAC36ΔK8.1 were subjected to Western blotting. K8.1A was probed with the 4C3 MAb, and gH was probed with MAb15. (B) Fluorescence microscopy analysis. The images are shown at ×100 magnification; the scale bar represents 100 μm. (C) Flow cytometry analysis. For each cell type, data are presented as percentages of cells infected (eGFP⁺) with each of the two viruses; the difference in the y axis scales for 293F versus MC116 reflects the inherent differences in infection susceptibilities of the cell lines, as previously reported for wild-type KSHV (15). The marked distinction between the relative effects of the K8.1 deletion on infection of 293F cells versus MC116 cells (a 17-fold difference) was consistently observed in 4 additional experiments (differences ranging from 4-fold to 17-fold).

Taken together, the immunochemical and mutational results mentioned above highlight the critical importance of the K8.1A glycoprotein selectively for KSHV infection of B cells but not other KSHV-susceptible target cell types.

Assessment of the K8.1A role in KSHV attachment to MC116 cells.

We measured KSHV attachment to cells using a quantitative-PCR (qPCR) assay that quantitates the relative numbers of viral genomes associated with cells after a 1-h incubation at 4°C, followed by incubation (1 min at 37°C) with or without trypsin to digest surface-bound virus and then extensive washing to remove unbound virus. Figure 5A shows virus binding to KSHV-permissive MC116 cells compared to other human B cell lines that are refractory to KSHV infection. For all cell types, the trypsin treatment reduced the qPCR signal to near background levels, indicating that the assay detects primarily surface-bound rather than internalized virus. Binding was comparable for the KSHV-permissive MC116 B cell line and the nonpermissive human B cell lines. Preferential virus binding to MC116 cells was not observed. We then tested the effects of the anti-K8.1A MAbs on KSHV binding to MC116 cells. Figure 5B shows that the neutralizing MAbs 4C3 and 4A4 had minimal effect on virus binding. Thus, neither the unusual KSHV permissiveness of MC116 cells compared to other human B cell lines nor the K8.1A requirement for KSHV infection of MC116 cells was manifested at the level of direct virus-cell binding.

FIG 5.

KSHV binding to human B cell lines. (A) Binding of rKSHV.219 virions to the indicated cell lines was determined by real-time PCR. The data are presented as bound KSHV copy number per 200 ng of input DNA. Samples were unexposed (− virus), exposed to KSHV-219 at 4°C for 1 h (+ virus), or equivalently exposed to KSHV-219 followed by treatment with 0.2 ml trypsin for 1 min at 25°C (+ virus, + trypsin). All the samples were then washed extensively in PBS. The relative amount of virus that remained was quantitated by real-time PCR. (B) rKSHV.219 virions were preincubated for 2 h at 4°C with bovine serum albumin (BSA) (10 μg/ml) (Control) or the indicated anti-K8.1 MAb (10 μg/ml), followed by incubation with MC116 cells at 4°C for 1 h. Samples were then washed in PBS extensively. The relative amount of virus remaining was quantitated by real-time PCR as for panel A and is presented as a percentage of the control. The data represent the means of triplicate samples ± SEM. The results shown are representative of 3 repeat experiments.

Assessment of the importance of the heparan sulfate binding activity of K8.1A for its role in MC116 cell infection.

HS binding (21, 22), the only known activity of the K8.1A glycoprotein, is thought to promote concentration of KSHV virion particles on the target cell surface, thereby facilitating virus entry (23). Previous findings demonstrated that the absence of HS on the surfaces of many human continuous B cell lines plays a major role in their refractoriness to KSHV infection (13). Therefore, we considered the possibility that the unusual KSHV permissiveness of the MC116 human B cell line might be explained by the presence of surface HS. We tested this by flow cytometry using as a probe the F58-10E4 murine IgM MAb, which reacts with an epitope on many types of native HS chains (36); this MAb has been employed as the HS probe in multiple KSHV studies, including a previous report describing the absence of surface HS on human B cell lines (13), as well as in analyses of many other viruses, including murine gammaherpesvirus 68 (MHV-68) (37–40). As verification of the HS specificity of the F58-10E4 MAb, we found that it reacted well with wild-type CHO-K1 cells (relative to two IgM isotype controls), but not with the pgsD-677 mutant, which specifically lacks HS (41) (Fig. 6A, CHO-K1 [top] and pgsD-677 [middle]). With MC116 cells, F58-10E4 reactivity was indistinguishable from that of the two IgM isotype controls (Fig. 6A, bottom). The absence of detectable HS on the surfaces of MC116 cells thus parallels the previously reported findings for other human B cell lines (13).

