Non-Antigen-Specific B-Cell Activation following Murine Gammaherpesvirus Infection Is CD4 Independent In Vitro but CD4 Dependent In Vivo

Abstract

The murine gammaherpesvirus MHV-68 multiplies in the respiratory epithelium after intranasal inoculation, then spreads to infect B cells in lymphoid germinal centers. Exposing B cells to MHV-68 in vitro caused an increase in cell size, up-regulation of the CD69 activation marker, and immunoglobulin M (IgM) production. The infectious process in vivo was also associated with increased CD69 expression on B cells in the draining lymph nodes and spleen, together with a rise in total serum Ig. However, whereas the in vitro effect on B cells was entirely T-cell independent, evidence of in vivo B-cell activation was minimal in CD4⁺ T-cell-deficient (I-A^b−/−) or CD4⁺ T-cell-depleted mice. Furthermore, the Ig present at high levels in serum was predominantly of the IgG class. Surprisingly, the titer of influenza virus-specific serum IgG in previously immunized mice fell following MHV-68 infection, suggesting that there was relatively little activation of memory B cells. Thus, CD4⁺ T cells seemed both to amplify a direct viral activation of B cells in lymphoid tissue and to promote new Ig class switching despite a lack of obvious cognate antigen.

Herpesvirus (HV) infections are often associated with non-antigen-specific B-cell activation (13, 14, 16, 21, 22). Although no definite role has been established for this process in viral pathogenesis, it is of particular interest in gammaherpesvirus (γ-HV) infections, since chronic B-cell stimulation may contribute to the oncogenesis (9, 15) associated with Epstein-Barr virus (EBV) and human herpesvirus 8 (HHV-8) infections. Infection with EBV activates B cells expressing the immunoglobulin (Ig) V_4-34 gene (4), which is also overrepresented in certain lymphomas (6, 25). EBV-activated V_4-34-expressing B cells can undergo somatic mutation and isotype switching, indicating a participation in normal germinal-center interactions (5). The latent membrane protein 1 (LMP-1) of EBV, which has intracellular signaling substrates similar to those of CD40 (12), and LMP-2A, which can trigger lymphocyte activation (2), may both contribute to this process. However, analysis of lymphocyte interactions in vivo has not been possible with the human γ-HVs.

The murine γ-HV-68 (MHV-68) is a natural γ-HV of small rodents that is related to EBV (8) and to HHV-8 (33). After intranasal (i.n.) infection of conventional mice, the virus spreads from the lung to the lymphoid tissue (29) and then persists in B lymphocytes (28) and in epithelial cells (27). This persistent infection is associated with an infectious mononucleosis-like illness (7, 20) characterized by a CD4-dependent splenomegaly and an increase in viral load (31). In BALB/c mice, MHV-68 causes an acute and apparently non-antigen-specific rise in total serum IgG (26). The virus-specific serum antibody response is, in contrast, relatively slow in onset and does not reach plateau levels until 2 to 3 months after infection (26). MHV-68-infected C57BL/6J (B6) mice have more IgG⁺ cells and fewer IgM⁺ cells in the spleen (18) than uninfected controls, but to what extent this represents normal immunity is unclear.

There is evidence (3) of MHV-68 infection in splenic germinal centers, and both the non-antigen-specific B-cell activation and the CD4-dependent increase in viral load may reflect an exploitation by the virus of normal germinal-center function. The present analysis defines the need, or lack thereof, for CD4⁺ T-cell help to drive B-cell activation following in vitro or in vivo exposure to MHV-68.

MATERIALS AND METHODS

Mice, virus infection, and sampling.

The B6, (B6 × 129)F₁, CD40 ligand-deficient (CD154^−/−) (10), and interleukin 6 (IL-6)-deficient mice (IL-6^−/−) (17) were purchased from Jackson Laboratories (Bar Harbor, Maine). The major histocompatibility complex (MHC) class II-deficient mice (I-A^b−/−) that lack CD4⁺ T cells (11) were bred at St. Jude Children’s Research Hospital. Except for i.n. infection with 600 PFU of MHV-68, all mice were kept under specific-pathogen-free conditions. Virus stocks were grown in owl monkey kidney cells (29), were free of contamination with lipopolysaccharide (LPS) (final concentration, <0.1 pg/ml) as determined by enzyme-linked immunosorbent assay (ELISA) (BioWhittaker, Walkersville, Md.), and were negative for mycoplasma by PCR ELISA (Boehringer Mannheim, Indianapolis, Ind.). Serum samples were obtained either from the axillary artery after terminal anesthesia or from a tail vein. Bone marrow was harvested, where indicated, from both femurs and tibiae.

