Association of a functional CD19 polymorphism with susceptibility to systemic sclerosis




CD19 is overexpressed in B cells from patients with systemic sclerosis (SSc), and plays a crucial role not only for autoantibody production, but also for skin fibrosis in mouse models. We previously reported the association of a GT repeat polymorphism in the 3′-untranslated region (3′-UTR) of CD19 with human systemic lupus erythematosus. In this study, we examined whether CD19 polymorphisms are associated with genetic susceptibility to SSc.


A case–control association study was performed for CD19 polymorphisms, −499G>T in the promoter region and a GT repeat polymorphism in the 3′-UTR, in 134 patients with SSc and 96 healthy individuals recruited at Kanazawa University. CD19 expression levels in the peripheral blood naive and memory B cells from SSc patients were examined by 2-color flow cytometry.


Carrier frequencies of the −499T allele in the promoter (odds ratio [OR] 2.18, 95% confidence interval [95% CI] 1.31–3.86, P = 0.003) and of the (GT)14 allele in the 3′-UTR (OR 1.86, 95% CI 1.05–3.28, P = 0.03) were significantly increased in SSc patients compared with healthy controls. Association was particularly evident in patients with limited cutaneous SSc with anticentromere antibodies. These alleles were in linkage disequilibrium, but the −499T allele seemed to play a primary role. CD19 expression levels in peripheral blood B cells were significantly elevated in both naive (P = 0.0029) and memory (P = 0.0022) B cells from patients with SSc who had the −499T allele as compared with those without the −499T allele.


The CD19 −499G>T polymorphism is associated with higher CD19 expression in B cells, and with susceptibility to SSc.

Although the etiology of systemic sclerosis (SSc; scleroderma) remains elusive, epidemiologic data, including ethnic influences on prevalence (1) and familial aggregation (2), strongly suggest a role for genetic factors. To date, polymorphisms of genes such as HLA (3), fibrillin 1 (4), and interleukin-1α (IL-1α) (5) have been reported to be associated with SSc.

The pathogenesis of SSc is highly complex. The immune system plays a significant role, along with fibroblasts and the vascular system. Most patients with SSc have autoantibodies against intracellular components such as topoisomerase I or centromere. Hypergammaglobulinemia is often observed in SSc, and hyperreactivity of SSc memory B cells with augmented immunoglobulin production is also detected (8). These observations suggest the role of B cells in immunologic abnormalities associated with SSc. CD19, a B cell–specific signal transduction molecule, is a general “rheostat” that defines signaling thresholds critical for humoral immune responses and autoimmunity (6). We previously demonstrated that the expression level of CD19 was increased in peripheral blood B cells from patients with SSc by ∼20% compared with normal individuals (7). Overexpression of CD19 is observed both in memory and in naive B cells (8). CD19 overexpression induces autoantibody production in mice to an extent similar to that observed in patients with SSc (7). Moreover, CD19 tyrosine phosphorylation was constitutively augmented in the TSK/+ mouse, a genetic model for human SSc, and loss of CD19 significantly decreased skin fibrosis as well as hypergammaglobulinemia and autoantibody production, indicating a crucial role for CD19 in these mice (9).

The human CD19 gene is located at 16p11.2. The chromosomal region 16p13–16q13 encompassing CD19 has been reported to be linked with autoimmune diseases such as rheumatoid arthritis (10), systemic lupus erythematosus (SLE) (11), and Crohn's disease. In the latter, the major susceptibility gene in this region has been shown to be CARD15/NOD2 (12, 13). In the process of systematic polymorphism screening of immune system genes and association studies with human rheumatic diseases (14), we identified human CD19 gene polymorphisms and detected association of a GT repeat polymorphism within the 3′-untranslated region (3′-UTR) with susceptibility to SLE (15). Interestingly, the expression level of CD19 was significantly decreased in peripheral blood B cells from patients with SLE (7, 8), and the allele that was increased in SLE was indeed associated with a tendency toward reduced levels of CD19 messenger RNA (mRNA) (15).

