C1 esterase inhibitor-mediated immunosuppression in COVID-19: Friend or foe?

From asymptomatic to severe, SARS-CoV-2, causative agent of COVID-19, elicits varying disease severities. Moreover, understanding innate and adaptive immune responses to SARS-CoV-2 is imperative since variants such as Omicron negatively impact adaptive antibody neutralization. Severe COVID-19 is, in part, associated with aberrant activation of complement and Factor XII (FXIIa), initiator of contact system activation. Paradoxically, a protein that inhibits the three known pathways of complement activation and FXIIa, C1 esterase inhibitor (C1-INH), is increased in COVID-19 patient plasma and is associated with disease severity. Here we review the role of C1-INH in the regulation of innate and adaptive immune responses. Additionally, we contextualize regulation of C1-INH and SERPING1, the gene encoding C1-INH, by other pathogens and SARS viruses and propose that viral proteins bind to C1-INH to inhibit its function in severe COVID-19. Finally, we review the current clinical trials and published results of exogenous C1-INH treatment in COVID-19 patients.


Introduction
Late in 2019, unexplained pneumonia deaths in Wuhan, China were the first cases in what quickly evolved into a global pandemic of the novel coronavirus, SARS-CoV-2 infections and deaths [1] . The broadscoping signs and symptoms of infection with SARS-CoV-2 have been termed coronavirus disease 2019 (COVID-19) [2] .
The worldwide scientific community has united in its research efforts to understand SARS-CoV-2 and COVID-19. Clinical manifestations of COVID-19 vary, as some patients are asymptomatic, whereas others endure more serious manifestations such as cough, dyspnea, respiratory failure, and stroke [2] . Leveraging bioinformatic methods that use RNA and DNA sequencing, metabolomic and proteomic characterization by mass spectrometry has allowed scientists to gain insight into the pathophysiologic mechanisms of SARS-CoV-2 infection [3][4][5][6][7][8][9][10][11][12][13][14][15] . Because of the rapid pace of new bioinformatics-based COVID-19 studies, it is difficult  C1-INH is the major plasma inhibitor of activated Factor XII and plasma kallikrein, proteins that initiate the CS and bradykinin inflammation, respectively. Similarly, C1-INH suppresses complement activation that drives host antiviral defenses by inhibiting proteins that initiate the proteolytic cascades of the classical and lectin pathways. Further, C1-INH inhibits the alternate pathway by binding C3b. Crosstalk between complement and the CS is mediated through thrombin-mediated complement activation of C3 and C5.  C1-INH inhibits CS and complement pathways, and  these activities have been leveraged for clinical benefit through utilization of exogenous C1-INH (exC1-INH). Deficient C1-INH activity results in angioedema, and in patients with hereditary angioedema (HAE), this effect is largely attributed to unregulated bradykinin signaling, and treatment with exC1-INH is often used to manage disease-associated attacks. Animal models of ischemia/reperfusion injury (IRI) have shown that exC1-INH protects from inflammation and decreases immune cell recruitment. Clinical trial results from exC1-INH treatment in human solid organ transplant studies suggests that exC1-INH may decrease IRI and antibody-mediated rejection (AMR) in transplant recipients. Clinical use of recombinant C1-INH (rhC1-INH) is currently in development to treat or prevent bradykinin and cytokine storms in COVID-19 patients. Lines with arrows indicate activation, and lines with a perpendicular line indicate inhibition. system (CS) ( Fig. 1 ) [16] . C1-INH is encoded by the SERPING1 gene, extensively glycosylated, and induced by interferon (IFN) [17][18][19][20][21][22][23][24] . Powerful innate immune responses such as activation of complement and the CS would be harmful if left unchecked; thus, similar to many biological responses, IFN-mediated upregulation of C1-INH acts as negative feedback to regulate these processes [25] .
C1-INH directly inhibits complement, a proteolytic cascade involved in linking innate and adaptive immune responses ( Figs. 1 and 2 ). Activation of complement protects the host through several mechanisms, including direct virolysis and lysis of infected host cells, inducing an antiviral state in the host cells, as well as through the adaptive immune response [26] . Activated complement proteins bind viral antigen and interact with B cells and CD4 + and CD8 + T cells to promote the formation of viral-specific antibodies and T cell responses [27] .
C1-INH concomitantly inhibits complement and the CS ( Figs. 1 and 2 ). Crosstalk between activation of complement and the CS is intermediated by Factor XII (FXII), historically known as Hageman factor [33] . Contact with artificial surfaces, cell-free RNA/DNA, pathogens, including viruses, and other anionic surfaces induces a conformational change in the zymogen FXII to form activated FXIIa [16] . Further, in the plasma kallikrein-kinin system, FXIIa activates plasma prekallikrein (PK) into activated plasma kallikrein (PKa), which is in turn able to perpetuate FXIIa activation, and proteolytically cleave high-molecular-weight kininogen (HK), releasing bradykinin (BK) [ 34 , 35 ]. A major role of the kallikrein-kinin system-mediated release of bradykinin is to induce vascular permeability ( Fig. 1 ) [35] . Patients with either hereditary or acquired loss of C1-INH activity experience episodes of angioedema, largely attributed to unregulated bradykinin signaling ( Figs. 1 and 2 ) [ 36 , 37 ].
FXIIa activates classical complement as effectively as aggregated IgG in vitro [33] , and FXIIa activation of intrinsic coagulation results in thrombin-mediated complement activation through activation of C3 and C5 ( Fig. 1 ) [38] . As opposed to alpha-2-macroglobulin, C1-INH is the major plasma inhibitor of activated FXII and PKa [25] . Both complement and the CS have been implicated in the pathophysiology of COVID-19 [ 6 , 39-43 ].

