Two overrepresented B cell populations in HIV-infected individuals undergo apoptosis by different mechanisms

Abstract

Perturbations of B cells in HIV-infected individuals are associated with the overrepresentation of distinct B cell populations. Here we describe high extrinsic CD95 ligand (CD95L)-mediated apoptosis in CD10⁻/CD21^lo mature/activated B cells that likely arise from HIV-induced immune activation. In addition, high intrinsic apoptosis was observed in CD10⁺ immature/transitional B cells that likely arise as a result of HIV-induced lymphopenia. CD10⁺ B cells expressed low levels of Bcl-2 and Bcl-xL, consistent with their high susceptibility to intrinsic apoptosis. Higher levels of activated Bax and Bak were induced in CD10⁺ B cells compared with CD95L-treated CD10⁻ B cells, consistent with the greater involvement of mitochondria in intrinsic vs. extrinsic apoptosis. Of interest, both extrinsic apoptosis in CD95L-treated CD10⁻ B cells and intrinsic apoptosis in CD10⁺ B cells were associated with caspase-8 activation. Our data suggest that two distinct mechanisms of apoptosis are associated with B cells of HIV-infected individuals, and both may contribute to the depletion and dysfunction of B cells in these individuals.

HIV infection leads to numerous defects in immune competent cells. Among B cells, a large number of these defects have been associated with HIV-induced immune activation, as evidenced by hypergammaglobulinemia, increased expression of B cell activation markers, decreased responsiveness to B cell stimuli, and increased susceptibility to oncogenesis (1–3). Several studies have demonstrated that after reduction of HIV plasma viremia by antiretroviral therapy, B cell abnormalities subside, especially those thought to be directly or indirectly associated with virus-induced aberrant immune activation (4–6). We have demonstrated previously both in longitudinal and cross-sectional analyses that overrepresentation of a distinct B cell population, defined by reduced expression of CD21 and increased secretion of immunoglobulins (7), is likely to be responsible for several of the B cell defects that have been associated with ongoing HIV replication (8–10). Among the B cell alterations that correlate with HIV plasma viremia and decreased survival of B cells is an increased expression of CD95 accompanied by an increased susceptibility to CD95 ligand (CD95L)-mediated cell death (8).

Recently, we reported the expansion of immature/transitional B cells in HIV-infected individuals with advancing disease (11). The presence in the peripheral blood of these B cells, defined by the expression of CD10, was first reported to be associated with HIV disease progression almost 20 years ago (12), and more recently has been associated with several non-HIV immunodeficiency disorders (13, 14). The homeostatic compensation observed with CD4⁺ T cell lymphopenia has been reported to be associated with increased serum levels of IL-7 in HIV-infected individuals (15, 16). More recently, this IL-7 associated lymphocyte homeostasis has been shown to involve B cells in both HIV and non-HIV immunodeficiencies (11, 17). These data suggest that, in contrast to the influence of HIV-mediated immune activation on mature B cell differentiation, the appearance of immature/transitional B cells in HIV disease may be a consequence of HIV-induced lymphopenia.

The regulation of B cell development and function relies heavily on both intrinsic and extrinsic pathways of apoptosis (18). The extrinsic pathway, initiated by the ligation of death receptors such as CD95 and other members of the TNF superfamily of receptors, is generally involved with removal of activated immune cells after they have performed their function or as a result of incomplete or inappropriate activation (19). In the setting of HIV disease and high plasma viremia, activated B cells also express high levels of CD95, leading to high susceptibility to CD95L-mediated apoptosis (8). The intrinsic or mitochondrial pathway of apoptosis, which is controlled by members of the Bcl-2 family, plays an important role in the ontogeny of B cells (20). Patterns of expression for several anti- and proapoptotic members of the Bcl-2 family have been shown to dictate survival potential in maturing B cells in mice, with immature B cells displaying a high ratio of expression of pro- to antiapoptotic members of the Bcl-2 family (21). In humans, little is known regarding the developmental checkpoints of maturing B cells, although immature/transitional B cells were recently shown to express low levels of Bcl-2 (13).