FIG 6.

Independence of the K8.1A requirement from its known activity of binding cell surface heparan sulfate. (A) Absence of detectable surface heparan sulfate on MC116 cells. Surface heparan sulfate was analyzed by flow cytometry on wild-type CHO-K1 cells, heparan sulfate-deficient pgsD-677 cells, and MC116 cells. Murine IgM MAb F58-10E4 was used as the primary antibody; two murine IgM MAbs, TEPC 183 (blue) and G155-228 (red), served as isotype controls. A FITC-conjugated anti-mouse IgM antibody was used as the secondary antibody; detection was performed via flow cytometry (FACSCalibur) in the FL-1 channel (FL1-H). (B) The K8.1A-dependent component of CHO cell infection does not involve heparan sulfate. rKSHV.219 virions were preincubated for 2 h at 4°C with 0, 10, or 100 μg/ml of MAb 4C3. The mixtures were then added at high input or low input (10% volume) to CHO-K1 (wild-type) or pgsD-677 (mutant) cells. Infection at 3 days postexposure was assessed by flow cytometric quantitation of eGFP-positive cells. For each cell type, the data are presented as percentages of cells infected (eGFP⁺); the y axis scales are adjusted to allow comparison of the infection profiles between the two cell types. In panel B, the data represent the means of duplicate samples ± SEM. Significance at each KSHV input refers to the fractional 4C3-mediated neutralization for pgsD-677 compared to CHO-K1 at both 10 μg/ml and 100 μg/ml. The results shown are representative of >5 repeat experiments. NS, not significant. The error bars indicate SEM.

We then examined the effects of the 4C3 MAb on KSHV infection of the CHO-K1 pgsd-677 mutant. Fig. 6B shows the effects of the antibody at a dose that fully neutralized MC116 cells (10 μg/ml [Fig. 1, 2, and 3]) and an even higher dose (100 μg/ml). The wild-type CHOK1 cells were only partially neutralized, i.e., ∼20 to 25% at a high virus input (Fig. 6B, top) and ∼45% at a lower virus input (Fig. 6B, bottom). The pgsD-677 HS-deficient mutant displayed essentially the same neutralization pattern as the wild type. Thus, the K8.1A-dependent (i.e., 4C3-sensitive) component of the CHO-based cell lines was independent of their surface display of HS chains.

DISCUSSION

The results presented here demonstrate that the K8.1A glycoprotein is selectively required for KSHV infection of B lymphocytes. Our conclusion is based on studies with the KSHV-susceptible MC116 continuous B cell line, as well as with primary B cells from human tonsils. Our confirmation of the previous reports that K8.1A is dispensable for KSHV infection of primary human endothelial cells (21) and the 293 cell line (26), coupled with our demonstration that the glycoprotein is also not required for Vero cell infection, support the conclusion that K8.1A is a major determinant governing KSHV selective tropism for B cells, the cell type associated with KSHV persistent reservoirs and the two lymphoproliferative disorders etiologically linked to the virus (1, 2, 4, 42). We note that our analyses of the early stage of KSHV infection with the MC116 cell line and nonactivated tonsillar B cells may not reflect possible variations in K8.1A involvement at later stages of infection and with B cells at diverse stages of activation and differentiation.

