Are monoclonals the way forward to prevent or to treat ventilator-associated pneumonia in the intensive care unit?

KEYWORDS:

• Antibiotic resistance
• intensive care unit
• prevention
• monoclonal antibodies
• ventilator-associated pneumonia
• virulence

1. Introduction

Antibiotics are amongst the greatest medical invention of the past century. Their widespread use is responsible for the extension of life expectancy and decrease of morbidity and mortality for a wide range of infectious syndromes. However, their usage has reached limitations related to their ecological impact at the population level – selection of antimicrobial resistance, and at the individual level – dysbiosis of intestinal flora, which motivates current research for alternatives to fight bacterial infections.

Ventilator-associated pneumonia (VAP) is the most common hospital-acquired infection in the intensive care unit (ICU) setting, affecting 10–25% of patients under mechanical ventilation [1]. Attributable mortality of VAP remains difficult to estimate but can reach 13% [2]. The VAP incidence has changed only little over time, despite the implementation of preventive strategies, such as the VAP prevention bundle and improved microbiological identification of causative pathogens [1,2].

In the search for alternatives to antibiotics, the development of monoclonal antibodies (mAbs) has seen a revival in the field of infectious diseases in the past 20 years. Phase I and II randomized controlled trials (RCT) have shown promising results for the prevention and treatment of VAP. However, mAbs have already been applied long before the 2000s in other medical domains.

1.1. History of mAbs

After the discovery that specific antibodies could protect against bacteria and toxins in 1890, antibodies were obtained through the transfusion of plasma from an immunized animal or human donor. These were used as passive immunization against many infectious diseases, but frequently resulted in, often severe, hypersensitivity reactions. Their usage, declined rapidly after World War II, with the introduction of antimicrobial chemotherapy, and remained limited to the treatment of patients affected by venoms, toxins, or certain viral infections. In the 1970’s, the development of the murine hybridoma technology and the technical improvement of the production of monoclonal, fully human antibodies, re-popularized the use of these molecules. Since 1985, mAbs have been developed to treat patients in oncology, rheumatology, and infectious diseases [3]. Currently, 114 mAbs have received approval mainly for the treatment of auto-immune diseases (n = 36), solid organ cancer (n = 28), or hemato-oncology (n = 18). Only 10 molecules have been developed for use in the infectious diseases field, the three most recent being for the treatment of COVID-19 (Figure 1) [4].

Figure 1. Timelines of development of mAbs. Before the massification of antibiotic use during late 1940s, serum treatment was widely used in infectious diseases. From 1970, technical developments on the production of large quantities of monoclonal antibodies (mAb), were marked by the development of the hybridoma technology, the production of fully human antibodies (Ab) in humanized mice, the screening of Ab candidates in libraries created in modified yeasts or phages, and the modification of Ab fragments to increase their affinity for the antigen, create bi-specific Abs or increase their half-life. Figures created using BioRender.com [3]. The first mAbs were approved for clinical use in late 1980s and the number of approvals has increased ever since. Graph adapted from [4]. Painting ‘Die Gewinnung des Diphtherieserums aus Pferdeblut im Behringwerk zu Marburg,’ Fritz Gehrke, from U.S. National Library of Medicine, in the public domain. Advertisement for penicillin production from Life magazine. Science Museum, London. Attribution 4.0 International (CC by 4.0).

1.2. Mode of action of mAbs in bacterial infections

mAbs are designed to target highly conserved epitopes at bacterial species level. To ensure bioavailability of these epitopes, the targets of choice are often surface proteins or secreted toxins important for bacterial virulence. mAbs achieve their antibacterial function by directly binding and mechanically blocking these targets, thus decreasing pathogenicity. Indirectly, the Fragment crystallizable (Fc) part of the antibody, once attached to bacterial virulence factors, can activate the cellular (induction of opsonophagocytic killing) and humoral (activation of the complement system) innate immune responses [5].