Cell cultures.

Spleens from naive mice were homogenized to single-cell suspensions (2 × 10⁷/ml) in RPMI (Life Technologies, Grand Island, N.Y.) supplemented (complete medium) with penicillin (60 μg/ml), glutamine (2 mM), 10% fetal calf serum (HyClone, Logan, Utah), and 55 μM 2-mercaptoethanol and were exposed for 1 h at 37°C to infectious virus (0.1 PFU/cell) or to an equivalent quantity of virus that had previously been heated for 3 h at 56°C to abolish infectivity (PFU count per milliliter < 0.01% that of untreated virus). After infection, the cells were washed once and cultured (3 × 10⁶/ml) for 3 days in complete medium at 37°C with 5% CO₂. Control spleen cell populations were cultured in complete medium alone and in complete medium with 10 μg of LPS/ml (Sigma Chemical Co., St. Louis, Mo.).

Flow cytometry.

Lymphocyte populations were washed in ice-cold phosphate-buffered saline (PBS) with 0.01% azide and 0.1% bovine serum albumin, stained on ice for 30 min with monoclonal antibodies (MAbs) to CD19, B220, and CD69 (Pharmingen, San Diego, Calif.), washed once more, and analyzed on a FACScan with Cellquest software (Becton Dickinson, San Jose, Calif.). Purified B lymphocytes were obtained from unfractionated spleen cells by flow cytometric sorting of B220⁺ CD19⁺ cells on a FACStar Plus (Becton Dickinson). These cells (>99% B220⁺ CD19⁺) were stimulated and cultured as described for unsorted spleen cells.

Total Ig and virus-specific Ig assays.

Total IgG and IgM levels in serum and culture supernatants were assayed as described previously (26). Briefly, Nunc Maxisorp immunoplates (Fisher Scientific, Pittsburgh, Pa.) were coated overnight at 4°C with goat anti-mouse IgG or goat anti-mouse IgM μ chain (Sigma) at 1 μg/ml in PBS. Plates were washed five times with PBS-Tween (0.05%), blocked for 1 h with PBS-Tween (0.05%)-bovine serum albumin (1%), incubated with threefold dilutions of serum or culture supernatant for 1 h, and washed five times. Bound antibody was detected with alkaline phosphatase-conjugated goat anti-mouse IgG γ chain (Sigma) or alkaline phosphatase-conjugated goat anti-mouse IgM μ chain (Sigma), by using a nitrophenylphosphate substrate (Sigma) and reading the absorbance at 405 nm on a model 3550 Microplate Reader (Bio-Rad, Hercules, Calif.).

Virus-specific IgG was assayed as for total IgG, except that plates were coated with Triton X-100 (0.05%)-disrupted MHV-68 or influenza A/HKx31 virus. MHV-68-specific neutralizing antibody was detected by a plaque reduction assay (26). Briefly, twofold serum dilutions in minimal essential medium (Life Technologies) were incubated with 50 PFU of MHV-68 on ice for 1 h in 96-well plates. Then 3 × 10⁴ BALB/c-3T3 cells (American Type Culture Collection, Manassas, Va.) were added to each well and, after a 6-h adherence, overlaid with minimal essential medium containing 10% fetal calf serum and 0.75% carboxymethyl cellulose. After 4 days of culture the cells were fixed with methanol and stained with Giemsa solution (Sigma). The neutralization titer was defined as the highest serum dilution giving a >50% reduction in the number of viral plaques. Sera from uninfected mice had no effect on plaque formation. The IgG- and IgM-secreting plasma cells were detected with the same coating and detecting antibodies used for the respective ELISAs, but with nitrocellulose-bottom 96-well plates (Millipore, Bedford, Mass.). After anti-Ig-coated plates were blocked with complete medium for 1 h, cultured cells were washed three times in complete medium and incubated for 4 h at 37°C with 5% CO₂. Spots visualized with 5-bromo-4-chloro-3-indolylphosphate toluidinium (BCIP)–nitroblue tetrazolium (NBT) substrate (Sigma) were counted microscopically.

RESULTS

MHV-68 infection in vitro.

