Clonal B cells of HCV-associated mixed cryoglobulinemia patients contain exhausted marginal zone-like and CD21low cells overexpressing Stra13


Correspondence: Dr. Massimo Fiorilli, Department of Internal Medicine and Clinical Immunology, Sapienza University of Rome, Viale dell'Università 37, 00185 Rome, Italy

Fax: +39-06-4463877



A clonal population of B cells expressing a VH1-69-encoded idiotype accumulates in hepatitis C virus (HCV) associated mixed cryoglobulinemia (MC). These cells are phenotypically heterogeneous, resembling either typical marginal zone (MZ) B cells (IgM+IgD+CD27+CD21+) or the exhausted CD21low B cells that accumulate in HIV infection or in common variable immunodeficiency. We show that both the MZ-like and the CD21low VH1-69+ B cells of MC patients are functionally exhausted, since they fail to respond to TLR and BCR ligands. The proliferative defect of VH1-69+ B cells can be overcome by co-stimulation of TLR9 and BCR in the presence of interleukin(IL)-2 and IL-10. The MZ-like VH1-69+ B cells do not express the inhibitory receptors distinctive of CD21low B cells, but display constitutive activation of extracellular signal regulated kinase (ERK) and attenuated BCR/ERK signaling. These cells also express abundant transcripts of Stra13 (DEC1, Bhlhb2, Sharp2, Clast5), a basic helix-loop-helix transcription factor that acts as a powerful negative regulator of B-cell proliferation and homeostasis. Our findings suggest that MZ B cells activated by HCV undergo functional exhaustion associated with BCR signaling defects and overexpression of a key antiproliferative gene, and may subsequently become terminally spent CD21low B cells. Premature exhaustion may serve to prevent the outgrowth of chronically stimulated MZ B cells.


Hepatitis C virus (HCV) is associated witha spectrum of extrahepatic manifestations, the best characterized of which is type II mixed cryoglobulinemia (MC) [[1]]. MC is a benign monoclonal lymphoproliferative disorder of B cells producing rheumatoid factor IgM that, in turn, forms cryoprecipitable immune complexes leading to small vessel vasculitis in vivo [[1]]. In a large proportion of MC patients, the monoclonal rheumatoid factor bears an idiotype encoded by the VH1-69 heavy chain variable gene [[2, 3]]. Several lines of evidence suggest that HCV activates B cells via the cross-reactivity of the VH1-69-encoded idiotype with the E2 glycoprotein of the viral envelope [[4, 5]]. VH1-69+ B cells chronically stimulated by HCV accumulate in the blood of MC patients, and may constitute nearly the totality of circulating B cells [[2]]. Over time, 5–10% of MC patients develop B-cell non-Hodgkin lymphoma, most frequently low-grade splenic marginal zone lymphomas (SMZLs) that carry trisomy 3q and may regress after the eradication of HCV infection [[6-9]]. The molecular identity of premalignant and malignant B-cell clones strongly suggests that lymphoma cells originate from the B cells clonally expanded in MC [[10]].

The monoclonal VH1-69+ B cells that accumulate in the blood of MC patients are phenotypically heterogeneous. Two major populations of these cells can be distinguished [[2, 11-13]]: one resembling typical IgM+IgD+CD27+CD21+ marginal zone (MZ) B cells [[14]], and another phenotypically matching the unusual CD21low B cells that are found in human tonsillar tissues [[15, 16]] and are massively expanded in the blood of patients with common variable immunodeficiency (CVID) [[17-19]], rheumatoid arthritis [[18]], or HIV infection [[20]].

The CD21low B cells expanded in CVID, in HCV+ MC, and in HIV-infected patients as well as those found in the tonsil show signs of previous activation and proliferation, fail to proliferate in response to B-cell stimuli and are unable to flux calcium upon BCR cross-linking, although they are in general poised to secrete high levels of immunoglobulins [[11-13, 15-20]]. In addition, CD21low B cells express a peculiar array of homing and inhibitory receptors, the latter including CD22, CD72, CD32b, CD85j, CD85d, Fc receptor-like 4 (FCRL4), and sialic acid binding Ig-like lectin 6 (Siglec-6) [[11, 15-20]]. The contribution of these inhibitory receptors, and particularly of FCRL4 and Siglec-6, to the dysfunction of CD21low B cells is supported by the partial recovery of the proliferative capacity and of effector function upon silencing of these genes with siRNA [[21]].

