FormalPara Take home message

Several ARDS therapies show promise in clinical studies, while a growing pipeline of therapies is in preclinical testing. The history of unsuccessful clinical trials of promising therapies underlines the challenges to successful translation. Attention is now focused on identifying biologically homogenous subtypes within ARDS, to enable us to identify more specific ‘precision medicines’ for this severe syndrome.

Introduction

Acute respiratory distress syndrome (ARDS) is the commonest form of acute severe hypoxemic respiratory failure in the critically ill. First described in 1967, the management of ARDS remains supportive [1]. Despite considerable advances in our knowledge regarding the pathophysiology of ARDS, insights into the biologic mechanisms of injury and lung repair, and advances in supportive care, particularly ventilatory management, there remains no effective direct therapy for ARDS. Mortality and morbidity remain unacceptably high [2], underlining the need to continue to develop and test therapies for this devastating clinical condition. The lack of effective ARDS therapies has been further highlighted in the evolving COVID-19 pandemic, which causes severe acute respiratory failure and ARDS in 3–5% of infected patients. The prior disappointing experience with potentially promising therapies that have subsequently failed in large-scale clinical trials must also be borne in mind [3].

In this review, we assess the current status of promising emerging therapies for patients with ARDS. We focus on drugs that: (1) modulate the immune response, both via pleiotropic mechanisms and via specific pathway blockade effects, (2) modify epithelial and channel function, (3) target endothelial and vascular dysfunction, (4) have anticoagulant effects, and (5) enhance ARDS resolution. We also critically assess drugs that demonstrate potential in emerging reports from clinical studies in patients with COVID-19-induced ARDS.

Therapies in clinical trials for ARDS

Immunomodulatory therapies

A number of medications with a broad base of ‘pleiotropic’ immunomodulatory effects are in clinical trials for the treatment of ARDS or to prevent ARDS development (Figs. 1, 2).

Fig. 1
figure 1figure 1

Classification of therapies in clinical studies by biologic target

Full size image
Fig. 2
figure 2

Pharmacological therapies and their targets, in clinical testing for ARDS therapy

Full size image

Steroids

Steroids have long been studied as a potential therapy for both early and late phase ARDS, with some studies suggesting potential benefit, via suppression of the pro-inflammatory cytokine response, while other studies demonstrating potential risks due to immune suppression. A recent interesting open-label multicenter study examined the efficacy of high-dose dexamethasone regimen in patients with established moderate to severe ARDS (i.e., P/F ratio < 200 mmHg at 24 h following ARDS diagnosis). Although terminated early for low recruitment, it was found that the mean number of ventilator free days was 4.8 days higher and the number of patient deaths was lower (21% versus 36%) following early treatment with dexamethasone [4]. The authors highlight the dosing regimen and time of administration as key to the use of steroid therapy in ARDS. Additional studies, focused on this specific moderate to severe ARDS population (diagnosed within 24 h), will be required to confirm and extend these interesting findings.

Ulinastatin

Ulinastatin is a urinary glycoprotein and protease inhibitor with potent antioxidant and anti-inflammatory effects [5]. In a small phase 2 trial, patients (n = 40 per group) with ARDS treated with ulinastatin injection (12 hourly for 14 days) demonstrated improved lung oxygenation and function and reduced duration of mechanical ventilation and reduced hospital stays compared to standard care [5]. Ulinastatin therapy also significantly lowered inflammatory cytokines and increased antioxidant activities [5]. Another phase 2 trial of ulinastatin is currently enrolling, and a number of other protease inhibitors are in the preclinical stages of testing.

Vitamin C

Vitamin C is recognized for its antioxidant and reparative properties. In a phase 2 study of patients with sepsis-induced ARDS, vitamin C did not reduce SOFA scores, which was the primary outcome, nor did it have an effect on biomarkers, even at high doses [6]. Of the secondary outcomes, vitamin C did reduce 28-day mortality. The time delay between onset of shock and development of ARDS delayed the administration of Vitamin C infusion when compared to other studies in sepsis [6]. A phase 2 trial is currently recruiting SARS-CoV-2 patients for treatment with vitamin C (NCT04254533).