Infection of naive spleen cells in vitro with MHV-68 caused a dramatic activation of B cells, manifested on flow cytometry as an increase in forward scatter (cell size) and an up-regulation of the early activation marker CD69 on the CD19⁺ population (Fig. 1A). The effect of LPS treatment is shown for comparison. The MHV-68-induced activation profile was not due to contaminating endotoxin, since this was undetectable in the virus preparation and would not have been inactivated by the heat treatment regime (see Materials and Methods). There was also evidence of B-cell proliferation: cell numbers were typically increased two- to threefold by 2 days after infection, and >90% of the B220⁺ cells were in S phase as shown by propidium iodide staining (data not shown). However, by day 5 of culture, >90% of the B cells were shown to be apoptotic by propidium iodide staining (data not shown), and we were not able to grow transformed B-cell lines from these cultures.

FIG. 1.

B-cell activation after in vitro exposure to MHV-68. Live lymphocytes were gated initially on the basis of forward and side scatter, then on the CD19⁺ population. In each panel, the specifically treated population (bold line) is compared with control, untreated cells (lightface line). (A) Profile for spleen cells; (B) profile for parallel cultures of flow cytometrically sorted B cells (>99% B220⁺ CD19⁺). Equivalent results were obtained in five further experiments.

The in vitro infection of naive spleen cells had no effect on T-cell phenotype, as assessed by cell size, CD69 expression, or CD62L expression (data not shown), suggesting that B-cell activation was a direct viral effect rather than a result of lymphocyte interactions. The activation of flow cytometrically sorted B cells (>99% B220⁺ CD19⁺) by MHV-68 was equally marked (Fig. 1B), indicating that it was an entirely T-cell-independent phenomenon. The absence of any requirement for CD4⁺ T-cell help was further confirmed by a degree of B-cell activation in spleen cell populations from CD4⁺ T-cell-deficient (I-A^b−/−) or CD40 ligand-deficient (CD154^−/−) mice equivalent to that in normal controls (Table 1). Although IL-6 production is prominent in spleen cultures after MHV-68 infection (24), B-cell activation occurred to an equal extent in spleen cells from IL-6-deficient mice or in cultures incorporating a neutralizing MAb to IL-6 (Pharmingen), added daily at 10 μg/ml (Table 1). Thus, IL-6 did not play a central role in mediating the virus-induced B-cell activation. Furthermore, the absence of this cytokine has no obvious effect on the pathogenesis of infection (23). The MHV-68 genome is not known to encode a viral analogue of IL-6 (33).

TABLE 1.

CD69 expression on CD19⁺ B cells after in vitro MHV-68 infection^a

Genetic background	Valueb for:
	Uninfected cells	MHV-68-infected cells
B6
I-Ab+/+	3.8	43.2
I-Ab−/−	3.4	35.8
B6 × 129
CD154+/+ Il-6+/+	14.4	79.3
CD154−/−	9.1	71.4
IL-6−/−	11.9	76.4
B6
Untreated	16.8	185.6
+ Anti-IL6 MAb	ND	201.6
+ Isotype control MAb	ND	162.3

^aSpleen cells either were not infected or were infected with MHV-68 at 0.1 PFU/cell, then cultured (3 × 10⁶/ml) for 3 days. Live lymphocytes were gated by forward and side scatter, and B cells were identified by further gating on the CD19⁺ population. In all cases there was significant up-regulation of CD69 expression on the CD19⁺ population following MHV-68 infection (p < 0.0001 by Kolmogorov-Smirnov statistics). Each set of results is representative of two experiments. 

^bGeometric mean fluorescence intensity of anti-CD69–fluorescein isothiocyanate staining. ND, not done. 

The in vitro infection protocol also stimulated Ig production, which was detectable from 2 days after infection by ELISA of culture supernatants (Fig. 2A). The Ig detected was predominantly of the IgM class, and ELISpot assays (Fig. 2B) confirmed that almost all (>99%) of the antibody-forming cells (AFCs) produced IgM rather than IgG. Thus, although B cells were activated and differentiated to AFCs after exposure to MHV-68 in vitro, there was no significant Ig class switching, even in the presence of CD4⁺ T cells.

FIG. 2.

Ig production by B cells after exposure to MHV-68 in vitro. (A) ELISA absorbance values are shown for threefold dilutions of culture supernatants from LPS-treated (○), MHV-68-infected (•), and untreated (□) cultures of flow cytometrically purified B cells, by using plates coated with affinity-purified anti-IgG or anti-IgM Ig. Selective IgM secretion was also observed with unfractionated spleen cell cultures (data not shown). Equivalent results were obtained in five further experiments. (B) ELISpot assay results. Means ± SD of quadruplicates are shown for unfractionated spleen cell cultures, by using nitrocellulose-bottom wells coated with affinity-purified antisera specific for either mouse IgG or mouse IgM. The number of IgM-producing AFCs was significantly increased in the LPS- and MHV-68-treated populations (P < 0.0001 by t test). Equivalent results were obtained in two further experiments and also with flow cytometrically purified B cell cultures.

