Circulating cell adhesion molecules in systemic sclerosis: a systematic review and meta-analysis

Introduction Patients with systemic sclerosis (SSc) have an increased risk of endothelial dysfunction, atherosclerosis, and cardiovascular events compared to the general population. Therefore, the availability of robust circulating biomarkers of endothelial dysfunction and atherogenesis may facilitate early recognition and management of cardiovascular risk in SSc. We sought to address this issue by conducting a systematic review and meta-analysis of studies investigating various types of circulating cell adhesion molecules involved in endothelial dysfunction and atherogenesis (i.e., immunoglobulin-like vascular cell, VCAM-1, intercellular, ICAM-1, platelet endothelial cell, PECAM-1, neural cell, NCAM, Down syndrome cell, DSCAM, and endothelial cell-selective, ESAM, adhesion molecules, E-, L-, and P-selectin, integrins, and cadherins) in SSc patients and healthy controls. Methods We searched PubMed, Scopus, and Web of Science from inception to 1 May 2024. Risk of bias and certainty of evidence were assessed using validated tools. Results In 43 eligible studies, compared to controls, patients with SSc had significantly higher plasma or serum concentrations of ICAM-1 (standard mean difference, SMD=1.16, 95% CI 0.88 to 1.44, p<0.001; moderate certainty), VCAM-1 (SMD=1.09, 95% CI 0.72 to 1.46, p<0.001; moderate certainty), PECAM-1 (SMD=1.65, 95% CI 0.33 to 2.98, p=0.014; very low certainty), E-selectin (SMD=1.17, 95% CI 0.72 to 1.62, p<0.001; moderate certainty), and P-selectin (SMD=1.10, 95% CI 0.31 to 1.90, p=0.007; low certainty). There were no significant between-group differences in L-selectin concentrations (SMD=-0.35, 95% CI -1.03 to 0.32, p=0.31; very low certainty), whereas minimal/no evidence was available for cadherins, NCAM, DSCAM, ESAM, or integrins. Overall, no significant associations were observed between the effect size and various patient and study characteristics in meta-regression and subgroup analyses. Discussion The results of this systematic review and meta-analysis suggest that specific circulating cell adhesion molecules, i.e., ICAM-1, VCAM-1, PECAM-1, E-selectin, and P-selectin, can be helpful as biomarkers of endothelial dysfunction and atherogenesis in the assessment of cardiovascular risk in SSc patients. Systematic review registration https://www.crd.york.ac.uk/prospero/, identifier CRD42024549710.


