Hospital Universitario Marqués de Valdecilla, Santander, Spain
Laboratorio de Inmunología del Trasplante, Unidad de Investigación, Hospital Universitario Marqués de Valdecilla, Departamento de Medicina y Psiquiatría, Escuela de Enfermería 4a planta, Av. Valdecilla s/n, 39008 Santander, Spain
Protein deimination, a process to modify arginine residues to citrulline by the addition of a neutral oxygen group, is associated with apoptosis. The presence of autoantibodies recognizing citrullinated peptides is highly specific to rheumatoid arthritis (RA) and is therefore a useful marker for the early diagnosis of RA. In this study, we explored whether anti–cyclic citrullinated peptide (anti-CCP) autoantibodies are produced in several experimental models of autoimmune diseases in mice.
The levels of anti-CCP autoantibodies were analyzed by enzyme-linked immunosorbent assay in several lupus-prone strains of mice, in animals with type II collagen (CII)–induced arthritis, and after induction of neonatal tolerance to alloantigens.
We observed the production of these autoantibodies in 2 different lupus-prone mice, MRL-lpr/lpr and (NZW × B6)F1-hbcl-2 transgenic mice, characterized by the presence of abnormalities in the regulation of B cell apoptosis. Other genetic defects, determining autoimmune susceptibility, present in MRL and NZW mice were additionally required for anti-CCP autoantibody production. The induction of autoantibodies in normal BALB/c mice injected at birth with semiallogeneic spleen cells from (BALB/c × B6)F1-hbcl-2 transgenic mice suggested that these additional autoimmune defects may be related, at least in part, to the establishment of abnormal interactions between T cells and B cells. In addition, anti-CCP autoantibodies were not produced in the course of CII-induced arthritis, an experimental model of RA in mice.
Our study provides evidence for the association between defects in the regulatory cell death machinery of B lymphocytes and the production of certain autoantibody specificities.
The production of autoantibodies directed toward cyclic citrullinated peptides (CCPs) has gained high importance for clinical use in the early diagnosis of rheumatoid arthritis (RA) (1–3). Citrullination is the process by which peptidyl arginine deiminase enzymes deiminate a peptidyl arginine to a peptidyl citrulline (4). This residue contains a neutral oxygen group that is specifically recognized by anti-CCP autoantibodies in a peptide-independent manner (5). These autoantibodies are highly specific for RA, and their production has been postulated to be a consequence of the citrullination of several synovial proteins, especially fibrin, in the inflamed RA tissue that can induce an antigen-driven activation of anti-CCP–specific B cells (6, 7). There are very few examples of citrullinated proteins in healthy mammalian cells (8). However, citrullination is closely related to increased apoptosis in vivo (9) and in vitro (10, 11). Selective deimination of arginine residues has been described in neurodegenerative lesions, where apoptosis is increased (9), as well as in mouse and human leukocytes after in vitro induction of apoptosis (10, 11).
Abnormalities in apoptosis have been shown to be involved in the pathogenesis of a number of human diseases characterized by the production of autoantibodies (12), but surprisingly, anti-CCP autoantibodies have been found only in RA (1, 2). The reasons for this selectivity are unknown, but genetic, hormonal, and environmental factors, together with local or systemic inflammation, may contribute to the breakdown of self-tolerance and the development of anti-CCP autoantibodies. Possible causes for the breakdown of tolerance include changes in the genetic program that regulate lymphocyte survival, which are often followed by the production of multiple autoantibodies and the development of autoimmune diseases (13).
Interestingly, it has been demonstrated that synovial or bone marrow stromal cells from RA patients may prevent apoptosis of resting B cells (14). This is accomplished through the up-regulation of Bcl-xL by a CD49b/CD29-CD106–dependent mechanism (15). In addition, B lymphocytes from RA patients have shown increased cell death resistance to certain apoptotic stimuli (16). In the present study, we investigated the relationship between the production of anti-CCP autoantibodies and the inhibition of B lymphocyte apoptosis using a number of murine experimental models of spontaneous or induced autoimmune diseases. Our results demonstrate that defects in the regulation of B cell survival are crucial for the production of anti-CCP autoantibodies.
MATERIALS AND METHODS
Mice and treatments.