We recently suggested that the CD21+ MZ-like VH1-69+ B cells of MC patients also fail to proliferate in response to TLR9 ligation [[13]]. Here, we characterized the responses of these cells to B-cell stimuli and investigated inhibitory mechanisms. We show that MZ-like VH1-69+ B cells are functionally exhausted since they fail to respond to TLR and BCR ligands, although their proliferative defect can be overcome by co-stimulation of TLR9 and BCR in the presence of interleukin(IL)-2 and IL-10. In addition, MZ-like VH1-69+ B cells display increased constitutive and decreased BCR-induced phosphorylation of extracellular signal regulated kinase (ERK); this pattern, however, was also observed in a subpopulation of MZ B cells from healthy individuals. Finally, although the CD21+ MZ-like VH1-69+ B cells do not express the inhibitory receptors of CD21low B cells, they strikingly overexpress Stra13, a transcriptional repressor that acts as a key negative regulator of activation and cell cycle progression in B cells [[22-24]]. Our results indicate that the VH1-69+ MZ B cells activated by HCV undergo premature exhaustion associated with, and probably at least in part sustained by, overexpression of Stra13.


Patients and phenotyping of VH1-69+B cells

By making use of the G6 antibody we could identify, among 40 MC patients screened, 12 patients with high proportions of circulating VH1-69+ B cells. The characteristics of these patients are summarized in Table 1. As previously reported [[2, 11-14]], the VH1-69+ B cells of MC patients could be subdivided in two major populations, one resembling typical IgM+IgD+CD27+CD21high MZ B cells and the other displaying a peculiar CD21low phenotype (Fig. 1A). The CD21low VH1-69+ B cells of MC patients were mostly CD27+, and closely resembled the CD21low B cells expanded in CVID [[17-19]] and in HIV infection [[20]] by the expression of a peculiar pattern of inhibitory (FCRL4+, CD95+, CD22+) and homing (CD11c+, CXCR5low, CCR7low, CXCR3high) receptors. By contrast, the MZ-like VH1-69+ B cells did not express this receptor pattern (Supporting Information Fig. 1). The MZ-like and the CD21low VH1-69+ B cells were variably represented in different MC patients, with either balanced proportions or a marked predominance of one population over the other (Table 1).

Table 1. Clinical and immunological features of mixed cryoglobulinemia (MC) patients
CaseAge/sexDiagnosisCryocrit (%)Total B cells/μLVH1-69+ (percent of B cells)MZ-like VH1-69+ (percent of VH1-69+)
1134/MMC + SMZL1711,1429899
1279/FMC + SMZL2.57639772
Figure 1.

The VH1-69+CD27+ clonal B cells of HCV+ mixed cryoglobulinemia (MC) patients fail to proliferate in response to TLR stimulation. (A) Phenotyping of B cells from a representative MC patient shows that the VH1-69+CD20+IgM+CD27+B cells can be further subdivided into CD21lowCD11c+ cells and CD21highCD11cMZ-like cells. (B) Strategy for the analysis of the proliferative responses of VH1-69+ and VH1-69B cells is shown. At the end of a 5-day culture in the presence of stimuli, CFSE labeled peripheral blood mononuclear cells (PBMCs) were permeabilized and stained with antibodies to CD20, IgM, and VH1-69 (or with mouse IgG as control). VH1-69+ and VH1-69B cells were analyzed for CFSE fluorescence (histograms). Percentages in the histograms denote the percent of divided cells (number of precursor B cells that have undergone at least one division). The CD20low/neg cells in the VH1-69 gating region are mostly cIgMhigh plasmablast-like cells. (C, D) In an independent experiment, the proliferative responses of a patient's (C) VH1-69+ and (D) VH1-69B cells were analyzed after activation with CpG or with combinations of stimuli and cytokines. (E) Cumulative results of the proliferative responses of VH169+ and VH169B cells of MC patients and of whole B cells from healthy donors (HDs) to CpG (left), R848 (middle), or the indicated combination of stimuli and cytokines (right). VH1-69B cells could not be investigated in two patients (no. 11 and 12 in Table 1) because of their scarcity. Each symbol represents a patient studied once and bars denote the geometric means. *p < 0.005 and **p < 0.0001 using the Mann–Whitney test.