Introduction
Systemic sclerosis (SSc), an autoimmune condition primarily affecting women, is characterized by vascular dysfunction and progressive fibrosis of the skin and internal organs (1,2).The global incidence of SSc ranges between 8-56 new cases per million persons per year and the prevalence varies between 38-341 cases per million persons (3).The mortality in SSc patients is three-to four-fold higher than the general population due to cardiorespiratory complications, renal and gastrointestinal disease, cancer, and infections (4,5).Increasing evidence also suggests that atherosclerosis is a critical additional component of the pathophysiology of SSc.This has led to a shift in the focus of basic and clinical research studies which have convincingly reported several pro-atherosclerotic arterial abnormalities, e.g., endothelial dysfunction, increased intima-media thickness and arterial stiffness, in SSc (6)(7)(8)(9)(10).Such alterations are similar to those observed in rheumatoid arthritis, another autoimmune condition associated with atherosclerosis and cardiovascular disease (11,12).Epidemiological studies have also reported an increased risk of atherosclerotic cardiovascular events in SSc, particularly myocardial infarction and peripheral vascular disease (13,14).In these studies, the prevalence and/or severity of hypertension, diabetes, and dyslipidemia in SSc patients was similar to that in control groups (9,13).This suggests that conventional risk factors only partially account for the increased risk of atherosclerosis and cardiovascular disease in SSc.Therefore, a focus of current research is the identification of alternative, more robust biomarkers of atherosclerosis allowing early risk stratification and preventive treatment.
Functional and structural alterations of the endothelium, associated with the impaired synthesis of the critical endogenous messenger nitric oxide, represent the initial step in the pathogenesis of atherosclerosis (15).At a cellular and molecular level, these alterations involve the adhesion of leukocytes and lymphocytes to the endothelium (endothelial activation) and their consequent migration to the tunica intima, where they initiate a sequence of events leading to the formation of the atherosclerotic plaque (16,17).The process of cellular adhesion to the endothelium is mediated by several molecules, e.g., the immunoglobulin-like vascular cell adhesion molecule-1 (VCAM-1), the intercellular vascular adhesion molecule-1 (ICAM-1), the platelet endothelial cell adhesion molecule-1 (PECAM-1), the neural cell adhesion molecule (NCAM), the Down syndrome cell adhesion molecule (DSCAM), and the endothelial cell-selective adhesion molecule (ESAM) (18)(19)(20)(21).VCAM-1 is expressed in endothelial cells and macrophages and binds to integrin a 4 b 1 (22,23).ICAM-1 is upregulated during inflammation and binds to the leukocyte specific b 2 integrins (24,25).PECAM-1 is expressed in leukocytes, platelets, and endothelial cells, and exerts its effects through the translocation of integrin a 6 b 1 (26).NCAM is expressed in the brain, skeletal muscle, and hematopoietic system.In addition to regulating cell adhesion, it modulates brain and kidney development and plays a pathophysiological role in cancer, schizophrenia, and other neurodegenerative disorders (27).DSCAM is primarily expressed in the brain and regulates neural development (28).ESAM is expressed mainly in endothelial cells and is critical in modulating angiogenesis, endothelial integrity, leukocyte adhesion and transmigration (29).
To evaluate the possible role of cell adhesion molecules as biomarkers of endothelial activation, dysfunction, and atherosclerosis in SSc, we conducted a systematic review and meta-analysis of studies investigating their plasma or serum concentrations in SSc patients and healthy controls.Where possible, we investigated possible associations between the effect size of the between-group differences in cell adhesion molecules and pre-defined study and patient characteristics.
Each abstract was screened by two independent investigators and reviewed as full text if considered relevant.Any disagreement between the investigators throughout the screening process was resolved by a third investigator.The inclusion criteria were: (a) the measurement of soluble ICAM-1, VCAM-1, PECAM-1, ESAM, NCAM, DSCAM, E-selectin, L-selectin, P-selectin, integrins, and cadherins in plasma or serum; (b) the comparison between SSc patients and healthy controls in original case-control research studies; (c) the inclusion of patients aged ≥18 years; and (d) the availability of the full text of the publication in English language.The exclusion criteria were: (a) the investigation of other autoimmune or autoinflammatory conditions; (b) case reports and review articles; and (c) the inclusion of children and/or adolescents.References of reviewed articles were also searched to identify additional studies.
Two investigators independently extracted the following data from each article: year of publication, first author, study country and continent, sample size, age, male to female ratio, SSc type (diffuse or localized), disease duration, concentrations of individual cell adhesion molecules, and biological matrix assessed (serum or plasma).The data were then manually transferred to separate custom extraction forms created using Microsoft Excel.Any discrepancy between the extraction forms was resolved by a third investigator.
The risk of bias was assessed using the Joanna Briggs Institute Critical Appraisal Checklist for analytical studies (53).The certainty of evidence was evaluated using the Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) Working Group system (54).The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement (Supplementary Table 1) (55).The protocol was registered in an international repository (PROSPERO registration number, CRD42024549710).

Statistical analysis
Standardized mean differences (SMDs) and 95% confidence intervals (CIs) were calculated to generate forest plots and assess the differences in the concentrations of individual cell adhesion molecules between SSc patients and healthy controls.A p-value <0.05 was considered statistically significant.If required, data were extracted from graphs using the Graph Data Extractor software (San Diego, CA, USA).Means and standard deviations were calculated from medians and interquartile ranges or full ranges according to published methods (56).The heterogeneity of SMD across studies was evaluated using the Q-statistic (significance level at p<0.10) and classified as low (I 2 ≤25%), moderate (25%< I 2 <75%), or high (I 2 ≥75%) (57,58).Sensitivity analysis and assessment of publication bias were conducted according to established methods (59)(60)(61)(62).
Univariate meta-regression and subgroup analyses were conducted to investigate associations between the effect size and the following parameters: year of publication, study continent, number of participants, age, male to female ratio, SSc type (diffuse or localized), mean disease duration, and biological matrix assessed (serum or plasma).Statistical analyses were performed using Stata 14 (Stata Corp., College Station, TX, USA).

