Association of Immune Semaphorins with COVID-19 Severity and Outcomes

Semaphorins have recently been recognized as crucial modulators of immune responses. In the pathogenesis of COVID-19, the activation of immune responses is the key factor in the development of severe disease. This study aimed to determine the association of serum semaphorin concentrations with COVID-19 severity and outcomes. Serum semaphorin concentrations (SEMA3A, -3C, -3F, -4D, -7A) were measured in 80 hospitalized adult patients with COVID-19 (moderate (n = 24), severe (n = 32), critical, (n = 24)) and 40 healthy controls. While SEMA3C, SEMA3F and SEMA7A serum concentrations were significantly higher in patients with COVID-19, SEMA3A was significantly lower. Furthermore, SEMA3A and SEMA3C decreased with COVID-19 severity, while SEMA3F and SEMA7A increased. SEMA4D showed no correlation with disease severity. Serum semaphorin levels show better predictive values than CRP, IL-6 and LDH for differentiating critical from moderate/severe COVID-19. SEMA3F and SEMA7A serum concentrations were associated with the time to recovery, requirement of invasive mechanical ventilation, development of pulmonary thrombosis and nosocomial infections, as well as with in-hospital mortality. In conclusion, we provide the first evidence that SEMA3A, SEMA3C, SEMA3F and SEMA7A can be considered as new biomarkers of COVID-19 severity.


Introduction
COVID-19 presents with a wide range of clinical manifestations, from asymptomatic infection to pneumonia with acute respiratory distress syndrome (ARDS) and multipleorgan failure [1].The severity of disease appears to be primarily influenced by the host's immune response, which is extremely heterogeneous and complex, involving both innate and adaptive immunity [2,3].Understanding immunopathogenesis and finding new biomarkers for the development of severe COVID-19 is necessary since they could serve as target sites for immunomodulation, which is the current cornerstone of severe and critical COVID-19 treatment, however, with modest clinical responses [4].
Semaphorins (SEMA) are a large family of secreted and membrane-bound signaling proteins expressed in most tissues, and divided into eight subclasses based on their Sema domain located on the N-terminal region [5].They exhibit their effects through binding to their receptors, plexins (PLXN) and neuropilins (NRP), which can be modified by a variety of co-receptors (such as cell adhesion molecules, receptor tyrosine kinases), and transmembrane semaphorins themselves [6].Although the molecular mechanisms of semaphorin signaling are still far from clear, the semaphorin's interaction with its receptor complex alters the cell cytoskeleton's structure and cell adhesion, which regulates cellular morphology and motility, as reviewed in [6].
Semaphorins were initially discovered as axonal guidance molecules in the development of the nervous system [7].Subsequent research revealed their involvement in a variety of other physiological and pathophysiological processes, such as vascular growth [8], gene expression reprogramming of cancer cells [9], tumor progression [7], allergic diseases [10], cardiovascular diseases [11], metabolic disorders [12] or liver diseases [13,14].Recently, increased focus has been placed on "immune semaphorins" and their roles in regulating immune cell activation, differentiation, mobility and migration in autoimmune diseases [15].Several studies have demonstrated the potential of semaphorins as diagnostic and therapeutic targets in immune-mediated diseases [8,10,16,17].
However, the role of semaphorins in the immunopathogenesis of infections remains to be elucidated.Studies on mouse models of sepsis showed increased concentrations of several semaphorins in serum and tissues, and a blockade of semaphorins or their receptors led to a reduction in tissue damage and better survival rates [18,19].In contrast, an experimental study of LPS-induced ARDS revealed decreased SEMA3A concentrations in lung tissue, while SEMA3A overexpression led to less severe lung impairment [20].Depending on the secreting cells and receptors involved, each SEMA has different and often opposite functions.Notably, there are no studies in humans.
Here, we hypothesize that semaphorin concentrations, due to their key function in controlling immune responses, correlate with COVID-19 severity and outcomes.

Study Design and Patients
This study was part of a prospective, non-interventional cohort study that included consecutively hospitalized adult patients with COVID-19 at the University Hospital for Infectious Diseases Zagreb (UHID) in Croatia between April and December 2021 (part of the COVID-FAT trail, NCT04982328).At that time, the Delta (B.1.617.2 and AY lineages) SARS-CoV-2 variant predominated in Croatia (data were taken from the ECDC database on SARS-CoV-2 variants) [21].The delta SARS-CoV-2 variant was shown to cause more severe disease and an excessive number of younger people dying despite receiving vaccinations [22,23].Eighty patients with COVID-19 were included, and these patients have not been reported in previous studies.The sample size was selected according to power analysis for the Kruskal-Wallis test to achieve an 80% chance of detecting a difference in median semaphorin concentrations at a 5% significance level.
All included patients had bilateral pulmonary infiltrates on chest images.Patients who had a concomitant bacterial infection at the time of admission were excluded, as were those who began corticosteroid or antiviral medication prior to enrollment.Active malignant disease, pregnancy and immunosuppression (disease and/or current medical therapy including corticosteroids) were other exclusion factors.
Forty healthy, SARS-CoV-2-RNA-negative, age-and sex-matched healthcare workers were included as controls.
All participants provided written informed consent.The study followed the Declaration of Helsinki's ethical principles and was approved by the UHID Ethics Committee (code 01-673-4-2021).

