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

Spondyloarthritis (SpA) comprises a group of common inflammatory musculoskeletal diseases, including ankylosing spondylitis, psoriatic arthritis, arthritis associated with inflammatory bowel disease, reactive arthritis, and undifferentiated SpA. Patients with SpA share the characteristics of a high frequency of the HLA-B27 allele in the MHC I loci and disease affecting the entheses and joints of the axial skeleton [1]. In SpA, disease disability is the sum of the inflammatory burden and chronic changes. Thus, distinguishing between inflammatory activity and chronic problems is critical for choosing a proper clinical intervention [2, 3].

The inflammatory process in SpA involves innate immune cells such as macrophages, monocytes, and fibroblasts [4]. Monocytes are present in the blood and can be divided into the three subsets: CD14⁺⁺ CD16^- classic monocytes, CD14⁺⁺ CD16⁺ intermediate monocytes, and CD14⁺ CD16⁺⁺ non-classic monocytes [5]. The intermediate monocyte subset increases during an infection [6] and is elevated in RA and SpA [7–9]. Intermediate monocytes have been described to have both pro-inflammatory and anti-inflammatory activities [5], and the role of intermediate monocytes in arthritis is still elusive. Disease activity in SpA is associated with migration of leukocytes to inflamed joints and entheses [1]. However, the complexity of the inflammatory process in SpA is still poorly understood.

The β2 (CD18) family of integrins includes CD11a/CD18 (LFA-1), CD11b/CD18 (Mac-1 or complement receptor [CR] 3), CD11c/CD18 (CR4, p150,95), and CD11d/CD18 and is crucial in transendothelial migration by the binding of CD11a/CD18 to intercellular adhesion molecule 1 (ICAM-1) [10, 11]. In arthritis, the increased expression of ICAM-1 results in the migration of leukocytes to the joints [12]. Animal studies support a central role of CD18 in the development of SpA. The mouse PL/J strain carries the CD18 hypomorphic mutation, which reduces the expression of CD18 to 2% to 16% of wild-type levels [13]. These mice develop a skin disease that closely resembles human psoriasis together with a condition very similar to psoriatic arthritis dominated by enthesitis [14], and the inflammatory response critically involves macrophages and T helper 17 (Th17) cells [15, 16]. Tumor necrosis factor-alpha (TNFα) induces endothelial expression of ICAM-1 as well as activation of integrins for ligand binding [17]. In this way, the success of biological therapy with anti-TNFα in the treatment of SpA also indirectly suggests that the CD18 integrins play a role in human disease.

Emerging literature [18–21] has identified soluble complexes of CD11/CD18. These complexes bind ICAM-1 [18, 19] in a way that competes with the binding of cellular-expressed CD11/CD18, implying an anti-inflammatory role [18]. In humans, we found alterations in the soluble CD18 (sCD18) plasma concentrations in patients with rheumatoid arthritis (RA) or SpA whereas patients with osteoarthritis had no such changes [18]. Shedding of CD18 is increased by TNFα stimulation [18] and occurs in processes that involve matrix metalloproteinase (MMP)-3 and MMP-9 [22, 23]. MMP-3 is the activator of the MMP-9 zymogen [24], and dysregulation is associated with a wide variety of autoimmune disorders [25], including RA [26] and SpA [27]. Recently, comprehensive studies in transgenic animals showed that leukocyte migration is critically dependent on efficient CD18 shedding, notably in the case of macrophage efflux from zones of inflammation [22]. However, not much is known about the cellular sources and targets of sCD18, and the association of sCD18 levels in the blood with other disease activity parameters has never been studied in any disease.

We previously demonstrated alterations in sCD18 complexes in the blood and synovial fluid of patients with arthritis and showed that sCD18 binding to ICAM-1 could be an anti-inflammatory regulatory mechanism in the immune system. Here, we study the mechanisms for these alterations and their association with SpA disease activity. In SpA, low sCD18 concentrations were part of an aggravated disease activity state. The low sCD18 levels seemed to be a result of a combination of increased capture of sCD11/CD18 complexes onto ICAM-1 expressed on inflamed tissues and insufficient CD18 shedding from monocytes. Thus, the resulting increased availability of ICAM-1 molecules on the endothelium and in the synovium could facilitate increased leukocyte infiltration in the entheses and joints and increased disease activity. Hereby, our findings link regulation of leukocyte infiltration by innate immune cells with inflammation and disease activity in SpA.

