Low T cell production of TNFalpha and IFNgamma in ankylosing spondylitis: its relation to HLA-B27 and influence of the TNF-308 gene polymorphism

Abstract

Objective: To test the hypothesis that ankylosing spondylitis (AS) is a T helper cell type 2 polarised disease by quantifying the T cell cytokines interferon gamma (IFNgamma), interleukin 4 (IL4), tumour necrosis factor alpha (TNFalpha), and IL10 at the single cell level in patients with AS in comparison with healthy HLA-B27 negative and HLA-B27 positive controls.

Methods: Peripheral blood mononuclear cells from 65 subjects (25 HLA-B27 positive patients with active AS, 18 healthy HLA-B27 positive controls, and 22 healthy HLA-B27 negative controls) were stimulated with phorbol myristate acetate/ionomycin for six hours, surface stained for CD3 and CD8, intracellularly stained for the cytokines IFNgamma, TNFalpha, IL4, and IL10, and analysed by flow cytometry. TNFalpha production was related to the genotype of the TNFalpha promoter at the -308 and -238 polymorphisms.

Results: In peripheral blood the percentage of TNFalpha+ T cells was significantly lower in HLA-B27 positive patients with AS (median 5.1% for CD4+ T cells) than in healthy HLA-B27 negative controls (median 9.5%; p=0.008). Surprisingly, the percentage of TNFalpha+ T cells was also significantly lower in healthy HLA-B27 positive controls (median 7.48%) than in healthy HLA-B27 negative controls (p=0.034). Furthermore, the percentage of IFNgamma+ T cells was lower in patients with AS and in healthy HLA-B27 positive controls than in healthy HLA-B27 negative controls (p=0.005 and p=0.003, respectively). The percentage of IL10+/CD8+ T cells was higher in patients with AS than in both control groups. In HLA-B27 positive subjects, TNF1/2 heterozygosity at -308 (n=6) was associated with a higher percentage of TNFalpha+ T cells than TNF1/1 homozygosity (n=25; median 9.97% v 5.11% for CD4+ T cells; p=0.017). In contrast, in HLA-B27 negative controls (n=18) there was no such genotype/phenotype correlation (median 9.4% v 10.6%).

Conclusions: The lower T cell production of TNFalpha and IFNgamma shown at the single cell level in HLA-B27 positive patients with AS and healthy HLA-B27 positive controls may contribute to the increased susceptibility of HLA-B27 positive subjects to develop AS. Preliminary genotype-phenotype correlations suggest that in HLA-B27 positive subjects TNF2 at -308 or a linked gene results in higher TNFalpha production and, therefore, might be a marker for a protective haplotype.

Figure 1

Percentage of tumour necrosis factor α (TNFα) positive CD4+ (A) and CD8+ (B) T cells among patients with AS (all HLA-B27 positive), healthy HLA-B27 positive controls, and healthy HLA-B27 negative controls. The horizontal bars indicate the medians.

Figure 2

Percentage of interferon γ (IFNγ) positive CD4+ (A) and CD8+ (B) T cells among patients with AS (all HLA-B27 positive), healthy HLA-B27 positive controls, and healthy HLA-B27 negative controls. The horizontal bars indicate the medians.

Figure 3

Percentage of interleukin 10 (IL10) positive CD8+ T cells among patients with AS (all HLA-B27 positive), healthy HLA-B27 positive controls, and healthy HLA-B27 negative controls. The horizontal bars indicate the medians.

Figure 4

Percentage of tumour necrosis factor α (TNFα) positive CD4+ (A) and CD8+ (B) T cells in relation to genotypes at TNF-308 (TNF1/1 and TNF1/2) among HLA-B27 positive subjects (patients with AS and healthy HLA-B27 positive controls combined) and healthy HLA-B27 negative controls. TNF genotypes were available for 49 subjects. The horizontal bars indicate the medians.

Figure 5

Percentage of tumour necrosis factor α (TNFα) positive CD4+ (A) and CD8+ (B) T cells among TNF1/1 homozygous patients with AS (n=14), healthy HLA-B27 positive controls (n=11), and healthy HLA-B27 negative controls (n=15). The horizontal bars indicate the medians.