Impaired proliferative responses of MZ-like VH1-69+B cells

We investigated the proliferative responses of VH1-69+ B cells by labeling these cells with the G6 antibody at the end of cultures. The strategy of analysis in a representative patient is illustrated in Fig. 1B–D. The results show that VH1-69+ B cells, unlike autologous VH1-69 B cells, fail to proliferate in response to CpG; in addition, VH1-69+ B cells do not differentiate into CD20low/negcIgMhigh (where cIgMhigh is high cytoplasmic IgM content) plasmablasts (Fig. 1B). In a separate experiment, this patient's B cells were activated with different stimuli (Fig. 1C and D); stimulation with anti-Ig and CD40L did not induce significant proliferation (not shown), but cooperated with CpG in driving a low-level proliferative response of VH1-69+ B cells. The addition of IL-2 and IL-10 strikingly increased the proliferative response of VH1-69+ B cells stimulated with CpG, anti-Ig, and CD40L.

The proliferative responses of VH1-69+ B cells to different B-cell stimuli were investigated in a total of 12 MC patients (Fig. 1E). In 11 of these patients, VH1-69+ B cells responded very poorly to CpG whereas VH1-69 B cells proliferated as vigorously as the B cells of healthy donors (HDs); in a single patient, a case of MC and SMZLs, a significant proportion (25%) of VH1-69+ B cells entered proliferation upon stimulation with CpG. The VH1-69+ B cells from four patients (no. 4, 6, 10, and 12 in Table 1) had predominantly (72–92%) a MZ-like phenotype and, nevertheless, completely failed to proliferate in response to CpG. VH1-69+ B cells also failed to proliferate in response to the ligation of TLR7 with R848 but, surprisingly, autologous VH1-69 B cells responded less efficiently than HDs B cells. The reason for reduced responsiveness of bystander nonclonal B cells of MC patients to TLR7 stimulation is unclear.

We observed that co-stimulation with CpG, anti-Ig, and CD40L in the presence of the B-cell stimulatory cytokines IL-2 and IL-10 increased the proliferative responses of VH1-69+ B cells up to the levels seen with autologous VH1-69 B cells and with HDs B cells (Fig. 1E). We choose to investigate these cytokines because IL-10 was shown to cooperate with IL-2 in enhancing CpG-induced proliferation of naïve and memory B cells from HDs, suggesting a fine-tuning of microbe-induced B-cell activation by these cytokines [[25]]. In addition, responsiveness to IL-2 and IL-10 was investigated in CD21low FCRL4+ B cells derived from tonsillar tissue [[15]], and it was reported that these cells proliferated vigorously in response to cytokine stimulation but not BCR ligation. By contrast, IL-2 and IL-10 could not override the proliferative defect of CD21low B cells from HIV-infected patients [[20]]. It was suggested [[20]] that this difference might be ascribed to the expression of CD27 by tonsillar but not by HIV patients’ CD21low B cells. Our observation of the responsiveness of VH1-69+ CD27+ B cells of MC patients to IL-2/IL-10 is consistent with this possibility. Thus, increased production of IL-10 related to promoter polymorphisms [[26]] or to the immunoregulatory effects of cryoglobulins [[27]] might facilitate lymphoproliferation and the evolution of MC to lymphoma.

Two of our patients had HCV+ MC evolved to SMZLs. One of these patients had marked monoclonal lymphocytosis with 10,919 VH1-69+ B cells/μL; his lymphoma carried trisomy 3q and regressed after the eradication of HCV infection with antiviral therapy. Interestingly, the VH1-69+ B cells from this patient proliferated significantly (25% of divided cells) in response to CpG (Fig. 1E). However, the mean number of cell divisions performed by proliferating cells (proliferation index) was much lower than that seen with B cells from HDs (1.5 divisions, compared with a mean ± SD of 2.3 ± 0.3 divisions). This suggests that with evolution from MC to SMZLs VH1-69 B cells may become more efficient in entering proliferation, but their ability to undergo multiple divisions remains impaired.