Carbon monoxide

Carbon monoxide (CO) is a gas produced endogenously by heme oxygenase, which protects against oxidative stress, cell death and suppresses inflammation [7]. Preclinical lung injury studies have shown safety and promising efficacy of low-dose inhaled CO [8]. In an exploratory phase 1 study, eight patients with ARDS were treated with inhaled low-dose CO (100–200 parts per million), which was well tolerated with trends toward a difference in lung injury severity score and a trend toward improved SOFA scores in the treatment group [9]. A phase 2 efficacy study of CO in ARDS is currently recruiting.

Mesenchymal stromal cell (MSC) therapies

MSCs have immunomodulatory and pro-reparative effects and show efficacy in preclinical models of ARDS [10, 11]. A single IV infusion of allogeneic, bone marrow-derived human MSCs was well tolerated in nine patients with moderate to severe ARDS in a 2015 phase 1 dose escalation trial [12]. However, in the subsequent phase 2a study in 60 participants, MSC treatment did not improve outcomes [13]. MSC viability was variable and may have altered their efficacy, while the patient group that had received MSC therapy was more severely ill at baseline [13]. A phase 1 study of an umbilical cord derived MSC in moderate–severe ARDS showed safety and potentially interesting immunomodulatory effects [14]. A preliminary report from an unpublished phase 1/2 trial of MultiStem® (bone-marrow-derived human MSCs) suggested that MultiStem® therapy enhanced the number of ventilator-free days (VFDs) and ICU-free days and lowered mortality [15]. Another MSC trial using umbilical cord derived cells is currently recruiting (NCT03042143), and two others are ongoing (NCT02444455, NCT03608592).

Pathway-specific immunomodulators to prevent ARDS

Dilmapimod

The p38 mitogen-activated protein kinase (MAPK) pathway is activated during cellular stress and drives downstream production of inflammatory cytokines [16]. Dilmapimod is a specific p38MAPK inhibitor and potent anti-inflammatory. In a small dose response study in trauma patients at risk for ARDS development, a 24-h dilmapimod infusion was well tolerated and reduced the concentrations of the pro-inflammatory cytokines IL-6, IL-8 and soluble tumor necrosis factor receptor 1 (TNFR1) [16]. The incidence of ARDS was low overall and not different between the groups [16].

Anti-TNFR1

An anti-TNFR1 antibody selectively antagonizes TNF-α signalling through TNF receptor-1 (TNFR1), but not through TNFR2. In a volunteer study in 37 healthy humans challenged with a low dose of inhaled LPS, anti-TNFR1 attenuated pulmonary neutrophil infiltration, inflammatory cytokine release, and reduced evidence of endothelial injury [17]. Targeting TNFR1 may have potential in ARDS and requires further investigation.

Therapies targeting epithelial/endothelial dysfunction

ARDS is a disorder involving injury and dysfunction of the pulmonary epithelium and endothelium, with resultant dysfunction of the alveolar–capillary barrier leading to lung edema. Consequently, targeting epithelial ion channels/channel dysfunction and endothelial/vascular dysfunction in ARDS constitute an important therapeutic target.

AP-301

AP-301 (also termed Solnatide) is an activator of alveolar epithelial sodium channels. Nebulized AP-301 every 12 h for 7 days was recently shown to decrease extravascular lung water and reduce ventilation pressures in a small phase 2 (n = 20 per group) randomized blinded exploratory study in patients with early ARDS (< 48 h of diagnosis) stratified based on SOFA score (SOFA score ≥ 11) [18]. Another, larger phase 2 study of AP-301 for the treatment of pulmonary edema in patients with moderate–severe ARDS is currently recruiting (NCT03567577), while another is recruiting COVID-19 ARDS patients (EudraCT Number: 2020-001244-26).

Citrulline

This nonessential amino acid is a substrate for nitric oxide synthase (NOS) in the formation of nitric oxide (NO). Low levels of citrulline are seen in patients with ARDS [19]. Citrulline deficiency may cause NOS to produce harmful nitrites, while a drop in NO can induce vasodilation, leukocyte adhesion, and alter other important aspects of endothelial function [19]. A recently completed, small phase 2 study of lower-dose (n = 26) versus higher-dose (n = 24) citrulline for patients with sepsis-induced ARDS showed no effect over placebo (n = 22) on the primary outcome measure (vasopressor dependency index), but a full report has not been published (NCT01474863).

