B lymphocyte stimulator expression in pediatric ... - Wiley Online Library

ARTHRITIS & RHEUMATISM Vol. 60, No. 11, November 2009, pp 3400–3409 DOI 10.1002/art.24902 © 2009, American College of Rheumatology

B Lymphocyte Stimulator Expression in Pediatric Systemic Lupus Erythematosus and Juvenile Idiopathic Arthritis Patients Sandy D. Hong,1 Andreas Reiff,1 Hai-Tao Yang,2 Thi-Sau Migone,3 Christopher D. Ward,3 Katherine Marzan,1 Bracha Shaham,1 Wee Choo Phei,1 Judith Garza,1 Bram Bernstein,1 and William Stohl2 lated with disease activity. In contrast, plasma BLyS protein levels were normal in JIA despite blood leukocyte BLyS mRNA levels being elevated to degrees similar to those in pediatric SLE. Among JIA patients, neither BLyS parameter was correlated with disease activity. In both pediatric SLE and JIA, the BLyS expression profiles remained stable at 6 months. Conclusion. Our findings indicate that, as previously noted in adult SLE, plasma BLyS protein and blood leukocyte BLyS mRNA levels are elevated in pediatric SLE. The correlation of plasma BLyS protein levels with disease activity points to BLyS as a candidate therapeutic target in pediatric SLE. Contrary to previous observations in adults with rheumatoid arthritis, plasma BLyS protein levels are normal in JIA despite elevated blood leukocyte BLyS mRNA levels. The absence of correlation between either of the BLyS parameters and disease activity in JIA calls for circumspection prior to assigning BLyS as a candidate therapeutic target in this disorder.

Objective. To assess the expression of B lymphocyte stimulator (BLyS) in patients with pediatric systemic lupus erythematosus (SLE) or juvenile idiopathic arthritis (JIA). Methods. Blood samples collected from patients with pediatric SLE (n ⴝ 56) and patients with JIA (n ⴝ 54) at the beginning and end of a 6-month interval were analyzed for plasma BLyS protein levels by enzymelinked immunosorbent assay and for blood leukocyte full-length BLyS and ⌬BLyS messenger RNA (mRNA) levels by quantitative real-time polymerase chain reaction (normalized to 18S expression). Healthy siblings (n ⴝ 34) of these patients served as controls. Results. In pediatric SLE, plasma BLyS protein and blood leukocyte BLyS mRNA levels were each significantly elevated, and plasma BLyS protein levels, but not blood leukocyte BLyS mRNA levels, were correSupported in part by a grant from the Arthritis Foundation, Southern California Chapter. 1 Sandy D. Hong, MD (current address: University of Iowa, Iowa City), Andreas Reiff, MD, Katherine Marzan, MD, Bracha Shaham, MD, Wee Choo Phei, MS, Judith Garza, MD, Bram Bernstein, MD: Childrens Hospital Los Angeles, Los Angeles, California; 2 Hai-Tao Yang, PhD, William Stohl, MD, PhD: Los Angeles County ⫹ University of Southern California Medical Center, and University of Southern California, Los Angeles; 3Thi-Sau Migone, PhD, Christopher D. Ward, MS: Human Genome Sciences, Inc., Rockville, Maryland. Dr. Migone and Mr. Ward own stock or stock options in Human Genome Sciences, Inc. Dr. Stohl has received consulting fees, speaking fees, and/or honoraria from Human Genome Sciences, Inc. (less than $10,000). Address correspondence and reprint requests to Sandy D. Hong, MD, Division of Pediatric Rheumatology, University of Iowa, Iowa City, IA 52242 (e-mail: [email protected]); or to William Stohl, MD, PhD, Division of Rheumatology, University of Southern California, 2011 Zonal Avenue, HMR 711, Los Angeles, CA 90033 (e-mail: [email protected]). Submitted for publication February 25, 2009; accepted in revised form July 17, 2009.

A considerable body of evidence points to a central role of B cells in the pathogenesis of both systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA). In addition to the well-known associations between these disorders and specific autoantibodies (e.g., anti–double-stranded DNA [anti-dsDNA] with SLE and rheumatoid factor [RF] with RA), B cells likely contribute to disease development via autoantibodyindependent pathways. Neither pathologic nor clinical disease develops in SLE-prone MRL-lpr/lpr mice genetically rendered devoid of B cells. However, genetic reconstitution of these mice with B cells incapable of secreting Ig does partially restore disease development despite the complete absence of circulating autoanti3400

EXPRESSION OF BLyS IN PEDIATRIC SLE AND JIA

bodies (1,2). In RA, B cell depletion therapy is highly efficacious in a substantial percentage of patients despite the modest effects of such therapy on circulating autoantibodies (3). Given the crucial role of B cells in the pathogenesis of SLE and RA, it stands to reason that factors which promote B cell survival and/or function are also crucial. One such vital B cell survival factor is B lymphocyte stimulator (BLyS; trademark of Human Genome Sciences, Rockville, MD), also commonly known as BAFF, a member of the tumor necrosis factor (TNF) ligand superfamily (4,5). At least 2 isoforms of BLyS are expressed. The full-length BLyS messenger RNA (mRNA) isoform codes for the biologically active fulllength type II transmembrane protein, whereas the alternatively spliced ⌬BLyS mRNA isoform codes for a protein with a small peptide deletion (6). In mononuclear cells, cleavage of membrane BLyS by a furin protease from the cell surface results in release of a soluble, biologically active 17-kd molecule (5,7), whereas in neutrophils, processing of full-length BLyS to the soluble form routinely takes place intracellularly (8,9). In contrast, ⌬BLyS protein is not expressed on the cell surface, is not released from cells, and does not bind to surface-expressed BLyS receptors (6). Accordingly, ⌬BLyS is not present in the circulation and has no BLyS-like agonistic activity. Indeed, the physiologic roles of ⌬BLyS, if any, in normal (healthy) states and the pathophysiologic roles of ⌬BLyS, if any, in disease states remain unknown. Regardless, ⌬BLyS can form biologically inactive intracellular heterotrimers with full-length BLyS, thereby serving as a dominantnegative antagonist of BLyS activity. As a consequence, selective overexpression of the ⌬BLyS isoform leads to reduced BLyS activity in vivo (10). In mice, treatment with a BLyS antagonist can be beneficial in the context of either SLE or inflammatory arthritis (11–14). In humans, both SLE and RA are associated with elevated circulating levels of BLyS (15,16), and BLyS levels in synovial fluids from affected joints are routinely greater than those in the corresponding sera (17). In SLE, cross-sectional observations have documented significant correlations between BLyS expression and disease activity (18), and a large longitudinal study documented a significant correlation between changes in circulating BLyS levels and changes in disease activity (19). All studies to date regarding BLyS expression in human rheumatic diseases have focused on the adult population. Inasmuch as important phenotypic and epidemiologic differences exist between adult SLE and

