Decreased plasma levels of soluble CD18 link leukocyte infiltration with disease activity in spondyloarthritis

Abstract

Introduction: Spondyloarthritis (SpA) comprises a group of diseases often associated with HLA-B27 and characterized by inflammation of the entheses and joints of the axial skeleton. The inflammatory process in SpA is presumably driven by innate immune cells but is still poorly understood. Thus, new tools for monitoring and treating inflammation are needed. The family of CD18 integrins is pivotal in guiding leukocytes to sites of inflammation, and CD18 hypomorphic mice develop a disease resembling SpA. Previously, we demonstrated that altered soluble CD18 (sCD18) complexes in the blood and synovial fluid of patients with arthritis have anti-inflammatory functions. Here, we study the mechanisms for these alterations and their association with SpA disease activity.

Methods: Plasma levels of sCD18 in a study population with 84 patients with SpA and matched healthy controls were analyzed with a time-resolved immunoflourometric assay (TRIFMA). Binding of sCD18 to endothelial cells and fibroblast-like synoviocytes (FLSs) was studied with confocal microscopy. Shedding of CD18 from peripheral blood mononuclear cells (PBMCs) was studied with flow cytometry and TRIFMA.

Results: Plasma levels of sCD18 were decreased in patients with SpA compared with healthy volunteers (P <0.001), and the lowest levels were in the HLA-B27-positive subgroup (P <0.05). In a multiple regression model, the sCD18 levels exhibited an inverse correlation with the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) (P <0.05), the level of morning stiffness (P <0.05), the Bath Ankylosing Spondilitis Metrology Index (P <0.05), the physician global assessment score (P <0.01), and the sacroiliac magnetic resonance imaging activity score (P <0.05). The mechanisms for these changes could be simulated in vitro. First, sCD18 in plasma adhered to inflammation-induced intercellular adhesion molecule 1 (ICAM-1) on endothelial cells and FLS, indicating increased consumption. Second, CD18 shedding from SpA PBMCs correlated inversely with the BASDAI (P <0.05), suggesting insufficient generation. CD18 was shed primarily from intermediate CD14⁺⁺ CD16⁺ monocytes, supporting the view that alterations in innate immunity can regulate the inflammatory processes in SpA.

Conclusions: Taken together, the failure of patients with SpA to maintain adequate sCD18 levels may reflect insufficient CD18 shedding from monocytes to counterbalance the capture of sCD18 complexes to inflammation-induced ICAM-1. This could increase the availability of ICAM-1 molecules on the endothelium and in the synovium, facilitating leukocyte migration to the entheses and joints and aggregating disease activity.

Figure 1

Plasma concentrations of soluble CD18 (sCD18) in patients with spondyloarthritis (SpA) at the time of inclusion and at the 4-year follow-up and healthy controls (HCs). The median values of plasma sCD18 were 639.7 mU/mL (interquartile range 514.3-726.8 mU/mL) for patients with SpA at the time of inclusion, 606.1 mU/mL (interquartile range 458.4-691.5 mU/mL) for patients with SpA at the 4-year follow-up, and 767.2 mU/mL (interquartile range 628.1-839.7 mU/mL) for HCs. Bars indicate median and interquartile range. ***P <0.001, ****P <0.0001.

Figure 2

The soluble CD18 (sCD18) plasma concentration, HLA genotype, and spondyloarthritis (SpA) disease activity. (A) The sCD18 plasma level in HLA-B27-positive versus HLA-B27-negative SpA patients at the time of inclusion. The median values were 605.7 mU/mL (interquartile range 461.6-704.1 mU/mL) for HLA-B27-positive SpA patients and 670.0 mU/mL (interquartile range 591.8-760.1 mU/mL) for HLA-B27-negative SpA patients. Bars indicate median and interquartile range. (B) Correlation between the sCD18 plasma levels in HLA-B27-positive SpA patients and self-assessment and clinical scores. (C) Correlation between the sCD18 plasma levels in HLA-B27-positive SpA patients and clinical test results. Solid lines represent the best fit in linear regression. *P <0.05; **P <0.01; ***P <0.001; ****P <0.0001.