ACE2

Angiotensin II is a vasoconstrictor, which has been implicated in lung inflammation and pulmonary edema, and is inactivated by angiotensin-converting enzyme 2 (ACE2). Angiotensin (1–7), the product of ACE2, attenuates ventilator- or acid aspiration-induced lung injury and inflammation [20] and reduces post-injury lung fibrosis [21]. Recombinant ACE2 administration was well tolerated in a phase 1 dose escalation study, while in the subsequent phase 2a study of 39 ARDS patients with concomitant infection/sepsis, there were no differences in lung or SOFA scores between the treatment and placebo groups [22].

Anticoagulants and thrombolytic therapies

Dysfunction of coagulation in ARDS plays a key role in ARDS pathogenesis. Consequently, anticoagulants and thrombolytics have also received attention as therapies for ARDS.

ALT-836

Tissue factor (TF) is a glycoprotein that is upregulated in the lung during inflammation and leads to fibrin deposition which incites further inflammatory effects [23]. Studies have observed that increased TF in the serum of ARDS patients correlates with higher mortality [23]. The anti-TF drug, ALT-836, was found to be safe when administered to ARDS patients in a phase 1, randomized, placebo-controlled, dose escalation study [24]. A phase 2 efficacy study of ALT-836 in 150 septic patients with ARDS was completed in 2013, but these results have not been published.

Heparin

Both heparin and antithrombin have been shown to dampen inflammation and ALI in preclinical models without negatively impacting systemic coagulation [25]. Nebulized heparin reduced the need for mechanical ventilation in a small phase 2 study of 50 critically ill patients [26]. Prophylactic nebulized heparin enhanced alveolar perfusion and CO2 elimination in patients following cardiac surgery [27].

Streptokinase

Streptokinase binds plasminogen to form plasmin. Nebulized streptokinase improved oxygenation and lung compliance in a phase 3 trial in 60 patients with late phase (> 10 days) severe ARDS, suggesting promise as a rescue therapy for ARDS patients [28].

Potential therapies in preclinical ARDS studies

There are a substantial number of potential therapies in preclinical testing. We will concentrate on those demonstrating particular promise in each of the key therapeutic target areas (Table 1, Fig. 3).

Table 1 Classification of therapies in preclinical studies classified by biologic target
Full size table
Fig. 3
figure 3

Pharmacological therapies and their targets, in preclinical testing for ARDS therapy

Full size image

Pleiotropic immunomodulators

Elafin

Elafin is an endogenous and immunomodulatory protease inhibitor produced by lung epithelial cells among others. Low levels of elafin, due to dysregulated cleavage, are associated with high mortality in ARDS [29,30,31]. One study showed that a functional variant of elafin that was more resistant to degradation had enhanced therapeutic benefit in a mouse model of LPS-induced ALI [30]. Specifically, it dampened immune cell infiltration into the lung and lowered monocyte chemoattractant protein (MCP)-1 levels [30].

Alpha 1-antitrypsin

Alpha 1-antitrypsin (AAT) is an endogenous protease inhibitor of several pro-inflammatory cytokines associated with ARDS including interleukin-6, IL-1β, and TNF-α. AAT inactivation has been demonstrated in infected lung lobes in community-acquired pneumonia [32]. AAT significantly improved oxygenation, decreased pulmonary edema and BAL protein levels and inflammatory cytokines, and inhibited cell apoptosis in a dual-hit mechanical ventilation and LPS-induced ALI rodent model [33]. Another study using the same dual-hit injury model in the rat (and a single-hit murine model) found no therapeutic benefit with AAT treatment [34], suggesting that additional studies are needed to further understand its therapeutic potential.

Pathway-specific immunomodulators

Imatinib

The tyrosine kinase inhibitor imatinib has potent antioxidant and anti-inflammatory effects in vivo and has been shown to ameliorate lung injury and mortality in single- and dual-hit ARDS preclinical models [35, 36]. There is also an ongoing ‘first-in-human study’ examining the effects of imatinib in healthy volunteers exposed to LPS with no results available yet (NCT03328117).

Bevacizumab

Bevacizumab, a human monoclonal antibody against vascular endothelial growth factor (VEGF), has been investigated in a model of high-permeability pulmonary edema in mice, which was induced by VEGF overexpression [37]. Bevacizumab was shown to reduce lung fluid and BAL protein levels [37]. Currently, there is a phase 2/3 trial recruiting patients with SARS-CoV-2 pneumonia for treatment with bevacizumab (NCT04275414).