3401

pediatric SLE and between adult RA and juvenile idiopathic arthritis (JIA), a blind extrapolation of findings from the adult diseases to their respective pediatric counterparts may lead to erroneous conclusions with serious consequences. Given the interest in the therapeutic targeting of BLyS in both SLE and RA (20–22), it stands to reason that similar targeting could be applied to pediatric SLE and/or JIA. In order to establish whether BLyS dysregulation is truly a feature of pediatric SLE and/or JIA, we have conducted the first study of BLyS expression in any pediatric rheumatic disease population, focusing our attention on pediatric SLE and JIA. PATIENTS AND METHODS Study approval. This study was approved by the Childrens Hospital Los Angeles Institutional Review Board. All subjects or, in the case of minors, their legal guardians gave written informed consent and assent. Subjects, blood collection, and clinical evaluation. Pediatric patients (n ⫽ 56 with pediatric SLE and n ⫽ 54 with JIA, who were further characterized as having systemic [n ⫽ 20], polyarticular [n ⫽ 20], or oligoarticular [n ⫽ 14] JIA) seen at Childrens Hospital Los Angeles and their healthy siblings who did not carry a diagnosis of any chronic disease (n ⫽ 34) were recruited for this study. Of the healthy siblings, 12 were genetically related to patients with pediatric SLE, 20 were genetically related to patients with JIA, and 2 were not genetically related to any of the patients. All pediatric SLE and JIA patients met American College of Rheumatology pediatric criteria for their respective diseases with disease onset prior to 16 years of age (23,24). Venous blood was drawn from all patients upon study entry and again 6 months later; single blood samples were collected from the controls. JIA patients were evaluated using the Childhood Health Assessment Questionnaire (C-HAQ; a measure of functional disability) (25), and pediatric SLE patients were evaluated using the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) (26). These instruments have previously been validated in both pediatric SLE and JIA (25,27). Medical charts were reviewed for clinical laboratory results (e.g., complete blood cell count and C-reactive protein [CRP] level). Plasma BLyS protein and blood leukocyte BLyS mRNA determinations. Whole venous blood was centrifuged to yield plasma and buffy coat (leukocyte) fractions. Plasma BLyS protein levels were determined by enzyme-linked immunosorbent assay (ELISA), and blood leukocyte full-length BLyS and ⌬BLyS mRNA levels were determined by quantitative reverse transcriptase–polymerase chain reaction as previously described (28). (Since the protein product of ⌬BLyS is neither expressed on the cell surface nor released into the circulation, the ELISA detects only products derived from the full-length BLyS protein.) Due to the inherent inaccuracy of the BLyS protein ELISA at concentrations ⬍0.78 ng/ml (the lower limit of quantification for this batch of assays), any sample with a BLyS protein concentration ⬍0.78 ng/ml was arbitrarily assigned a value of 0.5 ng/ml.

3402

HONG ET AL

Table 1. Baseline characteristics of the study cohorts* Control subjects (n ⫽ 34)

Pediatric SLE patients (n ⫽ 56)

JIA patients All JIA (n ⫽ 54)

Systemic (n ⫽ 20)

Polyarticular (n ⫽ 20)

Oligoarticular (n ⫽ 14)

Age, median (range) years 12.6 (3.0–21.0) 15.7 (10.0–20.1) 10.7 (2.2–20.0) 9.3 (2.2–19.0) 13.4 (6.2–20.0) 7.2 (2.2–15.9) Sex, no. male/female 14/20 13/43 14/40 8/12 1/19 5/9 Race, no. of patients Hispanic 24 42 36 13 17 6 Asian 3 7 1 1 0 0 African American 0 5 0 0 0 0 White 7 1 15 5 3 7 Mixed 0 1 2 1 0 1 Disease duration, median (range) years NA 2.4 (0.0–8.7) 2.6 (0.1–11.1) 4.2 (0.1–11.1) 1.8 (0.1–8.3) 1.3 (0.1–8.7) 3 3 NA 5.8 (2.2–16.7) 6.7 (4.0–15.4) 7.0 (4.0–15.4) 6.6 (4.0–11.1) 7.1 (5.2–11.2) WBCs, median (range) ⫻10 /mm † Lymphocytes, median (range) NA 0.22 (0.03–0.72) 0.37 (0.09–0.69) 0.37 (0.09–0.69) 0.38 (0.15–0.64) 0.37 (0.29–0.53) ⫻103/mm3† Medications, no. (%) of patients Prednisone ⱖ1 mg/kg/day NA 3 (5) 3 (6) 3 (15) 0 (0) 0 (0) Prednisone ⬍1 mg/kg/day NA 32 (57) 6 (11) 6 (30) 0 (0) 0 (0) Cyclophosphamide NA 11 (20) 0 (0) 0 (0) 0 (0) 0 (0) Mycophenolate mofetil NA 11 (20) 0 (0) 0 (0) 0 (0) 0 (0) Azathioprine NA 16 (29) 0 (0) 0 (0) 0 (0) 0 (0) Methotrexate NA 3 (5) 35 (65) 13 (65) 15 (75) 7 (50) Hydroxychloroquine NA 53 (95) 0 (0) 0 (0) 0 (0) 0 (0) Anakinra NA 0 (0) 5 (9) 5 (25) 0 (0) 0 (0) Etanercept NA 0 (0) 18 (33) 2 (10) 11 (55) 5 (36) Adalimumab NA 0 (0) 6 (11) 1 (5) 5 (25) 0 (0) Other biologic agent NA 3 (5) 6 (11) 6 (30) 0 (0) 0 (0) * SLE ⫽ systemic lupus erythematosus; JIA ⫽ juvenile idiopathic arthritis; NA ⫽ not applicable; WBCs ⫽ white blood cells. † In peripheral blood.