In the present study, we delineate the mechanisms of apoptosis in two distinct B cell populations that are expanded in the setting of HIV disease.

Results

Phenotypic Distinctions Between Two B Cell Populations That Are Overrepresented in HIV-Infected Individuals with Active Disease.

We described previously B cells from HIV-infected individuals that exhibited increased susceptibility to CD95L-mediated apoptosis; these cells reflected an overrepresented population of activated CD95⁺ B cells expressing reduced levels of CD21 (8). More recently, we identified a second overrepresented population of CD10⁺ B cells in HIV-infected individuals with advancing disease (11), a fraction of which also expressed reduced levels of CD21 (Fig. 1A). These two B cell populations represented close to 50% of the total B cells in the peripheral blood (Fig. 1A), compared with <15% in HIV-negative and -infected aviremic individuals and >50% in HIV-infected individuals with more advanced disease (7, 11). To determine which of the two overrepresented B cell populations was responsible for the increased susceptibility to CD95L-mediated apoptosis (8), levels of CD95 were measured on each population shown in Fig. 1A. In HIV-infected individuals with active disease, levels of CD95 were significantly higher in the CD10⁻/CD21^lo B cells, compared with CD10⁻/CD21^hi B cells [Fig. 1B and supporting information (SI) Table 1]. CD10⁻/CD21^hi B cells account for the vast majority of B cells in HIV-negative and -infected aviremic individuals (7, 11). Expression of CD95 was lowest on CD10⁺ B cells (Fig. 1B), consistent with their immature/transitional status (11, 14). To verify that increased expression of CD95 was also associated with increased susceptibility to CD95L-mediated apoptosis, B cells of HIV-infected individuals with active disease were incubated in the presence or absence of CD95L for 2 h, and levels of apoptosis were measured by Annexin V staining relative to the expression of CD21 and CD10. As shown in Fig. 2A, the majority of the CD95L-mediated apoptosis was confined to the CD10⁻/CD21^lo B cells, whereas CD10⁺ and CD10⁻/CD21^hi B cells demonstrated minimal susceptibility to CD95L-mediated apoptosis.

Fig. 1.

Three phenotypically distinct B cell populations identified in HIV-infected individuals with active disease. (A) Delineation of three B cell populations in peripheral blood of a representative HIV-infected individual with active disease: CD10⁺ immature/transitional, CD10⁻/CD21^hi mature, and CD10⁻/CD21^lo mature/activated B cells. Percent expression within each B cell population was determined for cell surface CD95 (B), and intracellular Ki-67 (C), Bcl-2 (D), and Bcl-x_L (E).

Fig. 2.

High extrinsic apoptosis in CD10⁻/CD21^lo B cells and high intrinsic apoptosis in CD10⁺ B cells. (A) Annexin V binding in B cells after 2 h in culture in the absence or presence of CD95L reveals that extrinsic CD95L-mediated apoptosis is restricted to CD10⁻/CD21^lo B cells. (B) Annexin V binding in CD10⁻ and CD10⁺ B cell fractions after 16 h in culture reveals significantly higher intrinsic apoptosis in CD10⁺ vs. CD10⁻ B cells. (C) Correlation between level of apoptosis after 6 h incubation and level of immaturity within the CD10⁺ B cell fraction, as measured by percentage of CD10⁺⁺/CD21^lo B cells within this fraction.

Considering that increased expression of CD95 on B cells was previously associated with HIV-induced immune activation and increased cell turnover (8), we also measured levels of Ki-67 relative to each B cell population shown in Fig. 1A. As depicted in Fig. 1C and SI Table 1, levels of Ki-67 were significantly higher in the CD10⁻/CD21^lo B cells, compared with both CD10⁻/CD21^hi and CD10⁺ B cells. The forward scatter of the various B cell populations was also very distinctive, with the activated and cycling CD10⁻/CD21^lo B cells clearly much larger than the small and more resting CD10⁺ B cells (our unpublished data).

