Antibody Production and In Vitro Behavior of CD27-Defined B-Cell Subsets: Persistent Hepatitis C Virus Infection Changes the Rules

Abstract

There is growing interest in the tendency of B cells to change their functional program in response to overwhelming antigen loading, perhaps by regulating specific parameters, such as efficiency of activation, proliferation rate, differentiation to antibody-secreting cells (ASC), and rate of cell death in culture. We show that individuals persistently infected with hepatitis C virus (HCV) carry high levels of circulating immunoglobulin G (IgG) and IgG-secreting cells (IgG-ASC). Thus, generalized polyclonal activation of B-cell functions may be supposed. While IgGs include virus-related and unrelated antibodies, IgG-ASC do not include HCV-specific plasma cells. Despite signs of widespread activation, B cells do not accumulate and memory B cells seem to be reduced in the blood of HCV-infected individuals. This apparent discrepancy may reflect the unconventional activation kinetics and functional responsiveness of the CD27⁺ B-cell subset in vitro. Following stimulation with T-cell-derived signals in the absence of B-cell receptor (BCR) engagement, CD27⁺ B cells do not expand but rapidly differentiate to secrete Ig and then undergo apoptosis. We propose that their enhanced sensitivity to BCR-independent noncognate T-cell help maintains a constant level of nonspecific serum antibodies and ASC and serves as a backup mechanism of feedback inhibition to prevent exaggerated B-cell responses that could be the cause of significant immunopathology.

Exploration of the conditions that stimulate and modulate B-cell proliferation and differentiation is critical to both the understanding of normal B-cell function and the detection of alterations leading to humoral or neoplastic B-cell disorders. In humans, this exploration is confined to in vitro models in which combinations of surrogate signals mimic the range of physiological stimulations that may occur in vivo (12, 27). Most of these signals are relatively well understood. They include B-cell receptor (BCR) cross-linking with antigen (9, 45), cell contact-mediated interactions with T cells (6, 11, 18, 26, 46), and secretion of soluble regulatory cytokines (2, 3, 13, 15, 36, 56, 57). The current consensus is that BCR engagement followed by cognate T-cell help drives the proliferation of antigen-specific naive B cells and their differentiation into memory B cells and plasma cells (7, 54). Whereas plasma cells are mitotically quiescent and terminally differentiated antibody-secreting cells (ASC) (62), memory B cells may be consecutively stimulated, expanded, selected, and turned into effector cells. When reexposed to antigen, they rapidly proliferate and differentiate. Their memory lineage is thus preserved, and large numbers of plasma cells are concomitantly generated (4). Unlike naive B cells, which are totally dependent on BCR signaling, memory B cells may be activated by bystander T-cell help without BCR triggering (10, 23, 68). This ability to respond to such a polyclonal stimulus, in the absence of cognate interaction, seems essential for maintenance of their serological memory.

The functional specialization of naive B cells, memory B cells, and plasma cells is instrumental in driving new and anamnestic antibody-mediated immune responses but seems critical in some chronic inflammatory conditions, including persistent viral infections. For instance, patients with chronic hepatitis C virus (HCV) are often hypergammaglobulinemic. They produce autoantibodies, have circulating immune complexes with cryoprecipitating properties, and display an increased risk of B-cell tumors (25, 52, 58, 63, 76). This is paradoxical considering the lack of antiviral function associated with anti-HCV antibodies (22) and considering that B cells seem not to be direct targets for productive HCV replication (35). Several investigators, including ourselves, have suggested that continued and indiscriminate virus-driven polyclonal stimulation is a plausible mechanism whereby abnormal clonal B-cell proliferation and antibody production are maintained throughout HCV infection (17).

Based on this model and the roles of naive B cells, memory B cells, and plasma cells within humoral response kinetics, the following predictions have frequently been put forward: (i) the frequency of B cells should be increased in patients with chronic hepatitis C; (ii) this expanded population should be enriched in memory B cells; (iii) the expanded memory B-cell subset should be polyclonal; and (iv) the level of serum antibody should be proportional to the frequency of memory B cells responding to polyclonal activators. The truth of these predictions began to be questioned when Ni et al. (44) clearly demonstrated that peripheral B cells from chronically HCV-infected patients show a naive, resting phenotype, thus challenging the idea of antigen-driven activation and proliferation.

On the strength of this observation, we set out to determine whether the natural behavior of B cells is affected by HCV persistence and to what extent their different levels of responsiveness provided an explanation of these biological discrepancies. Because immunoglobulin (Ig) secretion is the outcome of the several steps of B-cell activation, we first looked for nonspecific and virus-specific antibodies in the sera of patients with chronic HCV and then worked back to a detailed characterization of ASC and B-cell subsets, ending with an extensive assessment of the proliferative and secretive behavior of CD27⁻ and CD27⁺ B cells across diverse in vitro stimulations. Our results provide a complex picture of B-cell maturation, homeostasis, and antibody production. They reveal how HCV persistence changes the biological sensitivity of memory B cells and shifts the balance between cell survival and death.

MATERIALS AND METHODS

Study subjects.

Three groups of individuals were investigated: patients persistently infected with HCV (PI) (PI1 to PI20), patients who spontaneously recovered from HCV infection (spontaneous resolvers [SR]) (SR1 to SR8), and normal controls (healthy donors [HD]) (HD1 to HD20). The first group included individuals with chronic hepatitis who were antiviral therapy naive at the time of sampling. All but one of them (PI18) underwent diagnostic percutaneous-needle liver biopsy. The second group was composed of individuals who had spontaneously cleared HCV RNA more than 3 years earlier. The normal controls were healthy blood donors. All subjects were hepatitis B surface antigen negative and antibody negative to human immunodeficiency virus. All participants gave their prior written and informed consent. The study protocol conformed to the ethical guidelines of the 1975 Helsinki Declaration.