Statistical analysis. All analyses were performed using SigmaStat software (SPSS, Chicago, IL). Results that did not follow a normal distribution were log-transformed to achieve normality. Parametric testing between 2 matched or unmatched groups was performed using the paired or unpaired t-test, respectively. When neither the raw nor the logtransformed data were normally distributed or when the equal variance test was not satisfied, nonparametric testing was performed using the Mann-Whitney rank sum test or Wilcoxon’s signed rank test. Parametric testing among 3 or more groups was performed by analysis of variance (ANOVA). If the normality or equal variance tests failed, nonparametric testing was performed by ANOVA on ranks. Whenever the data sets included plasma BLyS levels ⬍0.78 ng/ml (for which accurate numerical values could not be assigned), nonparametric testing was performed. Correlations were determined by Pearson’s product-moment correlation for interval data and by Spearman’s rank order correlation for ordinal data or for interval data which did not follow a normal distribution. Nominal data were analyzed by chi-square analysis-of-contingency tables or by Fisher’s exact test.

RESULTS Baseline demographic and clinical characteristics. In order to increase the likelihood that our study patients would be representative of the pediatric SLE and JIA populations at large, no restrictions on race, sex,

disease duration, or medications were placed on subjects for entry into this study. In light of the Hispanic predominance in our catchment area and the global female predominance of both pediatric SLE and JIA, each of our 3 subject cohorts (pediatric SLE, JIA, and control) was predominantly Hispanic and female, with no significant differences in race (Hispanic versus non-Hispanic) or sex distribution among the cohorts (Table 1). No significant difference in disease duration between patients with pediatric SLE and patients with JIA was noted, although the pediatric SLE patients were older than both the JIA patients and the healthy controls (P ⬍ 0.001). Not surprisingly, there were marked differences between pediatric SLE and JIA patients with respect to treatment, with more pediatric SLE patients taking corticosteroids (P ⬍ 0.001), cyclophosphamide (P ⫽ 0.002), mycophenolate mofetil (P ⫽ 0.002), azathioprine (P ⬍ 0.001), and hydroxychloroquine (P ⬍ 0.001) and more JIA patients taking methotrexate (P ⬍ 0.001), anakinra (P ⫽ 0.061), and anti-TNF agents (etanercept or adalimumab; P ⬍ 0.001). Based on initial clinical presentation, patients with JIA can be classified as having the systemic, polyarticular, or oligoarticular subtype (24). Not only were


important differences noted between the JIA cohort as a whole and the pediatric SLE and control cohorts, but considerable heterogeneity was also noted across the individual JIA subtypes. The percentage of Hispanic patients among the group with oligoarticular JIA was lower than that among the group with systemic or polyarticular JIA (P ⫽ 0.036), and the percentage of female patients among the group with polyarticular JIA was higher than that among the group with systemic or oligoarticular JIA (P ⫽ 0.026). Patients with polyarticular JIA were older than patients with systemic or oligoarticular JIA (P ⫽ 0.002), and patients with systemic JIA had a longer disease duration than patients with polyarticular or oligoarticular JIA (P ⫽ 0.029). More patients with systemic JIA were taking corticosteroids (P ⬍ 0.001), anakinra (P ⫽ 0.005), and other biologic agents (P ⫽ 0.002) than were patients with polyarticular or oligoarticular JIA, and more patients with polyarticular JIA were taking anti-TNF agents than were patients with systemic or oligoarticular JIA (P ⬍ 0.001). BLyS expression in pediatric SLE. In adult SLE, circulating levels of BLyS protein are elevated, and blood leukocyte full-length BLyS and ⌬BLyS mRNA levels are increased (15,16,18,19). An identical BLyS expression profile was observed in pediatric SLE. At study entry, the median plasma BLyS protein level in pediatric SLE patients was more than twice that in the control group (1.16 ng/ml versus 0.50 ng/ml; P ⫽ 0.005) (Figure 1A). No difference in plasma BLyS protein levels was detected between siblings of patients with pediatric SLE and siblings of patients with JIA, so the results from all 34 siblings were pooled. When the analysis was limited to only those subjects (pediatric SLE patients and controls) whose plasma BLyS protein levels were in the quantifiable range (ⱖ0.78 ng/ml), the median plasma BLyS protein level was 56% greater in pediatric SLE patients than in controls (1.56 ng/ml versus 1.00 ng/ml, P ⬍ 0.001) (Figure 1B). Of note, the global increase in plasma BLyS protein levels among pediatric SLE patients persisted over time, as evidenced by the fact that repeat plasma samples collected 6 months after the original samples were collected showed similar degrees of elevation in BLyS protein concentrations (median plasma BLyS protein levels of 0.99 ng/ml versus 0.50 ng/ml for all pediatric SLE patients and control subjects, respectively [P ⫽ 0.003] and median plasma BLyS protein levels of 1.75 ng/ml versus 1.00 ng/ml for pediatric SLE patients and control subjects with quantifiable values, respectively [P ⬍ 0.001]) (Figures 1A and B). Importantly, increased plasma BLyS protein lev-

3403

Figure 1. B lymphocyte stimulator (BLyS) expression profile in pediatric systemic lupus erythematosus (pSLE). A and B, Plasma BLyS protein levels in all healthy controls (n ⫽ 34) and in all patients with pediatric SLE (n ⫽ 56) at study entry (pSLE [1]) and after 6 months (pSLE [2]) (A) and in controls (n ⫽ 16) and patients (n ⫽ 34 at study entry and n ⫽ 35 after 6 months) with reliably quantifiable results (B). C, Blood leukocyte full-length BLyS mRNA levels in all controls and patients. D, Blood leukocyte ⌬BLyS mRNA levels in all controls and patients. Each symbol indicates an individual subject. The composite results are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the 10th and 90th percentiles.

els in pediatric SLE could not be attributed to demographic differences between the pediatric SLE patients and control subjects. Among females and Hispanic subjects (the predominant sex and race, respectively, among both pediatric SLE patients and control subjects), plasma BLyS protein levels were higher in pediatric SLE patients than in control subjects (P ⬍ 0.001 for each comparison). Although plasma BLyS protein levels were inversely correlated with age among the control subjects (r ⫽ ⫺0.503, P ⫽ 0.003), no correlation between plasma BLyS protein levels and age was observed among the pediatric SLE patients. Since the control subjects were collectively younger than the pediatric SLE patients at the time of testing, it is plausible that (older) controls more closely matched by age to the pediatric SLE patients would have had even lower plasma BLyS protein levels, thereby further accentuating the difference between controls and patients with pediatric SLE. In addition, no consistent correlations between plasma BLyS protein levels and blood leukocyte, neutrophil,