NZB, NZW, DBA, BALB/c, C57BL/6 (B6), MRL-lpr/lpr, MRL/++, B6-lpr/lpr, BXSB, and B6-SV40-Eμ-human bcl-2 transgenic (B6-hbcl-2 transgenic) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The different F1 hybrid mice used in this study were produced in our animal facilities. The expression of hbcl-2 transgene in the experimental mice was assessed in peripheral blood B cells by flow cytometry using the 6C8 monoclonal antibody (mAb) specific for hBcl-2 conjugated to fluorescein isothiocyanate (PharMingen, San Diego, CA).
To evaluate the role of CD4+ T cells in the production of anti-CCP autoantibodies, (NZW × B6)F1-hbcl-2 transgenic mice were treated from birth (during the first 24 hours of life) until the age of 7 months with an anti-CD4 mAb (GK-1.5 [rat IgG2b]), as previously described (17). The efficiency of this treatment was evaluated monthly by flow cytometry. Mice were bled from the retro-orbital plexus, and the resulting sera were stored at −20°C until used. Animals were maintained in specific pathogen–free conditions, and all in vivo experiments with live animals were performed in compliance with the International League of Associations for Rheumatology Guide for the Care and Use of Laboratory Animals (18).
Induction of neonatal tolerance to alloantigens.
Neonatal tolerance was induced by intraperitoneal injection of 0.7 × 108 spleen cells from (BALB/c × B6)F1-hbcl-2 transgenic hybrid mice or 108 spleen cells from (BALB/c × B6)F1 nontransgenic hybrid mice into newborn BALB/c mice within the first 24 hours after birth, as previously described (19). These mice were bled every 3 weeks until 15 weeks of age.
Induction and assessment of arthritis.
Bovine type II collagen (CII) (provided by Dr. Marie Griffiths, University of Utah, Salt Lake City) was dissolved at a concentration of 2 mg/ml in 0.05M acetic acid at 4°C overnight and emulsified with the same volume of Freund's complete adjuvant containing 4 mg/ml of Mycobacterium tuberculosis (Chondrex, Seattle, WA). For the introduction of collagen-induced arthritis (CIA), (DBA × B6)F1 hybrid mice overexpressing hBcl-2 or not in B cells were injected once at the base of the tail with 150 μg of antigen in a final volume of 150 μl. Arthritis was assessed by inspecting the paws once a week. Clinical severity was quantified according to a graded scale of 0–3 as follows: 0 = no inflammation (normal joint); 1 = detectable swelling and erythema; 2 = swelling in >1 joint and pronounced inflammation; 3 = swelling of the entire paw and/or ankylosis. Each paw was graded, and the scores were summed (with a maximum possible score 12 per mouse). Serum samples were collected from these mice every 2 weeks and stored at −20°C until used.
Titers of IgG anti-CCP autoantibodies in sera were measured by enzyme-linked immunosorbent assay (ELISA) on plates coated with CCP (Immunoscan RA; Euro-Diagnostica, Utrecht, The Netherlands). The CCPs used in this assay were highly purified synthetic peptides deduced from known complementary DNA sequences of human profilaggrin. These peptides were made cyclic by substituting serine residues with cysteine in linear peptides (1, 5). As a secondary antibody, a goat anti-mouse IgG polyclonal antibody conjugated to phosphatase alkaline (Sigma, St. Louis, MO) was used. The cutoff point was considered to be 3 SD above the mean absorbance units obtained from the sera of 8-month-old B6 mice. Results were calculated as cutoff indexes obtained from the absorbance unit sample-to-cutoff ratio.
The presence of IgG anti-DNA autoantibodies was determined in sera by ELISA, and the results were expressed in titration units (TUs) in reference to a standard curve obtained from a serum pool from 6–8-month-old MRL-lpr/lpr mice (19).
Levels of anti-CII antibodies were quantified by ELISA on 96-well plates coated with CII (4 μg/well), as previously described (20). Titers were expressed as TUs, determined from a standard curve established with a positive control serum pool.
Statistical analysis of the differences between groups of mice was performed using the Mann-Whitney U test. Correlation coefficients were calculated by Spearman's rank test. P values less than 0.05 were considered significant.
Production of anti-CCP antibodies by MRL-lpr/lpr but not MRL/++, BXSB, and (NZB × NZW)F1 lupus-prone mice.