These lines of evidence led us to a hypothesis that genetic susceptibility to SSc may be associated with polymorphisms of CD19 that result in overexpression of CD19. We performed a case–control association study to examine the association of CD19 polymorphisms with susceptibility to SSc, and also with CD19 expression levels on B cells.


Patients and controls.

One hundred thirty-four patients with SSc (117 women and 17 men, mean ± SD age 52.5 ± 13.3 years) and 96 healthy individuals (38 women and 58 men, mean ± SD age 27.8 ± 6.4 years) were recruited for this study at the Department of Dermatology, Kanazawa University. All subjects were unrelated Japanese, and most of them were residents of the north central part of Japan. All patients fulfilled the criteria for SSc proposed by the American College of Rheumatology (formerly, the American Rheumatism Association) (16). The average disease duration in the patients was 7 years. Fifty-two patients (38.8%) were categorized as having diffuse cutaneous SSc (dcSSc), and 82 (61.2%) were categorized as having limited cutaneous SSc (lcSSc) (17). Antibodies to topoisomerase I, centromere, and U1 RNP were detected in 36 (26.8%), 61 (45.5%), and 11 (8.2%) of the patients, respectively. Thirty-nine patients (29.1%) were receiving corticosteroids at the time of the study. This study was reviewed and approved by the Research Ethics Committees of Kanazawa University and the University of Tokyo.

Detection of antinuclear antibodies.

Antinuclear antibodies were determined by indirect immunofluorescence using HEp-2 cells as the substrate, and autoantibody specificities were further assessed by enzyme-linked immunosorbent assay (ELISA) and immunoprecipitation. Specifically, anticentromere antibody was determined by the presence of a discrete speckled pattern on indirect immunofluorescence and was confirmed by ELISA using human recombinant CENP-B as antigen (Medical and Biological Laboratories, Nagoya, Japan). Anti–topoisomerase I antibody was determined by ELISA using human recombinant topoisomerase I as antigen (Medical and Biological Laboratories) and was further confirmed by immunoprecipitation when the titer was low in ELISA. Similarly, anti–U1 RNP antibody was determined by ELISA using human recombinant 70K, A, and C proteins as antigen (Medical and Biological Laboratories) and was further confirmed by immunoprecipitation when the titer was low on ELISA.

Genotyping and case–control association study.

Among the CD19 polymorphisms identified in our previous study (15), the GT repeat polymorphism in the 3′-UTR and the −499G>T single-nucleotide polymorphism (SNP), the only promoter region polymorphisms with a minor allele frequency of >0.05, were genotyped, because these polymorphisms were considered more likely than other polymorphisms to be associated with altered expression levels of CD19. Genomic DNA was purified from peripheral blood leukocytes using a QIAamp DNA Blood Mini kit (Qiagen, Hilden, Germany).

Nested polymerase chain reactions (PCRs) were used to genotype −499G>T SNP, followed by single-strand conformation polymorphism (SSCP) analysis or direct sequencing (15). The first primer set (CD19promoterF [CTCAAGTTTCCAGCCTCAATC] and CD19promoterR [GCACTCAACCATGGGTGTCT]) was designed to amplify −1351 to 136 using long PCR. Amplification conditions consisted of initial denaturation at 96°C for 10 minutes, followed by 35 cycles of denaturation at 96°C for 1 minute, annealing at 58°C for 30 seconds, and extension at 72°C for 3 minutes, using a thermal cycler (Biometra, Göttingen, Germany). The second amplification was performed using the primer set CD19promoterBF (TCAAGCGATCCTTCTACCTC) and CD19promoterBR (CCTGTAATCCCAGCTACTTA). Amplification conditions consisted of initial denaturation at 96°C for 10 minutes, followed by 35 cycles of denaturation at 96°C for 30 seconds, annealing at 60°C for 30 seconds, and extension at 72°C for 30 seconds, using a thermal cycler. The amplified DNA was analyzed using either SSCP analysis (25°C, 90 minutes) or direct sequencing.