The inflammatory process in SpA is still poorly elucidated, resulting in difficulties in monitoring disease activity and choosing treatment strategy. Several investigations in experimental animal models already indicate a central role of CD18 integrins in SpA-like disease [14–16, 37]. We previously demonstrated alterations in the sCD18 complexes in the blood and synovial fluid of patients with arthritis and the ability of sCD18 to inhibit leukocyte binding to ICAM-1. This study proposes sCD18 as a novel player in SpA disease pathogenesis. The sCD18 plasma level was decreased in patients with SpA compared with HCs, and levels correlated inversely with SpA disease activity scores. The failure of patients with SpA to maintain normal sCD18 levels seems to reflect insufficient CD18 shedding from monocytes to counterbalance the capture of sCD18 complexes to inflammation-induced ICAM-1. In this way, our findings on sCD18 link increased leukocyte infiltration potential with disease activity in SpA.

As noted by Feldmann and colleagues [38], inflammation is regulated by the balance between pro- and anti-inflammatory mechanisms. In this sense, the inflammatory response may originate as a consequence of over-activity of pro-inflammatory mechanisms or mechanistic failures in anti-inflammatory responses. Previously, it was rather speculative to assign the production of sCD18 to either one of the two modes of regulating the inflammatory response. Yet the ability of sCD11/CD18 complexes to compete with cellular-expressed CD11a/CD18 in ICAM-1 binding [18] indicates that the soluble complexes may aid in resolving cellular adhesion and hence the formation of inflammatory foci. We now found at least two indications to support such anti-inflammatory functions of sCD11/CD18.

First, we demonstrated the ability of the sCD11/CD18 complexes to bind ICAM-1 in the cell membrane of a human umbilical vein cell line and primary FLSs established from a site of major inflammation in SpA. ICAM-1 expression is increased on the lining and sublining cells of the synovium and on adjacent endothelial cells in patients with SpA and is regulated by TNFα [39, 40]. Our immunocytological staining for ICAM-1 showed a speckled pattern very similar to previous findings [35, 36]. Carman and colleagues [35, 36] revealed that the origin of this morphology is caused at least partly by the formation of ICAM-1-enriched cell membrane protrusions. These protrusions are critical in the pro-active role of the endothelium in the transmigration of leukocytes [35, 36]. The deposition of sCD11/CD18 complexes on the cells investigated in our experiment was not uniform and appeared in a way that could well reflect the selective binding to such ICAM-1-enriched zones. Indeed, the deposition was critically dependent on TNFα treatment of the cells, increasing the ICAM-1 expression. The requirement of a dense expression of ICAM-1 is furthermore consistent with the biochemical evidence reported earlier [18] that multimers of sCD11/CD18 (for example, 2 × CD11/CD18 and 4 × CD11/CD18) participate in a polyvalent interaction with ICAM-1 molecules, enabling a high avidity in the interaction [18, 41]. In this way, our findings support that sCD11/CD18 selectively binds ICAM-1-rich features important for the interaction of leukocytes with the synovial lining and endothelium of blood vessels. Likewise, the clinical investigations made in our study suggest that insufficient saturation of these features with sCD11/CD18 captured from the plasma pool may aggravate inflammation, probably by permitting undue leukocyte adhesion.

Second, it is an important observation that the negative correlation between the decreased plasma level of sCD18 and disease activity in SpA could be simulated in cell cultures in vitro. Thus, CD18 shedding from SpA patient-derived PBMCs correlated negatively with the BASDAI score. This indicates that inflammation may be downregulated by the ability of leukocytes to deliver an elevated production of sCD18. In contrast, failure by the leukocytes to compensate the capture of sCD11/CD18 onto ICAM-1-expressing surfaces appears to augment inflammation. Taken together, these data also indicate that a compound resembling sCD18 could be beneficial as a therapeutic drug in SpA.

Our investigations identified monocytes as the predominant contributors to CD18 shedding among PBMCs in cultures in vitro. It is well established that the monocyte population can be separated into subsets with distinct regulatory capabilities [5, 42, 43]. Intermediate monocytes have both pro-inflammatory and anti-inflammatory functions. Thus, they are precursors of osteoclasts in psoriatic arthritis [7] and promote the expansion of the Th17 subset in RA [8]. However, intermediate monocytes also show the highest expression of the anti-inflammatory cytokine IL-10 and are tolerant to stimulation during infection [6, 44]. Also, lack of CD16⁺ monocytes may be related to auto-inflammatory disease [45]. In line with previous studies, we found a high expression of CD18 integrins in the membrane of monocytes and especially the intermediate monocyte subset [46, 47]. Furthermore, the concentration of sCD18 in PBMC supernatants correlated with the percentage of intermediate monocytes among all monocytes. This suggests that intermediate monocytes are the primary producers of sCD18 in the PBMC subset. In this way, our study proposes a novel role for intermediate monocytes in the blood as regulators of leukocyte migration. In SpA this regulatory role seems to be inadequate. Thus, in HLA B27-negative SpA only, the percentage of intermediate monocytes among all monocytes was increased, resulting in sufficient upregulation of CD18 shedding and low disease activity. However, this increase in intermediate monocytes was not seen in HLA B27-positive SpA, resulting in insufficient CD18 shedding and higher disease activity.