Clinical assays.

Plasma HCV RNA levels were determined with a Roche Amplicor assay (Roche Diagnostics, Branchberg, NJ) and standardized to international units (IU). HCV genotypes were determined with a line probe assay (Innogenetics, Zwijnaarde, Belgium) and classified as detailed by Simmonds et al. (64, 65). Serum gamma globulin, IgG, and IgM concentrations were measured by nephelometry. Rheumatoid factor (RF) was determined by standard assays (16, 73). Liver biopsies were evaluated by pathologists according to well-accepted standards for the documentation of fibrosis (50) and inflammatory activity (8).

Cell preparations.

Thirty to forty milliliters of venous blood from each subject was drawn into heparin-coated tubes. Peripheral blood mononuclear cells (PBMC) were separated by Ficoll-Hypaque (Pharmacia Biotech, Uppsala, Sweden) density gradient centrifugation. Plasma samples were collected and stored at −70°C until analysis. Total B lymphocytes were isolated from PBMC by magnetic cell sorting with anti-CD19 microbeads and MACS columns (Miltenyi Biotec, Bergisch Gladbach, Germany). Both the positive selected cell fraction and the flowthrough (nonbinding) cells were retained for further use. CD19⁺ cells were fractionated further into CD27⁻ and CD27⁺ populations by use of anti-CD27 microbeads (Miltenyi Biotec).

CFSE labeling.

Freshly purified CD27⁻ and CD27⁺ B cells were labeled with 5(6)-carboxyfluorescein diacetate succinimidyl ester (CFSE) (Molecular Probes, Eugene, OR), according to the original method (24, 39). Briefly, cells were resuspended at 1 × 10⁷/ml in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (BSA). CFSE, dissolved in dimethyl sulfoxide, was added at a final concentration of 0.5 μM. After a 10-min incubation at 37°C, labeled cells were washed with cold PBS containing 0.1% BSA (Sigma-Aldrich, St. Louis, MO) and resuspended in culture medium.

Immunostaining, flow cytometry, and epifluorescence microscopy.

For cell surface marker analysis, cells were stained with different combinations of the following directly conjugated monoclonal antibodies: anti-CD19 peridinin chlorophyll protein, anti-CD27 fluorescein isothiocyanate (FITC), anti-CD21 phycoerythrin (PE), and anti-CD38 PE (all from Pharmingen-BD Biosciences, San José, CA). Incubations were conducted on ice for 20 min. Washings were performed twice with 200 μl cold PBS containing 0.1% BSA. For plasma membrane asymmetry analysis, B cells were resuspended in annexin V binding buffer and stained with FITC-labeled annexin V and anti-CD27 PE monoclonal antibodies (both from Pharmingen-BD Biosciences) by following the manufacturer's specifications for apoptosis detection kit 1 (Pharmingen-BD Biosciences). Stained cells were analyzed without delay with a BD FACScan flow cytometer by using CellQuest software (Becton Dickinson, San José, CA). FITC-, PE-, and peridinin chlorophyll protein-conjugated isotype controls were used in all analyses. At least 5,000 events were acquired for each sample. For mitochondrial membrane potential analysis, CD27⁺ B cells were stained with the fluorescent cationic dye 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1) by use of a mitochondrial membrane potential detection kit (Stratagene, La Jolla, CA) in accordance with the manufacturer's instructions. For nuclear and chromatin analysis, B cells were stained with 4′,6′-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich) at 0.1 mg/ml. Stained cells were examined under a Nikon TE2000 inverted microscope (Nikon Instruments S.p.A., Sesto Fiorentino, Italy) equipped with an epifluorescence source.

Recombinant HCV antigens.

Recombinant HCV-1a antigens were kindly provided by B. Phelps (Chiron, Emeryville, CA). Recombinant HCV-1b, -2a, and -2c antigens were synthesized by Biochemical Test Systems GmbH Laboratories (Reutlingen, Germany). Diphtheria toxoid (DT) and tetanus toxoid (TT) were purchased from Sigma-Aldrich.

Cell cultures.

CFSE-labeled CD27⁻ and CD27⁺ B cells were cultured in triplicate at 10⁴ cells/well in flat-bottomed microplates (Falcon; BD Biosciences) in 100 μl RPMI 1640 medium completed with 10% heat-inactivated fetal bovine serum, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mmol/liter l-glutamine (all from Sigma-Aldrich), and one of the following stimulatory conditions: (i) 2 μg/ml F(ab′)₂ fragments of polyclonal goat anti-human IgG and IgM (Jackson ImmunoResearch Laboratories, West Grove, PA); (ii) 500 ng/ml recombinant human CD40 ligand (CD40L) (Alexis Biochemicals, Lausen, Switzerland); (iii) 500 ng/ml recombinant human CD40L and 100 ng/ml interleukin-4 (IL-4) (PeproTech, Rocky Hill, NJ); (iv) 500 ng/ml recombinant human CD40L and 50 ng/ml IL-10 (PeproTech); (v) 10 μg/ml HCV core recombinant protein; or (vi) 10 μg/ml HCV NS3 recombinant protein. Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂ with the appropriate stimuli for 5 days. Aliquots of supernatant were removed and frozen at −70°C until batched analysis.

Measurement of antibodies and antibody-secreting cells.

Total IgG and IgM concentrations in culture supernatants were measured using a human IgG and IgM enzyme-linked immunosorbent assay (ELISA) quantitation kit (Bethyl Laboratories, Montgomery, TX), according to the manufacturer's specifications. Anti-DT and anti-TT serum levels were determined using a tetanus and diphtheria ELISA IgG test kit (Genzyme Virotech GmbH, Rüsselsheim, Germany) and standardized to IU.