The present findings assume added interest when viewed in the context of other gammaherpesviruses, all of which have genomic positional homologs of K8.1. Within the rhadinovirus genus, of which KSHV is a member, the Old World primate rhadinovirus RV2 expresses a homolog designated R8.1 that is derived from a spliced transcript and is displayed on the virion surface (43–45). Also among the rhadinoviruses, gp150 of MHV-68 (46), gp180 of bovine herpesvirus 4 (BoHV-4) (47), and Orf51 of herpesvirus saimiri (48) are K8.1 positionally homologous glycoproteins on the corresponding virions. Within the lymphocryptovirus genus, for which the human pathogen Epstein-Barr virus (EBV) is the prototype, the K8.1 positional homolog is the virion glycoprotein gp350/220 (49, 50); homologous glycoproteins are also present in the lymphocryptovirus of rhesus macaques (51) and marmosets (52). Like K8.1A for KSHV, the positional homologs of several of these gammaherpesviruses have been shown to be critical determinants of cell tropism, providing an enhancing/essential function for some target cell types and dispensable/inhibitory activity for others. For the extensively studied EBV, gp350/220 is required for efficient attachment and infection of B cells (by binding to surface CD21) but inhibits infection of epithelial cells (which express little or no CD21) (53–55). Similarly, gp150 of MHV-68 (37, 56, 57) and gp180 of BoHV-4 (47, 58) are involved in complex mechanisms regulating virus tropism for B cells and epithelial cells. We note that, in addition to studies focusing on the early steps of virus attachment and entry, K8.1A of KSHV (59) and the positional homologs of other gammaherpesviruses (60) have been implicated in virion egress required for transmission.

How might K8.1A function in selectively mediating infection of B cells? Our unexpected results suggest that this activity does not involve binding to HS, the only activity reported for the glycoprotein. The ligand-binding interactions of positional homologs from some other gammaherpesviruses allow speculation about the possible alternative K8.1A binding partners involved in B cell tropism. Orf51 of the herpesvirus saimiri virion does display heparan sulfate binding (48), but this activity is minimal for MHV-68 gp150 (57) and BoHV-4 (47) and has not been described for EBV gp350/220. Instead as noted above, EBV gp350/220 binding to CD21 facilitates the attachment step of B cell entry, although it is not required for subsequent membrane fusion (61). In the absence of CD21, gp350/220 can bind to an alternate protein receptor (CD35) to mediate B cell infection (62). These findings raise the possibility that the HS-independent function of K8.1A in B cell infection might involve its interaction with an as yet unidentified specific cellular receptor, perhaps a surface protein. Indeed, the possible interaction of K8.1A with non-HS surface components has been suggested based on binding studies with a soluble construct of the glycoprotein (22).

Might previously reported KSHV receptors on B cells serve as K8.1A binding partners? As for the well-studied members of the family Herpesviridae (24, 25), exceeding complexity has emerged for the functional interactions of KSHV-encoded glycoproteins with specific cell surface molecules governing the initial stages of infection (23). Distinct classes of receptors mediate virion binding versus internalization/fusion and the downstream signaling events associated with efficient entry. Importantly, the marked differences reported for alternate target cell types highlight the need to define pathways relevant to the different KSHV-associated pathologies. For B cell infection, specific identified receptors stand out. The C-type lectin DC-SIGN was reported as a surface receptor essential for KSHV binding and endocytic infection of activated blood and tonsillar B cells (33); the activity involves DC-SIGN binding to high-mannose glycans on the viral glycoprotein gB (63). DC-SIGN is an unlikely requirement for KSHV infection of MC116 cells, since an antibody against DC-SIGN that strongly blocks infection of Raji/DC-SIGN cells has minimal effect on MC116 cell infection (S. J. Dollery and E. A. Berger, unpublished data). Another cell surface component of potential interest is the ephrin receptor (Eph) family. EphA2 was implicated first as a KSHV entry receptor on endothelial and epithelial cells (64), through its interaction with the viral gH/gL glycoproteins (64, 65). Related findings for B cell infection have been made for the rhesus rhadinovirus (RRV) with the report that gH/gL binds to multiple Eph proteins and that soluble Eph decoys strongly block entry of rhesus rhadinovirus into B cells (and endothelial cells, but not fibroblasts and epithelial cells) (66). The relationships of EphA2 and other Eph family receptors, and identification of the relevant viral binding partners, remain to be determined for KSHV infection of B cells. Thus, with no candidate binding ligands for K8.1A other than HS, a major focus for mechanistic understanding of KSHV B cell tropism involves assessing the requirement for, and establishing the identity of, potential alternate K8.1A receptors on the B cell surface.