Study selection
The flow chart of study selection is illustrated in Figure 1.After initially identifying 1,542 articles, 1,486 were excluded because they were either irrelevant or presented duplicate data.A full-text review of the remaining 56 articles led to the exclusion of one study because of duplicate data, two studies because of missing information, four studies because their design was not case-control, and six studies because they included patients under 18 years old.Therefore, 43 studies were included in the analysis (Table 1) (63-105).The risk of bias was low or moderate in all studies except one which was assessed as having high risk (84) (Supplementary Table 2).The initial level of the certainty of evidence was considered low because of the case-control design of the selected studies (level 2).
There was significant publication bias (Begg's test, p=0.002;Egger's test, p=0.003).The "trim-and-fill" method was consequently used to address and correct this bias (62).This method operates under the assumption that the results of some studies, often those with null or negative findings, might be missing, leading to an asymmetric distribution in the funnel plot.Consequently, it estimates the number of such missing studies and adds them to the funnel plot to create a symmetrical distribution.This adjustment helps in evaluating the impact of publication bias on the overall results.The meta-analysis is then recalculated to include these additional studies.In this case, the "trim-and-fill" method identified seven missing studies to be added to the left side of the funnel plot to ensure symmetry (Supplementary Figure 2).This adjustment led to an attenuation of the resulting SMD however the effect size remained significant (SMD=0.76,95% CI 0.44 to 0.90, p<0.001).This suggests that while there was evidence of publication bias, the overall effect size of ICAM-1 concentrations remained robust.
The overall level of certainty was upgraded to moderate (level 3) after considering the low-moderate risk of bias in most studies (no change), the high but partially explainable heterogeneity (no change), the lack of indirectness (no change), the large effect size (SMD=1.16;upgrade one level) (106), and the presence of publication bias which was addressed with the "trim-and-fill" method (no change).

VCAM-1
Twenty-three studies, including 24 group comparisons, reported VCAM-1 concentrations in 1,413 SSc patients (mean age The risk of bias was considered low or moderate in all studies except one, which was assessed as having high risk (84) (Supplementary Table 2).
There was significant publication bias (Begg's test, p<0.001;Egger's test, p=0.001).The "trim-and-fill" method identified seven missing studies to be added to the left side of the funnel plot to ensure symmetry (Supplementary Figure 4).The resulting SMD was attenuated but remained significant (SMD=0.51,95% CI 0.08 to 0.94, p=0.021) No significant associations were found between the effect size and age (t=-0.96,p=0.35), male to female ratio (t=-1.59,p=0.13), publication year (t=-1.07,p=0.30), sample size (t=-1.13,p=0.27), or   The overall level of certainty was upgraded to moderate (level 3) after considering the low-moderate risk of bias in most studies (no change), the high but partially explainable heterogeneity (no change), the lack of indirectness (no change), the large effect size (SMD=1.09;upgrade one level) (106), and the presence of publication bias which was addressed with the "trim-and-fill" method (no change).