3404

lymphocyte, or monocyte counts were observed among the pediatric SLE patients (data not shown), so it is highly unlikely that differences in blood cell composition between pediatric SLE patients and control subjects could explain the elevated plasma BLyS protein levels in pediatric SLE. Results for blood leukocyte full-length BLyS and ⌬BLyS mRNA levels paralleled (and showed an even greater difference than) the plasma BLyS protein findings. At study entry, geometric mean full-length BLyS and ⌬BLyS mRNA levels in pediatric SLE were 22-fold and 24-fold greater, respectively, than the corresponding control values (0.062 versus 0.0028 and 0.0097 versus 0.00040; P ⬍ 0.001 for each comparison) (Figures 1C and D). As with plasma BLyS protein levels, no differences in full-length BLyS or ⌬BLyS mRNA levels were detected between siblings of pediatric SLE patients and siblings of JIA patients, so the results from all 34 siblings were pooled. Similar to the findings for plasma BLyS protein levels, the increases in full-length BLyS and ⌬BLyS mRNA levels persisted over time, with the geometric mean levels in the blood leukocyte samples collected at 6 months being 41-fold and 39-fold greater, respectively, than the corresponding control levels (0.12 versus 0.0028 and 0.015 versus 0.00040; P ⬍ 0.001 for each comparison). As with plasma BLyS protein levels, full-length BLyS and ⌬BLyS mRNA levels were not correlated with sex, race, age, or blood cell composition (data not shown). Although, as in adult SLE, blood leukocyte full-length BLyS mRNA levels were strongly correlated with blood leukocyte ⌬BLyS mRNA levels in pediatric SLE (r ⫽ 0.841, P ⬍ 0.001), no significant correlations between plasma BLyS protein levels and blood leukocyte full-length BLyS or ⌬BLyS mRNA levels were detected in pediatric SLE, in marked contrast to the positive correlations observed in adult SLE (data not shown) (18). Association between BLyS expression and disease activity in pediatric SLE. Among pediatric SLE patients, BLyS expression was independent of corticosteroid treatment. Unexpectedly, the median plasma BLyS protein levels among pediatric SLE patients with quantifiable levels were greater in patients taking a cytotoxic agent than in those not taking a cytotoxic agent at the first time point (1.47 ng/ml versus 0.85 ng/ml; P ⫽ 0.050). Although this pattern persisted at the 6-month time point, the difference in median plasma BLyS protein levels was no longer significant (1.12 ng/ml versus 0.90 ng/ml; P ⫽ 0.293). These counterintuitive results with regard to plasma BLyS protein levels may reflect differences in disease activity among patients

HONG ET AL

Figure 2. Correlation between B lymphocyte stimulator (BLyS) expression and disease activity in pediatric systemic lupus erythematosus (SLE). A and B, Correlation between the SLE Disease Activity Index (SLEDAI) and plasma BLyS protein levels at study entry (0 months) (A) and after 6 months (B). Only values that were reliably quantifiable were plotted. C and D, Lack of correlation between SLEDAI and blood leukocyte full-length BLyS mRNA levels at study entry (C) and after 6 months (D). E and F, Lack of correlation between SLEDAI and blood leukocyte ⌬BLyS mRNA levels at study entry (E) and after 6 months (F).

receiving versus those not receiving a cytotoxic agent. Indeed, disease activity was greater in the former set of patients than in the latter set of patients (median SLEDAI score 4 versus 2 at 0 months [P ⫽ 0.097] and median SLEDAI score 6 versus 2 at 6 months [P ⫽ 0.009]). It is also possible that reductions in B cell loads among patients receiving cytotoxic agents may have resulted in the elevated circulating BLyS protein levels, as has previously been demonstrated in adult RA or SLE patients treated with rituximab (29,30). We observed a correlation between disease activity and circulating (plasma) BLyS protein levels in pediatric SLE despite the limited number of patients available to us for study (Figure 2A). Importantly, this


correlation appeared to persist over time (Figure 2B). In contrast, no correlations between disease activity and either full-length BLyS or ⌬BLyS mRNA levels were observed at any time point tested (Figures 2C–F). Of note, these findings in pediatric SLE contrast with those in adult SLE, in which disease activity correlates not only with circulating BLyS protein levels, but also with blood leukocyte full-length BLyS and ⌬BLyS mRNA levels (18,19), raising the possibility that there are subtle, yet important, pathogenetic differences between adult SLE and pediatric SLE. BLyS expression in JIA. The BLyS expression profile in adult RA differs from that in adult SLE. Although circulating levels of BLyS protein are elevated in both adult RA and adult SLE (albeit to a lesser degree in RA than in SLE), blood leukocyte full-length BLyS and ⌬BLyS mRNA levels are normal in adult RA (15,16,18,31). In contrast to the similar BLyS expression profiles of adult SLE and pediatric SLE patients, the BLyS expression profile of JIA patients differed markedly from that of adult RA patients. When either the entire JIA and control cohorts were analyzed or only those JIA and control subjects with quantifiable plasma BLyS protein levels were analyzed, no differences between JIA and controls were observed in samples collected at either study entry or after 6 months (Figures 3A and B). (Since neither plasma BLyS protein levels, blood leukocyte full-length BLyS mRNA levels, nor ⌬BLyS mRNA levels differed significantly among the JIA patients with systemic, polyarticular, and oligoarticular subtypes at either the initial or the 6-month determinations, aggregate results for JIA in toto are presented.) Indeed, the median quantifiable plasma BLyS protein level in JIA patients at study entry was only 64% of that in pediatric SLE patients (1.00 ng/ml versus 1.56 ng/ml; P ⬍ 0.001) and remained only 51% of that in pediatric SLE patients at 6 months (0.90 ng/ml versus 1.75 ng/ml; P ⬍ 0.001). Not only did the plasma BLyS protein expression profile differ between JIA and adult RA patients (with normal expression in the former and elevated expression in the latter), but the blood leukocyte BLyS mRNA profiles of JIA patients also differed from those of adult RA patients. In sharp contrast to the normal blood leukocyte full-length BLyS and ⌬BLyS mRNA levels in adult RA patients, each level was markedly elevated in JIA patients, in both the 0-month and 6-month samples (geometric mean 20-fold and 14-fold increases in fulllength BLyS mRNA levels [0.056 versus 0.0028 and 0.040 versus 0.0028] in JIA patients versus controls at 0 and 6 months, respectively, and 39-fold and 19-fold