Constitutive activation of ERK and attenuated BCR signaling in MZ-like VH1-69+B cells

B cells constantly exposed in vivo to antigen become anergic because continuous engagement of BCR delivers tolerogenic signaling through ERK phosphorylation [[28, 29]]. Thus, we investigated whether VH1-69+ B cells, which are chronically exposed to antigenic stimulation by HCV, display a similar pattern of constitutive and induced ERK phosphorylation. For these studies, sorting of VH1-69+ B cells before testing could not be done because indirect staining with G6 antibody would cross-link the BCR; in addition, G6 staining of cells fixed/permeabilized by PhosFlow Protocol 1 after BCR stimulation was, in our hands, unreliable. Therefore, for pERK experiments we used negatively purified whole B cells from MC patients with a large predominance of VH1-69+ B cells (70–98% of B cells, Table 1).

Representative PhosFlow analyses of pERK expression by B cells from a MC patient and from a HD is illustrated in Fig. 2A. Collectively, the B cells from MC patients expressed more basal pERK than the B cells from HDs (Fig. 2B). BCR cross-linking with anti-Ig induced a mean fivefold increase of pERK mean fluorescence intensity (MFI) in normal B cells and a mean twofold increase (p < 0.0001) in MC B cells (Fig. 2C). High constitutive pERK expression and attenuated BCR signaling were also observed in VH1-69+ B cells from two patients with MC-associated SMZLs (Fig. 2B and C). In four of the eight MC patients studied, clonal B cells had predominantly (72–99% of VH1-69+ B cells) an MZ-like phenotype. Thus, these findings provide evidence for deregulated BCR/ERK signaling both in the MZ-like and in the CD21low VH1-69+ B cells of MC patients.

Figure 2.

VH1-69+B cells of mixed cryoglobulinemia (MC) patients display high constitutive and attenuated BCR-induced expression of pERK. (A) Peripheral blood mononuclear cells (PBMCs) of a representative MC patient, with 87% pure VH1-69+B cells, and of a representative healthy donor (HD), were left unstimulated or stimulated with anti-Ig. Cells were then co-stained with anti-CD20 and with either anti-pERK or control antibody. Unstimulated and stimulated CD20+B cells were analyzed for pERK expression (histograms). Open histograms denote staining with anti-pERK antibody, and gray-filled histograms staining with control antibody. (B–E) The cumulative results on the expression of pERK by unstimulated and anti-Ig-stimulated B cells from eight MC patients with large predominance (more than 70% of B cells) of VH1-69+B cells, and by (B, C) whole B cells or (D, E) B-cell subsets from healthy donors (HDs) are shown. Open triangles denote MC patients with MC and splenic marginal zone lymphomas (SMZLs). Each symbol represents a single patient studied once and bars denote the geometric means. *p < 0.01 and **p < 0.0001 using the Mann–Whitney test.

Interestingly, the BCR-induced pERK fluorescence profile in B cells from HDs was bimodal (Fig. 2A), suggesting that ERK phosphorylation was differently regulated in distinct B-cell subsets. By contrast, pERK fluorescence in stimulated MC B cells was unimodal, seemingly reflecting the large predominance of homogeneous clonal B cells in these samples. By electronic gating, we observed that the CD27+ B cells from HDs had higher constitutive pERK expression than naive CD27 B cells (Fig. 2D). Upon BCR cross-linking, the fold increase of pERK was, on average, similar in naïve and in CD27+ B cells of HDs (Fig. 2E). Thus, we further dissected BCR/ERK signaling in switched (IgMCD27+) and MZ (IgM+CD27+) B cells of HDs (Fig. 3A), and found that both cell types had constitutively high pERK levels (Fig. 3B and Supporting Information Fig. 2). Upon BCR cross-linking MZ B and switched B cells, but not naïve B cells, displayed two distinct populations, one with no or minimal increase and the other with marked increment of pERK fluorescence (Fig. 3B and Supporting Information Fig. 2). These findings suggest that BCR/ERK signaling is attenuated in subpopulations of MZ and memory B cells. Thus, the abnormal pattern of BCR/ERK signaling in VH1-69+ B cells of MC patients may be similar to that observed in a proportion of normal resting MZ B cells.

Figure 3.

Subpopulations of MZ and switched B cells from healthy donors (HDs) display attenuated BCR/ERK signaling. (A) Peripheral blood mononuclear cells (PBMCs) of a representative HD (out of 12 studied), left unstimulated or stimulated with anti-Ig, were co-stained with antibodies to CD20, IgM and CD27, and with either anti-pERK or control antibody. CD20+ naïve (IgM+CD27), MZ (IgM+CD27+), and switched (IgMCD27+) B-cell subsets were analyzed for pERK expression. (B) pERK expression by unstimulated (top) or anti-Ig-stimulated (bottom) B-cell subsets. Open histograms denote staining with anti-pERK antibody, and gray-filled histograms staining with control antibody. Data are representative of 12 patients.