MHV-68 infection in vivo.

Giving MHV-68 i.n. to B6 mice caused a sustained, approximately fourfold increase in total serum Ig titers (Fig. 3A and B). In contrast to the situation in vitro, IgG was the major component of the increased serum Ig levels (Fig. 3B). At day 16 after infection, t tests showed significant rises in both serum IgG (P < 0.0001) and IgM (P < 0.0005) levels in the B6 mice relative to uninfected controls. There was no significant rise in serum IgG in the I-A^b−/− mice (P > 0.1), but the relatively small rise in IgM was significant (P < 0.01). Virus-specific serum Ig (Fig. 3C and D) was undetectable in the I-A^b−/− mice and showed a rather gradual rise in the B6 mice. The results for B6 mice were consistent with those described previously for the BALB/c strain (26). The extent of B-cell activation after MHV-68 infection of B6 mice was also evident from the massive increase in CD69 expression on this lymphocyte population in the draining lymph nodes and, to a lesser degree, in the spleen (Fig. 4). This effect was only minimally evident in the CD4⁺ T-cell-deficient I-A^b−/− mice.

FIG. 3.

Virus-specific and total serum antibody levels after MHV-68 infection. Each point shows the mean titers (± SD) for six individual B6 (○) or I-A^b−/− (•) mice. The total serum IgM, total serum IgG, and total virus-specific serum IgG were determined by ELISA, while the neutralizing virus-specific antibody titer was determined by plaque inhibition (see Materials and Methods). All measurements were made with reference to a standard immune serum, and the titers are expressed as arbitrary units.

FIG. 4.

Flow-cytometric analysis of B-cell phenotypes from spleens and mediastinal lymph nodes (MLN) of B6 and I-A^b−/− mice after MHV-68 infection. Live lymphocytes were gated by forward and side scatter, and B cells, further gated on the basis of B220^hi staining, were divided into CD69^lo (○) or CD69^hi (•) subsets. Both CD69^hi B220^hi and CD69^lo B220^hi cells were uniformly CD19⁺ (data not shown). Cells were pooled from one to three mice for each time point. The numbers of each cell type per mouse were calculated from the total cell counts and the proportions stained specifically by flow cytometry. Each point shows the mean ± SD of two to six pools.

It is known that EBV uses HLA-DR as a cofactor for entry into B cells (19). However, the relative lack of B-cell activation after MHV-68 infection of I-A^b−/− mice was shown not to result from the absence of MHC class II glycoprotein on the B cells, by using adult thymectomized B6 mice that were depleted of CD4⁺ T cells by MAb treatment prior to infection. These CD4⁺ T-cell-depleted mice showed no rise in total serum IgG levels compared to uninfected controls (Fig. 5).

FIG. 5.

Serum IgG levels in intact and CD4-depleted mice. Adult thymectomized mice were depleted of CD4⁺ T cells by five injections of ascitic fluid containing the GK1.5 MAb at days −4 to +4 relative to i.n. virus challenge (1). Flow-cytometric staining of spleen cells at the time of sampling, 19 days after MHV-68 infection, indicated that depletion was >95%. Mean absorbance values (± SD) are shown for threefold serum dilutions from three individual thymectomized mice, either undepleted and MHV-68 infected (•), undepleted and uninfected (□), or CD4 depleted and MHV-68 infected (○). An equivalent CD8⁺ T-cell depletion protocol did not reduce the elevation of total serum IgG levels (data not shown). These results were reproduced in a repeat experiment.

Mice that had been immunized i.n. 2 to 3 months previously with the influenza A/HKx31 virus were infected with MHV-68 to determine the effect on established B cells that had already switched to IgG production. Surprisingly, there was a significant fall in the titer of influenza virus-specific IgG in serum after MHV-68 infection (Fig. 6), suggesting that IgG-switched memory B cells were not contributing to the acute rise in total serum IgG. No significant loss of influenza A/HKx31 virus-specific plasma cells was apparent in the bone marrow of the MHV-68-infected mice: in one of three equivalent experiments, the numbers of influenza A/HKx31 virus-specific AFCs/10⁵ bone marrow cells (means ± standard deviations [SD]; six mice per group) were 146 ± 44 with influenza A/HKx31 virus alone and 125 ± 53 with influenza A/HKx31 virus followed by MHV-68.