3405

Figure 3. B lymphocyte stimulator (BLyS) expression profile in juvenile idiopathic arthritis (JIA). A and B, Plasma BLyS protein levels in all healthy controls (n ⫽ 34) and in all patients with JIA (n ⫽ 54) at study entry (JIA [1]) and after 6 months (JIA [2]) (A) and in controls (n ⫽ 16) and patients (n ⫽ 32 at study entry and n ⫽ 21 after 6 months) with reliably quantifiable results (B). C, Blood leukocyte full-length BLyS mRNA levels in all controls and patients. D, Blood leukocyte ⌬BLyS mRNA levels in all controls and patients. Control values are identical to those shown in Figure 1. Each symbol indicates an individual subject. The composite results are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the 10th and 90th percentiles.

increases in ⌬BLyS mRNA levels [0.016 versus 0.00040 and 0.0077 versus 0.00040] in JIA patients versus controls at 0 and 6 months, respectively; P ⬍0.001 for each comparison) (Figures 3C and D). Indeed, full-length BLyS and ⌬BLyS mRNA levels in JIA patients were notably similar to those in pediatric SLE patients, with the geometric mean levels of each isoform in JIA and pediatric SLE being within a factor of 2.9 of each other at each time point tested. Increased full-length BLyS and ⌬BLyS mRNA levels were observed when JIA patients and control subjects were divided into Hispanic (P ⫽ 0.002 and P ⬍ 0.001) and non-Hispanic (P ⬍ 0.001 and P ⫽ 0.007) or female (P ⬍ 0.001 and P ⬍ 0.001) and male (P ⫽ 0.077 and P ⫽ 0.009) subgroups. Moreover, full-length BLyS and ⌬BLyS mRNA levels did not differ significantly among JIA patients with respect to treatment with corticosteroids, methotrexate, or biologic agents (data not shown). Similar to findings in pediatric SLE, blood

3406

HONG ET AL

leukocyte full-length BLyS mRNA levels in JIA patients were strongly correlated with blood leukocyte ⌬BLyS mRNA levels, but no significant correlations between plasma BLyS protein levels and blood leukocyte fulllength BLyS or ⌬BLyS mRNA levels were detected (data not shown). Of note, plasma BLyS protein levels tended to be higher, and full-length BLyS and ⌬BLyS mRNA levels tended to be lower, in the 10 RF-positive patients with polyarticular JIA (whose disease, among all the JIA subtypes, most resembles adult RA) than in the 10 RF-negative patients with polyarticular JIA (data not shown), which was consistent with a difference in BLyS expression profile between “adult RA–like” JIA and “non–adult RA–like” JIA. Lack of association between BLyS expression and disease activity in JIA. In contrast to the correlation between the plasma BLyS protein concentration and disease activity (SLEDAI) among pediatric SLE patients, no correlation between the plasma BLyS protein concentration and disease activity (according to the C-HAQ) was detected among JIA patients (data not shown). No correlations between C-HAQ scores and blood leukocyte full-length BLyS or ⌬BLyS mRNA levels or between circulating erythrocyte sedimentation rate (ESR) or CRP levels and the BLyS parameters were detected (data not shown). DISCUSSION The many phenotypic similarities between pediatric SLE and adult SLE and between JIA and adult RA notwithstanding, it is well appreciated that important differences do exist between these respective sets of disorders. For example, pediatric SLE patients exhibit higher rates of neuropsychiatric and renal disease than do their adult counterparts, whereas the female-biased distribution is less extreme among pediatric SLE patients than among adult SLE patients (32,33). Moreover, elevated levels of RF are common in adult RA but are largely restricted to a subgroup of polyarticular JIA patients, and joint erosions routinely develop much later in the course of JIA than in the course of adult RA. Thus, given the compelling evidence for a salutary effect of BLyS antagonism in murine SLE and inflammatory arthritis models (11–14) coupled with the ongoing interest in therapeutic targeting of BLyS in human adult SLE and RA (20–22), it was imperative to directly measure BLyS expression in pediatric SLE and JIA, although the BLyS expression profiles in adult SLE and adult RA

have previously been characterized by several laboratories (15–19). Whereas 3 previous studies have focused on BLyS expression in various pediatric populations (34– 36), none has involved patients with rheumatic disease. This first study of BLyS expression in any pediatric rheumatic disease population has demonstrated both important similarities and important differences in BLyS expression profiles between pediatric SLE and adult SLE and between JIA and adult RA. With regard to the former set of disorders, circulating (plasma) BLyS protein levels in pediatric SLE and adult SLE share a BLyS expression profile characterized by elevated plasma BLyS protein levels and elevated blood leukocyte fulllength BLyS and ⌬BLyS mRNA levels. These observations provide evidence of BLyS overproduction in pediatric SLE and, as is the case for adult SLE, identify BLyS as a rational candidate therapeutic target. Given the apparent efficacy of belimumab, an anti-BLyS monoclonal antibody, in the “seropositive” (serum antinuclear antibody titer ⱖ1:80 and/or anti-dsDNA ⱖ30 IU/ml) subset of adult SLE patients (37), belimumab or other BLyS antagonists may be efficacious in a similar subset of pediatric SLE patients. It must be stressed that elevations in plasma BLyS protein levels and blood leukocyte BLyS mRNA levels reflect different, albeit overlapping, biologic phenomena. Elevated plasma BLyS protein levels reflect a balance between the egress of BLyS protein generated at sites of overproduction (e.g., secondary lymphoid tissue in adult SLE and pediatric SLE; affected joints in adult RA and JIA) into the circulation and its consumption (by binding to B cells that express BLyS receptors). Elevated blood leukocyte BLyS mRNA levels reflect the entry into the circulation of cells (predominantly neutrophils and/or monocytes) in which transcription of the BLYS gene has been up-regulated. These 2 parameters need not necessarily parallel each other. Indeed, in the present study, plasma BLyS protein levels were not correlated with either of the blood leukocyte BLyS mRNA levels in pediatric SLE despite such significant correlations in adult SLE (18). Moreover, disease activity in pediatric SLE was correlated only with plasma BLyS protein levels and not with blood leukocyte BLyS mRNA levels, whereas the correlation between disease activity and blood leukocyte BLyS mRNA levels in adult SLE was even stronger than that between disease activity and plasma BLyS protein levels (18). These disparities may reflect differences between adult SLE and pediatric SLE in terms of blood cell composition and/or recirculation patterns of activated BLyS-producing cells