Low Expression of Antiapoptotic Bcl-2 Proteins in CD10⁺ B Cells Is Associated with High Susceptibility to Intrinsic Apoptosis.

While investigating susceptibility to extrinsic apoptosis, we observed that CD10⁺ B cells, although being refractory to CD95L-induced apoptosis, exhibited a higher background apoptosis ex vivo compared with CD10⁻/CD21^hi B cells (Fig. 2A). To gain insight into the mechanisms of cell death in immature/transitional B cells in HIV-infected individuals, we measured intracellular levels of various members of the Bcl-2 family. In agreement with reports on both human and murine immature B cells (13, 22, 23), the CD10⁺ immature/transitional B cells of HIV-infected individuals with active disease expressed low to undetectable levels of antiapoptotic Bcl-2 family members Bcl-2 and Bcl-x_L (Fig. 1 D and E); these levels were significantly lower compared with both of the CD10⁻ mature B cell populations (SI Table 1). These data predict that CD10⁺ B cells would manifest a higher propensity to intrinsic apoptosis compared with mature B cell populations because of low expression of antiapoptotic members of the Bcl-2 family.

To evaluate the survival potential of CD10⁺ B cells ex vivo, B cells of HIV-infected individuals with active disease were separated into CD10⁺ and CD10⁻ fractions and cultured at 37°C for 16 h. As shown in Fig. 2B, a median of 86% of B cells in the CD10⁺ fraction were Annexin V⁺ after 16 h in culture, a significantly higher level compared with 53% in the CD10⁻ fraction (P < 0.01). A more immature B cell subset within the immature/transitional B cell population was defined previously by the expression of high-intensity CD10 and low-intensity CD21 (CD10⁺⁺/CD21^lo) and was associated with more advanced HIV disease (ref. 11 and our unpublished data). When data on levels of apoptosis in CD10⁺ B cell fractions were compiled on a group of HIV-infected individuals with varying levels of disease, a direct and highly significant correlation was observed between the percentage of CD10⁺⁺/CD21^lo B cells within the CD10⁺ compartment and Annexin V staining (Fig. 2C). This finding is consistent with another report demonstrating that the most immature B cells within the transitional B cell population were most susceptible to cell death (14).

To confirm the apoptotic nature of the cell death detected by Annexin V staining, the B cell fractions were also stained with antibodies directed against cleaved products of caspase-3 and D4-GDI, both of which are generated during the effector phase of apoptosis (24). Levels of cleaved caspase-3 and D4-GDI were substantially higher in CD10⁺ compared with CD10⁻ B cells (SI Fig. 6), and whereas cells that were positive for both cleavage products did not account for all Annexin V⁺ cells (our unpublished data), they nonetheless represented the vast majority of dying cells. These data strongly suggest that the high propensity of CD10⁺ B cells to die spontaneously ex vivo is mediated by apoptosis.

The proapoptotic members of the Bcl-2 family Bax and Bak play an essential role in the intrinsic apoptotic pathway by permeabilizing the mitochondrial membrane (25). The conformational changes that occur in Bak and Bax as they translocate into the mitochondrial outer membrane can be detected with activation-specific antibodies (26, 27). Whereas intracellular staining for activated Bak and Bax is technically difficult to combine with most markers of apoptosis, particularly Annexin V, a good surrogate of dying B cells is loss of CD21 cell surface expression (28). Cell surface levels of CD21 as well as intracellular levels of Bak and Bax were measured in unfractionated B cells before (Fig. 3, 0h) and in CD10⁺ and CD10⁻ B cell fractions after incubation at 37°C for 16 h. As shown in Fig. 3A for one representative HIV-infected individual with active disease, the percentage of B cells expressing activated Bak increased from 6.2% at 0 h to 58.5% in CD10⁺ and 47% in CD10⁻ B cells after 16 h at 37°C. In addition, levels of activated Bax increased from 0.5% at 0 h to 60% in CD10⁺ B cells and 16% in CD10⁻ B cells after 16 h at 37°C (Fig. 3B). Of note, the presence of B cells expressing low levels of CD21 at 0 h and loss of CD21 expression in dying cells were not confounding factors in these analyses given very little activated Bax and Bak were expressed at 0 h (Fig. 3). Thus, the increases in the levels of activated Bax and Bak occurred predominantly in B cells that were expressing low levels of CD21, consistent with these events occurring in dying cells.