PECAM-1
Two European studies reported PECAM-1 concentrations in 100 SSc patients and 41 healthy controls (83, 88) (Table 1).Measurements were conducted in serum in one study (83) and plasma in the other (88).The risk of bias was low in one study (83) and moderate in the other (88) (Supplementary Table 2).
Pooled analyses showed that PECAM-1 concentrations were significantly higher in SSc patients compared to controls (SMD=1.65,95% CI 0.33 to 2.98, p=0.014;I 2 = 89.0%,p=0.003; Figure 4).Assessment of the risk of bias, meta-regression, and subgroup analyses could not be performed because of the small number of studies.The overall level of evidence was downgraded to very low (level 1) because of the high and unexplained heterogeneity and the lack of assessment of publication bias.
No publication bias was observed after removing the study by Iversen et al. (91) (Begg's test, p=0.26;Egger's test, p=0.53).The "trim-and-fill" did not identify any missing studies to be added to the funnel plot to ensure symmetry (Supplementary Figure 7).
Meta-regression analysis did not show any significant associations between the effect size and age (t=0.86,p=0.41), male to female ratio (t=0.70,p=0.50), year of publication (t=0.17,p=0.87), sample size (t=0.96,p=0.35), or SSc duration (t=1.15,p=0.27).In sub-group analysis, there were non-significant differences (p=0.91) in pooled SMD between studies conducted in Europe (SMD=1.21,95% CI 0.53 to 1.89, p<0.001;I 2 = 95.9%, Forest plot of studies investigating PECAM-1 concentrations in SSc patients and controls.The overall level of certainty was upgraded to moderate (level 3) after considering the low-moderate risk of bias in most studies (no change), the high but partially explainable heterogeneity (no change), the lack of indirectness (no change), the large effect size (SMD=1.17;upgrade one level) (106), and the absence of publication bias (no change).
A ll s tu dies had a low or mod er ate risk of bias (Supplementary Table 2).
Pooled analyses showed that L-selectin concentrations were nonsignificantly different between SSc patients and controls (SMD=-0.35,95% CI -1.03 to 0.32, p=0.31;I 2 = 87.4%,p<0.001; Figure 6).Forest plot of studies investigating E-selectin concentrations in SSc patients and controls.The pooled SMD values were stable in sensitivity analysis, ranging between -0.56 and -0.29 (Supplementary Figure 8).Assessment of publication bias, meta-regression and sub-group and analysis could not be performed because of the small number of studies.
The overall level of evidence was downgraded to very low (level 1) because of the high and unexplained heterogeneity and the lack of assessment of publication bias.
The risk of bias was considered low or moderate in all studies except one which was assessed as having high risk (84) (Supplementary Table 2).
The overall level of certainty remained low (level 2) after considering the low-moderate risk of bias in most studies (no change), the high but partially explainable heterogeneity (no change), the lack of indirectness (no change), the large effect size (SMD=1.10;upgrade one level) (106), and the lack of assessment of publication bias (downgrade one level).

NCAM, DSCAM, ESAM, and integrins
No studies investigating these cell adhesion molecules in SSc and healthy controls were identified.