For the GT repeat polymorphism within the 3′-UTR, a 256-bp fragment encompassing the repeat was amplified using Fam- or Hex-labeled CD19-3′-UTRF (AGAGGGAACAGGGTTCCTAG) and CD19ex15R (AGGAATACAAAGGGGACTGG) primers. Amplification conditions were the same as those for the second PCR for the promoter SNP. The amplified product was analyzed with a DNA sequencer using GeneScan software (Applied Biosystems, Foster City, CA). Association of the genotype with disease susceptibility or clinical characteristics was statistically analyzed by chi-square test or, when 1 or more of the variables were ≤5, by Fisher's exact test, using StatView for Windows (version 5.0; SAS Institute, Cary, NC). Linkage disequilibrium parameters were estimated from typing results using the EH program (18).

Flow cytometric analysis.

Surface expression intensity of CD19 in peripheral blood memory and naive B cells was examined in 40 patients with SSc, of whom 15 had the −499 G/G, 23 had the G/T, and 2 had the T/T genotype, by 2-color flow cytometry using phycoerythrin-conjugated anti-CD19 (B4; Beckman Coulter, Miami, FL) and fluorescein isothiocyanate–conjugated CD27 monoclonal antibodies (M-T271; BD PharMingen, San Diego, CA), as described elsewhere in detail (8). The difference in the mean fluorescence intensity (MFI) of CD19 staining between patients with and those without the −499T allele was compared separately in CD27med (memory) and CD27 (naive) B cell subpopulations. Statistical significance was examined by Student's t-test, using StatView.


Association of CD19 polymorphisms with SSc.

Genotype frequencies at the −499G>T SNP in the patients with SSc were deviated significantly from those in the controls (P = 0.008). Allele carrier frequency of −499T (T/T and G/T genotypes combined) was significantly increased in SSc (odds ratio [OR] 2.18, 95% confidence interval [95% CI] 1.31–3.86, P = 0.003), as was the allele frequency of −499T (OR 1.69, 95% CI 1.09–2.62, P = 0.019) (Table 1).

Table 1. Association of CD19 polymorphisms with systemic sclerosis (SSc)*
 SSc (n = 134)Controls (n = 96)POR (95% CI)
  • *

    Values are the number (%). OR = odds ratio; 95% CI = 95% confidence interval; SNP = single-nucleotide polymorphism; NS = not significant; 3′-UTR = 3′-untranslated region.

  • (T/T + G/T) versus G/G.

  • (G/G + G/T) versus T/T.

  • §

    Chi-square analysis using 2 × 2 contingency table.

  • [(GT)14/(GT)14 + (GT)14/(GT)x] versus (GT)x/(GT)x, when x is any allele other than (GT)14.

Promoter SNP (−499G>T)    
 Genotype frequency    
  T/T8 (6.0)6 (6.3)0.0081.35 (0.44–4.13)
  G/T63 (47.0)26 (27.1) 2.45 (1.39–4.35)
  G/G63 (47.0)64 (66.6) 1
 Allele frequency    
  T79 (29.5)38 (19.8)0.0191.69 (1.09–2.62)
  G189 (70.5)154 (80.2) 1
 Allele carrier frequency    
  T +71 (53.0)32 (33.3)0.0032.18 (1.31–3.86)
  G +126 (94.0)90 (93.8)NS 
3′-UTR (GT)n    
 Allele frequency§    
  101 (0.4)1 (0.5)NS 
  12179 (66.8)135 (70.3)  
  1320 (7.5)20 (10.4)  
  1456 (20.9)28 (14.6)  
  153 (1.1)3 (1.6)  
  189 (3.4)5 (2.6)  
 Allele carrier    
  frequency: 14 +53 (39.6)25 (26.0)0.031.86 (1.05–3.28)

In the 3′-UTR, the frequency of (GT)14 allele carriers was marginally increased among patients with SSc (OR 1.86, 95% CI 1.05–3.28, P = 0.03) (Table 1). This allele was in linkage disequilibrium with −499T both in the controls (D′ = 0.86, r2 = 0.51, P = 4.3 × 10−23) and in the patients (D′ = 0.83, r2 = 0.43, P = 6.0 × 10−27). To examine whether either one of the polymorphic sites plays a primary role, 2-locus analysis was performed (Table 2). Individuals carrying −499T but who did not carry (GT)14 (group b) demonstrated a significantly increased risk for SSc compared with those carrying neither of these alleles (group d). In contrast, the risk among individuals carrying both alleles (group a) was not different from that among those with only the −499T allele (group b). These results strongly suggested that the −499G>T SNP has the primary role in the association with SSc.