(e.g., neutrophils, monocytes). How these differences may relate to disease pathogenesis and/or maintenance is uncertain, and additional experimentation will be required to address these issues. In comparison with the modest differences in BLyS expression profiles between pediatric SLE and adult SLE, the differences in BLyS expression profiles between JIA and adult RA were much more striking. Whereas adult RA is characterized by elevated plasma BLyS protein levels and normal blood leukocyte BLyS mRNA levels (15,16,18,31), the opposite pattern was observed in JIA. Although the mechanistic basis for this unexpected BLyS expression profile in JIA remains elusive at present, its stark contrast to the BLyS expression profile in adult RA may highlight a pathogenetic difference between JIA and adult RA, a point to keep in mind when considering BLyS antagonists in the treatment of JIA. Whereas BLyS antagonism has shown some efficacy in adult RA (38), similar efficacy might not necessarily be realized in JIA. Another notable feature of the results for JIA in the present study is the absence of significant correlations between any of the BLyS parameters and the C-HAQ or circulating ESR/CRP levels. Although the most straightforward explanation for this observation is a dissociation between BLyS expression and disease activity in JIA, this may not necessarily be the case. C-HAQ is influenced not just by ongoing disease activity but by extant disease morbidity as well. Circulating ESR/CRP levels can be influenced by a myriad of variables that have nothing to do with JIA disease activity per se. Thus, whether any of the BLyS parameters would correlate with a better measure of disease activity than the C-HAQ or circulating ESR/CRP remains speculative and will require further investigation. An important caveat to our findings is that the control subjects in the present study were not randomly selected healthy subjects from the population at large, but rather, were clinically healthy siblings of the pediatric SLE or JIA patients. This was necessary from a practical standpoint, inasmuch as we found it nearly impossible to recruit clinically healthy children from an outbred population unrelated to the patients being studied. Whereas healthy adults can give informed consent on their own behalf, minor children cannot do so, and we found the parents/legal guardians at our institution to be highly reluctant to permit phlebotomy of a child without any perceived direct benefit to the child. However, parents/legal guardians of our pediatric SLE or JIA patients were frequently more agreeable to phlebotomy of one or more of their healthy children

3407

(i.e., siblings of the patients) in the context of a perceived ultimate benefit to the control donor’s sibling. Since our control subjects were not randomly selected, the possibility of biologically important bias must be considered. Polymorphisms within the BLYS gene have been reported to be associated both with changes in circulating BLyS protein levels and with changes in BLyS mRNA levels (39–41). Moreover, polymorphisms in each of the 3 BLyS receptors (BCMA, TACI, and BAFF-R/BLyS receptor 3) have also been reported (42–47), which in principle, could affect the expression of BLyS at the protein and/or mRNA level. We did not screen any of our patients for polymorphisms in their BLYS, BCMA, TACI, or BAFFR genes, so it is possible that our control subjects may have included individuals that had a polymorphism in one or more of these genes and, thereby, may have potentially skewed the putative “normal” values. Furthermore, the control subjects in our study may not necessarily be immunologically “normal,” even if none of them harbor a BLYS, BCMA, TACI, or BAFFR polymorphism. Multiple immunologic abnormalities, including the presence of circulating autoantibodies, are quite prevalent in otherwise healthy firstdegree relatives of patients with immune-based systemic rheumatic diseases (48). Since BLyS expression is regulated by numerous cytokines, such as interferon-␥ (IFN␥), IFN␣, and transforming growth factor ␤ (7,8,49,50), differences in the respective cytokine milieus between outbred healthy controls and healthy firstdegree relatives of pediatric SLE or JIA patients may lead to differences in BLyS expression. Nevertheless, we feel confident that any bias introduced by using the patients’ siblings as control subjects was minimal at most. The plasma BLyS protein and blood leukocyte BLyS mRNA values for the pediatric controls in the present study are notably similar to those previously reported for adult controls (18). Since the respective assays in each study were performed in the same laboratories, the similarities in the control results in the 2 studies argue strongly against any technical differences having spuriously given rise to nearly identical values. Thus, the overall similarity in BLyS expression profiles between pediatric SLE and adult SLE is consistent with a role for BLyS overproduction in the development/maintenance of disease in both and identifies BLyS as a rational therapeutic target. The striking dissimilarity in BLyS expression profiles between JIA and adult RA raises questions regarding the precise role of BLyS in these disorders and, by inference, in other inflammatory arthritides. Nevertheless, until

3408

HONG ET AL

BLyS expression profiles are determined in greater numbers of JIA patients with each of the individual subtypes (systemic, polyarticular, and oligoarticular) and confirm our present findings, it would be imprudent to dismiss BLyS as a candidate therapeutic target in JIA. ACKNOWLEDGMENTS

12.

13.

14.

The authors thank all of the subjects and their parents/ legal guardians for their participation in the study. 15.

AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Stohl had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Hong, Reiff, Yang, Shaham, Stohl. Acquisition of data. Hong, Reiff, Yang, Migone, Ward, Marzan, Shaham, Garza, Bernstein. Analysis and interpretation of data. Hong, Reiff, Yang, Migone, Ward, Marzan, Choo Phei, Stohl.

16.

REFERENCES

19.

