It has been reported that approximately 200,000 cases occur in the United States annually (8), with poor clinical outcomes with a pooled mortality rate of approximately 40% (8). Twenty-five percent of ARDS cases are classified as mild and 75% as moderate or severe. However, a third of the mild cases can progress to moderate or severe disease (9).

ARDS represents a stereotypic response to various etiologies. As shown in Table 1 , its pathological changes have been described as three overlapping phases: exudative (acute), proliferative (sub-acute) and fibrotic (chronic) phases (11, 12). The hallmark of the acute phase is alveolar-capillary barrier dysfunction driven by immune-cell-mediated destruction of alveolar epithelium and vascular endothelium (11, 13), and is characterized by bilateral infiltration (14) ( Figures 1A, B ), interstitial and alveolar edema, capillary congestion, intra-alveolar hemorrhage, inflammatory cell infiltration, diffuse alveolar damage (DAD), intra-alveolar and intravascular fibrin deposition, hyaline membrane formation, and microthrombi (14, 15) ( Figures 1C-G ). These alterations lead to hypoxia, hyperinflammation, metabolic acidosis, pulmonary hypertension, and increased mortality. Traditionally, ARDS is recognized to be a neutrophil-driven disease. However, accumulating evidence has demonstrated the involvement of innate immune cells such as macrophages and platelets, and the adaptive immune system in the pathogenesis of ARDS. The proliferative phase occurs early (within the initial three days) and attempts to reduce the damage caused during the acute phase and restore pulmonary function with the regeneration of alveolar type II cells, differentiation of alveolar type II cells to alveolar type I cells, and the proliferation of fibroblasts and myofibroblasts. The fibrotic phase develops due to a failure of removal of alveolar collagen, which results in an increased dead-space fraction, high minute ventilation requirement, pulmonary hypertension, and further reduction of lung compliance. A hallmark of the damage seen in ARDS is that it is not uniform.

ARDS has many risk factors with heterogeneous etiologies. Generally, ARDS is classified into two major types: direct lung injury affecting the pulmonary epithelium and indirect lung injury disrupting the pulmonary vascular endothelium. Greater understanding of the differences between direct and indirect lung injury may help development of new therapies targeted at specific subgroups of patients with ARDS. The common causes of ARDS are summarized in Table 2 . There are certain common pathophysiological differences between direct and indirect lung injury (10) ( Table 2 ).

No single diagnostic test confirms or refutes a diagnosis of ARDS. As shown in Table 3 , the current diagnosis of ARDS is based on the Berlin Consensus Criteria in 2012 according to the timing, hypoxemia, positive end-expiratory pressure (PEEP), chest radiograph, and origin of edema. Safe, inexpensive, and point-of-care tools such as ultrasound imaging and pulse oximetry, can be implemented as diagnostic modalities for monitoring pulmonary infiltration and measuring SpO₂/FiO₂ ratio, respectively. In 2016, investigators proposed alternative criteria (SpO₂/FiO₂ ratio and lung ultrasound imaging) for ARDS diagnosis-the so-called Kigali modification of the Berlin Criteria. There is no biomarker for ARDS recommended in clinical practice. With the evolution of clinical care and increasing recognition of the global burden of ARDS, and the emerging new methodologies, the advent of these new methods (multiomics, big data, systems biology, and multiplexed point-of-care testing) combined with insights into biomarker strategies will likely yield robust information that will address the dynamic and complex interaction between risk factors, phenotypes, endotypes, and expression modulators and facilitate the development of precision medicine for ARDS patients through the endotype-/theratype-driven approach ( Table 3 ).

ARDS requires urgent intubation and ventilation, careful assessment and treatment of the underlying causes and continual reassessment for potential complications. The major treatments are divided into supportive managements and specific managements. Supportive therapies include venous thromboembolism (VTE) prophylaxis, nutrition, mobilization, and sedation. Specific managements comprise lung protective mechanic ventilation, fluid-conservative management, neuromuscular blockade, prone positioning, and extracorporeal membrane oxygenation (ECMO) (24). Despite these standards of care, ARDS is still associated with poor clinical outcomes with a mortality rate of approximately 40%. The lack of effective pharmacotherapies presents a continuing challenge in the field.

The complement system has key roles in both the innate and adaptive immune responses. It is composed of nearly 60 complement proteins that interact in a proteolytic cascade. As shown in Table 4 , complement proteins can be divided into two broad categories: complement components that are involved in activation and complement regulators that inhibit complement activation. Traditionally, three different initiation pathways are known to activate the ComC. Complement can also be activated by extrinsic proteases including coagulation cascade (CoaC) proteases and neutrophil-/macrophage-derived proteases. The complement pathways are tightly regulated by complement inhibitors presented in the soluble form (blood) and the solid form (membrane bound).