Table 2. Primary role of −499T for the susceptibility to SSc*
Group −499T/(GT)14 statusSSc (n = 134)Controls (n = 96)POR (95% CI)
  • *

    P values, OR, and 95% CI were calculated in relation to individuals carrying neither −499T nor (GT)14 (group d). Values are the number (%). See Table 1 for definitions.

a (+/+)47 (35.1)22 (22.9)0.0092.28 (1.23–4.23)
b (+/–)24 (17.9)10 (10.4)0.0222.56 (1.15–5.75)
c (–/+)6 (4.5)3 (3.1)0.322.14 (0.24–18.8)
d (–/–)57 (42.5)61 (63.5) 1

With respect to the clinical characteristics, the −499T allele was found to be significantly associated with the lcSSc, anticentromere antibody–positive, anti–topoisomerase I–negative patient subgroups (Table 3).

Table 3. Association of CD19 −499T with limited cutaneous SSc and anticentromere antibody positivity*
−499G > TLimited/diffuse cutaneous SScAnticentromere antibodyAnti–topoisomerase antibodyControls (n = 96)
Limited (n = 82)Diffuse (n = 52)Positive (n = 61)Negative (n = 73)Positive (n = 37)Negative (n = 97)
  • *

    Values are the number (%). See Table 1 for definitions.

  • Genotype frequencies were compared between each SSc clinical subset and the control group by chi-square test (2 × 3 contingency table).

  • Allele carrier frequencies were compared between each SSc clinical subset and the control group by chi-square test (2 × 2 contingency table).

 T/T4 (4.9)4 (7.7)3 (4.9)5 (6.8)3 (8.1)5 (5.2)6 (6.3)
 G/T44 (53.6)19 (36.5)31 (50.8)32 (43.8)14 (37.8)49 (50.5)26 (27.1)
 G/G34 (41.5)29 (55.8)27 (44.3)36 (49.3)20 (54.1)43 (44.3)64 (66.6)
Allele carrier       
  frequency: T+48 (58.5)23 (44.2)34 (55.7)37 (50.7)17 (45.9)54 (55.7)32 (33.3)
 OR (95% CI)2.82 (1.54–5.17)1.59 (0.79–3.17)2.51 (1.31–4.85)2.06 (1.10–3.83)1.70 (0.79–3.68)2.51 (1.41–4.48) 

Association with CD19 expression level.

We next examined whether −499G>T SNP is associated with CD19 expression levels on the surface of peripheral blood B cells in SSc patients. A representative flow cytometric profile is shown in Figure 1a. The results are summarized in Figure 1b. The MFI of CD19 staining was significantly stronger in B cells in the patients carrying the −499T allele compared with those without −499T, both in the naive B cells (mean ± SD MFI 443.6 ± 70.4 versus 373.8 ± 61.4 [18.7% increase]; P = 0.0029) and the memory B cells (502.6 ± 86.6 versus 421.9 ± 50.4 [19.2% increase]; P = 0.0022). Among the 40 patients examined, 17 were taking low-dose prednisolone (2–15 mg/day); however, no correlation was observed between the CD19 expression level and the dosage of prednisolone (r = 0.22, P = 0.90) (Figure 1c).

Figure 1.