Regulation of SERPING1/C1-INH by pathogens
Out of over 22,000 possible genes in a human respiratory virus challenge study, SERPING1 was found to be a member of a 30-gene transcriptional signature that was commonly induced and accurately diagnosed viral infection prior to symptom onset [44] . In an impressive 5-year prospective follow-up study, nine different naturally contracted respiratory viral infections were detected, including coronavirus [45] . The previously characterized 30-gene signature, which included SERPING1, accurately predicted viral infection up to three days prior to viral shedding and symptom onset [45] .
Additional studies have identified SERPING1 as a genetic marker of viral infection. Herberg et al. characterized a 38-gene signature out of 8,565 differentially expressed genes to distinguish bacterial from viral infection in febrile children [46] . In this study, SERPING1 significantly distinguished patients with a viral versus bacterial infection and those with a viral infection versus healthy controls [46] . These studies imply that SERPING1 maybe a specific biomarker of viral infections; however, Mycobacterium tuberculosis (Mtb) induces host interferon-mediated responses in a manner similar to a viral infection [47] , and when compared to latent tuberculosis (TB), increased SERPING1 transcript levels correlated with subclinical and active TB in 15 of 16 independent studies [ 48 , 49 ].
Gordon et al. reported significant overlap in the interactomes of human protein-pathogen protein interactions between SARS-CoV-2, Mtb, and human immunodeficiency virus (HIV), among others [10] . Interestingly, in dually infected HIV + Mtb patients, SERPING1 was significantly increased in the blood cells of HIV + subclinical TB patients that progressed to active TB compared to those patients whose TB did not progress [48] . C1-INH protein levels do not consistently correlate with SERPING1 transcript levels in active TB patients [ 48 , 49 ]; however, C1-INH complexed with serine protease targets are rapidly cleared from circulation, which may be one possible explanation for this discrepancy [50] .
HIV infection of cultured immature dendritic cells resulted in increased SERPING1 levels post-infection [51] . Likewise, monocytes from HIV + patients showed significantly higher expression of SERPING1, and transcript abundance positively correlated with viral load and was significantly increased in patients with high compared to low viral load [52] . HIV-1 viral protein components upregulate C1-INH [51] and cleave the highly glycosylated N-terminus of C1-INH [53] .
The consensus among these studies is that increased levels of SER-PING1 correlate with viral infection, is associated with active disease, and can be directly regulated by viruses. Further, viral proteins interact and regulate C1-INH. It is likely context dependent whether increased SERPING1 translates to increased C1-INH activity.