Fig. 3.

High intrinsic apoptosis in CD10⁺ B cells accompanied by high levels of activated Bak and Bax. Extracellular levels of CD21 and intracellular levels of activated Bak (A) and Bax (B) were measured in unfractionated B cells at time 0 (0h) and in CD10⁻ and CD10⁺ B cell fractions incubated at 37°C for 16 h. Data shown for one HIV-infected individual with active disease are representative of data collected on eight subjects.

Kinetics of Extrinsic and Intrinsic Apoptosis.

To further investigate the high propensity of CD10⁺ B cells to undergo intrinsic apoptosis, we performed kinetic analyses of intrinsic apoptosis and compared these to extrinsic apoptosis mediated by CD95L. Accordingly, B cells from HIV-infected individuals with active disease were fractionated into CD10⁺ and CD10⁻ populations. Considering that the susceptibility of B cells to CD95L-mediated apoptosis resides almost exclusively within the CD10⁻ B cell population (Fig. 2A), kinetic analyses of CD95L-mediated apoptosis were restricted to the CD10⁻ B cell fraction. As shown in Fig. 4 for a representative HIV-infected individual with advanced disease (CD4⁺ T cell count of 98 cells/μl and HIV plasma viremia of 402,137 RNA copies/ml), the induction of intrinsic apoptosis was rapid and extensive in the CD10⁺ B cells, with 75% of cell death after 2 h and >90% by 16 h. Although intrinsic apoptosis in the CD10⁻ B cells was substantially slower and lower compared with its CD10⁺ counterpart, the extrinsic CD95L-mediated apoptosis in CD10⁻ B cells was rapid and extensive and closely paralleled the intrinsic apoptosis observed in the CD10⁺ B cells (Fig. 4).

Fig. 4.

Extrinsic apoptosis in CD10⁻/CD21^lo and intrinsic apoptosis in CD10⁺ B cells have similar kinetics of Annexin V binding but not of activation of Bak and Bax. Kinetics of Annexin V binding and activation of Bax and Bak measured in CD10⁺, CD10⁻, and CD95L-treated CD10⁻ B cells of an HIV-infected individual with advanced disease. The profiles shown are representative of data collected on five individuals.

Whereas intrinsic apoptosis relies on a cascade of events that proceed through the mitochondria and involve the activation of Bak and Bax, extrinsic apoptosis can depend less on the mitochondrial pathway (29). Measurement of activated Bak and Bax revealed that for similar kinetics of intrinsic and extrinsic apoptosis in CD10⁺ and CD10⁻ B cells, respectively (Fig. 4 Left), activation of Bak and Bax was more rapid and extensive in the CD10⁺ B cells (Fig. 4 Center and Right). Levels of activated Bak and Bax were lowest in CD10⁻ B cells cultured in the absence of CD95L, consistent with their lowest levels of cell death.

Role of Caspase-8 in Extrinsic and Intrinsic Apoptosis.