Anti-IFN-γ

Interferons appear to play a complex role in ARDS, with variable effects reported depending on the specific interferon, whether type I, II or III, and ARDS etiologic agent. Interferon-β1α (Type I interferon), which has anti-viral, anti-inflammatory, and anti-fibrotic functions demonstrated promise in a phase 2a study, but the subsequent phase 3 study did not show efficacy in ARDS [38]. In contrast, certain interferons may worsen influenza-induced ARDS, as evidenced by the finding that a monoclonal antibody to IFN-γ (Type II interferon) reduced the severity of murine H1N1 influenza-induced ARDS, reduced inflammation, and improved mortality [39]. Interestingly, a recent study by Ziegler et al. showed that IFN-γ upregulates ACE2 expression in lung epithelial cells and hence could aid SARS-CoV-2 viral entry [40]. Anti-IFN-γ therapy may have potential as a therapy for COVID-19.

NLRP3 inflammasome inhibitors

The NLRP3 inflammasome is important in innate immunity and causes caspase 1 activation and the release of pro-inflammatory cytokines such as IL-1β [41]. Pirfenidone, a NLRP3 inflammasome inhibitor, was shown to suppress oxidative stress and apoptosis in vitro [42]. In a LPS-induced ALI mouse model, pirfenidone reduced lung injury scores, lung cell infiltration, and lung permeability, while also limiting caspase activation, inflammatory IL-1β release and profibrotic, TGF-β release [42]. In a recently published abstract, tetracycline, another NLRP3 inflammasome inhibitor, was shown to reduce mortality, vascular leakage, and neutrophil infiltration in a murine LPS ALI model [43]. Caspase activation and pro-inflammatory cytokine release were also diminished [43]. Currently, pirfenidone is under phase 3 clinical investigation in the treatment of SARS-CoV-2 (NCT04282902).

Targeting epithelial/endothelial dysfunction

TRPV4 inhibitors

The transient receptor potential vanilloid 4 (TRPV4) channel is a mechano-sensitive and immuno-sensitive calcium transport channel which functions to maintain pulmonary epithelial cell homeostasis. Increased TRPV4 channel activity has been implicated in ARDS pathology particularly in the context of lung stiffness [44, 45], leading to alveolar epithelial and endothelial barrier dysfunction, activation of innate immune cells, and potentiation of pro-inflammatory cytokine release, oxidative stress, and extracellular matrix deposition [45, 46]. TRPV4 −/− mice are protected against VILI [47] and chemically induced ALI [44], while TRPV4 channel inhibitors GSK2220691 and GSK2337429A also reduced ALI [44]. The TRPV4 inhibitors, GSK634775 and GSK1016790, attenuated acid instillation or chlorine gas-induced lung injury, decreasing lung edema, improving oxygenation, and attenuating immune cell infiltration and pro-inflammatory cytokine release [48]. However, a recent first-in-human study of TRPV4 inhibitor, GSK2798745, in volunteers receiving inhaled LPS was terminated early for inefficacy (NCT03511105). The effect of TRPV4 appears cell and injury specific, affecting its utility as a therapeutic target, as recently, macrophage TRPV4 activity has been shown to enhance macrophage phagocytosis and to confer protection against Pseudomonas aeruginosa infection in mice [49].

Adenosine A2A receptor agonists

Adenosine A2A receptors which are expressed on many cell types have been shown to regulate fluid transport as well as inflammation in the lung [50]. The adenosine A2A receptor agonist GW328267C enhanced alveolar fluid clearance in models of acid instillation, LPS, and live E.coli-induced lung injury [50]. Another adenosine A2A receptor agonist, CGS-21680, improved lung compliance and reduced neutrophil infiltration and pro-inflammatory cytokine release in a rat VILI model [51].

RAGE inhibitors

The receptor for advanced glycation end-products (RAGE) is expressed primarily in alveolar type-1 epithelial cells and is a regulator of epithelial barrier transport. Plasma-soluble RAGE concentrations constitute a marker of epithelial lung injury, are increased in ARDS patients, and can predict ARDS development in ‘at risk’ patients [52]. RAGE appears to drive lung injury also, as evidenced by the finding that blockade of RAGE (using peptides, monoclonal antibodies, or soluble RAGE decoy receptors) reduced acid-induced lung injury in mice [53] and piglets [54].

