Description of the study population
A total of 17 patients with active oligoarticular JIA, undergoing therapeutical synovial fluid aspiration, were included in the study. The patient cohort and samples are described in Table 1.
Synovial fluid neutrophils have an aged and activated phenotype
Blood neutrophils were primarily of a mature phenotype (CD16hi, CD62Lhi), with very few immature (CD16mid, CD62Lhi) or aged (CD62Llow) neutrophils. However, in synovial fluid, most neutrophils had an aged phenotype (range 41-99 %, median 76 %) and no immature neutrophils were observed (Figure 1A-D). Synovial fluid neutrophils also had higher levels of the age-related marker CD10 compared to neutrophils in the circulation (Figure 1E).
Neutrophils in synovial fluid had an activated phenotype, demonstrated by increased levels of activation markers CD11b and CD66b (Figure 1F-G), and low levels of CD62L, which is shed upon neutrophil activation and/or migration. Despite the activated phenotype, S100A8/A9, a protein released from activated neutrophils, was only present on the surface of less than 3 % of both synovial fluid and blood neutrophils (Figure 1H).
Synovial fluid neutrophils express monocyte/macrophage-related surface markers
Synovial fluid neutrophils had increased CD14 levels compared to the circulating neutrophils (Figure 1I). Mannose receptor CD206, not commonly known as a neutrophil receptor, but expressed by monocytes, macrophages and dendritic cells, was found on a significant proportion of synovial fluid neutrophils (range 6-92 %, median 56 %), while being low or non-existent on circulating neutrophils (Figure 1J). The increased expression of both CD14 and CD206 might suggest that the neutrophils gain a monocyte/macrophage related phenotype.
CD206-expressing neutrophils are found in tissue biopsies
To confirm the finding of CD206+ neutrophils in synovial fluid we investigated whether neutrophils migrating through synovial tissue express CD206. Synovial tissue biopsies obtained from four patients were stained for CD206 and myeloperoxidase (MPO). Neutrophils co-expressing both MPO and CD206 were observed in tissue biopsies from patient 4, 8 and 18 (Figure 2, arrows), demonstrating that CD206-expressing neutrophils can be present in synovial tissue. In all biopsies there were also neutrophils without CD206 expression (Figure 2, arrow heads). In two of biopsies we observed synovial blood vessels, and the circulating neutrophils within the vessels had no or low CD206 expression (Figure 2B-C). The biopsy from patient 17 did not contain any area with neutrophils (data not shown).
Circulating neutrophils are similar during inactive disease and flares
Patients included in the study did not exhibit systemic symptoms although previous literature suggest that circulating neutrophils in JIA are activated during flares (10,12,29). We therefore compared blood neutrophil phenotypes during flares and inactive disease. We did not observe systemic activation of neutrophils during flares; neutrophil activation markers CD66b and CD11b were mostly unchanged, CD62L levels slightly increased and cell surface S100A8/A9 was present on less than 3 % of the neutrophils (Figure 3A-D). CD14 levels were markedly increased during flares (Figure 3E), while levels of CD16 and CD10 were not significantly altered (Figure 3F-G). CD206 expression, which was low or non-existent on blood neutrophils during flares, was equally low during inactive disease (Figure 3H).
The synovial neutrophil phenotype is not dependent on transmigration or synovial fluid alone
To investigate if the synovial fluid neutrophil phenotype found in the JIA patients could be explained by exposure to synovial fluid, healthy blood neutrophils were treated with synovial fluid in vitro. Exposure to synovial fluid did not induce any major phenotype shift on healthy blood neutrophils (Figure 4A). Most surface markers remained unchanged except CD11b, which, in opposite to the findings in patients, was significantly lower after exposure to synovial fluid. There was a very modest increase of CD206 (Figure 4A). These results suggests that the phenotype shift between circulating and synovial fluid neutrophils cannot be explained by exposure to synovial fluid alone.
To investigate if tissue transmigration towards other sites would result in a phenotype shift similar to synovial fluid neutrophils, we investigated transmigrated neutrophils from the oral cavity of healthy controls. The phenotype shifts between healthy blood- and oral cavity neutrophils were compared with JIA blood- and synovial fluid neutrophils. Neutrophils in both oral cavity and joint were activated compared to their blood counterparts (Figure 4B-C). CD11b surface expression was elevated in the transmigrated neutrophils at both sites but more pronounced in synovial fluid compared to oral neutrophils (Figure 4C). On the other hand, neutrophil surface S100A8/A9 was markedly increased on oral cavity neutrophils but not on synovial fluid neutrophils (Figure 4D). Levels of CD62L were lower in oral neutrophils than synovial fluid neutrophils, compared to their circulating counterparts (Figure 4E). Oral neutrophils had dramatically decreased levels of CD16 while synovial fluid neutrophils had a small increase in this marker (Figure 4F). Both oral and synovial fluid neutrophils had increased levels of CD14 and CD206 (Figure 4G-H). CD10 levels were more increased in synovial compared to oral neutrophils, despite not reaching statistical significance (Figure 4I). Taken together, transmigration towards both sites induced a shift in neutrophil phenotypes, but with distinctly different surface marker patterns.
Synovial fluid neutrophils have impaired phagocytosis and oxidative burst
Neutrophil effector functions, phagocytosis and oxidative burst, were evaluated in blood and synovial fluid. Synovial fluid neutrophils had a significantly decreased ability to phagocytose opsonized E. coli compared to circulating neutrophils (Figure 5A). The impaired phagocytosis was not due to the presence of synovial fluid in the assay. Dilution of healthy control blood in cell-free JIA synovial fluid led to an increase in neutrophil phagocytic ability, in contrast to addition of JIA serum which led to a slight impairment in phagocytosis (Figure 5B). Upon stimulation with PMA, there was a trend towards impaired ROS production in synovial fluid neutrophils compared to circulating neutrophils, although not reaching statistical significance (Figure 5C). ROS production upon stimulation with opsonized E. coli was similar in synovial and circulating neutrophils (Figure 5D). Neutrophil capacity of phagocytosis and oxidative burst was not influenced by patient age (Supplemental Figure 2).
Proportion of CD206+ neutrophils correlates with impaired phagocytosis and oxidative burst
Monocytes typically have lower capacity for both phagocytosis and oxidative burst compared to neutrophils, and we therefore hypothesized that CD206 expression on synovial fluid neutrophils might be associated with altered effector functions. Indeed, we observed a negative correlation of the proportion of CD206+ neutrophils with both phagocytosis of opsonized E. coli (Figure 5E) and ROS production (Figure 5F-G) in synovial fluid neutrophils, supporting the hypothesis that synovial fluid neutrophils have attained a more monocyte-like phenotype.