Overexpression of Stra13 by MZ-like VH1-69+B cells

Stra13 is a basic helix-loop-helix transcriptional repressor that suppresses proliferation in a variety of cell types [[30, 31]], including B cells [[23, 24]]. Interestingly, Stra13 is rapidly upregulated in activated MZ but not in follicular B cells [[32]]. Thus, we investigated the expression of this gene in MZ-like VH1-69+ B cells of MC patients. Since, to our knowledge, the expression of Stra13 by human peripheral blood B cells was not previously investigated, we preliminarily examined its expression in fluorescence activated cell sorter (FACS) purified B-cell subsets derived from the peripheral blood of two HDs. We found that naïve (IgM+CD27), MZ (IgM+CD27+), and switched memory (IgMCD27+) B cells expressed similar levels of Stra13 transcripts (not shown). Thus, we compared Stra13 transcripts in whole peripheral blood B cells from five HDs and in VH1-69+ B cells (more than 80% pure) from five MC patients (Fig. 4). We found that Stra13 transcripts were, on average, 6.4-fold more abundant in VH1-69+ B cells than in HDs B cells. In four of these five MC patients, VH1-69+ B cells had predominantly (72% to 99%) an MZ-like phenotype, whereas in one patient 60% of them were CD21low. Noteworthy, VH1-69+ B cells from the two patients with MC and SMZLs expressed less abundant Stra13 transcripts than those from patients with uncomplicated MC.

Figure 4.

Real-time PCR analysis of Stra13 transcripts in pooled B cells from five healthy donors (HDs) and in enriched (more than 80% pure) VH1-69+B cells from five mixed cryoglobulinemia (MC) patients. Triangles denote patients with MC and SMZLs. Each symbol represents a single patient and bars represent the geometric means; data are representative of two separate experiments.


Recently, an expanded population of functionally exhausted CD21low B cells has been characterized in patients with CVID [[17-19]] or HIV infection [[20]] and in the normal tonsillar tissue [[15, 16]]. CD21lowB cells display a peculiar pattern of homing and inhibitory receptors [[15-20]], and it is believed that their dysfunction largely depends on the expression of the latter receptors since their silencing partially restores proliferative responses [[21]]. The precursors of CD21low B cells have not yet been identified, and it is not known whether these cells are a unique and stable population or if they represent a phenotypic subset of a larger population of exhausted B cells.

We provide evidence that the MZ-like VH1-69+ B cells of MC patients are functionally exhausted much like their CD21low counterparts. Importantly, dysfunction of the VH1-69+ MZ-like B cells cannot be ascribed to the inhibitory receptors typically expressed by CD21low B cells. Our data suggest that attenuated BCR signaling and overexpression of Stra13 may contribute to dampening the responsiveness of MZ-like VH1-69+ B cells.

Two recent papers [[11, 12]] described anergy of CD21low B cells from HCV+ MC patients. Anergy of these cells was reflected by decreased calcium mobilization, reduced upregulation of activation markers, and failure to proliferate in response to BCR triggering. At variance with our findings, both papers reported that the CD21high MZ-like B cells from these patients were not anergic according to these criteria, and one paper [[12]] reported that the CD21low B cells remained responsive to TLR9 stimulation. Differences in the experimental design might account for these discrepancies. In both studies, functional assays were done with cells sorted or analyzed according to their CD21low or CD21+ phenotype without discriminating for VH1-69 expression. Thus, significant amounts of nonclonal B cells could have contaminated the CD21+ MZ-like cell fractions in calcium mobilization experiments [[13]]. The claim of Terrier et al. [[12]] for preserved TLR9 responsiveness of CD21low clonal B cells is intriguing, since the similar CD21low B cells that seed the human tonsil and are expanded in patients with CVID or HIV infection fail to proliferate in response to CpG [[15, 17, 19-21]]. The conclusion for preserved TLR9 responsiveness in CD21low B cells of MC patients was based on the observation of a more efficient upregulation of activation markers after stimulation with CpG compared with BCR triggering. However, the proliferative responses of CD21low B cells to CpG were not reported while, in keeping with our findings, it was shown that these cells proliferated to some extent upon combined stimulation with anti-IgM, CD40L, and CpG. Thus, lack of data on the proliferative responses to CpG alone in that study precludes, in our opinion, firm conclusions on TLR9 function in CD21low B cells of MC patients.