Discussion
The results of this systematic review and meta-analysis have highlighted the presence of significant elevations in the concentrations of specific cell adhesion molecules, markers of endothelial activation, dysfunction, and atherogenesis, in patients with SSc.Such elevations were particularly evident in studies investigating ICAM-1, VCAM-1, PECAM-1, E-selectin, and Pselectin.The results were stable in sensitivity analysis, and the effect size of the observed between-group differences was generally not associated with individual study and patient characteristics.In particular, the lack of significant associations with SSc duration supports the proposition that elevations in cell adhesion molecules are already present in SSc patients with early disease, further supporting their potential clinical utility in assessing atherosclerotic burden.By contrast, no between-group differences were observed with L-selectin and minimal/no evidence was available for NCAM, DSCAM, ESAM, cadherins, and integrins.Epidemiological studies have highlighted the increased risk of atherosclerotic cardiovascular events in SSc.The largest, conducted in Denmark, used data from administrative sources between 1995 and 2015 to identify patients with SSc and age-and sex-matched controls in a 1:5 ratio (14).Over a follow-up period of 8.9 years, SSc patients (n=2778) had a significantly increased risk of myocardial infarction (hazard ratio, HR=2.08, 95% CI 1.65 to 2.64), ischemic stroke (HR=1.28,95% CI 1.04 to 1.58), and peripheral vascular disease (HR=5.73,95% CI 4.63 to 7.09).These associations were maintained, except for ischemic stroke (HR=1.13,95% CI 0.90 to 1.42), after adjusting for co-morbidities and medications.Our analyses suggest that measuring specific cell adhesion molecules might be helpful in evaluating atherosclerotic burden, stratifying cardiovascular risk, and facilitating the initiation of preventive strategies in SSc patients.Pending further research, assessing cell adhesion molecules may be particularly useful to demonstrate early endothelial dysfunction in absence of overt clinical evidence of vascular damage and atherosclerosis.However, an important issue to be addressed in further studies is whether such alterations in cell adhesion molecules might reflect endothelial dysfunction not only in the microcirculation, a vascular territory primarily affected in SSc (107,108), but also in middle-size and large arteries, typically affected by the atherosclerotic process (7,(109)(110)(111).
Several studies have reported significant alterations in surrogate markers of endothelial function and arterial stiffness and an increased atherosclerotic burden in SSc (6)(7)(8)(9)(10).Endothelial cell injury, the main promoter of these alterations, stimulates adhesion and transmigration of leukocytes and monocytes into the tunica media of the arterial wall, initiating a sequence of events leading to the formation of the atherosclerotic plaque (8).A number of factors have been proposed as triggers of endothelial cell injury in SSc, including viruses (e.g., cytomegalovirus and Epstein Barr virus) (112,113), cytotoxic CD4+ and CD8+ T-cells (114), autoantibodies against endothelial cells (115), and oxidative stress (116).However, further research is warranted to determine their role in vivo.
The capacity to restore endothelial function in SSc pharmacologically has been reported with dihydropyridine calcium channel blockers, statins, nitrate, endothelin-1 receptor antagonists, phosphodiesterase-5 inhibitors, soluble guanylate cyclase activators, prostacyclins, and cyclophosphamide (8),.Notably, dihydropyridine calcium channel blockers (117), statins (118,119), endothelin-1 receptor antagonists (120), phosphodiesterase-5 inhibitors (121), soluble guanylate cyclase activators (122), and prostacyclins (123) have also been shown to downregulate several cell adhesion molecules in vitro and in vivo.Furthermore, systematic reviews and metaanalyses on the effects of statins on cell adhesion molecules in other patient populations have shown an effect size (SMD) on VCAM-1 and ICAM-1 between -0.28 and -0.75 and on E-, L-, and P-selectin between -0.39 and -0.73 (118,119).The magnitude of these effects suggests that statins may be effective in reducing the circulating concentrations of cell adhesion molecules in SSc.Regardless, further research should investigate whether the measurement of cell adhesion molecules in SSc patients receiving these therapies may reflect a state of improved endothelial function and reduced atherosclerotic burden.
Strengths of our study include the comprehensive assessment of a wide range of cell adhesion molecules, the robust evaluation of the certainty of evidence for each cell adhesion molecule investigated, the evaluation of specific study and patient characteristics associated with the effect size by means of meta-regression and subgroup analysis, and the generalizability of our findings to different geographical areas although most studies were conducted in European countries.One important limitation is represented by the generally high heterogeneity observed which, however, could be partially explained for some cell adhesion molecules (ICAM-1 and VCAM-1: diffuse/localized disease patient ratio; E-selectin: study continent, biological matrix assessed, and diffuse/localized disease patient ratio; P-selectin: diffuse/localized disease patient ratio).
In conclusion, our study has shown significant elevations of specific cell adhesion molecules, i.e., ICAM-1, VCAM-1, PECAM-1, E-selectin, and P-selectin in SSc, which reflects a state of endothelial activation, dysfunction, and atherogenesis in this patient group.Pending the results of further prospective studies in patients with subclinical and overt atherosclerosis, which also investigate the microcirculation and the effect of specific therapies, cell adhesion molecules may assist in cardiovascular risk stratification in SSc.

FIGURE 2
FIGURE 2Forest plot of studies investigating ICAM-1 concentrations in SSc patients and controls.

FIGURE 3 Forest
FIGURE 3Forest plot of studies investigating VCAM-1 concentrations in SSc patients and controls.

FIGURE 6 Forest
FIGURE 6Forest plot of studies investigating L-selectin concentrations in SSc patients and controls.

FIGURE 7 Forest
FIGURE 7Forest plot of studies investigating P-selectin concentrations in SSc patients and controls.

TABLE 1 Main
characteristics and results of the studies included in the meta-analysis.