Haptoglobin

Plasma-free hemoglobin causes the formation of reactive oxygen species and is elevated in clinical pneumonia or sepsis. Scavengers of plasma-free hemoglobin such as haptoglobin reduced iron availability, oxidative injury, and lung injury and increased survival in a preclinical model of S. aureus pneumonia [55]. Transgenic mice overexpressing haptoglobin were also protected from hemoglobin-included lung injury [56].

Pro-resolution effects

Lipoxin A4

Lipoxin A4, which is an endogenous pro-resolving lipid mediator, enhanced alveolar epithelial wound repair, promoted differentiation of alveolar type II (ATII) cells to type I cells, promoted ATII proliferation and limited apoptosis in vitro [57]. In a murine LPS-induced ALI model, lipoxin A4 enhanced alveolar epithelial type II cell proliferation, decreased apoptosis by limiting caspase 3 activation and limited epithelial–mesenchymal transition as evidenced by immunofluorescent staining [58]. Lipoxin A4 warrants further investigation in other preclinical ARDS models.

Emerging therapies for COVID-19-induced ARDS

The lack of proven therapies for COVID-19 ARDS has prompted a vast research effort to identify new targets or repurpose existing drugs to treat COVID-19-induced ARDS (Table 2). There are two distinct strategies being pursued, namely strategies that are targeted at the virus itself (reducing replication, ACE-2 receptor binding, etc.) and strategies that modulate the host immune response to the virus infection (targeted or nonspecific immune-modulating drugs). Much of the data available in studies to date come from clinical case series, retrospective analyses, or uncontrolled clinical trials, and so definitive proof of efficacy for interventions is lacking. Nevertheless, given the urgent need for information on which to base treatment decisions, we have included such studies where better designed studies are lacking.

Table 2 Emerging therapies for SARS-CoV-2
Full size table

A positive effect of this global focus on severe COVID-19 disease should be the acceleration of multiple potential therapies into clinical testing. Given the rapidly evolving nature of COVID-19 research, we indicate where have cited unpublished and/or un-reviewed reports in this section.

Antiviral therapies/strategies

Remdesivir

Remdesivir, a broad-spectrum antiviral originally investigated as an anti-Ebola drug [59], is an analogue of adenosine that disrupts viral RNA polymerase and viral replication [60]. Remdesivir inhibits MERS-CoV and SARS-CoV in vitro and in vivo [60]. A recent study showed that remdesivir was particularly effective against SARS-CoV-2 infection in vitro [61]. A study of compassionate remdesivir use in 61 patients with SARS-CoV-2 infection observed clinical improvement in 68% of cases with improved oxygenation and a decrease in patients requiring mechanical ventilation [62]. An unpublished recent report suggesting that remdesivir shortened recovery times but did not impact mortality rates has led to the drug being licensed for use in COVID-19 patients in the USA. A recently completed phase 3 study of 237 COVID-19 patients in China showed no significant improvement in clinical outcomes although there was a trend for enhanced recovery time with remdesivir treatment [63]. Most recently, a randomized, blinded, placebo controlled trial in over 1000 patients demonstrated that Remdesivir shortened the time to recovery in adults hospitalized with COVID-19 and evidence of lower respiratory tract infection [64]. The results of several other phase 2/3 remdesivir clinical trials are awaited (Table 2).

Favipiravir

Favipiravir is a broad-spectrum antiviral RNA polymerase inhibitor, already approved for use in influenza A and B [65]. A recent, open-label, control study, showed that favipiravir exhibited significant improvements in chest CT scans and viral clearance in COVID-19 patients [66]. Several other clinical studies are underway with one examining the potential of favipiravir in combination with tocilizumab.

Lopinavir/ritonavir

Lopinavir/ritonavir are HIV protease inhibitors and are generally used as part of combination therapies. A recently concluded, open-label trial of lopinavir/ritonavir in 199 severe COVID-19 patients unfortunately showed no clinical improvement, although the mortality rate was slightly lower in the treatment group (19.2% vs. 25%) [67]. Potential explanations include lopinavir/ritonavir use in late COVID-19 infection, its use as a single agent, and in relatively lower doses, which should be addressed in ongoing studies [67]. Of relevance, another recently completed phase 2 study showed that early combined treatment of lopinavir/ritonavir with IFN-β1β and ribavirin reduced viral shedding and shortened hospital stays compared to lopinavir/ritonavir alone in mild–moderate COVID-19 patients [68].