At first glance, our findings of constitutive activation of ERK and attenuated BCR signaling in VH1-69+ B cells are reminiscent of the molecular signature of anergy induced by continual BCR occupancy [[28, 29]], in keeping with the notion that VH1-69+ B cells are chronically exposed to the antigenic pressure of HCV. However, some features clearly distinguish VH1-69+ B cells from murine B cells made anergic by tolerogenic BCR/ERK signaling. First, tolerant murine B cells fail to secrete antibody but retain intact proliferation in response to CpG [[28]], a pattern opposite to VH1-69+ B cells. Furthermore, MZ-like VH1-69+ B cells express high levels of surface IgM whereas surface IgM is low in several models of B-cell anergy [[33]]. Finally, constitutive activation of ERK and defective BCR signaling appear not to be unique to pathologic MZ B cells, since we observed it in MZ B cells from HDs. Constitutive phosphorylation of ERK was previously described [[34, 35]] in resting murine B1a B cells, which resemble VH1-69+ B cells by the expression of CD5 [[2]]. Defective BCR signaling and reduced proliferation in response to anti-IgM were demonstrated in murine B1a B cells, but these cells proliferated in response to the stimulation of TLR4 with lipopolysaccharide (LPS) [[36]]. More recently, it has been shown that putative human B1 B cells display tonic intracellular signaling through phosphorylated PLC-γ2 and Syk [[37]]. It has been suggested that B1 B cells are programmed to maintain BCR signaling in a condition of anergy, which can be overcome by alternative pathways such as TLR and CD40, especially in the context of microbial infections [[38]].

Our findings indicate that the exhausted MZ-like VH1-69+ B cells of MC patients differ both from tolerant B cells [[28, 29]] and from normal B1 B cells [[34-38]] because, although like those cells they have tonic constitutive and attenuated BCR-induced signaling, they fail to proliferate after TLR9 ligation. They also differ from exhausted CD21low B cells because they lack the inhibitory receptors typical of these cells [[15-21]]. Thus, alternative inhibitory mechanisms should play a role in the dysfunction of MZ-like VH1-69+ B cells. We choose to investigate a possible role for Stra13, a negative regulator of B cells, because this gene is rapidly upregulated in murine MZ B cells activated by immunization with Streptococcus pneumoniae [[32]].

Stra13 is a basic helix-loop-helix transcription factor that negatively regulates B-cell proliferation [[23, 24]] by a mechanism involving histone deacetylase [[30]] and cyclin D1 [[31]], and is involved in the control of apoptosis [[39, 40]]. Stra13 is a key regulator of lymphocyte homeostasis, since Stra13−/– mice develop autoimmune disease and accumulation of activated T and B cells [[22]], whereas Stra13 transgenic mice have impaired development of B and T cells [[23]]. Stra13 mRNA is upregulated in murine pro-B, immature and splenic B cells, while it is downregulated in pre-B and in germinal center B cells [[23]]. In vitro activation of splenic B cells with a variety of stimuli including CD40L, anti-IgM, and LPS rapidly down-regulates Stra13 expression [[24]]. Our observation that normal human naïve, MZ, and switched memory B cells expressed similar levels of Stra13 is in keeping with a resting condition of these mature B-cell populations. Interestingly enough, MZ B cells appear to have an opposite behavior since they upregulate Stra13 within 1 h after intravenous immunization of mice with S. pneumoniae [[29]]. The functional relevance of upregulation of Stra13 in activated murine MZ B cells has not been investigated so far. Our finding of an association of increased Stra13 expression with reduced proliferative responses in MZ-like B cells may help to highlight a role for Stra13 in human MZ B-cell regulation.