HCV-specific antibody levels in sera and culture supernatants were determined by a standard in-house ELISA. Briefly, 96-well microtiter plates (Falcon; BD Biosciences) were coated overnight with either HCV core or NS3 recombinant protein at a concentration of 10 μg/ml in 0.05 M carbonate-bicarbonate buffer. Plates were blocked with 50 mM Tris, 0.14 M NaCl, 1% BSA for 30 min at room temperature and washed with 50 mM Tris, 0.14 M NaCl, 0.05% Tween 20. Culture supernatants were added to the wells and incubated for 2 h at room temperature. Plates were washed, and horseradish peroxidase-conjugated goat anti-human IgG (Bethyl Laboratories) was added at a 1:10,000 dilution in 50 mM Tris, 0.14 M NaCl, 1% BSA, 0.05% Tween 20. Following incubation for 1 h at room temperature, plates were washed and then developed using 0.4 g/liter 3,3′,5,5′-tetramethylbenzidine (TMB) in citrate buffer with 0.02% H₂O₂. The reaction was quenched with 2 M H₂SO₄ after 15 min. Plates were read immediately at 450 nm with a Bio-Rad 3550 plate reader.

ASC were detected by an enzyme-linked immunospot assay (ELISPOT assay) whose original protocol (61) was adapted to enumerate IgG-, IgM-, DT-, TT-, and HCV-specific ASC. Briefly, MultiScreen 96-well filtration plates (Millipore, Bedford, MA) were coated with goat anti-human Ig (Bethyl Laboratories, Montgomery, TX) at 10 μg/ml for total IgM-secreting cells (IgM-ASC) and IgG-ASC, HCV core recombinant protein at 10 μg/ml for anti-HCV core-specific ASC, HCV NS3 recombinant protein at 10 μg/ml for anti-HCV NS3-specific ASC, TT at 5 μg/ml for anti-TT-specific ASC, or DT at 5 μg/ml for anti-DT-specific ASC. After overnight incubation at 4°C, plates were washed with PBS, blocked with PBS-1% BSA for 1 h at room temperature, and then washed with complete medium. Appropriate numbers of freshly purified total (CD19⁺) B cells, CD27⁻ B cells, or CD27⁺ B cells resuspended in complete medium were added to duplicate wells and incubated overnight at 37°C. Plates were then washed with PBS, followed by PBS containing 0.05% Tween 20. Horseradish peroxidase-conjugated goat anti-human IgG or anti-human IgM (Bethyl Laboratories) in PBS containing 0.05% Tween 20-1% BSA was added, and the plates were incubated overnight at 4°C. Plates were again washed with PBS and then developed using 3-amino-9-ethylcarbazole (AEC) (Sigma-Aldrich). Developed plates were rinsed in tap water and allowed to dry before spots were counted under a dissecting microscope.

Molecular analysis of Ig heavy chain CDR3 gene regions.

Amplification, cloning, sequencing, and examination of B-cell DNA encoding the Ig heavy chain CDR3 gene region were performed as previously described (19, 53).

Data analysis.

Flow cytometric data were analyzed with WinMDI 2.8 software. Statistical analyses were performed using Prism (GraphPad Software, San Diego, CA). Nonparametric statistics were used because much of the data was not normally distributed. Tests included the Kruskal-Wallis analysis of variance and the Mann-Whitney test for comparisons of groups and Spearman's rank test for correlations.

RESULTS

Circulating nonspecific and virus-specific antibodies are increased in persistently HCV-infected patients.

Antibodies are usually the measurable correlate of B-cell immune responses. To evaluate both the nonspecific and virus-specific B-cell responses associated with persistent HCV infection, levels of total and antigen-specific Ig in the sera of infected and uninfected individuals were measured by ELISA. As expected, total Ig concentration was significantly higher in persistently infected patients than in healthy donors and spontaneous resolvers (P = 0.0013) (Fig. 1A). This hypergammaglobulinemia was polyclonal, as assessed by isoelectric focusing (data not shown), and primarily attributable to large amounts of IgG (P = 0.0133) (Fig. 1A). Despite the fact that most persistently infected patients were RF⁺ (Table 1), their IgM levels (mean, 186 mg/dl; standard deviation, 133 mg/dl) were comparable to those of uninfected individuals (Fig. 1A).

FIG. 1.

Circulating nonspecific and virus-specific antibody levels in HD, PI, and SR. Sera were analyzed by either nephelometry or ELISA. (A) Ig concentrations (means ± standard errors of the means [SEM]). (B) Correlation between total Ig concentration and plasma viral load in PI. (C) Correlation between IgG concentration and plasma viral load in PI. (D) Plasma viral loads in PI with negative or positive RF test. (E and F) Diphtheria and tetanus antitoxin IgG concentrations (means ± SEM). (G and H) Endpoint titers of IgG against HCV core and NS3 proteins. (I and L) Correlation between titers of IgG against core and NS3 proteins and plasma viral load in PI.

TABLE 1.