Despite their importance in the diagnosis of RA in humans, the presence of serum anti-CCP autoantibodies has not been analyzed in murine models of systemic or local autoimmune diseases. Since the production of those autoantibodies is likely to be related to the combination of immunologic, genetic, hormonal, and environmental factors (21), we studied the presence of circulating IgG anti-CCP autoantibodies in several strains of mice where the combination of such factors promotes the development of an autoimmune syndrome resembling human systemic lupus erythematosus (SLE): MRL/++ and MRL-lpr/lpr mice, BXSB male, and (NZB × NZW)F1 female mice (22–24). The production of IgG anti-DNA autoantibodies was used as a serologic marker of disease activity (25). As previously described, both MRL/++ and MRL-lpr/lpr mice exhibited high titers of serum anti-DNA autoantibodies at 9 and 3 months of age, respectively, although such levels were significantly higher in MRL-lpr/lpr mice (Figure 1). Likewise, 8-month-old BXSB males and (NZB × NZW)F1 female mice produced levels of anti-DNA autoantibodies similar to those observed in MRL/++ mice (Figure 1).
When looking at the production of anti-CCP autoantibodies in the groups of mice described previously, almost all MRL-lpr/lpr mice analyzed had increased titers of IgG anti-CCP autoantibodies (Figure 1). In contrast, the majority of MRL/++, (NZB × NZW)F1 female, and BXSB male mice failed to produce anti-CCP autoantibodies (Figure 1). The production of anti-CCP autoantibodies in MRL-lpr/lpr mice did not parallel that of anti-DNA autoantibodies since there was no significant correlation between the two parameters (r = 0.143, P = 0.736). These results indicated that despite the presence of an active lupus-like syndrome in lupus-prone MRL/++, BXSB, and (NZB × NZW)F1 mice, only MRL-lpr/lpr mice, which have intrinsic genetic defects in the apoptotic program secondary to mutations in the Fas receptor (26), produce anti-CCP antibodies.
Production of anti-CCP autoantibodies in (NZW × B6)F1 mice is associated with defects in the regulation of B cell death and is dependent on the presence of CD4+ T cells.
To explore in detail whether defects in the regulation of lymphocyte apoptosis are required for the production of anti-CCP autoantibodies, we analyzed the levels of these autoantibodies in (NZW × B6)F1 hybrid mice bearing an hbcl-2 transgene. (NZW × B6)F1-hbcl-2, but not (NZW × B6)F1 nontransgenic littermates, overexpressed hBcl-2 in B cells at all stages of differentiation, and developed an aggressive lupus-like autoimmune syndrome characterized by the production of several autoantibodies and the presence of a lethal glomerulonephritis (Diez MA, et al: unpublished observations). In addition, B lymphocytes from these hbcl-2 transgenic mice displayed increased in vitro survival (ref.27, and data not shown). Accordingly, 8-month-old (NZW × B6)F1-hbcl-2 transgenic mice had high levels of IgG anti-DNA autoantibodies in their sera (Figure 2). The majority of (NZW × B6)F1-hbcl-2 transgenic mice produced increased levels of IgG anti-CCP autoantibodies, at titers comparable to those of MRL-lpr/lpr mice (Figure 2). Again, no correlation between anti-DNA and anti-CCP autoantibody production was observed (r = 0.142, P = 0.351). It should be noted that (NZW × B6)F1-hbcl-2 transgenic mice, in contrast to MRL-lpr/lpr mice (28), failed to develop clinical signs of arthritis (during a period of 18 months; data not shown). These results clearly indicate that the production of anti-CCP autoantibodies is associated at least with alterations in the genetic cell death program of B cells.
To explore whether the production of anti-CCP autoantibodies in (NZW × B6)F1-hbcl-2 transgenic mice required the presence of CD4+ T cells, these mice were treated from birth until the age of 7 months with an anti-CD4 mAb. This treatment induced a complete and persistent depletion of CD4+ T cells, as documented by flow cytometric analysis (data not shown). The absence of CD4+ T cells completely inhibited the production of IgG anti-CCP autoantibodies as well as anti-DNA autoantibodies (Figure 2).
Inhibition of lymphocyte apoptosis by itself is not sufficient to generate anti-CCP antibodies.