Murine MZ B cells, together with B1 B cells, play an important role in the early defense from blood-borne infections and, upon encounter with particulate microbial antigens, produce massive waves of plasmablasts [[41]]. Interestingly, infection with Ehrlichia muris is accompanied by the expansion of an unusually large population of CD11c+ plasmablasts [[42]], reminiscent of human CD21lowCD11c+ B cells. Although the propensity of MZ to expand rapidly and robustly is functional to the early defense against microorganisms, it may also predispose to neoplasia [[43]]. This risk is particularly relevant when the microbial stimulus persists for a long time, as it occurs with chronic HCV or with Helicobacter pylori infection [[44]]. Thus, the early expression by activated human and murine MZ B cells of Stra13, a powerful negative regulator of cell proliferation, may serve to constrain their outgrowth and to attenuate the risk of neoplastic transformation.

In conclusion, we provide evidence that the MZ-like and the CD21low VH1-69+ clonal B cells expanded in HCV-associated MC represent different phenotypic aspects of a single population of exhausted cells. Phenotypic characteristics suggest that prematurely exhausted MZ-like VH1-69+ B cells take subsequently the shape of terminally spent CD21low B cells [[13]]. Exhaustion of MZ-like VH1-69+ B cells is not caused by the inhibitory receptors typical of CD21low B cells, but could be at least in part related to the overexpression of Stra13. The significance of attenuated BCR/ERK signaling in MZ-like VH1-69+ B cells, which is also observed in normal resting MZ and B1 B cells, needs to be clarified. It is possible that also in other human immunological disorders, such as HIV infection and CVID, CD21low B cells represent one phenotypic subset within an expanded pool of exhausted MZ B cells. In this regard, it would be of interest to investigate whether in those conditions there is accumulation of exhausted B cells with an MZ-like phenotype.

Materials and methods

Study subjects

Forty MC patients, all HCV-positive by serology and viremia, were screened by staining with the G6 antibody for an expansion of circulating VH1-69+ B cells. Twelve patients, two of whom had MC complicated by SMZLs, had increased proportions of VH1-69+ B cells (10– 98% of total B cells, median 83%) compared with HDs (less than 3%), and were used for this study. Seven of the patients, including the two patients with SMZLs, were untreated, whereas five patients were nonresponders to antiviral therapy with pegylated interferon and ribavirin, which was terminated at least 2 years before the study. All the patients had clinical signs of small vessel vasculitis. Forty-two HDs were studied as controls. All subjects provided informed consent, in accordance with the Institutional Review Board of the Sapienza University of Rome and with the declaration of Helsinki.

Cells and immunophenotyping

Peripheral blood mononuclear cells (PBMCs) were obtained by density-gradient centrifugation. Whole B cells were negatively purified (more than 97%) from PBMCs by immunomagnetic sorting (Dynal); B-cell subpopulations were purified by FACS sorting using a FACSAria instrument (Becton-Dickinson Biosciences, Franklin Lakes, NJ). B-cell phenotyping was performed with combinations of fluorochrome-labeled monoclonal antibodies (Becton-Dickinson Biosciences). FCRL4 was detected either with an unconjugated antibody (kindly provided by G.R.A. Ehrhardt, Atlanta, GA), or with a fluochrome-conjugated antibody (BioLegend, San Diego, CA). We used the G6 and G8 antibodies (kindly provided by R. Jefferis, Birmingham, UK), which recognize distinct epitopes of the VH1-69-encoded protein [[45]], to identify VH1-69+ B cells by flow cytometry. Initial experiments demonstrated that G6 and G8 identified identical populations of B cells in MC patients and, therefore, subsequent studies were done using only the G6 antibody. Unlabeled G6 and anti-FCRL4 were counterstained with FITC- or PE-conjugated goat anti-mouse IgG (Becton-Dickinson Biosciences), using mouse IgG as control. Flow cytometric analyses were done with a FACSCalibur instrument (Becton-Dickinson Biosciences) using the CellQuest (Becton-Dickinson Biosciences) and FlowJo (Tree Star, Ashland, OR) software.