Demographic and clinical parameters of patients with persistent HCV infection

Patient no.	Age (yr)	Sexa	HCV genotype	HCV RNA (IU/ml)	Presence of anti-HCV Absb	Grade (1-4)	Stage (1-4)	RF status	Gamma globulin level = (g/dl)
1	52	F	1b	310,000	+	2	3	−	2.01
2	61	M	1b	730,000	+	2	3	+	1.90
3	46	M	1b	530,000	+	2	2	−	1.86
4	59	F	1b	374,600	+	2	2	+	1.70
5	65	F	1b	162,900	+	3	2	+	1.60
6	58	F	2a/2c	15,000	+	2	2	−	1.90
7	61	M	1b	266,000	+	2	4	−	1.79
8	47	F	1b	86,200	+	3	2	−	1.20
9	29	M	2a/2c	16,000	+	2	2	+	0.80
10	35	F	1b	1,320,000	+	3	3	+	2.30
11	65	F	1b	180,000	+	2	2	−	1.42
12	40	F	2c	980,000	+	2	3	+	2.10
13	56	M	1b	26,300	+	3	4	−	2.20
14	59	F	1b	456,000	+	1	2	+	1.80
15	38	F	1b	1,227,000	+	1	4	+	2.11
16	52	M	1b	570,000	+	3	2	−	2.01
17	67	F	2c	75,000	+	3	2	−	1.90
18	44	F	1b	890,000	+	NDc	ND	+	2.43
19	63	M	1a	140,000	+	2	2	+	0.90
20	46	F	1b	600,000	+	2	1	+	2.20

^aF, female; M, male.

^bAbs, antibodies.

^cND, not determined.

A positive correlation between plasma viral load and both total Ig (P = 0.0053; r = 0.5983) and IgG (P = 0.009; r = 0.5681) levels in the persistently infected patients was found (Fig. 1B and C). No such correlation existed for IgM levels. However, comparison of the RF⁺ patients with the RF⁻ patients showed that the HCV RNA levels of the RF⁺ patients were significantly higher (P = 0.0403) (Fig. 1D). A trend towards statistical significance (P = 0.0599) between histologic stage of liver fibrosis and IgG level was observed. In keeping with the Italian vaccination program, all subjects in the study had been vaccinated with DT and TT. We therefore determined levels of serum antibodies against those antigens to evaluate the status of serologic memory. Surprisingly, both anti-DT and anti-TT IgG titers were significantly higher for the persistently infected patients than for the healthy donors and spontaneous resolvers (P = 0.0246 and P = 0.0026 for anti-DT and anti-TT IgG titers, respectively) (Fig. 1E and F). Age and gender were comparable in the three groups.

To verify whether the observed increase in Ig concentration also affected HCV-specific antibody responses, we determined endpoint antibody titers to HCV antigens in the sera of persistently infected patients and spontaneous resolvers. Recombinant HCV core and NS3 proteins were chosen as model antigens for structural and nonstructural viral proteins. At the lowest serum dilution tested, all samples were reactive against both HCV proteins. As samples were diluted further, it became evident that persistently infected patients had significantly higher antibody titers against both core (P = 0.0131) and NS3 (P = 0.0153) proteins than the spontaneous resolvers (Fig. 1G and H). The antibody titers of the persistently infected patients were negatively correlated with their viral loads (P = 0.0497 and P = 0.0341 for core and NS3 protein results, respectively) (Fig. 1I and L).

Circulating virus-nonspecific plasma cells are increased in persistently HCV-infected patients.

Differentiation into plasma cells is the culmination of the sequence of events that follow B-cell activation. Plasma cells are terminally differentiated cells that spontaneously secrete copious amounts of antibody and are detectable in peripheral blood on the way to the bone marrow. To determine whether the different antibody levels detected in uninfected and HCV-infected individuals reflected a diverse frequency of circulating plasma cells, PBMC samples were examined ex vivo for nonspecific and virus-specific ASC by use of an ELISPOT assay. The frequency of circulating IgG-secreting plasma cells in persistently infected patients was significantly elevated compared with the frequency in healthy donors and that in spontaneous resolvers (P = 0.0019), though the range was very broad because of a considerable patient-to-patient variation (Fig. 2A).

FIG. 2.

Circulating nonspecific and virus-specific plasma cell frequencies in HD, PI, and SR. Freshly isolated PBMC were assayed by ELISPOT assay. (A and B) IgG- and IgM-ASC. (C) IgM-ASC in PI with negative or positive RF test. (D) Correlation between IgG-ASC number and plasma viral load in PI. (E and F) Proportions of DT- and TT-specific ASC among total IgG-ASC (means ± standard errors of the means). (G and H) Proportions of HCV core- and NS3-specific ASC among total IgG-ASC in PI and SR.

By contrast, IgM-secreting plasma cell levels were not significantly elevated in persistently infected individuals (Fig. 2B), though they were significantly higher in RF⁺ patients than in RF⁻ patients (P = 0.0301) (Fig. 2C). Comparison of the IgG-secreting plasma cell frequencies with plasma HCV RNA levels revealed a positive correlation between these parameters (P = 0.0346; r = 0.4744) (Fig. 2D). Up to 12.5% and 10% of circulating IgG-secreting plasma cells were, respectively, TT and DT specific in persistently infected patients. Percentages were significantly lower for healthy donors (P = 0.0368) and spontaneous resolvers (P = 0.0114) (Fig. 2E and F). No plasma cells specific for HCV core or NS3 protein were detected ex vivo in the persistently infected patients. In contrast, four of the eight spontaneous resolvers had HCV-specific plasma cells in their peripheral blood (Fig. 2G and H).

Circulating B cells are not increased in persistently HCV-infected patients.

The number of B cells traveling in the blood usually reflects the dynamic equilibrium between proliferation and death, as well as retention and emigration of B cells within and from lymphoid tissues (48). In view of the diverse amounts of antibodies and ASC in blood samples from HCV-infected and uninfected individuals, we set out to specifically assess the frequency of peripheral B cells. Whole-blood samples were stained for the coreceptor molecules CD19 and CD21 and analyzed by flow cytometry. Percentages and absolute numbers of CD19⁺ and CD21⁺ cells in persistently infected patients were more heterogeneous than but did not statistically differ from those in healthy donors and spontaneous resolvers. Generally, CD21⁺ cells were slightly fewer in number than CD19⁺ cells. The highest CD21 frequency was found in the HCV-infected patient with the highest CD19 frequency (Fig. 3A and B). Electrophoretic analysis of the Ig heavy chain CDR3 DNA, amplified from PBMC by PCR, did not reveal dominant bands in any of the persistently infected patients. None of 12 randomly picked PCR clones from one selected patient displayed a deduced amino acid sequence similar to those of human antibodies with HCV specificity (Table 2).