To analyze whether the inhibition of B lymphocyte apoptosis by itself is enough to induce the production of anti-CCP autoantibodies, we measured the levels of circulating anti-CCP autoantibodies in 8-month-old B6-lpr/lpr and B6-hbcl-2 transgenic mice, which do not carry the genetic abnormalities of MRL and NZW mice. The presence of the lpr mutation induced the production of anti-DNA autoantibodies in the majority of B6 mice (28 of 42 animals), whereas only 8 of 42 B6-lpr/lpr mice developed significant titers of IgG anti-CCP autoantibodies (Figure 3). In contrast to B6-lpr/lpr mice, B6-hbcl-2 transgenic mice hardly developed significant levels of IgG anti-DNA and anti-CCP autoantibodies (Figure 3).
Overexpression of hBcl-2 in B cells is not able to promote anti-CCP autoantibody production during CIA.
Because of the association of anti-CCP autoantibodies and RA in humans (1–3), we explored whether IgG anti-CCP autoantibodies were produced in the course of CIA in mice, an experimental model of RA. The production of anti-CCP autoantibodies was analyzed in mice that did and did not overexpress hBcl-2 in B lymphocytes. To this end, 8-week-old (DBA/1 × B6)F1-hbcl-2 and nontransgenic mice were immunized at the base of the tail with bovine CII emulsified in Freund's complete adjuvant. Both types of F1 mice developed clinical signs of inflammatory joint disease and at a similar extent from the fourth week after immunization; the mean (±SD) grade of arthritis after 8 weeks of immunization was 7.0 ± 2.6 in (DBA/1 × B6)F1-hbcl-2 transgenic mice (n = 9) and 7.9 ± 2.4 in (DBA/1 × B6)F1 nontransgenic mice (n = 9). In addition, all animals produced significant titers of circulating IgG anti-CII antibodies 8 weeks after immunization (Figure 4). However, despite the development of an aggressive arthritis, both (DBA/1 × B6)F1-hbcl-2 transgenic and nontransgenic mice failed to produce IgG anti-CCP autoantibodies (Figure 4). It should also be stated that none of the (DBA/1 × B6)F1-hbcl-2 transgenic and nontransgenic mice produced IgG anti-DNA autoantibodies (data not shown).
Overexpression of hBcl-2 in B cells can induce the production of anti-CCP autoantibodies in the context of an abnormal semiallogeneic T cell–B cell interaction.
From the studies described above, it can be concluded that alterations in the regulation of B cell apoptosis promote the production of anti-CCP autoantibodies only in susceptible lupus-prone mice, but not in mice lacking a lupus background, even during the course of CIA. Interestingly, both lupus-prone mice and human RA patients develop a polyclonal B cell activation resulting in the production of multiple autoantibodies (29, 30); this was not the case in (DBA/1 × B6)F1-hbcl-2 transgenic mice with CIA. Accordingly, we hypothesized that the nature of the genetic background present in mice producing anti-CCP autoantibodies in the presence of apoptotic B cell defects can be, at least in part, related to the establishment of abnormal polyclonal T–B cell interactions.
To analyze this hypothesis, we used the experimental model of host-versus-graft disease observed after the induction of neonatal tolerance to alloantigens (31). In this model, the injection of semiallogeneic spleen cells from (BALB/c × B6)F1 mice (H-2d/b) into newborn BALB/c mice (H-2d) promotes a self-limited production of various types of autoantibodies (31). It has clearly been demonstrated that the polyclonal activation of donor semiallogenic F1 B cells by host CD4+ T cells is the mechanism of autoantibody production in these animals (32). We also demonstrated that the injection of spleen cells from (BALB/c × B6)F1 mice overexpressing hBcl-2 in B cells into newborn BALB/c mice results in the induction of a chronic and fatal lupus-like disease (19).
As previously reported (19, 31, 32), BALB/c mice injected at birth with spleen cells from either (BALB/c × B6)F1-hbcl-2 transgenic or nontransgenic mice produced substantial titers of IgG anti-DNA autoantibodies at 6 weeks of age (Figure 5). The levels of these autoantibodies were much higher in mice receiving (BALB/c × B6)F1-hbcl-2 transgenic spleen cells than in mice injected with (BALB/c × B6)F1 nontransgenic spleen cells. In addition, all the mice injected with F1 transgenic spleen cells overexpressing hBcl-2, but none of those receiving F1 nontransgenic spleen cells, produced significant levels of IgG anti-CCP autoantibodies at 6 weeks of age (Figure 5). However, the levels of IgG anti-CCP autoantibodies in these mice were limited compared with the levels observed in both MRL-lpr/lpr and (NZW × B6)F1-hbcl-2 transgenic mice (P < 0.001 in both cases), which indicated the contribution of other genetic factors to the high-level production of anti-CCP autoantibodies.