SARS and SERPING1/C1-INH expression and regulation
Prior to the COVID-19 pandemic, SERPING1 expression had been characterized in SARS-CoV-1 patients and found to be increased in whole blood RNA when compared to healthy controls [54] . And again, when compared to healthy controls SARS-CoV-2 infection was associated with increased SERPING1 transcript in whole blood RNA [ 9 , 54 , 55 ]. However, SERPING1 was found to be decreased in mRNA purified from bronchoalveolar lavage fluid from COVID patients when compared to a mixed cohort of subjects and controls from a previous study of obesity and asthmatic disease severity [56] .
Proteomics of COVID patient sera have afforded valuable insight into C1-INH levels in circulation. Increased C1-INH levels in COVID-19 patients have been described in several studies. Interrogation of the online resource, covid-omics.app developed by Overmyer et al., showed that there was a significant association between increased leukocyte SERP-ING1 and plasma C1-INH in COVID-19 hospitalized patients versus non-COVID-19 hospitalized patients [15] . In a study that correlated protein abundance in sera versus IL-6 levels, C1-INH levels were found more abundant than controls in patients with low, mid, and high IL-6 levels [11] .
COVID proteomics has played an important role in machine learning model development to predict outcomes. Suvarna et al. engaged deep proteomics of COVID-19 patient sera to develop a machine learning model that used 20 proteins including increased C1-INH to classify severe from non-severe COVID [14] . By performing proteomics and metabolomics on COVID patient sera, Shen et al. developed a random forest machine learning model based on the prioritization of 29 molecules (22 proteins and 7 metabolites) to identify non-severe from severe COVID-19, and upregulated C1-INH was one of the top 10 proteins with the highest importance to the accuracy of the model [6] . Further, C1-INH was in the top 25 of 57 proteins in COVID plasma with the highest relevance in a machine learning model used to predict survival of severely ill COVID patients [57] . Despite the inclusion of C1-INH in other machine learning models, in an ultra-high-throughput proteomics analysis of COVID patient sera and plasma, C1-INH was not one of the 27 proteins determined to be biomarkers that correlated with COVID severity [13] .
In contrast to plasma and sera, C1-INH was found to be less abundant in urine from mild, severe, and COVID-recovered patients compared with healthy controls [12] .
Hadjadj et al. reported that when compared to patients with mild/moderate or severe symptoms, type I IFN signaling was impaired in critical COVID-19 patients, despite significantly increased plasma SARS-CoV-2 viral load [9] . SERPING1 levels were increased in all three of the COVID-19 severity groups described in Hadjadj et al. when compared to healthy control subjects.
SARS-CoV-2 may not elicit strong type I IFN responses; however, the initial type I response seems to be critical for curbing life-threatening COVID-19 symptoms [ 58 , 59 ]. Recently published studies have shown that loss-of-function mutations in proteins critical to type I IFN responses and auto-antibodies against type I IFNs are associated with COVID-19 biomarkers and disease severity [ 58 , 59 , 64 , 65 , 67 ]. Moreover, preexisting antibodies against cytokines and IFNs may predispose autoimmune disease patients to severe COVID-19 pneumonia [66] Thorne et al. determined that type I IFN induction and sensitivity to IFN mediated inhibition were decreased with the SARS-CoV-2 Alpha variant. Whether caused by auto-antibodies, genetic defects, SARS-CoV-2 variants, or a heretofore-unknown mechanism, lack of sufficient type I IFN signaling does not seem to negatively influence SERPING1/C1-INH induction and correlates with severity. One possible explanation is that SERPING1 is most potently induced by IFN-, a type II IFN, and to a lesser extent by type I IFNs [ 21 , 24 ]; thus, increased SERPING1 may occur as a result of natural killer cells secreting IFN-in response to IL-1 and IL-18 release from SARS-CoV-2 infected cells during cytokine/bradykinin hyperinflammation ( Fig. 3 ) [68] .
Women show stronger type I IFN responses, while men are more susceptible to COVID-19, and interestingly, SERPING1 was identified as a gene with potential gender susceptibility of infection and severity in men [69] .
Many studies also described an increase in SERPING1-encoded C1-INH, thus there seems to be a disconnect between the current tenets of hypercoagulability and over-activation of complement in COVID-19 patients and C1-INH activity. Intriguingly, patients suffering from hereditary angioedema with a loss of expression or function of C1-INH show augmented coagulation and fibrinolysis during angioedema attacks, similar to COVID-19 patients [43] .
C1-INH function may be inhibited by binding of virally expressed proteins as seven of SARS-CoV viral proteins bind C1-INH [70] and by sequence similarity these proteins are predicted to bind to C1-INH during SARS-CoV-2 infection [71] . This supports the idea that while C1-INH appears to be increased in COVID-19 patients, it's function may be compromised ( Fig. 3 ). In support, C1-INH-serpin protease complexes are rapidly cleared from circulation [50] implying that the observed increase in C1-INH concurrent with high complement activation and hypercoaguability may be due to the inability of C1-INH-viral protein aggregates to be cleared from circulation ( Fig. 3 ).