FIG. 3.

Circulating B-cell subsets in HD, PI, and SR. Whole-blood samples were immunostained and examined by flow cytometry. Lymphocytes were distinguished by forward and orthogonal light scatter characteristics. (A and B) Percentages of CD19⁺ and CD21⁺ cells among gated lymphocytes. (C) Representative flow cytometry analyses of CD19/CD27-stained PBMC. (D) CD27 expression on B cells. The numbers assigned to the patients are shown at left. (E) Absolute numbers of CD19⁺, CD19⁺ CD27⁻, and CD19⁺ CD27⁺ B cells (means ± standard errors of the means). (F) Correlation between CD27⁺ B-cell percentage and plasma viral load in PI.

TABLE 2.

Deduced amino acid sequence and theoretical specificity of Ig heavy chain CDR3 DNA PCR clones from one patient with persistent HCV infection

Clonotype no.	Ig heavy chain CDR3 deduced amino acid sequence	NCB1 database sequence with highest homology (gi no.)	Specificity
1	CARGRGKLYYYGLDVWGQGTLV	56966692	HLA-A1:mage-A1
2	CAKMGSSGWPYFDYWGQGTLV	4836338	Capsular polysaccharide of Neisseria meningitidis
3	CARAEDIVVPAAMAFDYWGQGTLV	No match
4	CAKEFGSSGWYVDYWGQGTLV	25056062	Platelet glycoprotein Ia/IIa
5	CARAVGYHRSSAGYFDYWGQGTLV	58221716	TT
6	CARDFGGDTAMVYWYFDLWGQGTLV	37014219	Grass pollen
7	CARGTMVRTCWYFDLW	507339	HIV-1a and Escherichia coli β-galactosidase
8	CVGYCSGGSCYSDVAYWGQGTLV	86840	Pr2 erythrocyte
9	CAKVSNYDSSTYYLPYYYYYGMDVWGQGTLV	4836350	Capsular polysaccharide of Neisseria meningitidis
10	CAREFQGDYDSPETLSYFDYWGQGTLV	563649	DNA
11	CARDWEYCSSTSCFFDYWGQGTLV	No match
12	CARDWEYCSSTSCFFDYWGQGTLV	1730301	DNA

^aHIV-1, human immunodeficiency virus type 1.

Circulating CD27⁺ B cells are reduced in persistently HCV-infected patients.

The recent finding that expression of CD27 is restricted to B cells with abundant cytoplasm, high immunoglobulin productivity, enhanced ability to differentiate into plasma cells, and somatically hypermutated V region genes has facilitated the distinction between antigen-inexperienced naive B cells and antigen-experienced memory B cells (1, 33, 69). Once formed, memory B cells reside within discrete regions of secondary lymphoid tissue or reenter and become detectable in the bloodstream.

Since memory B cells serve to replenish the pool of plasma cells, we looked to see whether increased antibody production during persistent HCV infection is sustained by memory B cells. Whole-blood cells from infected and uninfected individuals were double stained for CD19 and CD27 and analyzed by flow cytometry. As shown in Fig. 3C, analysis of a representative individual from each group, expression of CD27 on B cells of the persistently infected patient was reduced compared with results from the other two patients. When expressed either as percentages of the circulating total B cells or as absolute numbers, the frequencies of CD27⁺ B cells in persistently infected patients were significantly lower than those in healthy donors and spontaneous resolvers (P < 0.0001 and P = 0.008, respectively) (Fig. 3D and E). Overall, CD27⁺ B cells contained both IgM⁺ IgD⁺ cells and isotype-switched IgA- and IgG-expressing B cells. In contrast, CD27⁻ B cells were mainly IgM-expressing cells (data not shown).

Generally, patients with higher plasma viral loads had lower percentages of CD27⁺ B cells, suggesting that high viral replication is associated with a reduction of CD27⁺ B cells. This negative correlation was almost significant (P = 0.0566; r = −0.4328) (Fig. 3F). Both CD27⁺ and CD27⁻ B cells were mostly CD21⁺ for all infected individuals (data not shown).

CD27⁺ B cells of persistently HCV-infected patients exhibit reduced proliferation following in vitro noncognate T-cell help.

Naive and memory B cells do not spontaneously secrete antibodies and require in vitro stimulation in order to divide and differentiate into ASC. Two types of stimulation trigger B-cell proliferation and differentiation: that mimicking BCR engagement by antigen and that providing T-cell-derived signals, such as CD40L and cytokines. As already remarked in the introduction, stimulation with CD40L alone or in combination with cytokines reproduces B cells receiving T-cell help in the absence of specific antigen recognition (10) and may play a crucial role during persistent viral infections. Under conditions in which high antigen concentrations are generated, unselective B cells may be permitted to undergo major histocompatibility complex class II loading of viral peptides independent of BCR specificities (29).

To determine whether the observed disproportion between CD27⁺ and CD27⁻ B cells in persistently infected patients had a functional explanation, we measured the capacities of these two populations to proliferate after in vitro stimulation. B cells isolated from infected and uninfected individuals were immunomagnetically separated into CD27⁺ and CD27⁻ fractions. A representative flow cytometry analysis of the purified cell populations is illustrated in Fig. 4A for PI6, indicating that the CD27-enriched and CD27⁻ fractions showed a B-cell purity, as measured by CD19 staining, above 95%. The purified cells were then stimulated with F(ab′)₂ antibody fragments to human Ig for BCR engagement or recombinant CD40L, with and without cytokines for noncognate T-cell help. Proliferation was explored using the division-tracking dye CFSE (36, 37). The number of cell divisions was evaluated by flow cytometry at the end of a 5-day culture period.