MATERIALS AND METHODS

Cells and anti-KSHV antibodies.

The human B cell lymphoma line MC116 was obtained and propagated as previously described (15). Primary HUVEC and HMVEC (obtained from Lonza) were cultured in endothelium growth medium (Lonza). Vero cells obtained from the American Type Culture Collection (ATCC) (Manassas, VA) were cultured in Dulbecco’s modified Eagle medium (DMEM) (Life Technologies) supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. 293F cells (Freestyle 293F subclone [Invitrogen] adapted to grow as adherent cultures) were cultured in the medium used for Vero cells further supplemented with 1% nonessential amino acids. Variants of the Chinese hamster ovary cell lines CHO-K1 (wild type) and pgsD-677 (HS deficient) (41) were obtained from the ATCC. For experiments with human tonsillar B cells, human lymphocyte aggregate cultures were prepared from fresh human tonsils from explant surgeries (provided by the Cooperative Human Tissue Network). Briefly, the tonsils were minced in medium (RPMI, with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin). Lymphoid cells were isolated from the tissue by repeated mincing and washing and then passed through a 70-μm cell strainer to remove debris. The cells were used immediately or cryogenically frozen for storage; following thawing, live cells were enriched using Ficoll-Hypaque (GE Healthcare).

Hybridomas for MAbs 4C3 and 4A4, previously reported to react with K8.1A in various assay formats (18), were the generous gift of Bala Chandran (Rosalind Franklin University of Medicine and Science). The corresponding purified antibodies were produced from the hybridomas by Covance Immunology Services, Denver, PA.

KSHV recombinant viruses.

Most experiments employed the recombinant virus rKSHV.219 (30) containing the eGFP gene linked to the EF-1α promoter (active during both latent and lytic infection), the red fluorescent protein (RFP) gene linked to the KSHV lytic PAN promoter (active only during lytic infection), and the puromycin resistance gene (for selection). Cell-free virus was generated from the stably infected doxycycline-inducible cell line iSLK.219 clone 10 (67) (generously provided by Jinjong Myoung and Donald Ganem, Novartis Vaccines and Diagnostics). Briefly, the cells were induced with 0.2 μg/ml doxycycline and 1 mM sodium butyrate. Virus was harvested from culture supernatants by centrifugation and filtration to remove debris; concentrated by ultracentrifugation as previously described (15); and then resuspended in RPMI medium supplemented with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin and stored at −80°C.

BACs containing the genomes of KSHV WT (BAC36) and the K8.1 deletion mutant (BAC36ΔK8.1) (kindly provided by G. Kousoulas, Louisiana State University) were used to generate the corresponding recombinant viruses as reported previously (26). Briefly, 293F cells were transfected with BAC36 and BAC36-ΔK8.1A DNAs using Effectene (Qiagen) according to the manufacturer’s instructions. Episome-containing cells were selected by a combination of fluorescence-activated cell sorting for eGFP-positive cells (FACSAria; BD Biosciences) and hygromycin selection (the medium was the same as for 293F cells plus 100 μg/ml hygromycin). Restriction enzyme digestion patterns and DNA sequence analysis (Macrogen) around the regions of interest matched previous reports (26). Virus was induced via sodium butyrate induction, and virus in the cell supernatant was concentrated by a previously described method (68) that proved optimal for low viral yields. The titers of virus preparations from both iSLK.219 clone 10 cells and BAC-transfected 293F cells were determined on 293F cells as previously described (15).

KSHV infection and antibody inhibition assays.