Clinical use of C1-INH treatment
Although seemingly self-evident based on function and current clinical use, exogenous C1-INH-facilitated immunosuppression has been shown in many studies ( Fig. 2 ). In animal models of ischemia/reperfusion injury (IRI) in brain, heart, and muscle, C1-INH treatment decreases tissue injury, presumably because of lower levels of reactive oxygen species and inflammation and a decrease in immune cell recruitment [ 25 , 72-74 ]. C1-INH-mediated protection may have originated from its protease inhibitor activity, as well as anti-inflammatory properties independent of its role in complement and CS, by disrupting leukocyte/endothelial adhesion and decreasing neutrophil infiltration [ 72 , 75 , 76 ]. C1-INH treatment improved solid organ transplant outcomes by decreasing IRI and ameliorating antibody-mediated rejection in the recipient [77] , which showed that C1-INH treatment acts as an immunosuppressant.
Cancer cell survival often depends on immunosuppressive mechanisms, and with the exclusion of lymphoproliferative cancers, C1-INH levels have been shown to be increased in many diverse cancer types, including squamous cell lung carcinoma, glioblastoma, and pancreatic cancer [78][79][80][81][82][83] . Increased C1-INH levels were characterized as an independent prognosticator for gastric cancer and were found to be associated with decreased survival in colorectal cancer patients [ 81 , 82 ].

Clinical use of C1-INH in COVID-19 patients
In an uncontrolled trial with five patients treated with the FDA-approved, recombinant C1-INH (rhC1-INH) concentrate Conestat alfa (Ruconest, Pharming Group/Salix Pharmaceuticals) showed very promising results in severe COVID-19 patients ( Fig. 2 ) [84] . A larger interventional clinical trial, where Conestat alfa was administered upon hospital admission to COVID-19 patients with respiratory involvement, at least one risk factor for progression to mechanical ventilation, and symptom onset within the last 10 days, was terminated due to enrollment and changes to the standard of care [ 85 , 86 ]. Another clinical trial with a similar protocol and from the same sponsor based in the United States is on-going and currently recruiting patients [87] . An additional trial of Conestat alfa is aimed at improving neurological symptoms associated with post-acute COVID-19 syndrome [88] . Long-term consequences observed in survivors of SARS-CoV-1 and MERS and now COVID-19 patients are varied and may negatively affect pulmonary, hematologic, cardiovascular, neuropsychiatric, and other body systems [89] .
A clinical trial aimed at targeting pulmonary sequelae due to overactive bradykinin in COVID-19 infections tested C1-inhibitor administered either alone or concurrently with Icatibant, a specific bradykinin B2 receptor antagonist, against the placebo control arm [90] . This trial was completed with a total of 44 participants, but results have yet to be published. A recently published Brazilian clinical trial compared standard of care with patients treated with Icatibant or a human plasma-derived C1 esterase/kallikrein inhibitor (Berinert®, CSL Behring LLC.) [ 91 , 92 ]. Neither treatment group showed decreases in the time to clinical improvement when compared to standard of care; however, both drugs were safe and improved lung CT scores [92] . It will be interesting to see if C1-INH administered concurrently with Icatibant produced more positive results. Further, the timing of C1-INH administration in the course of COVID-19 may be critical to its therapeutic success.

Concluding statements
We review evidence that when compared to healthy controls, SARS-CoV-2 infection results in an increase in SERPING1 transcript and C1-INH levels in plasma/sera, including mild cases of COVID-19 ( Fig. 3 ). We speculate that because of sufficient type I IFN signaling, SARS-CoV-2 virus is neutralized by innate immune responses, and the early induction of C1-INH functionally inhibits hyperinflammation, hypercoagulability, and aberrant complement activation, resulting in mild COVID-19.
Studies have shown that increasing levels of C1-INH correlate with COVID-19 severity, and this presents an apparent disconnect between C1-INH function in the face of overactivated complement and hypercoagulation that is observed in severe COVID-19 patients. SARS-CoV-2 viral proteins may bind to C1-INH inhibiting its function and clearance from circulation that results in detection of increased C1-INH, albeit non-functional.
C1-INH is an immunosuppressant protein that inhibits pathways in innate and adaptive immunology. If asymptomatic SARS-CoV-2 infected persons also have increased circulating functional C1-INH levels compared to healthy controls, we speculate that this may be a contributing factor to the lower abundance of antibodies to SAR-CoV-2 and the extended viral shedding observed in the asymptomatic patient population [ 93 , 94 ]. Further, increased functional C1-INH may contribute to reinfection risk, as has been observed [ 5 , 95 , 96 ] or reduced durability of SARS-CoV-2 vaccine-derived immunity [97] .
As more bioinformatic datasets are published, we may find more novel connections and findings, such as the observation of increased C1-INH levels in COVID-19 patients, and this may contribute to our understanding of COVID-19 and its detrimental effects.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data Sharing
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.