The high degree of overlap between plasma levels of sCD18 in patients with SpA and HCs implies that measuring sCD18 will not be of particular value in the diagnosis of SpA. However, measuring sCD18 could assist in discriminating between inflammatory activity and broader problems due to chronic changes in SpA. From a clinical viewpoint, discrimination between ongoing inflammation and chronicity is thus critical for choosing the proper clinical intervention [2]. To best measure the level of inflammation, many composite scores, such as the BASDAI, BASFI, and physician global assessment score, have been developed. However, these measures are only moderately objective and are difficult to compare across individuals. Also, BASDAI does not correlate well with objective measures such as CRP and MRI findings or erythrocyte sedimentation rate [48]. Nevertheless, in lieu of better measures of disease activity, the BASDAI is currently regarded as the gold standard regarding treatment initiation and response [49]. This underlines the need for better objective measures, especially because the role of CRP in SpA is debatable, as it is induced mainly by IL-6 and anti-IL-6 has little effect in SpA [50, 51]. In agreement, CRP showed no correlations with self-assessment scores, clinical findings, or MRI activity scores in our study population (data not shown). By contrast, the decrease in plasma levels of sCD18 found here correlated with a number of single and composite inflammatory measures, including BASDAI, level of morning stiffness, and physician global assessment score and the objective measures BASMI and SIJ MRI activity score in patients with SpA. These associations were greater when adjusting for differences in age, disease duration, HLA-B27 status, and treatment at time of inclusion, making the study population more homogeneous. Additionally, no correlation was observed between sCD18 and MRI chronicity scores or patient global score. Although MRI is not considered the best measure of structural damage, the MRI chronicity scores used in this study were based on MRI features described previously to correlate with radiographic changes [30, 31]. Thus, this adds to the validity of sCD18 as a marker of the inflammatory component of SpA disability. Importantly, the association with disease activity remained after correction for CRP.

Recently, the serum concentrations of both MMP-3 and MMP-9 have been found to be associated with disease activity in ankylosing spondylitis [52, 53]. Shedding of CD18 occurs in processes that involve MMP-3 and MMP-9 [22, 23]. With the observation that CD18 is a substrate for MMP-9 [23], measurement on sCD18 may provide an indication of enzymatic activity in vivo, which is central to pathogenic mechanisms of these diseases [25]. In this way, sCD18 could be a novel mechanistic biomarker of SpA disease activity involved in the disease pathogenesis and reflecting inflammatory processes [54, 55].

No inflammatory disease controls were assessed in this study. Thus, the specificity of our findings in SpA is uncertain. Interestingly though, the plasma levels of sCD18 were decreased in plasma from HLA-B27-positive SpA patients compared with HLA-B27-negative patients. Additionally, shedding of CD18 from PBMCs was increased among HLA-B27-negative patients only. HLA-B27 positivity in patients with ankylosing spondylitis predicts a poorer prognosis and a better response to anti-TNFα treatment [56]. It is unclear whether this reflects a different pathogenesis in HLA-B27-positive versus HLA-B27-negative SpA. HLA-B27-positive SpA patients also had increased BASMI, CRP, and MRI activity scores compared with HLA-B27-negative SpA patients in our study population (data not shown). From the established correlations, the decreased levels of sCD18 in HLA-B27-positive SpA may thus be explained by the higher degree of disease activity in this group. However, mechanistic couplings between HLA-B27 and shedding of CD18 cannot be excluded. Validation studies are needed to clarify whether measuring the sCD18 plasma level can be used as a marker of inflammatory activity in SpA.

The clinical impact of CD18 deficiency has been long known from the several types of leukocyte adhesion deficiencies [57]. Our work now argues that more subtle changes in CD18 functions are sufficient to affect disease activity in SpA through alterations in the systemic concentration of sCD18. The influences appear to involve the failure of monocyte subsets to maintain a sufficient concentration of sCD11/CD18 complexes. Our work suggests that a delicate balance exists between cellular-expressed and sCD18 integrins, which may be disturbed by the well-characterized changes in MMP activity and ICAM-1 expression associated with a vast range of illnesses involving chronic inflammation. With the central role of CD18 integrins in supporting leukocyte migration and immunological synapse formation, it encourages the hypothesis that such disturbances are involved in the disease-causing mechanisms in SpA. This points to the level of sCD18 as a potential marker of inflammatory activity or a compound resembling sCD18 as a therapeutic drug.