1. Shlomchik MJ, Madaio MP, Ni D, Trounstein M, Huszar D. The role of B cells in lpr/lpr-induced autoimmunity. J Exp Med 1994;180:1295–306. 2. Chan OT, Hannum LG, Haberman AM, Madaio MP, Shlomchik MJ. A novel mouse with B cells but lacking serum antibody reveals an antibody-independent role for B cells in murine lupus. J Exp Med 1999;189:1639–47. 3. Edwards JC, Szczepanski L, Szechinshi J, Filipowicz-Sosnowska A, Emery P, Close DR, et al. Efficacy of B-cell-targeted therapy with rituximab in patients with rheumatoid arthritis. N Engl J Med 2004;350:2572–81. 4. Moore PA, Belvedere O, Orr A, Pieri K, LaFleur DW, Feng P, et al. BLyS: member of the tumor necrosis factor family and B lymphocyte stimulator. Science 1999;285:260–3. 5. Schneider P, MacKay F, Steiner V, Hofmann K, Bodmer JL, Holler N, et al. BAFF, a novel ligand of the tumor necrosis factor family, stimulates B cell growth. J Exp Med 1999;189:1747–56. 6. Gavin AL, Ait-Azzouzene D, Ware CF, Nemazee D. ⌬BAFF, an alternate splice isoform that regulates receptor binding and biopresentation of the B cell survival cytokine, BAFF. J Biol Chem 2003;278:38220–8. 7. Nardelli B, Belvedere O, Roschke V, Moore PA, Olsen HS, Migone TS, et al. Synthesis and release of B-lymphocyte stimulator from myeloid cells. Blood 2001;97:198–204. 8. Scapini P, Nardelli B, Nadali G, Calzetti F, Pizzolo G, Montecucco C, et al. G-CSF-stimulated neutrophils are a prominent source of functional BLyS. J Exp Med 2003;197:297–302. 9. Scapini P, Carletto A, Nardelli B, Calzetti F, Roschke V, Merigo F, et al. Proinflammatory mediators elicit the secretion of the intracellular B-lymphocyte stimulator pool (BLyS) that is stored in activated neutrophils: implications for inflammatory diseases. Blood 2005;105:830–7. 10. Gavin AL, Duong B, Skog P, Ait-Azzouzene D, Greaves DR, Scott ML, et al. ⌬BAFF, a splice isoform of BAFF, opposes full-length BAFF activity in vivo in transgenic mouse models. J Immunol 2005;175:319–28. 11. Gross JA, Johnston J, Mudri S, Enselman R, Dillon SR, Madden

17.

18.

20.

21.

22.

23. 24. 25.

26.

27.

28.

K, et al. TACI and BCMA are receptors for a TNF homologue implicated in B-cell autoimmune disease. Nature 2000;404:995–9. Kayagaki N, Yan M, Seshasayee D, Wang H, Lee W, French DM, et al. BAFF/BLyS receptor 3 binds the B cell survival factor BAFF ligand through a discrete surface loop and promotes processing of NF-␬B2. Immunity 2002;17:515–24. Wang H, Marsters SA, Baker T, Chan B, Lee WP, Fu L, et al. TACI-ligand interactions are required for T cell activation and collagen-induced arthritis in mice. Nat Immunol 2001;2:632–7. Gross JA, Dillon SR, Mudri S, Johnston J, Littau A, Roque R, et al. TACI-Ig neutralizes molecules critical for B cell development and autoimmune disease: impaired B cell maturation in mice lacking BLyS. Immunity 2001;15:289–302. Zhang J, Roschke V, Baker KP, Wang Z, Alarcon GS, Fessler BJ, et al. Cutting edge: a role for B lymphocyte stimulator in systemic lupus erythematosus. J Immunol 2001;166:6–10. Cheema GS, Roschke V, Hilbert DM, Stohl W. Elevated serum B lymphocyte stimulator levels in patients with systemic immunebased rheumatic diseases. Arthritis Rheum 2001;44:1313–9. Tan S-M, Xu D, Roschke V, Perry JW, Arkfeld DG, Ehresmann GR, et al. Local production of B lymphocyte stimulator protein and APRIL in arthritic joints of patients with inflammatory arthritis. Arthritis Rheum 2003;48:982–92. Collins CE, Gavin AL, Migone T-S, Hilbert DM, Nemazee D, Stohl W. B lymphocyte stimulator (BLyS) isoforms in systemic lupus erythematosus: disease activity correlates better with blood leukocyte BLyS mRNA levels than with plasma BLyS protein levels. Arthritis Res Ther 2006;8:R6. Petri M, Stohl W, Chatham W, McCune WJ, Chevrier M, Ryel J, et al. Association of plasma B lymphocyte stimulator levels and disease activity in systemic lupus erythematosus. Arthritis Rheum 2008;58:2453–9. Dall’Era M, Chakravarty E, Wallace D, Genovese M, Weisman M, Kavanaugh A, et al. Reduced B lymphocyte and immunoglobulin levels after atacicept treatment in patients with systemic lupus erythematosus: results of a multicenter, phase Ib, double-blind, placebo-controlled, dose-escalating trial. Arthritis Rheum 2007; 56:4142–50. Furie R, Stohl W, Ginzler EM, Becker M, Mishra N, Chatham W, et al. Biologic activity and safety of belimumab, a neutralizing anti-B-lymphocyte stimulator (BLyS) monoclonal antibody: a phase I trial in patients with systemic lupus erythematosus. Arthritis Res Ther 2008;10:R109. Tak PP, Thurlings RM, Rossier C, Nestorov I, Dimic A, Mircetic V, et al. Atacicept in patients with rheumatoid arthritis: results of a multicenter, phase Ib, double-blind, placebo-controlled, doseescalating, single- and repeated-dose study. Arthritis Rheum 2008; 58:61–72. Ferraz MB, Goldenberg J, Hilario MO, Bastos WA, Oliveira SK, Azevedo EC, et al. Evaluation of the 1982 ARA lupus criteria data set in pediatric patients. Clin Exp Rheumatol 1994;12:83–7. Petty RE, Southwood TR, Baum J, Bhettay E, Glass DN, Manners P, et al. Revision of the proposed classification criteria for juvenile idiopathic arthritis: Durban, 1997. J Rheumatol 1998;25:1991–4. Lam C, Young N, Marwaha J, McLimont M, Feldman BM. Revised versions of the Childhood Health Assessment Questionnaire (CHAQ) are more sensitive and suffer less from a ceiling effect. Arthritis Rheum 2004;51:881–9. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH, and the Committee on Prognosis Studies in SLE. Derivation of the SLEDAI: a disease activity index for lupus patients. Arthritis Rheum 1992;35:630–40. Brunner HI, Silverman ED, Bombardier C, Feldman BM. European Consensus Lupus Activity Measurement is sensitive to change in disease activity in childhood-onset systemic lupus erythematosus. Arthritis Rheum 2003;49:335–41. La DT, Collins CE, Yang H-T, Migone T-S, Stohl W. B lympho-


29.