COVID-19 Disease Severity Classification
According to the National Institute of Health, COVID-19 severity was classified based on clinical symptoms, the oxygen level at admission and level of care (pandemic department or intensive care unit) [24].Briefly, the severity of COVID-19 was classified into three categories: moderate (bilateral pneumonia with SpO2 > 93% on room air), severe (bilateral pneumonia with SpO2 ≤ 93% on room air, dyspnea and/or tachypnea > 24/min), and critical (intensive care unit, ARDS criteria, high-flow nasal cannula oxygen therapy (HFNC), non-invasive (NIV)/invasive mechanical ventilation (IMV)) [24].

Data Collection
At admission, demographic and comorbidity data were collected, including the presence of cardiovascular disease (CVD), arterial hypertension, chronic pulmonary disease (asthma, chronic obstructive pulmonary disease), chronic renal failure (CRF), diabetes mellitus, dyslipidemia, gastritis or gastroesophageal reflux disease (GERD) and chronic medications.All patients had their body mass index (BMI) measured.
The patients were treated according to the standard of care at the time, which included anticoagulants, remdesivir, tocilizumab, baricitinib and dexamethasone, at the discretion of the supervising physician.
Clinical monitoring, including oxygen requirements, invasive and non-invasive ventilation and complications, were assessed daily and collected in a standardized form.

Statistical Analysis
Clinical, laboratory and demographic data were analyzed and reported descriptively as frequencies and medians with interquartile ranges.To compare two groups, Fisher's exact test and the Mann-Whitney U test were used.To compare three or more groups, the Kruskal-Wallis test with Dunn's multiple comparisons test was used.All tests were two-tailed, with a statistically significant p-value of 0.05.Spearman's rank correlation coefficient was used to examine correlations, which were then summarized in a correlation matrix.A receiver operating characteristic (ROC) analysis was used to compare the discriminatory performance of the laboratory variables under consideration.Time to hospital discharge or readiness for discharge stratified by biomarker levels was evaluated using the Kaplan-Meier method and hazard ratios (HR) with 95% confidence intervals (95% CI) and p-values were calculated by the log-rank test.Risk factors associated with critical COVID-19 were investigated using a univariate and subsequently multivariable logistic regression analysis.The strength of association was expressed as an odds ratio (OR) and its corresponding 95% CI.GraphPad Prism Software version 10 (San Diego, CA, USA) was used for statistical analyses.
COVID-19 severity was categorized as moderate in 24 (30%), severe in 32 (40%), and critical in 24 (30%) patients.As shown in Table 1, there were no differences in the age, gender, comorbidities, chronic medications or duration of symptoms before admission between the groups.Patients with critical COVID-19 had a higher BMI than those with severe or moderate COVID-19.As expected, the severity of clinical symptoms and disease severity scores, including MEWS, SOFA, PSI and the 4C mortality score, differed significantly between groups (Table 1).Patients in the critical group had lower peripheral oxygen saturation (82%, IQR 78-87 vs. 88%, IQR 82-89 vs. 95% IQR 93-96, p = 0.0001) and a lower PaO2/FiO2 ratio (126 IQR 76-174 vs. 183 IQR 137-250 vs. 347 IQR 323-428, p = 0.0001) on admission.As illustrated in Table 2, patients with critical and severe COVID-19 had significantly higher serum concentrations of CRP and IL-6, glucose, urea, AST, LDH and CK at the time of admission.There were no differences in other routine laboratory findings.

Correlation of Semaphorin Serum Concentrations with COVID-19 Severity
As shown in Figure 2 and Table 4, serum concentrations of SEMA3A and SEMA3C were negatively correlated with disease severity, with the lowest concentrations in the most severely ill patients.SEMA4D serum concentrations showed no correlation with disease severity.In contrast, serum concentrations of SEMA3F and SEMA7A positively correlated with COVID-19 severity, with the highest levels in patients with critical COVID-19 (Figure 2, Table 3).