The activated complement provides a critical link between the innate and adaptive immune responses. As shown in Table 4 , the cascade has 4 initiating arms: the classical, lectin, alternative and extrinsic (driven by activated coagulation/fibrinolytic/kinin proteolytic enzymes) pathways. Each pathway is activated in a distinct mechanism. The classical pathway (CP) is initiated by binding pattern recognition proteins (PRPs) to immune complexes (ICs), C-reactive protein (CRP), serum amyloid P (SAP), DAMPs, and/or PAMPs, whereas the lectin pathway (LP) is homologous to the classical pathway and triggered by binding PRPs to exposed aberrant carbohydrates on foreign, damaged or necrotic cells. Unlike the CP and LP, the alternative pathway (AP) amplifies the initial response and maintains a low level of activity by a mechanism known as “tickover” (25). However, the AP is always primed to respond quickly and vigorously to pathogens or injury, and accounts for as much as 80% of C5a and C5b-9 production.

In addition, the extrinsic pathway that activates ComC has been described by Huber-Lang et al. (26). Traditionally, the plasma cascades, including the ComC, coagulation cascade (CoaC), fibrinolytic cascade (FibC), and kinin cascade (KinC), are described as separate cascades. Nevertheless, these systems belong to complex inflammatory networks (27) and exhibit similar characteristics regarding specialized functions of their activators and inhibitors. Expanding evidence indicates multiple interactions among these cascades under physiological and pathological conditions. It has been documented that active components of the CoaC (factor IIa, IXa, Xa, XIa, XIIa) (26, 28), FibC (plasmin) (29–31), and kinC (kallikrein) (32) can cleave and/or activate proteins (C3, C5) of the ComC, leading to the generation of ComC activation products (C3a, C3b, C5a, and C5b-9).

Appropriate activation of the ComC play an important role in host hemostasis, defense against infection, immune complexes and damaged cell clearance, priming adaptive immunity, regeneration, and metabolic reprogramming. However, excessive ComC activation could trigger systemic inflammatory response syndrome, cellular damage, and alveolar vascular/epithelial hyperpermeability leading to ARDS and multiple-organ failure ( Figures 2 , 3 ).

The ComC is part of the innate sensor and effector systems that are considered the host’s ´first line of defense´ with a number of biological functions. An intact ComC plays an integrative role in host defense and homeostasis by interacting with microorganisms, removing damaged cells, crosstalking with toll-like receptors (TLRs), bridging with innate and adaptive immune cells, interplaying with the CoaC, promoting tissue repair, and regulating inflammatory resolution. The pulmonary airway is exposed to inhaled pathogens and allergens on a daily basis: Airway epithelial cells and macrophages play a central role in this defense against pathogens and allergens via various host mediators including the ComC. Although the liver is the primary site for the synthesis of circulating complement proteins (except C1q, C7, and factor D), various studies in both animals and humans have demonstrated that alveolar epithelial cells, airway epithelial cells and alveolar macrophages synthesize a number of complement components. In particular, alveolar type II epithelial cells produce complement proteins C2, C3, C4, C5, C6, C7, C8, C9, factor B, factor H, factor I, and C1s inhibitor (33, 34), whereas human bronchiolar epithelial cells [C3, Decay-accelerating factor (DAF), membrane cofactor protein (MCP), and CD59] and lung fibroblasts (C1s, C3, C4, C5, C6, C8, and C9) can synthesize complement proteins (34, 35). Alveolar macrophages have the potential to synthesize the complement proteins of the functional alternative and terminal pathways (36), whereas the alveolar macrophages under a disease condition produced more complement components than their healthy counterparts (37). Furthermore, alveolar macrophage-derived serine proteinases cleaved local synthesized C5 into C5a, which initiates pulmonary inflammation (38). This local production not only significantly contributes to the systemic pool of complement, but also forms fully functioning complement pathways in the pulmonary niche thereby influencing local complement processes. The pulmonary synthesis of complement proteins in the microenvironment plays an important role in host defense against infection, clearance of ICs and damaged cells, and regulation of metabolism. Researchers reported that C1q and mannan-binding lectin (MBL) bound to and enhanced apoptotic cell uptake by alveolar macrophages in vivo (39). Chang et al. revealed that MBL was present in the lung of healthy wild type mice and MBL deficiency increased susceptibility to influenza A virus infection (40). However, excessive local ComC activation increases pulmonary vascular permeability, initiates thromboinflammation, and causes pneumonic cell damage. Our previous studies have demonstrated that excessive pulmonary complement synthesis and ComC activation contribute to pulmonary inflammation and ALI in a swine model of traumatic hemorrhage (41–44). Furthermore, one would expect prolonged activation by C3a and C5a under such pathological conditions because of the absence of carboxypeptidase N (a major anaphylatoxin inactivator) in the pulmonary interstitial space (45).