Infectivity by rKSHV.219 (wild type and mutant) was quantitated by flow cytometry analysis (FACSCalibur or FACSCanto-II; BD Biosciences) of eGFP expression on live cells (based on forward and side light scatter); data were analyzed with FlowJo software (version 6.3; Tree Star). The indicated titers of KSHV stocks (in infectious units [IU] per milliliter) were calculated based on infection of MC116 target cells. Infectivity assays were performed in 24-well cell culture cluster flat-bottom plates (Costar, Corning, NY); stable cell lines and primary endothelial cells were plated at 10⁵ cells per well; human lymphocyte aggregate cultures from tonsil tissue were plated at 10⁶ total cells per well. In experiments with human lymphocyte aggregate cultures, B cells were identified based on CD20 positivity (anti-CD20-allophycocyanin [APC]; clone 2H7) and CD3 negativity (anti-CD3-peridinin chlorophyll protein [PerCP]-Cy5.5; clone UCHT1); the antibodies were purchased from BD Biosciences and used as recommended. For qualitative photographic analyses, fluorescence images were obtained using a Leica DM IRB fluorescence microscope equipped with a SPOT insight digital camera and SPOT software (version 4.5.9.3); images were processed using Adobe Photoshop (version 8.0).

For antibody inhibition assays, the virus input was optimized to be in the linear range of infection for the cell types tested. Virions were preincubated for 2 h at 4°C with the indicated anti-K8.1A MAb at the indicated concentrations; for no-antibody controls, phosphate-buffered saline (PBS) only was volumetrically matched to the MAb samples. The cell-antibody mixtures (0.2 ml) were then added to the indicated target cells (0.5 ml) in 24-well plates, and the cultures were maintained in 5% CO₂ at 37°C for 3 days.

Western blot analysis.

BAC-transfected/selected 293F cells induced with sodium butyrate were lysed in buffer containing 1% NP-40. Aliquotes were mixed with SDS sample buffer containing β-mercaptoethanol, placed in a heating block for 10 min, cooled, and centrifuged at 16,000 × g for 5 min to remove debris. Equal volumes of samples were added to a NuPAGE 4% to 12% bis-Tris gel (Life Technologies, ThermoFisher, Washington, DC) for polyacrylamide gel electrophoresis. Transfer to nitrocellulose membranes was performed with an iBlot gel transfer system. The membranes were blocked and incubated with the indicated anti-KSHV MAbs (1 μg/ml for 4C3; 4 μg/ml for MAb15). Following overnight incubation, the membranes were washed and incubated with horseradish peroxidase-conjugated goat anti-mouse antibody (Pierce/Thermofisher, Washington, DC), and detection was achieved with enhanced chemiluminescence substrate (Pierce/Thermofisher) followed by exposure to X-ray film (Kodak, Rochester, NY).

KSHV/cell binding assays.

Binding assays were performed in a fashion to parallel the infection assays. Virus samples (in 0.2 ml) were added to the indicated target cells (in 0.5 ml, preincubated on ice for 15 min) in 24-well plates. The cell-virus mixtures were gently rocked at 4°C for 1 h and then washed 5 times in PBS to remove unbound virus. Replicate samples were subject to a 1-min incubation at 37°C in the absence or presence of 0.25% trypsin and then washed 5 times. Relative amounts of bound virus were determined by quantitating the KSHV genomic DNA copy number using real-time PCR of the LANA gene, as described previously (15). Briefly, DNA was prepared using a DNeasy Blood and Tissue kit (Qiagen) according to the manufacturer’s instructions. Analyses were performed using a Step One Plus real-time PCR instrument (Applied Biosystems), a TaqMan probe to the LANA nucleic acid sequence generated by Invitrogen (Life Technologies), and the ABsolute qPCR mix ROX (Thermo Fisher, Atlanta, GA).

Staining for surface heparan sulfate.

Cell surface heparan sulfate was measured by flow cytometry using as the primary antibody 0.2 μg/ml MAb F58-10E4 (Amsbio, Cambridge, MA); as negative controls, the murine IgM MAbs TEPC 183 (purified immunoglobulin; Sigma-Aldrich) and G155-228 (anti-TNP; BD Biosciences) were employed. Fluorescein isothiocyanate (FITC)-labeled MAb DS-1 (BD Biosciences) was the second antibody used for detection.

Statistical analysis.

Statistical significance was calculated using a two-tailed unpaired t test.