FIG. 4.

In vitro proliferation of CD27⁻ and CD27⁺ B cells from HD, PI, and SR. B cells were separated into CD27⁺ and CD27⁻ fractions, labeled with CFSE, cultured with the indicated stimuli for 5 days, and analyzed by flow cytometry. (A) Representative immunomagnetic purification of CD27⁺ and CD27⁻ B cells. (B) Representative CFSE profiles. Sequential peaks of decreased fluorescence intensity identify subsequent generations of proliferating daughter cells. (C) Percentages of CD27⁺ B cells undergoing one or more divisions after stimulation (means ± standard errors of the means).

Figure 4B illustrates representative division profiles for each group. Proliferation ability of the CD27⁻ B-cell population was similar for infected and uninfected individuals: CD27⁻ B cells underwent significant rounds of clonal expansion when stimulated with anti-Ig but remained almost undivided after incubation with CD40L alone or in combination with IL-4 or IL-10. By contrast, proliferation ability of the CD27⁺ B-cell population was different for infected and uninfected individuals: CD27⁺ B cells of persistently infected patients underwent approximately two divisions fewer than those of healthy donors or spontaneous resolvers following incubation with T-cell-derived stimuli. Calculation of the overall proportion of cells that had undergone one or more divisions after stimulation proved that the reduced proliferative capacity of CD40-activated CD27⁺ B cells of persistently infected patients reached statistical significance (P < 0.05) in the presence of cytokines and was particularly enhanced by IL-4 (Fig. 4C). Five-day in vitro stimulation with recombinant HCV core or NS3 protein failed to induce considerable proliferation of CD27⁻ or CD27⁺ B cells purified from either persistently infected patients or spontaneous resolvers (data not shown).

CD27⁺ B cells of persistently HCV-infected patients exhibit enhanced antibody production following in vitro noncognate T-cell help.

Antibody secretion is the hallmark of terminal B-cell differentiation and hence can be experimentally exploited as a readout of this process. To evaluate the capacity of naive and memory B cells to differentiate into ASC following in vitro stimulation, the amounts of secreted Ig in 5-day culture supernatants of immunomagnetically purified CD27⁺ and CD27⁻ B cells were determined. Standard ELISAs were used to quantitate either nonspecific or virus-specific antibodies.

Total Ig levels in cultures of CD27⁻ B cells of infected and uninfected individuals were similar: negligible concentrations of IgG and IgM were detected upon stimulation with CD40L, either alone or in combination with IL-4. Low, but detectable, concentrations of IgM were detected upon stimulation with CD40L plus IL-10; substantial concentrations of IgG and IgM were detected upon stimulation with anti-Ig. Total Ig levels in cultures of CD27⁺ B cells were several orders of magnitude greater than those detected in CD27⁻ B-cell cultures. IgG levels were generally more abundant than IgM levels, irrespective of the culture conditions. However, IgG levels produced by CD40-activated CD27⁺ B cells of persistently infected patients were higher than levels for healthy donors and spontaneous resolvers. This difference became statistically significant in the presence of cytokines (P < 0.05) and was drastically amplified by IL-10 (Fig. 5A). HCV-specific IgG concentrations were detected only in supernatants of CD27⁺ B-cell cultures and in larger quantities after stimulation with CD40L and IL-10. In particular, concentrations of core-specific IgG produced by cells of persistently infected patients were about two- to threefold higher than those produced by spontaneous resolvers (Fig. 5B).

FIG. 5.

In vitro antibody secretion of CD27⁻ and CD27⁺ B cells from HD, PI, and SR. Purified CD27⁺ and CD27⁻ B cells were cultured with the indicated stimuli. Five days later, supernatants were analyzed by ELISA. (A) Nonspecific Ig concentrations. (B) HCV-specific antibody levels (optical density [OD]). The dashed horizontal line indicates the cutoff.

CD27⁺ B cells of persistently HCV-infected patients lose CD27 and are more prone to activation-induced cell death following in vitro noncognate T-cell help.

Strict control of B-cell survival is essential to permit timely production of antibodies while preventing immunopathology. Because CD27⁺ B cells of persistently HCV-infected patients showed impaired T-cell-dependent proliferation, we sought to establish whether this population was more prone to activation-induced cell death than the corresponding populations of healthy donors and spontaneous resolvers.

To monitor CD27⁺ B-cell survival over time, we used annexin V, a sensitive probe that binds phosphatidylserine on the surface of apoptotic cells (71). Purified CD19⁺ cells were stained with annexin V and anti-CD27 antibody before and after 2 and 5 days of stimulation with CD40L and IL-4, the stimulatory combination that provided the most powerful effect on cell proliferation. As demonstrated in Fig. 6, by comparison of flow cytometry analyses of a representative individual from each group, the proportion of CD27⁺ cells from persistently infected patients drastically declined over time. This decrease paralleled an increase in the CD27⁻ cell proportion and was already evident on day 2, when the number of annexin V⁺ cells was still negligible and the viable cell count was comparable to that on day 0 (95% ± 1% on day 2 versus 94% ± 1% on day 0). Thus, loss of memory B cells was due to a downregulation of CD27 expression and not to cell death. On day 5, there was a dramatic increase in the frequency of apoptotic cells in cultures from persistently infected patients compared to those for the two uninfected groups receiving identical stimuli. Apoptotic cells came mostly from the CD27⁻ subset of B cells, which became overrepresented at this time of the culture period.

FIG. 6.