Discussion
In this study, we provide the first evidence that COVID-19 patients have different semaphorin serum concentrations as compared to healthy controls.While SEMA3A was decreased, SEMA3C, SEMA3F and SEMA7A were increased in COVID-19.Furthermore, we showed an association of semaphorin levels with COVID-19 severity; SEMA3F and SEMA7A were higher in critical COVID-19, while SEMA3A and SEMA3C negatively correlated with COVID-19 severity and were lower in the critical group.Semaphorins showed equal or better accuracy in predicting disease severity than the widely used CRP, IL-6 or LDH.
Firstly, we found decreased serum concentrations of SEMA3A that further decreased with COVID-19 severity.Class 3 semaphorins, specifically SEMA3A, have immunosuppressive and regulatory effects that include neutrophil migration, induce the shift of activated macrophages (M1) to the resolution phase phenotype (M2) and negatively control the T cell-mediated response predominantly via the activation of regulatory T cells (Tregs) [29].In patients with autoimmune diseases (e.g., SLE, rheumatoid arthritis, systemic sclerosis and allergic diseases), a reduced expression of SEMA3A correlated with T-cell-mediated inflammation and disease severity [10,15].The role of SEMA3A in the pathogenesis of infections might depend on the receptor utilized, which has been the subject of several experimental studies.Inhibition of the SEMA3A/PLXNA4 complex attenuates Toll-like receptor (TLR) pathways, which was associated with a decreased septic response and improved survival rates [30].In contrast, inhibition of the SEMA3A/NRP-1 complex demonstrated increased production of proinflammatory cytokines and higher mortality [31].In the transcriptome study (GEO dataset GSE57011), SEMA3A was identified as the most downregulated gene in ARDS patients [20], and overexpression of SEMA3A in the lipopolysaccharides (LPS)-induced ARDS model alleviates oxidative stress and inflammation by suppressing activation of the extracellular signal-regulated kinase/Jun-N-Terminal Kinase (ERK/JNK) signaling pathway in rat pulmonary microvascular endothelial cells [20].
Since the fine-tuned immune response is vital in determining the outcome of the SARS-CoV-2 infection [32], we can hypothesize that decreased serum concentrations of SEMA3A in COVID-19 patients and its negative correlation with disease severity might result in the lack of anti-inflammatory and immunosuppressive effects of SEMA3A, which might lead to an uncontrolled inflammatory cascade.
Next, we found that SEMA3C is increased in moderate, but not in critical, COVID-19.Furthermore, SEMA3C was increased in a subgroup of patients diagnosed with pulmonary thrombosis.Less is known about the immunoregulatory role of SEMA3C, but it was shown that SEMA3C regulates fibrosis, vascular development, pathological angiogenesis and the migration of tumor cells.The expression of SEMA3C is related to tumor progression and poor prognosis in lung cancer, prostate, breast cancer, gastric cancer and ovarian cancer, which makes it a potential therapeutic target for malignant diseases [33,34].SEMA3C regulates extracellular matrix composition through increased expression of IL-6, transforming growth factor-β (TGF-β) and connective tissue growth factor (CTGF), and was implicated in the development of liver fibrosis [13,14,35].Recently, it was shown that SEMA3C plays an important role in the development of murine acute kidney injury by promoting vascular permeability, interstitial edema, leukocyte infiltration and tubular injury [36].After administration of SEMA3C in murine models, systemic and renal hemodynamics changed: mean arterial pressure decreased and vascular resistance was reduced [36].SEMA3C has a pivotal role in vascular smooth muscle cell migration and cardiovascular system development [37,38].Interestingly, SEMA3C might regulate pathological angiogenesis, where SEMA3C exerted potent inhibiting effects and was suggested as a potent and selective inhibitor of pathological retinal angiogenesis [39].Since severe COVID-19 is characterized by significant distortion of the lung angioarchitecture, small vessel vasculitis and microthrombosis, all associated with worse prognosis and increased mortality [40], we can theorize that SEMA3C might play a role in aberrant angiogenesis associated with the development of severe ARDS that is still to be investigated.
Similarly, SEMA3F and SEMA7A concentrations were increased and positively correlated with COVID-19 disease severity and were identified as predictors of COVID-19 outcomes, including the need for mechanical ventilation, development of pulmonary thrombosis and nosocomial infections.Both SEMA3F and SEMA7A are secreted by activated immune cells and have emerged as regulators of neutrophil migration, vascular permeability and cytoskeletal remodeling, and the initiation of an inflammatory signaling cascade in models of acute lung injury.An increased number of neutrophils in bronchoalveolar fluid with a high expression of SEMA3F and NRP2 was observed in a murine model after the LPS challenge, and neutrophil-specific loss of SEMA3F resulted in more rapid neutrophil recruitment and clearance from the lungs [41].By interacting with various receptors and organ systems, SEMA7A has opposing effects: in interactions with integrin receptors, SEMA7A has a protective and anti-inflammatory effect, while via the plexin C1 receptor it promotes extravascular neutrophil migration and the release of pro-inflammatory cytokines [42].In the murine model, SEMA7A causes transendothelial migration of neutrophils into lung tissue [43,44], and a blockade of SEMA7A in in vivo and in vitro models showed reduced injury-induced neutrophil influx, correlating with reduced lung injury along with reduced cytokine response [18].To summarize, the SEMA3F and SEMA7A in COVID-19 might regulate vascular permeability and cytoskeletal remodeling along with neutrophil migration and retention in inflamed tissue, which leads to the amplification of inflammation.
In our study, there were no differences in SEMA4D serum concentrations between healthy controls and COVID-19 patients.SEMA4D regulates immune activation and inflammatory responses by modulating cytoskeleton reorganization through its principal receptor, CD72, located on T cells, B cells, macrophages and dendritic cells [45].In animal models of autoimmune diseases such as multiple sclerosis and autoimmune encephalomyelitis, SEMA4D correlated with disease severity [46], as in patients with psoriasis and rheumatoid arthritis [47].SEMA4D is most extensively studied in oncology and it is currently considered a promising target for antitumor therapy for breast cancer [48].The impact of class IV semaphorins on COVID-19 outcomes should be further examined.
Recent research has shown that micro RNAs (miRNA) control SEMA signaling in immune, cardiovascular and nervous systems, and malignancies, directly through their receptors, or indirectly by modulating the molecules that regulate the expression of semaphorins [49].Similarly, there are reports that specific miRNA signatures in blood or respiratory samples can distinguish COVID-19 disease severity or COVID-19 patients from healthy people [50].Some of them are linked with semaphorin signaling such as miR-17-5p [51], miR-142-5p [52], miR-126-3p [53], miR-19b-3p [54], miR-92a-3p [55] and miR-320a [56], thus highlighting the potential importance of semaphorin signaling in COVID-19.Interestingly, SARS-CoV-2-encoded miRNAs were shown not only to target viral genomes and alter viral fitness, but can also be transported to host cells during viral infection and bind to host miRNAs and genes and alter immune responses [57].However, their role in regulating semaphorins remains to be elucidated.
This study should be viewed within its limitations; since this was an observational study, causality could not be determined; a relatively small number of participants in COVID-19 severity subgroups limits statistical analysis and should be confirmed in a larger population; the impact of comorbidities on semaphorins and other inflammatory markers was not evaluated; the concentrations of semaphorins were determined at a single time point, and dynamic variations related to clinical outcomes were not examined.Nevertheless, we studied a well-defined cohort of patients and report the first data examining the semaphorins' profile in patients with COVID-19.Additional studies are needed for a better understanding of the complex underlying immunopathological mechanisms including their effect on development and activation of B and T cells, and how semaphorins contribute to COVID-19 progression.