Umifenovir

Umifenovir (also known as arbidol), an antiviral approved for influenza that can affect viral interaction and binding via ACE2, was recently shown to enhance viral clearance in comparison with lopinavir/ritonavir treatment, in a retrospective study of 50 COVID-19 patients [69]. An un-reviewed preprint reporting an open-label, multicenter trial comparing arbidol with favipiravir in 240 COVID-19 patients, with recovery at day 7 as the primary outcome measure, found no differences between these two treatments [70]. A number of studies are currently examining the safety and efficacy of arbidol in patients with COVID-19.

Chloroquine and hydroxychloroquine

The antimalarial drugs, chloroquine and its hydroxylated version, hydroxychloroquine, disrupt ACE2 binding and hence viral entry and also affect endosomal and lysosomal pH, which can inhibit the virus from merging with host cells [71]. These drugs also suppress pro-inflammatory cytokine release [72]. Chloroquine has specifically been shown to inhibit influenza A H5N1 virus-induced lung injury in preclinical models [73] and SARS-CoV-2 infection in vitro [61]. A small clinical study recently showed that hydroxychloroquine in combination with azithromycin reduced viral load in 20 patients with SARS-CoV-2 infection [74]. Conversely, concerns have been raised regarding potential adverse effects (e.g., cardiotoxicity) with chloroquine and hydroxychloroquine, particularly at high doses and when used in combination with azithromycin, in COVID-19 patients [75, 76]. A major observational study in 14,888 COVID-19 patients treated with either hydroxychloroquine or chloroquine, alone or in combination with a macrolide, found that these patients had an increased risk of mortality and an increased risk of de novo ventricular arrhythmia [77]. A number of other and larger clinical investigations of chloroquine and hydroxychloroquine, alone or in combination with other antivirals, are underway (Table 2).

TMPRSS2 inhibitor

SARS-CoV-2 viral entry into lung epithelial cells is dependent on the ACE2 receptor, while priming of the viral spike protein is dependent on the host serine protease TMPRSS2 [78]. A protease inhibitor of TMPRSS2 blocked viral entry in vitro and may be a promising therapeutic option [78]. Clinical studies investigating the efficacy of TMPRSS2 inhibitor, camostat mesilate, are currently recruiting.

Baricitinib

Another drug which may inhibit viral entry via ACE2 receptor-mediated endocytosis is baricitinib, a JAK inhibitor, that also disrupts the cytokine cascade and dampens inflammation [79] and is an approved drug for rheumatoid arthritis. Baricitinib with its anti-inflammatory and antiviral potential was identified using a data search with the BenevolentAI drug discovery platform. A recent study of 12 patients with moderate COVID-19 observed that baricitinib administered at 4 mg/day for 14 days was well tolerated and improved outcome in these patients when compared to patients receiving standard care [80]. Other larger trials evaluating baricitinib for COVID-19 are underway.

Convalescent plasma

Hoffmann et al. showed that SARS-CoV-1 serum from convalescent patients offered protection from SARS-CoV-2 infection, and this option may perhaps be effective if used prophylactically [78]. Convalescent plasma has also been shown to reduce viral load and mortality in critically ill H1N1 patients [81] and most recently has been shown to reduce viral load and improve outcome in a series of 5 cases of critically ill SARS-CoV-2 patients [82]. An un-reviewed preprint of the results from a large expanded access trial of 5000 COVID-19 patients treated with convalescent plasma (NCT04338360) showed that treatment was well tolerated [83]. Other trials assessing the safety and efficacy of anti-SARS-CoV-2-inactivated convalescent plasma in COVID-19 patients are underway.

Angiotensin II

SARS-CoV-2 binds to the ACE receptor on lung epithelial cells, which is a key step in virus infection of these cells. This also leads to a decrease in ACE2 and an increase in detrimental angiotensin II. Losartan, which is an angiotensin II receptor antagonist, is currently under investigation in SARS-CoV-2 patients (NCT04328012).