In vitro survival kinetics of B cells from HD, PI, and SR. Purified CD19⁺ cells were cultured with CD40L plus IL-4, stained with annexin V and anti-CD27 antibody, and analyzed by flow cytometry at different time points. Five thousand events were acquired for each sample, and dot plots represent one of eight individuals from each group. Numbers indicate the percentage of cells in each quadrant.

The sensitivity of CD27⁺ B cells to apoptosis was more rigorously explored by use of JC-1, a cationic dye that signals the loss of mitochondrial potential in apoptotic cells which do not yet exhibit morphological changes (66), and DAPI, a DNA-specific probe (49). Purified CD27⁺ B cells were cultured with CD40L plus IL-4, stained with either JC-1 or DAPI, and analyzed by epifluorescence microscopy. Unlike those from uninfected individuals, CD27⁺ B cells from persistently infected patients were already committed into apoptosis after 2 days and displayed chromatin condensation and nuclear fragmentation after 5 days (Fig. 7A and B). As further revealed by flow cytometry analysis of JC-1-stained cells, the percentage of apoptotic cells showing primary green fluorescence on day 2 was 47% ± 12.5% (data not shown).

FIG. 7.

Apoptotic signals in cultured CD27⁺ B cells from HD, PI, and SR. Purified CD27⁺ cells were cultured with CD40L plus IL-4, stained with either JC-1 or DAPI, and examined by epifluorescence microscopy. (A) Representative micrographs of cells stained with JC-1 (top and middle rows) and DAPI (bottom row) after 2 and 5 days of stimulation, respectively. JC-1 fluoresces red (under green excitation) and yellow-orange (under blue excitation) in cell areas with high mitochondrial potential and green (under blue excitation) in areas of lower potential. DAPI forms blue fluorescent complexes with double-stranded DNA. Apoptotic cells can be identified by a decrease in JC-1 red fluorescence and an increase in JC-1 green fluorescence in the cytoplasm. Apoptotic nuclei can be identified by a condensed chromatin gathering at the periphery of the nuclear membrane or a total fragmented morphology of nuclear bodies. (B) High magnification of apoptotic CD27⁺ B cells from PI. Cytoplasmic accumulation of green monomeric JC-1 (left panel) and nuclear fragmentation (right panel) are shown.

CD27⁺ B cells of persistently HCV-infected patients are mostly committed to the ASC lineage.

CD38 is a marker for human bone marrow plasma cells (4) and has been used extensively to monitor in vitro generation of ASC from memory B cells. Having found that CD27⁺ B cells from persistently infected patients increased Ig production over 5 days but began to die on day 2, it was of interest to investigate the temporal relationship between their ASC commitment and apoptosis initiation. To this end, purified CD27⁺ B cells were stained with anti-CD38 antibody before and after 2 days of stimulation with CD40L alone or in combination with cytokines.

As demonstrated by the flow cytometry values presented in Table 3, the percentage of cells expressing CD38 was almost 50% (48.8% ± 2.4%) at the time of isolation from infected patients (day 0) and reached 71.2% ± 6.3% after 2 days of culture with CD40L and IL-10. This was not the case for healthy donors and spontaneous resolvers, whose percentages of cells expressing CD38 were, respectively, 13.4% ± 2.4% and 12.8% ± 2.4% of the purified CD27⁺ cells at the time of isolation (day 0) and never exceeded 30% (26.8% ± 1.6% and 31.6% ± 2.3%, respectively) after 2 days of culture with CD40L and IL-10. Overall, the proportion of CD38-expressing CD27⁺ B cells was consistent with the frequency of IgG-ASC detected ex vivo by ELISPOT assays (Fig. 3).

TABLE 3.

Effect of T-cell-dependent stimulation on CD38 expression of CD27⁺ B cells

Time point and culture	% CD27+ cells (mean ± SEM) expressing CD38a
	HD (n = 20)	PI (n = 20)	SR (n = 8)
Before stimulation (day 0)	13.4 ± 2.4	48.8 ± 12.2	12.8 ± 3.6
After 2 days of stimulation
CD40L	18.2 ± 4.6	52.1 ± 9.2	22.0 ± 2.2
CD40L + IL-4	19.2 ± 2.2	62.4 ± 10.2	28.6 ± 0.8
CD40L + IL-10	26.8 ± 1.6	71.2 ± 6.3	31.6 ± 2.3

^aP was < 0.05 by one-way analysis of variance comparisons of PI, HD, and SR.

DISCUSSION

This study overcomes several apparent inconsistencies related to in vivo antibody production in chronic HCV infection and in vitro B-cell behavior. We sampled only the peripheral blood to avoid invasive procedures. Even so, the effects of viral persistence on the cells of the peripheral blood may mirror, at least qualitatively, its effects on cells in other tissues, such as liver, lymph nodes, and bone marrow.

A 10 to 20% increase in serum Ig is commonly observed to occur during chronic HCV infection. This increase is polyclonal and is determined primarily by increased levels of IgG, despite the presence of circulating IgM with RF activity in more than half of infected patients. Serum IgGs include both HCV-specific antibodies associated with the extent of viral replication and nonspecific antibodies. Although we tested only two antigens (HCV core and NS3), this finding is consistent with the emergence of HCV escape mutants (21) and may indicate that the antibody pool mounted against HCV contains binding antibodies but not virus-neutralizing antibodies. HCV recovery diminishes HCV-specific antibodies, thus confirming the report that antibodies can completely disappear 10 to 20 years after spontaneous viral clearance (67). Nonspecific antibodies comprise antibodies to unrelated recall antigens to which the patients are immune, such as DT and TT, two common vaccination antigens. Nonspecific antibodies increase in parallel with viral load, but HCV recovery resolves serum Ig elevation.