Accumulating evidence is showing that overwhelming or deregulated complement activation may fuel a cytokine storm, endotheliopathy and neutrophil extracellular traps (NETs)-driven thromboinflammation leading to collateral tissue damage. In patients, complement activation has been detected in plasma (46–49) and bronchoalveolar lavage (BAL) fluid (47, 50, 51), as demonstrated by C3a, C3c, C5a and C5b-9 increase. Besides C5a, C3a and C3adesArg (a variation of C3a that plays a central role in the metabolism of adipose tissue) were also elevated in the plasma and BAL fluid obtained from patients with ARDS compared to that obtained from patients that did not develop ARDS (6, 47, 52). However, it seems that C5adesArg was not significantly changed during ARDS (6). In addition, Zilow and colleagues reported a more sensitive assay by monitoring the C3a/C3 ratio in the plasma and BAL fluids to predict the risk to ARDS at an early stage (53, 54). Furthermore, in a particular subtype of ARDS patients, such as trauma related patients, complement activation was also detected via increased levels of C3a and C4a in plasma (4, 55). Fosse et al. suggested that different types of ARDS may have different patterns of complement activation (5). In trauma associated patients, complement activation occurred immediately after the injury and correlated strongly with severity scores of lung injury, whereas other types of ARDS (such as in sepsis patients), complement activation occurred at a relatively later stage (5). A recent study demonstrated widespread complement activation characterized by excessive C3a generation, C5b-9 formation, and C3d, C4d, MASP-2 deposition in lung tissues as well as prominent increase in blood C5a in proportion to the severity and high expression of C5aR in blood and pulmonary myeloid cells in patients with severe COVID-19 (56–59). These insights suggest that blockade of the ComC could be used to prevent the neutrophil/monocyte infiltration, endotheliopathy, and thromboinflammation associated with ARDS.

However, the mechanism of ComC activation in ARDS/ALI is largely unknown. Recently, compelling evidence reveals that DAMPs and PAMPs are activators of the ComC in other diseases and conditions. Ratajczak et al. demonstrated that high mobility group box protein 1 (HMGB1) and S100 trigger the complement lectin pathway activation via binding to the MBL during mobilization of hematopoietic stem/progenitor cells (60). Similarly, HMGB1 release activated the classical pathway by binding to C1q in a mouse model of brain ischemia/reperfusion injury (61). Furthermore, pattern-recognition molecules such as MBL, ficolins, and collectins binding to invading pathogens or damaged cells initiate ComC activation via the lectin pathway (62–65). Therefore, we postulate that the release of DAMPs and/or PAMPs after ARDS/ALI may activate the ComC. Another important factor in regards to ComC activation in the setting of ARDS/ALI is respiratory acidosis-induced ComC activation. Low extracellular pH favors the activation of the ComC and CoaC (66–73). Respiratory acidosis and metabolic acidosis (ischemic conditions) significantly lower the local and/or systemic pH and therefore might lead to complement activation.

To study the pathogenesis of ARDS, animal models are an appropriate approach. No single model can mimic all the pathophysiological changes that are observed in patients (74), but complement activation associated with ARDS development have been explored in several experimental animal models, including related to conditions of sepsis (75), IgG immune complex (76), blunt chest trauma (77, 78), transfusion-related acute lung injury (TRALI) (79) and influenza H5N1 viral infection induced-ALI/ARDS (80). Besides the correlation of complement activation with ARDS development, several other studies involving complement-mediated ARDS pathogenesis have been conducted in animals. Till et al. demonstrated that intravenously injected cobra venom factor (CVF) induces lung injury in rats (81). While direct administration of C5a has been shown to induce alveolar inflammations in guinea pigs (82), rats (76) and rabbits (83). Furthermore, using genetic manipulation of ComC, studies have showed reduced lung damage via complement-mediation: SARS-COVID-infected C3^-/- mice had reduced pulmonary infiltration of neutrophils and monocytes, reduced systemic inflammation, and less respiratory dysfunction (3); C5^-/- mice exhibited a reduction of lung injuries in an ICs-induced ALI model (84); C5aR^-/- and C5L2^-/- mice had attenuated lung injuries in three models of acute lung injury induced by LPS, IgG immune complexes, C5a, and bacterial polyphosphates (85, 86); and significantly increased survival, increased release of interferon- γ, and decreased production of IL-10 accompanied by improved pathogen clearance were observed in mild-moderate septic models of C5aR1-deficient mice (87). Additionally, pharmacological inhibition of complement levels and complement activation in preclinical and clinical ALI/ARDS demonstrated beneficial effects: C3 inhibition with AMY-101 in ex vivo whole blood infection models reduced IL-6 release (88); anti-C5 antibody (eculizumb) treatment attenuated ALI in a non-human primate septic model (89); C3 inhibitor (AMY-101) was safe and associated with a favorable clinical outcome in a patient with COVID-19 pneumonia with systemic hyperinflammation (90); anti-C5a antibody (IFX-1) treatment led to increased lung oxygenation and decreased systemic inflammation in COVID-19 ARDS patients (58, 59), attenuated ALI in patients with influenza H7N9 viral infection (91); and administration of IFX-1 in patients with severe sepsis and septic shock selectively neutralized C5a without detectable safety issues (92). E culizumab treatment in severe COVID-19 patients with pneumonia/ARDS led to a significant drop in inflammatory markers and duration of disease (93). Taken together, these studies demonstrated that complement activation is associated with ARDS development in both patients and in animal models.