Immunomodulatory—pleiotropic effects

Methylprednisolone

The role of steroids indications for COVID-19 patients is unclear, with effects reported that might be harmful or beneficial depending on the specific clinical context [84,85,86]. Some evidence suggests that steroid use may hinder viral clearance in MERS coronavirus infection [87]. However, the effects of steroids in COVID-19 appear to depend on the dose and the degree of ‘hyperinflammation’ present, the stage of infection, and the presence of ARDS [85, 86]. A recent single-center, retrospective study of 46 patients with COVID-19 published as an un-reviewed preprint showed that early, low-dose, and short-term administration of methylprednisolone improved chest CT and clinical outcome in the treatment group [88]. Another larger retrospective study of 201 COVID-19 patients showed that methylprednisolone treatment in those with ARDS reduced the risk of death [89]. Currently, there are a number of phase 2/3 clinical trials investigating the efficacy and safety of methylprednisolone in patients with COVID-19 ARDS. Hopefully, these studies should provide clarity on the role of steroids in these patients. The recent press release suggesting a mortality benefit for Dexamethasone in COVID-19 patients in the RECOVERY trial is of particular interest in this regard.

Thalidomide

Thalidomide, an immunomodulatory drug that acts to enhance apoptosis, inhibits IL-6 and promotes T cell responses and has been shown to lead beneficial effects in preclinical bacterial- and viral-induced ARDS [90, 91]. Clinical phase 2 investigations of thalidomide for therapy against SARS-CoV-2 infection are underway.

Interferons

As discussed earlier, type I interferon, interferon-β1α, was ineffective as a sole agent in a recent phase 3 ARDS trial ARDS [38]. However, type I interferons have been shown to respond with different inhibitory potencies toward MERS and SARS [92] and, as such, interferons have been investigated, as an adjunct to antivirals, in these viral infections [93]. A recent study published as an un-reviewed preprint has observed that SARS-CoV-2 infection is potentially sensitive to type I interferons [94]. As mentioned previously, a recent phase 2 study of triple therapy with lopinavir/ritonavir, ribavirin, and IFN-β1β enhanced the recovery of patients with SARS-CoV-2 infection compared to lopinavir/ritonavir alone [68]. There are a number of other phase 2/3 clinical trials investigating the efficacy of both type I or type III interferons (including REMAP-CAP, DisCoVeRy, and SOLIDARITY), either as sole agents or as co-therapies in patients with SARS-CoV-2 (Table 2).

Mesenchymal stromal cell (MSC) therapies

The immunomodulatory effects of MSCs have generated considerable interest as a potential therapeutic for COVID-19 ARDS. A recent study of 7 COVID-19 patients observed that a single dose of ACE2-/- MSCs (10 million cells/kg) was well tolerated and improved pulmonary function, reduced TNF-α release while enhancing IL-10 release in comparison with the placebo [95]. A number of other trials are investigating the effects of MSCs and MSC-derived exosomes in patients with SARS-CoV-2 infection (Table 2).

Immunomodulatory—pathway specific

A subgroup of severely ill COVID-19 patients develop a ‘cytokine storm’ profile with rapid and sustained elevations in cytokines such as IL-6, and fulminant organ failure with features in common with secondary hemophagocytic lymphohistiocytosis (HLH) [96]. This has led to interest in specific anti-cytokine therapies.

Tocilizumab and sarilumab

Tocilizumab and sarilumab are human monoclonal antibodies that block the IL-6 receptor. IL-6 inhibition has been shown to be therapeutic in patients with adult-onset Still’s disease complicated with SIRS and ARDS [97, 98]. One recent non-controlled retrospective study of 21 patients with COVID-19, published as an un-reviewed preprint, suggested that tocilizumab treatment may have decreased white cell counts and improved CT lung opacity and lung oxygenation [99]. There are currently several phase 2/3 trials investigating tocilizumab and/or sarilumab for COVID-19 patients, with reports expected imminently.

Anakinra

Anakinra is a recombinant IL-1 receptor antagonist that neutralizes the biologic activity of IL-1a and IL-1b by competitively inhibiting their binding to interleukin-1 type I receptor and is widely used in rheumatic diseases. Anakinra did not improve mortality in patients with sepsis and septic shock in large phase 3 studies [100,101,102]. However, in a post hoc analysis anakinra improved survival in the subgroup of sepsis patients with features of HLH (ferritin elevation in excess of 2000 ng/ml, coagulopathy, and liver enzyme elevations) [103]. Anakinra is being trialled in the ‘COVID domain’ of the REMAP-CAP study (NCT02735707).

Other potential therapies

Heparin

Disordered coagulation, specifically, pulmonary microvascular thrombosis is increasingly implicated in the pathogenesis of severe COVID-19 respiratory failure. Other thrombotic complications including deep venous thrombosis are also reported. Anticoagulant therapy, mainly with low molecular weight heparin, has been associated with better prognosis in severe COVID-19 patients with evidence of coagulation activation such as markedly elevated D-dimers [104]. Consequently, heparin has been recommended by some expert consensus groups; however, its efficacy remains to be proven. Intravenous heparin is being trialled in the REMAP-CAP study (NCT02735707). Studies of nebulized heparin, such as the CHARTER study, are also in progress [105].