Figure 3

Confocal microscopy analysis of the ability of sCD11/CD18 complexes to bind intercellular adhesion molecule 1 (ICAM-1) expressed on the human umbilical vein cell line EA.hy926. (A) Illustration of the cellular incubations. In step 1, adherent cells were stimulated with 10 ng/mL tumor necrosis factor-alpha (TNFα), which increased the ICAM-1 expression. In step 2, a source of CD11/CD18 was added (that is, either normal human serum (NHS) or supernatant from synovial fluid mononuclear cell culture). In step 3, biotinylated antibody recognizing ligand-binding activated CD11/CD18 (KIM127) was added followed by addition of fluorochrom-labelled streptavidin for detection using confocal microscopy. Binding of sCD18 to ICAM-1 expressed on EA.hy926 cells incubated with (B) or without (C) TNFα. Red staining indicates the binding of sCD18, further indicated with white arrows. The positions of cell nuclei were located by 4′,6-diamidino-2-phenylindole (DAPI) staining, indicated in blue. The staining was distinctly localized to small foci on the cell membrane on 10% to 15% of the cells. Expression of ICAM-1 on EA.hy926 cells incubated with (D) or without (E) TNFα. ICAM-1 is indicated with a green staining, and cell nuclei are indicated with a blue staining.

Figure 4

Spontaneous shedding of CD18 from peripheral blood mononuclear cells (PBMCs) cultured in vitro. (A) Levels of spontaneous shedding of CD18 from spondyloarthritis (SpA) and healthy control (HC) PBMCs. The median values were 22.0 mU/mL (interquartile range 14.0-51.8 mU/mL) for HLA-B27-positive SpA patients, 51.7 mU/mL (interquartile range 21.5-57.0 mU/mL) for HLA-B27-negative SpA patients, and 11.2 mU/mL (interquartile range 8.59-12.3 mU/mL) for HCs. PBMCs from 5 HLA-B27-positive SpA patients, 5 HLA-B27-negative patients, and 5 HCs were used. Bars indicate median and interquartile range. (B) Correlation between PBMC donor BASDAI score and the spontaneously shed CD18 in the PBMC culture supernatant. (C) CD18 spontaneous shedding capacity and correlation with matrix metalloproteinase-9 (MMP-9) production. PBMCs from 10 patients with SpA were used. Hatched horizontal lines connect identical measurements of sCD18. Solid black lines represent the best fit in linear regression. *P <0.05.

Figure 5

Cellular-expressed and shed CD18 attributed to peripheral blood mononuclear cell (PBMC) source. (A) The cell membrane expression of CD18 on T cells, natural killer (NK) cells, and monocytes using PBMCs from a representative healthy control (HC) with the gating strategy indicated. (B) The cell membrane expression of CD18 on T cells, NK cells, and monocytes using PBMCs from 10 HCs and 2 patients with spondyloarthritis (SpA). Levels of CD18 expression were measured by the median fluorescence intensity (MFI). Bars indicate median and interquartile range. (C) The concentration of sCD18 in supernatants from cultured T cells, NK cells, and monocytes derived from PBMCs from 5 HCs. Bars indicate median and interquartile range. ***P <0.001.

Figure 6

CD18 expression and distribution of monocyte subsets in healthy controls (HCs) and patients with spondyloarthritis (SpA). (A) The cell membrane expression of CD18 on CD14⁺ CD16⁺⁺ (non-classic) monocytes, CD14⁺⁺ CD16⁺ (intermediate) monocytes, and CD14⁺⁺ CD16^- (classic) monocytes using peripheral blood mononuclear cells (PBMCs) from a representative HC with the gating strategy indicated. (B) The cell membrane expression of CD18 on non-classic monocytes, intermediate monocytes, and classic monocytes in PBMCs from 5 HCs and 10 patients with SpA. Bars indicate median and interquartile range. (C) The percentages of intermediate monocytes among all CD14⁺ monocytes in HLA-B27-positive SpA, HLA-B27-negative SpA, and HCs. PBMCs from the same 10 patients with SpA and 5 HCs were used. Bars indicate median and interquartile range. *P <0.05; **P <0.01; ****P <0.0001.

Figure 7

Correlation between monocyte subsets and soluble CD18 (sCD18) in the peripheral blood mononuclear cell (PBMC) culture supernatant. The percentage of non-classic (A), intermediate (B), or classic (C) monocytes among all monocytes in a sample of PBMCs was correlated against the sCD18 concentration in supernatant from culture of the PBMCs. Samples were analyzed from PBMCs derived from a total of 5 healthy controls (HCs) and 10 patients with spondyloarthritis (SpA).