Another feature that distinguishes extrinsic from intrinsic apoptosis is the activation of caspase-8, essential in the induction of the extrinsic pathway but not thought to contribute to the intrinsic pathway (30). The role of caspase-8 in an apoptotic pathway can be inferred by the reversal of apoptosis in the presence of peptide-based inhibitors that bind to caspase-8 or by the appearance of caspase-8 cleavage products. In preliminary experiments, the caspase-8 inhibitor Z-IETD-fmk induced a strong inhibitory effect on both intrinsic apoptosis in CD10⁺ B cells and extrinsic CD95L-mediated apoptosis in CD10⁻ B cells, with a median percent inhibition of apoptosis at 2 h of 73% and 77% respectively (data not shown). The role of caspase-8 in an apoptotic pathway can also be inferred by the appearance of caspase-8 cleavage products in apoptotic cells (31). Accordingly, when cells were costained with an active caspase-8-specific antibody and an antibody directed against cleaved D4-GDI (a downstream target of activated caspase-3 and a marker of apoptosis; see SI Fig. 6), activated caspase-8 was observed in cells undergoing apoptosis (Fig. 5A). Rapid induction of apoptosis, as illustrated by progressive increase in levels of cleaved D4-GDI (Fig. 5B), was observed in untreated CD10⁺ B cells and in CD95L-treated but not untreated CD10⁻ B cells. The cleavage of caspase-8 was delayed in untreated CD10⁺ B cells relative to CD95L-treated B cells but eventually reached similar levels by 16 h of incubation (Fig. 5C). In addition to the differences in kinetics shown in Fig. 5 B and C, the dot plots in Fig. 5A at 2 h and 6 h incubation revealed two populations within the CD10⁻ B cell population that expressed cleaved caspase-8, one expressing higher and the other lower intensities of cleaved caspase-8. The high-intensity cleaved caspase-8 was not observed in CD10⁺ B cells (Fig. 5A).

Fig. 5.

Kinetics of caspase-8 cleavage reveal differences between extrinsic and intrinsic apoptosis. (A) FACS plots depicting kinetics of cleavage of caspase-8 and D4-GDI in untreated CD10⁻ and CD10⁺ B cells, and in CD95L-treated CD10⁻ B cells of a representative HIV-infected individual with active disease. Data shown in A were graphed to illustrate changes over time in cleaved D4-GDI (B) and cleaved caspase-8 (C). Data are representative of four experiments. 0h, time 0.

Discussion

B cells of HIV-infected individuals become increasingly perturbed with advancing disease. In previous studies, we identified two distinct populations of B cells that are overrepresented in HIV disease, namely immature/transitional CD10⁺ B cells that likely arise as a result of homeostatic compensation in response to the presence of lymphopenia (11), and mature/activated CD10⁻/CD21^lo B cells that likely arise as a result of HIV-induced immune activation (7). We also demonstrated that HIV viremia is associated with a decline in absolute B cell counts and an increased susceptibility of B cells to CD95L-mediated apoptosis that may in part explain this decline (8). In the present study, we extend these findings by demonstrating that HIV disease is associated with two distinct mechanisms of apoptotic B cell death, both of which may contribute to the abnormal B cell profile of cellular dysfunction and propensity to apoptosis that results in declining B cell levels. On the one hand, we find immature/transitional CD10⁺ B cells to be highly susceptible to intrinsic apoptosis, consistent with the propensity of these developing cells to die unless they receive appropriate BCR signals (20). On the other hand, we find mature/activated CD10⁻/CD21^lo B cells, which we have described to be dysfunctional (8), to be highly susceptible to extrinsic apoptosis because of increased expression of CD95, likely the consequence of aberrant immune activation and HIV-induced cell turnover (19).