Finding ARDS therapies—future directions

Improved preclinical models

Understanding and, where relevant, addressing limitations to current preclinical models may help reduce future ‘translational failures’ of potential therapies for ARDS. Preclinical models are designed to be reliable and reproducible but, in achieving this, may poorly model the complexity of ARDS. More clinically relevant experimenal models can provide initial proof-of-principle; it allows ineffective strategies to be rapidly discarded.

Issues such as multiple or sequential insults, the timing of insults, the role of host factors such as age, sex, and premorbid conditions, and the usually prolonged duration of ARDS are not well reflected in current preclinical models. Testing promising therapies in more complex and diverse animal models, of varying age and species, employing multiple hits, and modeling longer durations of ARDS, while challenging, may be a useful step prior to embarking on clinical studies. Multicenter trials, incorporating randomization and blinding for preclinical studies, may minimize bias and improve robustness by increasing heterogeneity.

Other useful ‘intermediate’ steps for promising therapies prior to trials in ARDS patients may be the use of human models such as endotoxin inhalation in volunteers or testing in surgical populations, such as those undergoing one lung ventilation. Testing promising therapies in the ex vivo human lung perfusion model may provide proof of concept that the intervention can work in an acutely injured human lung.

Improved clinical trials

Improving our approach to clinical trial design and patient selection [106] may enhance the likelihood of finding effective therapies. One key issue relates to the heterogeneity of ARDS and the nonspecific nature of the ARDS clinical criteria, which may result in recruitment of patients who do not possess the underlying injury processes and biologic pathways characteristic of ARDS. ‘Practical enrichment’ involves careful selection of candidates who are likely to complete the intervention and survive the study period. ‘Prognostic enrichment’ aims to reduce the numbers required to detect a significant difference by enrolling patients who are most likely to experience the primary endpoint. ‘Predictive enrichment’ involves selecting patients based on pathobiological factors that will predispose them to a treatment response. This latter approach may offer most promise, by selecting for patients who have a strong likelihood for a response to the intervention (and by the same token, select ‘out’ those who are unlikely to respond). This would reduce study noise, sample size, and study-associated harm. In ARDS, this approach has already borne fruit: an important—positive—study of prone positioning randomized only patients who demonstrated an initial positive response to prone positioning. The use of adaptive clinical trial designs, which permits modifications of the trial and/or statistical procedures after its initiation, e.g., to favor recruitment to intervention arms where favorable outcome data appears to be emerging, may also enhance the potential to identify effective interventions. The REMAP-CAP trial is an example of such a trial in a relevant clinical population.

Targeting ARDS subtypes

Identifying patients more likely to respond to a specific pharmacologic intervention should increase chances of trial success. A key recent advance in our understanding of the pathobiology of ARDS has been the ability to divide ARDS into subgroups or sub-phenotypes. Latent class analysis identifies one-third of ARDS patients with a ‘hyper-inflammatory’ phenotype, and reanalysis of a large negative RCT of simvastatin in ARDS using this approach suggested benefit in the ‘hyperinflammatory’ group [107]. ARDS phenotyping based on the focal versus diffuse distribution of lung infiltrates is also potentially feasible [108], as are transcriptomics-based approaches [109]. While prospective trials are required to validate phenotyping approaches, and subsequently to test therapies in specific phenotypes, this approach offers considerable hope for the repurposing of drugs previously deemed to have ‘failed’ clinical translation.

Conclusions

There is a host of potential drug therapies demonstrating promise for ARDS, from drugs that modulate the immune response, specific inflammatory pathway blockers, epithelial and channel function modulators, endothelial and vascular dysfunction therapies, anticoagulant drugs, and therapies that aid resolution of ARDS. A promising pipeline of therapies is also progressing through preclinical testing. An important area of investigation is the potential for advances in our understanding of the pathobiology of ARDS and specifically the potential to identify biologically homogenous subtypes within ARDS, to enable us to target more specific therapies. It is hoped that the substantial number of studies globally investigating potential therapies for severe COVID-19 patients will help the identification of effective therapies for ARDS.