Excessive complement activation in ARDS can be detected very early in both plasma and pulmonary tissue (49, 94–97). However, the mechanism of initiation of complement activation in ARDS is unclear. Several possible mechanisms have been proposed: (1) classical pathway activation via C1q binding to the antigen-antibody complex, CRP, SAP and apoptotic cells (98); (2) lectin pathway activation through MBL interaction with PAMPs (for example, the spike protein of SARS-CoV-2, viral RNA, LPS, and lipoteichoic acid) (99); (3) alternative pathway activation by means of DAMPs-TLRs (88).

Upon activation, complement products induce a profound inflammatory response, of which the main contributors are anaphylatoxins C3a and C5a. In physiological conditions, C3a and C5a are rapidly degraded by enzymatic cleavage of the C-terminal arginine residue by plasma carboxypeptidase N and R to generate cleavage products C3adesArg and C5adesArg (100). These cleavage products are involved in ARDS pathogenesis as described by Huber-Lang et al. (6), although the anaphylactic activity of C5adesArg and C3adesArg seems much lower. In the setting of ARDS, C3a interacts with its receptors C3aR, while C5a engages with its two receptors, the C5aR1 (101) and C5L2 (102) to sense the DAMPs/PAMPs signals. These receptors are highly expressed on neutrophils, macrophages, and lymphocytes. All these inflammatory cells are significantly activated in the lungs and are effector cells which perform their biological roles via multiple mechanisms. Furthermore, excessive C5b-9 formation can directly damage pulmonary cells, including alveolar epithelial cells and endothelial cells during ARDS as described below.

Complement interventions for ALI/ARDS in preclinical studies are under active investigation. In the context of trauma-, hemorrhage-, sepsis-, virus- and IR-induced ALI/ARDS, the compiled evidence of preclinical experiments focusing on the role of complement in ALI/ARDS is shown in Table 5 .

Considering clinical trials of ALI/ARDS, 29 observational studies addressed complement activation after trauma, sepsis, viruses, and transplantation ( Table 6 ). These clinical observational studies (excluding one) further demonstrated that early complement activation plays a pivotal role in the initiation and development of ALI/ARDS, and is associated with mortality. They also elucidate that complement activation is useful for prognosis and diagnosis of ALI/ARDS ( Table 6 ). Further investigation should be carried out to test whether complement targeting pharmacotherapeutic strategies have any benefits in ALI/ARDS patients.

In clinical practice, there are only two classes of complement inhibitors available. Plasma-derived human C1 inhibitors [Cinryze (Shire), Berinert (SCL, Behring), and Cetor (Sanquin)] and recombinant C1 inhibitor (Ruconest, Parming) are already approved for hereditary angioedema (HAE) (153, 297). C5 inhibitor Eculizumab (Soliris, Alexion) is another FDA-approved drug for use in patients with paroxysmal nocturnal hematuria (PNH) (153, 297). Previous efforts in complement drug discovery were hampered by limited molecular insights, technical challenges, and safety concerns. After years of development, there are currently more than a dozen candidate drugs covering a wide range of the ComC being evaluated in clinical trials. They demonstrate a broad spectrum of therapeutic targets, in the use of molecular entities (small molecules, peptides, proteins, antibodies, and oligonucleotides), and expending indications and off-label use such as transplantation, SIRS, sepsis, inflammatory diseases, and chronic obstructive pulmonary disease (298).

With regard to targeting complement therapy in human studies of ALI/ARDS, there are currently seven interventional trials ( Table 7 ). All studies have been designed to address the safety and efficacy of the complement inhibition by targeting C1q/MBL, C3, or C5 in COVID-19-induced ALI/ARDS ( Table 7 ).