The expression of Ki-67 ex vivo has been shown to be a good surrogate marker of cell turnover in vivo, as measured by incorporation of BrdU (32). Levels of Ki-67 were highest in the CD10⁻/CD21^lo B cells and similar to levels reported for plasmablasts (33), thus consistent, as suggested (7), with these cells being en route to terminal differentiation as a result of HIV-induced immune activation. At the other end of the differentiation spectrum, the immature/transitional CD10⁺ B cells expressed Ki-67 at levels that were similar to what has been described for transitional B cells in healthy HIV-negative individuals and thought to reflect recently generated bone marrow emigrants (33). However, it should be noted that, although there are clear distinctions in activation and cell turnover within the B cell populations of HIV-infected individuals with active disease, comparisons of B cell populations between HIV-infected and -negative individuals may be somewhat confounded by evidence that ongoing HIV replication induces a certain level of activation on all B cell populations, including CD10⁻/CD21^hi B cells (our unpublished data).

The propensity of B cells and other lymphocytes to die by apoptosis is dictated by the overall balance of pro- and antiapoptotic members of the Bcl-2 family of proteins, a concept that has been used to establish apoptotic indices for several stages of B cell differentiation in mice (21). Of all of the B cell populations characterized in mice, immature and transitional B cells have been shown to express the lowest levels of the antiapoptotic proteins Bcl-2 (34), and Bcl-x_L (35, 36) and to have a high predisposition to undergo apoptosis ex vivo (23) and a short lifespan in vivo (37). Recently, immature/transitional B cells in humans with non-HIV immunodeficiencies were shown to express reduced levels of Bcl-2 when compared with more mature counterparts (13), indicating that observations made in mice may extend to humans. Here, we confirmed and extended these findings by demonstrating low levels of Bcl-2 and Bcl-x_L in CD10⁺ B cells of HIV-infected individuals that translated into high susceptibility to intrinsic apoptosis, along with induction of high levels of the proapoptotic proteins Bax and Bak. A recent study on human immature/transitional B cells also reported a high propensity to cell death, especially in the least mature of these B cells, although the data suggested a nonapoptotic mechanism (14). Our findings agreed with the observation of high cell death that was most accentuated in the most immature B cells. However, our findings strongly support an apoptotic pathway of cell death.

Intrinsic and extrinsic pathways of apoptosis can be distinguished by a prominent role for the mitochondria in the former and activation of caspase-8 in the latter pathway (29). The loss of mitochondrial potential that leads to the release of mediators of apoptosis has been shown to strictly rely on the activation of Bak and Bax (25). Our data were consistent with a dichotomous role of the mitochondria in extrinsic and intrinsic apoptosis in that high levels of activated Bak and Bax were observed in CD10⁺ B cells undergoing intrinsic apoptosis, whereas more modest levels of activated Bak and Bax were observed in CD10⁻ B cells undergoing similar levels of apoptosis by the extrinsic pathway. Thus, the mitochondria played a more prominent role in the intrinsic than in the extrinsic apoptotic pathway. In contrast, the role of caspase-8 was far less discriminating between the two pathways; both extrinsic and intrinsic apoptotic events were inhibited by caspase-8 inhibitors, and both were accompanied by cleavage of caspase-8. These observations either indicate a previously undescribed attribute of caspase-8 as a mediator of intrinsic apoptosis, as suggested in some studies (38–40), or reflect an endogenous extrinsic apoptotic pathway in CD10⁺ B cells that was mediated by an unidentified death receptor and its membrane-bound ligand. However, considering that we have found no evidence for the latter possibility (our unpublished data), yet clear evidence for differences between untreated CD10⁺ and CD95L-treated CD10⁻ B cells regarding response to caspase-8 inhibition and kinetics of caspase-8 cleavage, we believe the data are more consistent with an intrinsic apoptotic pathway in CD10⁺ B cells involving a unique form of caspase-8 activation.