HCV infection induces a substantial and selective increase in virus-nonspecific IgG-ASC numbers but not IgM-ASC numbers in the blood, in agreement with immunization studies demonstrating that IgM-ASC represent less than 2% of all circulating ASC (51). Ig elevation and increased frequency of nonspecific IgG-ASC are probably associated phenomena, since overproduction of IgG by circulating ASC tends to increase serum IgG concentrations. Unlike nonspecific ASC, HCV-specific ASC are not detectable ex vivo in the blood of persistently infected individuals and therefore do not reflect the frequency of serum anti-HCV antibodies.

Two questions arise considering these findings. (i) How does HCV stimulate virus-nonspecific ASC differentiation and IgG production by B cells? (ii) Why do circulating HCV-specific plasma cells make a negligible contribution to the total serum antibodies?

HCV might stimulate B cells in a BCR-independent and therefore unselective fashion. As elegantly demonstrated previously with lymphocytic choriomeningitis virus-infected mice (29), this might occur if specific CD4⁺ T helper cells recognize HCV peptides presented in the context of major histocompatibility complex class II molecules by B cells that have processed viral antigens, irrespective of BCR specificity. BCR-independent uptake and presentation of HCV antigens could theoretically happen through complement or Fc receptor, nonspecific pinocytosis, or cell infection. HCV's ability to productively infect B cells has not yet been fully demonstrated, but it is relatively common to find HCV proteins in B cells that do not harbor antigenomic RNA sequences (60). Among B cells, the most receptive to the BCR-independent stimulation would be the memory B cells because of their ability to proliferate and differentiate into plasma cells in response to T-cell help, even in the absence of cognate restriction (10). In addition, ASC continuously produced in response to HCV might migrate to the bone marrow and drive resident virus-nonspecific plasma cells into circulation. This is highly probable in light of recent data showing that during a secondary immune response, newly generated antigen-specific ASC leave the follicles of secondary lymphoid tissues as plasma blasts and successfully compete with preexisting long-lived plasma cells for occupation of a limited number of survival niches in the bone marrow (41, 47).

Circulating HCV-specific plasma cells make a negligible contribution to the total serum antibodies because most HCV-specific ASC might be confined in a biological compartment that is not in equilibrium with the blood and therefore differs from that sustaining the nonspecific-antibody production. This is not surprising in view of the fact that under chronic inflammatory conditions, B cells bearing antigen-specific receptors can be stimulated to proliferate and differentiate into ASC within ectopic sites (34, 38, 72, 74, 75), including the germinal centers of intraportal lymphoid follicles (42, 43, 53, 59).

There is no evidence of B-cell accumulation in the blood of individuals with persistent HCV infection. Moreover, the B-cell population expressing the CD27 antigen and representing the memory subset is shrunken and therefore does not reflect the overall level of serum antibodies and circulating ASC. Yet, how could a continuous and widespread stimulation process decrease the number of CD27⁺ B cells and, at the same time, increase the production of polyclonal immunoglobulins? The answer to this question comes from the observation of the in vitro behavior of CD27⁺ B cells. When stimulated by CD40 ligand plus cytokines, a mimic of bystander T-cell help in the absence of specific antigen recognition, CD27⁺ B cells from persistently HCV-infected patients do not expand. Instead, they step over terminal differentiation into ASC. Overall, the process is quite rapid and involves loss of surface CD27, prompt release of IgG, and direct progression towards cell death. Specifically, IgG secretion is supposed to occur within 2 days, since CD27 downregulation and apoptosis occur as early as 48 h after stimulation. Such a quick Ig-secreting capacity is supported by the CD38⁺ phenotype of most CD27⁺ cells, which are already committed to the ASC lineage at the time of isolation from patients.

These are probably the most interesting results of our study because they offer a hypothetical model to explain the in vivo predominance of CD27⁻ B cells over CD27⁺ B cells, despite the high levels of circulating antibodies and ASC. They strongly suggest that in the chronic phase of HCV infection the memory B-cell subset is functionally heterogeneous and includes cells which have been stimulated to different extents and have reached different levels in the differentiation path to the ASC lineage. They also suggest caution in confidently delineating bona fide naive B cells from hyperreactive memory B cells that have lost expression of CD27.

It is difficult to reconcile our detection of CD27 downregulation with the reports showing that expression of CD27 is increased on ASC that are generated in vitro and in vivo (5, 20, 28, 31, 70). However, CD27 ligation inhibits terminal differentiation of murine B cells into Ig-secreting plasma cells, and inhibition is more marked for T-cell-dependent stimulations (55). So, under conditions of persisting HCV antigenemia, memory B cells not receiving specific BCR triggering before having T-cell help would be pushed to enhance Ig production and rendered subject to apoptosis by downregulating expression of CD27.

Although we cannot rule out in vitro manipulation influencing cell death, apoptosis of memory B cells could prevent excessive cell accumulation and cull B-cell expansion subsequent to protracted antigenic stimulation. This mechanism would stop the potentially disastrous increase in B-cell proliferation that could be triggered by an endless supply of stimulatory signals. Chronically activated and rapidly proliferating B cells, especially memory B cells undergoing Ig-variable gene hypermutation (19, 30), are at risk of sustaining mutations in proto-oncogenes or tumor suppressor genes (37, 40) that may lead to the development of lymphoproliferative disorders. Chromosomal translocations have frequently been observed to occur in B cells of HCV-infected patients (32), and, interestingly, the majority of circulating B cells of patients with HCV-associated B-cell non-Hodgkin lymphoma or type II cryoglobulinemia have been shown to be CD27⁺ (14).

Heightened sensitivity of memory B cells to BCR-independent T-cell help sustains a constant level of nonspecific serum antibodies and ASC and serves to dampen HCV-specific humoral responses, with detrimental consequences for the production of neutralizing antibodies.