In the present study, we analyzed the production of anti-CCP autoantibodies in several experimental models of autoimmune disease in mice. We demonstrated the presence of anti-CCP autoantibodies in MRL-lpr/lpr mice, in (NZW × B6)F1-hbcl-2 transgenic mice, and in BALB/c mice tolerized at birth with semiallogeneic spleen cells from (BALB/c × B6)F1-hbcl-2 transgenic mice. However, anti-CCP autoantibodies were not found in other lupus-prone mice, such as MRL/++ mice, BXSB males, and (NZB × NZW)F1 females, and in mice with CIA. The autoimmune diseases with anti-CCP autoantibody production were characterized by the presence of genetic defects in the regulation of B cell survival and by the establishment of abnormal polyclonal T cell–B cell interactions. In addition, other unidentified factors modulate the levels of anti-CCP autoantibodies produced in the different pathologic conditions.
Anti-CCP autoantibodies have gained an important role in the clinical diagnosis of human RA, mostly due to their high specificity (1, 2). To date, anti-CCP autoantibodies have not been reported in other human autoimmune diseases (1, 2). We show here that anti-CCP autoantibodies are also observed in ∼70% of MRL-lpr/lpr mice at 5–6 months of age. These mice spontaneously develop synovial and periarticular inflammatory reactions very similar to those observed in human RA (28). However, the fact that the majority of (NZW × B6)F1-hbcl-2 transgenic mice produce anti-CCP autoantibodies in the absence of arthritis (data not shown) indicates that this autoantibody specificity is not exclusively associated with inflammatory joint disease. In addition, animals with CIA do not produce anti-CCP autoantibodies despite the development of an aggressive arthritis resembling RA in humans. These results also indicate that the development of anti-CCP autoimmune responses is not a consequence of the joint lesions that occur in RA.
One common characteristic observed in all strains of mice producing anti-CCP autoantibodies analyzed here is the existence of abnormalities in the regulation of lymphocyte apoptosis. The presence of anti-CCP autoantibodies in (NZW × B6)F1-hbcl-2 transgenic mice overexpressing hBcl-2 only in B cells and in BALB/c mice injected at birth with semiallogeneic spleen cells from (BALB/c × B6)F1-hbcl-2 transgenic mice indicates that defects in the regulation of B cell survival are crucial for the production of these autoantibodies. Suggestive of this are recent studies demonstrating that stromal cells from bone marrow or rheumatoid synovium obtained from patients with RA are able to increase the in vitro survival of B cells by promoting the expression of anti-apoptotic genes, such as Bcl-xL (15). The anti-apoptotic effect of RA stromal cells seems to be dependent on the interaction between CD106 (vascular cell adhesion molecule 1) on stromal cells and CD49d/CD29 (very late activation antigen 4) on B cells and is greater than that of stromal cells from patients with osteoarthritis or than that of skin fibroblast cell lines (15). This is likely due to the fact that stromal cells from RA patients constitutively express high levels of CD106 (14), suggesting the existence of intrinsic defects in stromal cells from RA patients that favor B lymphocyte survival.
However, we cannot exclude the possibility that abnormalities in T cell survival may also contribute to anti-CCP autoantibody production. In fact, in marked contrast to the complete absence of these autoantibodies in B6-hbcl-2 transgenic mice, a modest level of anti-CCP production is observed in approximately one-fifth of B6-lpr/lpr mice, in which both B and T cells are defective in Fas expression (26).