The data presented herein thus provide evidence for two apoptotic mechanisms that may contribute to the progressive dysfunction and depletion of B cells in HIV disease. We have demonstrated that CD10⁻/CD21^lo activated/mature B cells are likely to arise as a result of HIV-induced immune activation, rendering them susceptible to CD95L-mediated apoptosis. As for the overrepresentation of CD10⁺ immature/transitional B cells in HIV-infected individuals with advancing disease, it is more likely that they arise as a result of compensatory mechanisms of lymphopenia, because their expansion has also been described in non-HIV immunodeficiency disorders (13, 41) and transplant recipients (42). However, the overrepresentation of immature/transitional B cells in HIV-infected individuals with advancing disease suggests that, contrary to individuals who are transiently lymphopenic after transplantation (42), sustained immunodeficiency must somehow prevent immature and mature B cells from attaining an appropriate equilibrium. Studies on immature/transitional B cells in the setting of HIV and non-HIV immunodeficiency diseases (11, 41) suggest that the overrepresentation of these B cells may be associated directly with CD4⁺ T cell lymphopenia or indirectly by increased serum levels of IL-7 resulting from homeostatic compensation (11, 43). It is tempting to speculate that advanced HIV disease leads to increased B cell production in the bone marrow as a result of a homeostatic mechanism related primarily to lymphopenia, but that in the absence of survival and maturation signals, these cells quickly die by intrinsic apoptosis.

Methods

Peripheral Blood B Cells.

HIV-infected individuals with active disease, defined as having a CD4⁺ T cell count <350 cells/μl and plasma HIV viremia >10,000 copies HIV RNA per ml, were recruited to undergo leukapheresis. Peripheral blood mononuclear (PBMC) cells were prepared by density-gradient centrifugation, and B cells were isolated from PBMC by negative selection as described (8). The B cells were separated into CD10⁺ and CD10⁻ B cell fractions by using a biotin-based microbead enrichment system (Miltenyi Biotec, Auburn, CA), as described (11). Purity of each B cell fraction ranged from 80% to 95%. All study participants provided informed consent, in accordance with the Institutional Review Board of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.

FACS Analysis.

Cell-surface stains were performed with the following anti-human mAbs: anti-CD21-FITC (Beckman Coulter, Miami, FL), anti-CD10-allophycocyanin, and anti-CD95-phycoerythrin (PE) (BD Biosciences, San Jose, CA). For intracellular stains, cells were first stained for cell surface markers and then fixed (FACS Lysing Buffer, BD Biosciences) and permeabilized (Permeabilizing Solution 2, BD Biosciences). Antibodies for intracellular stains included anti-Bcl-2-PE (Caltag, Carlsbad, CA), rabbit anti-Bcl-x_L (SouthernBiotech, Birmingham, AL), rabbit anti-Bax (NT, Upstate Biotechnology, Lake Placid, NY), and mouse anti-Bak (Ab-1, Calbiochem, La Jolla, CA), and appropriate secondary PE-conjugated antibodies. Ki-67 staining was performed with anti-Ki-67-PE (clone B56; BD Biosciences). FACS analyses were performed on a FACScalibur flow cytometer (BD Biosciences) by using FlowJo software (Tree Star, Ashland, OR).

Determination of Apoptosis.

B cells and derived fractions were cultured in the presence or absence of CD95L at 500 ng/ml (Kamiya Biomedical, Seattle, WA) and evaluated for apoptosis by Annexin V staining, as described (8). Apoptosis was also measured by intracellular staining with FITC-labeled anticleaved D4-GDI mAb (Imgenex, San Diego, CA) and anticleaved caspase-3 rabbit antibody (Asp-175; Cell Signaling Technology, Beverly, MA) followed by a PE-labeled anti-rabbit antibody. Cleaved caspase-8 was measured by intracellular staining with rabbit mAb 18C8 (Asp-391; Cell Signaling Technology) followed by a PE-labeled anti-rabbit antibody. Response to caspase inhibitors was evaluated by incubating cells in the presence of DMSO (control), or 120 μM caspase-8 inhibitor Z-IETD-fmk (Kamiya Biomedical).

Statistical Analyses.

Differences between various B cell fractions or populations were compared by the two-tailed Wilcoxon signed-rank test. The Spearman rank correlation was used to determine the association between parameters.

Abbreviation

CD95L

CD95 ligand.