We recruited 116 patients with RA and 31 gender and age-matched healthy subjects. The demographics and clinical characteristics of the RA patients are presented in Table 1. Among 116 patients, 83 patients (71.6%) were female. The median age of RA patients was 66 years. The majority of RA patients received DMARDs, mostly methotrexate or MTX in combination with other synthetic DMARDs. Despite the treatments, a median-DAS28-ESR score of 2.8.
We measured serum levels of Gal-9 in RA patients using a specific ELISA assay. As shown in Figure. 1, serum Gal-9 concentrations in patients with RA were significantly higher compared to those in healthy subjects (median 7577 pg/ml; [interquartile range (IQR); 5570–10201] versus 4738 pg/ml [IQR; 4267–5630], p=0.001). Serum levels of Gal-9 were compared in the sub-grouped RA patients stratified by the disease durations. However, there was no significant difference in serum Gal-9 between RA patients with and without shorter disease durations ( less than 5 years, 7009 pg/ml, IQR[5134-4527] versus 5 years or more, 7886 pg/ml, IQR[6154-10692], p=0.40). Although there was no significant difference in serum levels of Gal-9 between RA patients with and without smoking history (Figure 2A, p=0.615), higher levels of serum Gal-9 were observed predominantly in RA patients with RA-ILD (9606 pg/ml [IQR: 8522–12167] versus 7078 pg/ml [IQR: 5225–9447], p<0.001) (Figure 2B). We also found a significant difference in ACPA titers between RA patients with and without RA-ILD (128.3 U/ml, IQR[24.7-896.0] versus 38.1 U/ml, IQR[2.5-215.2], p=0.014).
We investigated the relationship between serum Gal-9 and each parameters of RA patients (Figure 3A). Serum Gal-9 were significantly correlated with ESR (r=0.344, p<0.001), MMP-3 (r=0.234, p=0.004) and ACPA titers (r=0.275, p=0.002). Also serum Gal-9 was significantly correlated with RA disease activity scores, DAS28-ESR (r=0.269, p=0.005).
Although we investigated the correlation between rheumatoid factor and serum Gal-9, there was no significant correlation between rheumatoid factor and Gal-9 (r=0.16, p=0.09, data not shown).
To further evaluate the ability of serum Gal-9 to differentiate RA phenotype, we analyzed the distribution pattern of serum Gal-9 values in combination with ACPA titer (Figure 3A). The cut-off values of ACPA titers (200 U/ml) was determined according to the ability to differentiate the inverse correlation between Gal-9 and ACPA titer. When RA patients were grouped according to the presence of high titers ACPA (≧200U/ml), some correlations between circulating Gal-9 and clinical features were identified. In the two dimensional heat-nap consisting serum Gal-9 and ACPA titer, we identified two groups (Figure 3A). Group1 RA patients exhibited high ACPA titers (ACPA≧200 U/ml), and Group 2 RA patients exhibited moderate to low ACPA titers (ACPA<200 U/ml). There was a significant modest correlation between Gal-9 and ACPA titers in Group1 RA patients (r=0.508, p=0.002). Conversely, there was no correlation between Gal-9 and ACPA titer in Group 2 RA patients (r=0.211, p=0.060) suggesting that Gal-9 was not modulated by the status of ACPA titers (Figure 3B). Next we evaluated the correlations between Gal-9 and clinical parameters in subdivided Group 1 or Group 2 RA patients. A significant correlations between circulating Gal-9 and inflammatory markers, ESR (r=0.451, p<0.001) or DAS28-ESR (r=0.331, p=0.004) were identified in RA patients with low titers of ACPA (<200U/ml). However, there was no correlation between circulating Gal-9 and these parameters (r=0.203, p=0.249 and r=0.160, p=0.365) in RA patients with high titers of ACPA (≧200U/ml) (Figure 4A, 4B). Similarly, in RA patients with high titers of ACPA (≧200U/ml), there was no correlation between serum Gal-9 and MMP-3 (r=0.111, p=0.519), whereas serum Gal-9 levels were significantly correlated with serum MMP-3 (r=0.300. p=0.007) in RA patients with low titers of ACPA (<200U/ml) (Figure 4C). These findings suggest that serum Gal-9 were upregulated in link to autoimmune response in RA patients with high titer ACPA. In contrast, serum Gal-9 was upregulated in response to inflammatory mediators in RA patients with low titers of ACPA.
From a clinical point of view, the circulating Gal-9 were compared according to the presence or absence of clinical remission in these sub-grouped RA patients (Figure 5). In RA patients low titers of ACPA (<200U/ml), circulating Gal-9 were significantly higher in patients without clinical remission compared to those with clinical remission (8252 pg/ml [IQR: 5870–10996] versus 7103 pg/ml [IQR: 5328–8357], p=0.013). There was no significant difference in circulating Gal-9 between patients with and without clinical remission (10647 pg/ml [IQR: 6960–13367] versus 8635 pg/ml [IQR: 6372–10092], p=0.703) in RA patients with high titers of ACPA (≧200U/ml).
Finally, we subdivided RA patients according to the progressing of joint damage (Stage) and we evaluated the relationship between serum Gal-9 and progressive joint damage. As shown in Figure 6A, RA patients with advanced articular lesions (Stage Ⅱ-Ⅳ) had significantly higher levels of circulating Gal-9 compared to those without advanced articular lesions (Stage I) (8790 pg/ml [IQR: 5631–10953] versus 7103 pg/ml [IQR: 5882–8810], p=0.023). There was a significant difference in circulating Gal-9 between those with and without advanced articular lesions (Stage Ⅱ-Ⅳ) in RA patients low titers of ACPA (7367 pg/ml [IQR: 5931–10109] versus 7009 pg/ml [IQR: 4602–9450] p=0.004). In contrast, there was no significant difference in circulating Gal-9 between those with and without advanced articular lesions (Stage Ⅱ-Ⅳ) in RA patients with high titers of ACPA (10146 pg/ml [IQR: 8771–12257] versus 5882 pg/ml [IQR: 5053–8138], p=0.182) (Figure 6B).