FIG. 6.

Influenza virus-specific serum IgG titers in influenza virus-immune B6 mice after MHV-68 infection. Mice primed i.n. with 30 hemagglutination units of influenza A/HKx31 virus were challenged i.n. 2 months later with 600 PFU of MHV-68 (•) or were left untreated (○). Each point shows the mean titer (± SD) from six individuals per group. The same mice were bled (30 to 50 μl per sample) serially from a tail vein. A standard immune serum was included on each ELISA plate to allow for comparison between samples. The influenza A/HKx31 (H3N2) virus that was used to prime the mice 8 to 12 weeks prior to the challenge with MHV-68 was also used as the immunoabsorbent. The titers of influenza virus-specific serum IgG were significantly lower (P < 0.0002 by t test) in MHV-68-infected mice at days 21, 28, and 38 than in uninfected control mice. Equivalent results were obtained in two further experiments.

We then asked whether the relative fall in the influenza virus-specific Ig titer reflected simple dilution due to an increased rate of production of new Ig specificities. The decay in the titer of Sendai virus-specific IgG in serum was monitored over 30 days in mice given a bolus intravenous injection of a Sendai virus-specific IgG2a MAb (500 μg/mouse) at the same time as mock infection or infection with MHV-68. No difference (mean ± SD for six individuals per group) in IgG turnover (32) was observed between uninfected mice (half-life = 8.5 ± 3.2 days) and MHV-68-infected mice (half-life = 10.0 ± 1.6 days). A possible implication is that factors promoting continued Ig production by the influenza virus-specific memory B cells were in some way compromised by the MHV-68 infection.

DISCUSSION

Exposing splenic B cells to a relatively high dose of MHV-68 in vitro led to generalized B-cell activation, independent of a contribution from other cell types. In contrast, in vivo B-cell activation was considerably reduced in congenitally CD4⁺ T-cell-deficient or CD4⁺ T-cell-depleted mice compared with that in immunocompetent controls. This difference could indicate that the MHV-68 viral load is limiting in vivo and that CD4⁺ T-cell-dependent proliferation of the activated, virus-infected B cells is required to achieve a significant effect. It may also be the case that CD4⁺ T cells provide survival signals to the activated B cells, as the CD4⁺ T-cell-independent proliferative stimulus provided to B cells in vitro by MHV-68 infection rapidly led to apoptosis.

An interesting feature of the in vitro B-cell stimulation was a lack of clear dependency on virus dose, with equivalent activation being observed for a range of 0.01 to 10 PFU/cell (data not shown). Essentially all the B cells were activated, despite the presence of little infectious virus (<1 PFU/10⁴ cells by plaque assay), suggesting that uninfected B cells were also being induced, perhaps by a specific viral protein. Whatever the mechanism, the net effect seems to be that in vivo MHV-68 infection activates B cells such that they can interact with CD4⁺ T cells independently of normal antigenic stimulation. One obvious question concerns the source of the “helper” CD4⁺ T cells that might be involved. Although massive T-cell activation is a prominent feature of MHV-68 infection in vivo (7), the lack of an effect on CD4⁺ T cells in vitro and the absence of an obvious T-cell receptor Vβ bias in the CD4⁺ (as distinct from CD8⁺) population (30) argue against a superantigen effect. Instead, the fact that the in vivo interaction between CD4⁺ T cells and virus-infected B cells was not reproduced in cell culture implied that a distinct source of activated CD4⁺ T cells was required, perhaps even the normal, antigen-specific CD4⁺ T-cell response (26).

The Ig class switching presumably occurred at the same time as B-cell activation and may reflect a component of normal germinal-center function in an essentially virus-driven process. An important prediction of this model is that virus should be found in Ig class-switched, memory B cells that may also show evidence of somatic hypermutation. We have not, at this stage, identified an appropriate surface antigen that would allow us to separate such infected B cells for genetic analysis.

Since non-antigen-specific B-cell activation occurs with both B-cell-tropic (16, 22) and non-B-cell-tropic (13, 14, 21) HV infections, it may also represent a means exploited by the virus to inhibit the normal humoral response. It is notable in this regard that the MHV-68-specific serum Ig response (26) is slow to reach plateau levels (Fig. 3C and D). The experiments with influenza virus-immune mice (Fig. 6) suggested that plasma cell function is somehow suppressed during the polyclonal activation, possibly due to massive cytokine release. Clearly any such suppression could also affect the MHV-68-specific response, thus promoting viral dissemination and persistence.