The production of high levels of anti-CCP autoantibodies by MRL-lpr/lpr and (NZW × B6)F1-hbcl-2 transgenic mice and their marked reduction or absence in B6-lpr/lpr and B6-hbcl-2 transgenic mice, respectively, indicates that defects in the regulation of B cell survival are not, by themselves, sufficient to promote the production of these autoantibodies. In addition, the fact that B6-lpr/lpr mice exhibit IgG hypergammaglobulinemia and high levels of several autoantibodies, such as anti-DNA or rheumatoid factor autoantibodies (33, 34), clearly indicates that the production of anti-CCP autoantibodies in MRL-lpr/lpr and (NZW × B6)F1-hbcl-2 transgenic mice is not just an epiphenomenon related to the increased IgG production observed in these strains of mice. NZW and MRL mice bear several dominant genetic loci linked to lupus traits (35–38). Although the nature of these genetic abnormalities remains largely unknown, we suggest that they can be translated, at least in part, in the presence of self-reactive T cells able to deliver cognate T cell help to autoreactive B cells, as previously proposed (39). This abnormal T cell–B cell cooperation seems to be polyclonal, as deduced by the broad spectrum of autoantibodies of the IgG isotype observed in these animals.
The same scenario can also be applied to patients with RA in whom the production of anti-CCP autoantibodies is accompanied by the presence of multiple autoantibodies (4). Consistent with that, circulating anti-CCP autoantibodies are found in BALB/c mice injected at birth with semiallogeneic spleen cells from (BALB/c × B6)F1-hbcl-2 transgenic mice. In this experimental condition, the appropriated T cell help is provided by H-2d host CD4+ T cells recognizing alloantigens in all H-2d/b semiallogeneic donor B cells (32). In addition, the production of anti-CCP autoantibodies in (NZW × B6)F1-hbcl-2 transgenic mice requires the presence of CD4+ T cells. Conversely, the absence of anti-CCP autoantibodies in mice with CIA can be explained by the lack of such CCP-specific T cell help. In fact, these mice produce antibodies against only CII, but not other autoantibodies, such as anti-DNA autoantibodies. It should also be stressed that the levels of these autoantibodies in mice tolerized at birth with spleen cells from hbcl-2 transgenic mice are significantly lower than those of MRL-lpr/lpr and (NZW × B6)F1-hbcl-2 transgenic mice. This suggests that the production level of anti-CCP autoantibodies is positively regulated by other abnormalities present in lupus-prone MRL and NZW mice.
Based on the present data, we propose the following model to explain the production of anti-CCP autoantibodies in lupus-prone MRL-lpr/lpr and (NZW × B6)F1-hbcl-2 transgenic mice and in patients with RA. Citrullination has been proposed as one of a number of posttranslational modifications of proteins occurring during apoptosis (8). Thus, increased levels of citrullinated antigens should be found in the course of autoimmune reactions where apoptosis is enhanced (40). Although the deficiency in Fas expression in mice bearing the lpr mutation may result in the inhibition of certain, but not all (Fas-independent cell death), apoptotic cell death events, and evidence for abnormal peptide citrullination in MRL-lpr/lpr mice has not yet been reported, we speculate that these posttranslational peptide modifications may be enhanced in these animals secondary to the development of autoimmune tissue damage. A similar scenario may be applied to (NZW × B6)F1-hbcl-2 transgenic mice, where the defects in the control of apoptotic cell death are selectively restricted to B lymphocytes.
The enhanced production of CCP may induce T cell–dependent antibody responses only in the presence of CCP-specific CD4+ T cells. The presence of these cells in the periphery is controlled, at least in part, by unknown genetic defects present in these animals or patients. On the other hand, the defects in the regulation of cell death in B cells may alter censoring mechanisms that operate during B cell development, allowing the escape of autoreactive low-affinity anti-CCP B cell clones to the periphery. The increased survival may then facilitate the required subsequent interaction with CCP-specific CD4+ T cells and the production of autoantibodies. This possibility has been validated using different Ig transgenic models, in which mice have an increased frequency of autoreactive B cells in the presence of the autoantigen. In this model of transgenic mice, the overexpression of hBcl-2 increases the survival and the exit to the periphery of self-reactive low-affinity B cells (41, 42). However, these exported autoreactive B cells remain developmentally arrested unless stimulated by antigen-specific CD4+ T cells (41–43).
Finally, the close association between inhibition of lymphocyte apoptosis and anti-CCP autoantibodies shown here may be of interest for the diagnosis of human autoimmune diseases in which defects in lymphocyte survival have been observed. In this regard, it would be of interest to analyze the presence of these autoantibodies in patients with autoimmune lymphoproliferative syndrome due to mutations in Fas/Fas ligand–mediated apoptosis or different caspases (44).
We thank Marta Pen̈a and Maria Landeras for technical assistance.