30.

31.

32. 33.

34.

35.

36.

37.

38.

cyte stimulator expression in patients with rheumatoid arthritis treated with tumour necrosis factor ␣ antagonists: differential effects between good and poor clinical responders. Ann Rheum Dis 2008;67:1132–8. Cambridge G, Stohl W, Leandro MJ, Migone T-S, Hilbert DM, Edwards JC. Circulating levels of B lymphocyte stimulator in patients with rheumatoid arthritis following rituximab treatment: relationships with B cell depletion, circulating antibodies, and clinical relapse. Arthritis Rheum 2006;54:723–32. Cambridge G, Isenberg DA, Edwards JC, Leandro MJ, Migone T-S, Teodorescu M, et al. B cell depletion therapy in systemic lupus erythaematosus: relationships among serum B lymphocyte stimulator levels, autoantibody profile and clinical response. Ann Rheum Dis 2008;67:1011–6. Bosello S, Youinou P, Daridon C, Tolusso B, Bendaoud B, Pietrapertosa D, et al. Concentrations of BAFF correlate with autoantibody levels, clinical disease activity, and response to treatment in early rheumatoid arthritis. J Rheumatol 2008;35: 1256–64. Klein-Gitelman M, Reiff A, Silverman ED. Systemic lupus erythematosus in childhood. Rheum Dis Clin North Am 2002;28: 561–77. Shaham B, Bernstein B. The rheumatic diseases of childhood. In: Weisman MH, Weinblatt ME, Louie JS, editors. Treatment of the rheumatic diseases: companion to Kelley’s textbook of rheumatology. 2nd ed. Philadelphia: WB Saunders; 2001. p. 423–45. Jin R, Kaneko H, Suzuki H, Arai T, Teramoto T, Fukao T, et al. Age-related changes in BAFF and APRIL profiles and upregulation of BAFF and APRIL expression in patients with primary antibody deficiency. Int J Mol Med 2008;21:233–8. Pranzatelli MR, Tate ED, Hoefgen ER, Swan JA, Coliver JA. Therapeutic down-regulation of central and peripheral B-cellactivating factor (BAFF) production in pediatric opsoclonusmyoclonus syndrome. Cytokine 2008;44:26–32. Takeda T, Sakurai Y, Tatsumi K, Kato J, Kasuda S, Yoshioka A, et al. Elevation of B cell-activating factor belonging to the tumour necrosis factor family (BAFF) in haemophilia A patients with inhibitor. Thromb Haemost 2009;101:408–10. Wallace DJ, Stohl W, Furie RA, Lisse JR, McKay JD, Merrill JT, et al. A phase II, randomized, double-blind, placebo-controlled, dose-ranging study of belimumab in patients with active systemic lupus erythematosus. Arthritis Rheum 2009; 61:1168–78. McKay J, Chwalinska-Sadowska H, Boling E, Valente R, Limanni A, Racewicz A, et al. Belimumab (BmAb), a fully human monoclonal antibody to B-lymphocyte stimulator (BLyS), combined with standard of care therapy reduces the signs and symptoms

3409

39.

40.

41.

42.

43.

44. 45.

46.

47. 48.

49. 50.

of rheumatoid arthritis in a heterogeneous subject population. Arthritis Rheum 2005;52 Suppl:S710–1. Kawasaki A, Tsuchiya N, Fukazawa T, Hashimoto H, Tokunaga K. Analysis on the association of human BLYS (BAFF, TNFSF13B) polymorphisms with systemic lupus erythematosus and rheumatoid arthritis. Genes Immun 2002;3:424–9. Gottenberg J-E, Sellam J, Ittah M, Lavie F, Proust A, Zouali H, et al. No evidence for an association between the ⫺871 T/C promoter polymorphism in the B-cell-activating factor gene and primary Sjo ¨gren’s syndrome. Arthritis Res Ther 2006;8:R30. Novak AJ, Grote DM, Ziesmer SC, Kline MP, Manske MK, Slager S, et al. Elevated serum B-lymphocyte stimulator levels in patients with familial lymphoproliferative disorders. J Clin Oncol 2006;6: 983–7. Kawasaki A, Tsuchiya N, Fukazawa T, Hashimoto H, Tokunaga K. Presence of four major haplotypes in human BCMA gene: lack of association with systemic lupus erythematosus and rheumatoid arthritis. Genes Immun 2001;2:276–9. Salzer U, Chapel HM, Webster AD, Pan-Hammarstrom Q, Schmitt-Graeff A, Schlesier M, et al. Mutations in TNFRSF13B encoding TACI are associated with common variable immunodeficiency in humans. Nat Genet 2005;37:820–8. Castigli E, Wilson SA, Garibyan L, Rachid R, Bonilla F, Schneider L, et al. TACI is mutant in common variable immunodeficiency and IgA deficiency. Nat Genet 2005;37:829–34. Salzer U, Birmelin J, Bacchelli C, Witte T, Buchegger-Podbielski U, Buckridge S, et al. Sequence analysis of TNFRSF13b, encoding TACI, in patients with systemic lupus erythematosus. J Clin Immunol 2007;27:372–7. Losi CG, Silini A, Fiorini C, Soresina A, Meini A, Ferrari S, et al. Mutational analysis of human BAFF receptor TNFRSF13C (BAFF-R) in patients with common variable immunodeficiency. J Clin Immunol 2005;25:496–502. Warnatz K, Salzer U, Gutenberger S, Schlesier M, Grimbacher B, Peter HH, et al. Finally found: human BAFF-R deficiency causes hypogammaglobulinemia. Clin Immunol 2005;115:S20. Dudeney C, Shoenfeld Y, Rauch J, Jones M, Young CM, Tavassoli M, et al. A study of anti-poly(ADP-ribose) antibodies and an anti-DNA antibody idiotype and other immunological abnormalities in lupus family members. Ann Rheum Dis 1986;45:502–7. Litinskiy MB, Nardelli B, Hilbert DM, He B, Schaffer A, Casali P, et al. DCs induce CD40-independent immunoglobulin class switching through BLyS and APRIL. Nat Immunol 2002;3:822–9. Kim H-A, Jeon S-H, Seo G-Y, Park J-B, Kim P-H. TGF-␤1 and IFN-␥ stimulate mouse macrophages to express BAFF via different signaling pathways. J Leukocyte Biol 2008;83:1431–9.