Cell proliferation and differentiation

Cell proliferation was measured by the carboxyfluorescein diacetate succinimidyl ester (CFSE) dilution assay [[46]]. PBMCs were labeled with CFSE (Invitrogen, Life Technologies, Grand Island, NY) and cultured at 2 × 105 cells per well in 96-well U-bottom plates in the absence or presence of the TLR9 ligand CpG (Sigma Genosys The Woodlands, TX; 2.5 μg/mL) or of the TLR7 ligand R848 (Invitrogen; 0.25 μg/mL). Stimulation with R848 was routinely done in the presence of IFN-α (1000 U/mL; Schering-Plough, Kenilworth, NJ), in order to investigate TLR7 function independently on the activity of plasmacytoid dendritic cells. In some experiments, cell were costimulated with F(ab′)2 anti-human Ig (4 μg/mL; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), soluble CD40L (1 μg/mL plus 1 μg/mL enhancer; Alexis Biochemicals, San Diego, CA), IL-10 (10 ng/mL; BD PharMingen, San Diego, CA), and IL-2 (20 units/m; BD Pharmingen). Cell proliferation was measured at day 5 of culture by flow cytometry. The number of cells entering proliferation (percent of divided cells) and the mean number of cell divisions performed by proliferating cells (proliferation index) were calculated using the FlowJo software (Tree Star, Ashland, OR). Before flow cytometric analysis, cells were permeabilized (Permeabilizing-Solution 2; Becton-Dickinson Biosciences) and counterstained with anti-bodies to CD20, IgM, and VH1-69 (G6 or G8 mAb [where mAb is monoclonal antibody]), or with other antibodies according to the experimental design. Plasmablasts were identified as CD20low/neg cells with cIgMhigh.

RNA isolation and real time (RT)-PCR

The expression of the Stra13 gene (also known as DEC1, Bhlhb2, Sharp2, Clast5; reference sequence NM_003670.2) was measured by RT-PCR. Total RNA was isolated from ∼5 × 105 purified B cells using TRIzol (Invitrogen, Life Technologies, Grand Island, NY) according to the manufacturer's instructions. Reference RNA was obtained from whole B-cell populations of five HDs. RT-PCR was performed using the ImPro-IITM Reverse Transcription System (Promega, Madison, WI) following the manufacturer's instructions. Primers, designed with the Primer3 Input 0.4.0 software, were synthesized by Eurofins MWG Operon (Ebersberg, Germany). Quantitative RT PCR was performed using a Go Taq qPCR Master MIX (Promega) on an ABI PRISM 7000 Sequence Detection System (Applied Biosystem, Carlsbad, CA). Quantitative analysis was performed by the ΔCT method [[47]]. Primers used were: β-actin, 5′-CATCGAGCACGGCATCGTCA-3′ (forward) and 5′-TAGCACAGCCTGGATAGCAAC-3′ (reverse); Stra13, 5′-GTACCCTGCCCACATGTACC -3′ (forward) and 5′-GTGACCGGATTAACGAGTGC-3′ (reverse).

PhosFlow assay for pERK

The intracellular pERK content was measured by the BD PhosFlow system as per manufacturer's Protocol 1 (Becton-Dickinson Biosciences). Cells were split in two vials (106 cells each), resuspended in 100 μL of RPMI 1640 containing 5% fetal bovine serum (complete medium), and equilibrated at 37°C for at least 20 min. An equal volume of prewarmed complete medium, either alone (unstimulated control) or containing 20 μg/mL of F(ab′)2 anti-human Ig (Jackson Immunoresearch Laboratories), was then added, and the cells were returned to 37°C for 10 min. Cells were then fixed by the addition of 200 μL of prewarmed PhosFlow Fix Buffer I for 10 min at 37°C, washed twice in PhosFlow Perm/Wash Buffer I, split in two vials, and stained either with anti-pERK1/2-Alexa488 or with mouse IgG-Alexa488 as control. Samples were simultaneously stained with fluochrome-conjugated mAbs to CD20, CD27, and IgM, or with other mAbs as requested by the experimental design. The pERK-specific MFI was calculated by subtracting the MFI values obtained with control mouse IgG from those obtained with anti-pERK antibody.

Statistical analysis

Comparisons were done with the Kruskal–Vallis test or the Mann–Whitney test run by the SPSS PAWStatistics 18.0 software. A p value of less than 0.05 was considered significant.


We are grateful to the patients for their willingness to participate in our study. We thank Ezio Giorda (Ospedale Pediatrico Bambino Gesù, Rome) for FACS sorting, and Roy Jefferis (University of Birmingham, Birmingham, United Kingdom) for generously providing the G6 antibody. This study was supported by the Intramural Research Program of the Sapienza University of Rome.

Conflict of interest

The authors declare no financial or commercial conflict of interest and from AIRC 5xmille grant to GR.


high cytoplasmic IgM content


common variable immunodeficiency


healthy donor


mixed cryoglobulinemia


marginal zone


splenic marginal zone lymphoma