CD19 expression levels on peripheral blood B cells from patients with systemic sclerosis (SSc). a, Representative 2-color flow cytometry patterns in SSc patients with or without the CD19 −499T allele. Polygons indicate CD19+,CD27− naive B cells and CD19+,CD27med memory B cells. The horizontal dashed line is provided for reference. b, Cell surface CD19 levels on naive and B cells from SSc patients with and without the CD19 −499T allele. Values represent the mean fluorescence intensity of B cells stained for CD19. The mean ± SD of each group is indicated. c, Lack of correlation of prednisolone dose and expression levels of CD19 on memory B cells of SSc patients. Correlation was also not observed in naive B cells (data not shown). lcSSc = limited cutaneous SSc; dcSSc = diffuse cutaneous SSc.


In the present study, we demonstrated that an SNP in the upstream region of CD19, −499G>T, is genetically associated with susceptibility to SSc. Moreover, the −499T allele was associated with an 18–19% higher expression level of CD19 on the surface of B cells, which could, at least in part, account for our previous observations indicating enhanced CD19 expression in SSc patients (7). In a study of CD19-transgenic mice, we previously demonstrated that a 15–29% increase in CD19 expression level could lead to production of various autoantibodies (7). Those findings, taken together with our present observations, suggest that the SNP is indeed relevant to SSc pathogenesis. This study provides evidence that CD19 overexpression in SSc is at least partly based on genetic background.

At this time, the mechanism by which the −499G>T SNP leads to overexpression of CD19 is unclear. The nucleotide sequence encompassing this SNP was not predicted to be a binding site for transcription factors ( (19). Transcription factors such as BSAP (PAX5) (20), SP1 and Egr-1 (21), and EBF-like protein (22) have been shown to bind to the immediate proximal promoter region of human CD19 (up to ∼−300 bp); however, little is known about the role of the more-upstream promoter region for transcription regulation. Of interest, the presence of a negative regulatory cis element beyond the −400-bp upstream region has been suggested (21). Thus, our observations raise an intriguing possibility that the −499G>T SNP may alter affinity for negatively regulating transcriptional factors. This needs to be investigated in the future.

Another new finding in this study was that the association with −499G>T was observed preferentially in lcSSc patients with anticentromere antibody. Although this association needs to be confirmed in large-scale studies, a possible difference in the genetic backgrounds of lcSSc and dcSSc patients has already been suggested from HLA studies (23, 24). Consistent with this, each SSc subset is associated with specific organ involvement; for example, dcSSc patients frequently exhibit lung fibrosis and scleroderma renal crisis, while lcSSc patients show pulmonary hypertension (17). The lack of association of −499G>T with dcSSc, however, does not exclude the role of CD19, because expression of CD19 may be regulated not only by the promoter SNP, but also by other molecules at the transcriptional, posttranscriptional, and posttranslational levels. On B cells, CD19 is expressed in association with at least 3 other cell surface proteins: CD21, CD81, and Leu-13/CD225 (6). CD19 expression levels are ∼50% higher on B cells from mice deficient in CD21 expression (25). In addition, CD19–CD81 interactions have been shown to influence CD19 expression in a postendoplasmic reticulum compartment during B cell development (26).

This study has some limitations. The number of subjects, especially the regional controls, was rather small; thus, to decrease the possibility of Type I and Type II errors, association with disease susceptibility needs to be confirmed by independent studies. The sex ratio was substantially different between the patient and control groups; however, because the genotype frequencies were not significantly different between male and female controls (data not shown), essentially identical results would have been derived if adjustment for sex ratio were performed. The controls were younger than the patients, which could have made our analysis conservative. However, because of the low prevalence of SSc, it is unlikely that difference in age distribution had a substantial effect on the results.

In the case of SLE, extended GT repeat alleles in the 3′-UTR with 15–18 repeats were associated with lower expression levels of CD19 mRNA, and with disease susceptibility (15). Lower expression of CD19 at the protein level on the surface of B cells was also confirmed in SLE (7, 8). Thus, different CD19 polymorphisms with opposite functions are associated with 2 systemic autoimmune rheumatic diseases, which underscores the critical role of tight regulation of CD19 in the maintenance of normal immune responses. Further studies are required to disclose the mechanism that regulates the expression level of CD19, and eventually enable the controlling of CD19 expression levels as a therapeutic approach.