This report is the first to demonstrate the possibility of optimizing the selection between TNF-i and IL-17-i based on serum IL-22 concentrations.
PsA is a disease with very high heterogeneity, and its clinical symptoms vary among patients. Thus, we focused on serum cytokine concentrations as biomarkers that could contribute to simple and proper drug selection. In addition, we previously reported that drug selection based on phenotyping of peripheral blood lymphocytes leads to more frequent achievement of low disease activity as assessed by the SDAI (7). In the present study, we aimed to develop treatment strategies that would lead to achievement of DAPSA-REM, PASI90 or Minimal Disease Activity and more effective control of disease activity.
Since treatment was selected based on shared decision-making between the attending physicians and the patients in this non-randomized study, there were statistically significant differences in the prevalence of comorbid diabetes mellitus and serum IL-17A concentrations measured after the end of the observation period in cohort 1 (Supplementary Table S1). However, we continued the analyses described below, because we did not intend to directly compare the efficacy of TNF-i and IL-17-i. Instead, we aimed to identify cytokines predictive of responses to each of the TNF-i and IL-17-i therapies and to develop new treatment strategies based on the identified cytokines.
Univariate logistic analyses identified serum IL-22 concentrations as a factor contributing to the achievement of DAPSA-REM in IL-17-i therapy (Supplementary Table S4).
Patients in cohort 1 were stratified based on serum IL-22 concentrations, and the efficacy of IL-17-i and TNF-i therapies was evaluated. The results showed that significantly more patients with low IL-22 concentrations who were treated with IL-17-i achieved DAPSA-REM and Minimal Disease Activity (Table 1). Meanwhile, a comparison of the baseline characteristics showed significant differences in the concomitant use of MTX (Supplementary Table S6). Although we considered that low IL-22 concentrations may reflect the low inflammatory state due to MTX, no statistically significant differences were observed in the baseline values of CRP, or serum IL-22 concentrations between the 29 patients with concomitant use of MTX and the 18 patients without in cohort 1 (Supplementary Figure S3). Similarly, there was no significant difference in serum IL-17 concentrations between these two groups with and without concomitant use of MTX. Since serum concentrations of IL-17 and IL-22 did not correlate with CRP (Supplementary Figure S4), the levels of CRP and the concomitant use of MTX seemed unlikely to reflect the serum concentrations of IL-17 and IL-22. Thus, we focused on IL-22 concentrations as a simple biomarker for the proper selection of TNF-i and IL-17-i to develop treatment strategies.
Moreover, no factors contributing to the achievement of PASI90 were identified for either TNF-i or IL-17-i therapy (Supplementary Table S4), and no differences were observed in efficacy in terms of PASI during treatment (Table 1 and 3). This may have been due to the fact that the PASI was already favorable at baseline.
Similar to helper T (Th) 17 cells, Th22 expresses chemokine receptor (CCR) 4 and CCR6. In addition, IL-22 is a cytokine produced by Th22 cells, a relatively new Th cell subset that is positive for CCR10 and expresses the aryl hydrocarbon receptor as the master transcription factor. In addition, an increase in IL-22-producing CD4+ T cells in the peripheral blood, as well as an increase in IL-17-producing CD4+ T cells is characteristic of PS and PsA. In addition, serum IL-22 concentrations have been reported to be characteristically and significantly higher in PsA than in PS (9). In a study using an in vitro experimental system with human naïve T cells, we demonstrated that the addition of IL-1β, TNF-α, and IL-6 to T cell receptors stimulation induces CD3+CD4+CCR4+CCR6+CCR10+Th22 cells that selectively produce IL-22 (10). Moreover, when serum cytokine concentrations were compared among the HC, IL-22 high, and IL-22 low groups in the present study, serum IL-17 concentrations were significantly higher in both the IL-22 high and IL-22 low groups than in the HC group, whereas no significant difference was observed between the IL-22 high and IL-22 low groups. Alternatively, serum TNF-α concentrations were significantly higher only in the IL-22 high group than in the HC and IL-22 low groups (Figure 3). As described above, the differentiation of Th22 cells is dependent of TNF-α; however, the differentiation of Th17 cells is not (10).Therefore, we assumed that cell populations characterized by IL-17 production, such as Th17 cells and innate lymphoid cells (ILCs), may be mainly involved in the pathogenesis of patients with low IL-22 concentrations, and that, in addition to Th17 cells and ILCs, Th22 cells, which are characterized by the production of IL-17 and IL-22 and induced by cytokines including TNF-α, may be involved in the pathogenesis of patients with high IL-22 concentrations. Based on this assumption, we developed treatment strategies for initiating TNF-i therapy in patients with high IL-22 concentrations and IL-17-i therapy in patients with low IL-22 concentrations.
In cohort 2 (validation cohort), we compared treatment responses over 1 year between the strategic and mismatched treatment groups. The results showed that the rate of achieving DAPSA-REM and Minimal Disease Activity at 1 year was significantly higher in the strategic treatment group (Table 3). As we hypothesized, it was suggested that patients with PsA may include two types of patients. In the first type of patients, Th17 cells and ILCs, which are involved in IL-17 production, are mainly involved in pathogenesis, which is consistent with the conventional assumption. In the second type of patients, Th22 cells, which are involved in IL-22 production, are also involved in the pathogenesis of this disease. Future studies on the correlation between patients with high IL-22 concentrations and Th22 cell subsets in peripheral blood and on changes in serum concentrations of IL-22 and Th22 cells due to treatment are expected to advance the elucidation of the pathology with a focus on the Th22-IL-22 correlation.
This study excluded patients using IL-12/23 (p40) inhibitors, because no clinical studies have demonstrated that IL-12/23 (p40) inhibitors are comparable to TNF-i or IL-17-i. This study focused on the proper selection of only TNF-i and IL-17-i. Moreover, patients using IL-23 (p19) inhibitors, which were approved in Japan only shortly before this study, were also excluded due to an insufficient number of patients. Based on our hypothesis that there are patients whose main pathogenic factors are Th17 cells and ILCs, which are involved in IL-17 production, and patients in whom Th22 cells involved in IL-22 production are also involved in the pathogenesis, IL-23 (p19) inhibitors may be more effective in the former. In addition, the use of JAK-i has recently expanded to include the treatment of PsA (11, 12). Since JAK-i is expected to inhibit the differentiation of Th17 cells as well as Th22 cells, it appears to be a potential drug that is highly effective for a wide spectrum of PsA.