Conclusions
In conclusion, we have shown that patients with COVID-19 have different expressions of SEMA3A, SEMA3C, SEMA3F and SEMA7A than healthy controls, which correlate with disease severity and outcomes.Due to their role in regulating inflammation, cell migration, fibrosis and angiogenesis, which have already been explored in neoplastic and autoimmune diseases, semaphorins could be new diagnostic and prognostic biomarkers and potential therapeutic targets in COVID-19, which warrant further investigation.

Figure 1 .
Figure 1.(a) Serum concentrations of semaphorin measured by ELISA in healthy controls and in patients with COVID-19.(b) ROC curve analysis of serum semaphorins for determination of COVID-19.Shown are AUCs with corresponding 95% CI.

Figure 1 .
Figure 1.(a) Serum concentrations of semaphorin measured by ELISA in healthy controls and in patients with COVID-19.(b) ROC curve analysis of serum semaphorins for determination of COVID-19.Shown are AUCs with corresponding 95% CI.

Figure 2 .
Figure 2. Serum concentrations of semaphorin SEMA3A, SEMA3C, SEMA3F and SEMA7A in healthy controls (HC) and patients with COVID-19 stratified by disease severity (moderate, severe, critical).Data are presented as medians with interquartile ranges.The p-values are calculated by Kruskal-Wallis test with Dunn's multiple comparisons test.

Figure 4 .Figure 4 .
Figure 4. Spearman correlation correlogram.The strength of the correlation between two variables is represented by the color at the intersection of those variables.Colors range from dark blue (strong negative correlation; r = −1.0) to red (strong positive correlation; r = 1.0).Results were not displayed if p > 0.05.

Table 2 .
Laboratory findings at admission.

Table 3 .
Serum concentrations of semaphorins in healthy controls and COVID-19 patients and ROC analysis of sensitivity and specificity in differentiating COVID-19 patients from healthy controls.

Table 3 .
Serum concentrations of semaphorins in healthy controls and COVID-19 patients and ROC analysis of sensitivity and specificity in differentiating COVID-19 patients from healthy controls.
Data are presented as medians with interquartile ranges.The p-values are calculated by Kruskal-Wallis test with Dunn's multiple comparisons test.