Their mode of action makes mAbs an appealing solution to the problems caused by conventional antibiotic treatment. First, mAbs’ half-life (ranging from 2 weeks up to 100 days) allows for single dose use and thus increases treatment compliance. Second, the use of mAbs has no effect on the host’s normal microbiota, preventing secondary effects related to dysbiosis, including Clostridioides difficile infection, and their possible role in the alteration of the immune tonus or development of metabolic diseases [6]. Finally, to date, no resistance emergence has been observed. This may be explained partially by the current lack of widespread use, but it is more probably due to mAbs’ mechanism of action. In fact, current mAbs’ targets are virulence factors, which do not directly affect bacterial growth. Accordingly, they are not likely to induce selection pressure over the targeted bacteria [7]. However, natural resistance mutations may appear, as seen in SARS-CoV-2, which could, in the long term, compromise efficacy. Nevertheless, mutation frequency is lower in bacteria compared to RNA viruses, but it remains a possibility that needs to be evaluated in the long term.

1.3. mAbs in VAP

mAbs that have been developed and evaluated for VAP prevention and adjunctive treatment to date have targeted Staphylococcus aureus and Pseudomonas aeruginosa, the two pathogens most frequently associated with VAP (Table 1).

Table 1. mAbs evaluated in prevention or treatment of VAP.

Pathogen	P. aeruginosa				S. aureus
mAb	KB-001	Panobacumab	Gremubamab	Aerubumab	ASN100	Suvratoxumab	Tosatoxumab
Inclusion period	Apr 2008 – Jul 2009	Feb 2008 – Jul 2009	Mar 2016 – Dec 2019	Mar 2017-Apr 2019	Nov 2016 – Jul 2018 Terminated for futility	Oct 2014 – Apr 2018	May 2019 – Jul 2022
Type of Immunoglobulin	PEGylated mAb fragment	IgM	Bi-specific IgG1	IgG1	IgG1 cocktail of ASN1 and ASN2	IgG1 with modified Fc to increase its half-life	IgG1
Target	PcrV	LPS O11	Psl and PcrV	Alginate	Hla, HlgAB, HlgCB, LukED, LukSF and LukAB	Hla	Hla
Half-life	8.1–9.3 days	35–179 hours	5.6 days		7.5 days	80–112 days
Latest phase	I/II (NCT00691587)	I/II (NCT00851435)	II (NCT02696902)	II (NCT03027609)	II (NCT02940626)	II (NCT02296320), Currently running phase III	III (NCT03816956)
Intervention	Preventive	Adjunctive	Preventive	Adjunctive	Preventive	Preventive	Adjunctive
Patients enrolled	41 patients with LRT positive to P. aeruginosa (culture)	17 patients with P. aeruginosa O11 HAP/VAP (PCR or conventional microbiology and serology)	188 patients with LRT positive to P. aeruginosa (PCR)	158 patients with P. aeruginosa VAP (culture)	155 patients with LRT positive to S. aureus (culture)	211 patients with LRT positive to S. aureus (PCR)	174 patients with S. aureus VAP (culture, gram, PCR)
Randomization	3 or 10 mg/kg KB-001 or Placebo	NA	1500 mg Gremubamab or Placebo	20 mg/kg Aerubumab or Placebo	3600 mg ASN100 or Placebo	2000 or 5000 mg Suvratoxumab or Placebo	20 mg/kg Tosatoxumab or Placebo
Safety	No significant AE or severe AE				No significant AE or severe AE
Efficacy	Decreased VAP incidence in KB-001 arm vs. Placebo (30% vs. 60%)	All patients who received three doses (n = 13) achieved clinical resolution within 9 ± 2.7 days	No difference in VAP rates vs. placebo (22% vs. 18%, p = 0.491)	No difference in clinical cure rates vs. placebo at days 7 (23% vs. 27% p = 0.843), 14 (51% vs. 46% p = 0.356) or 21 (49% vs. 55% p = 0.615)	No difference in VAP rates (7% vs. 9% p = 0.547), days under mechanical ventilation (12 ± 7.5 vs. 10 ± 6.9) or deaths (39% vs. 33%)	Decreased VAP rate in 5000 mg arm vs. Placebo (18% vs. 26% p = 0.166)	Increased rate of clinical cure at day 21 vs. Placebo (69% vs. 58% p = 0.23)
Related publication	[8]	[9,10]	[11]	NA	[12]	[13]	[14,15]

Note: Abbreviations: mAb = monoclonal antibody; LRT = lower respiratory tract; HAP = hospital-acquired pneumonia; VAP = ventilator-associated pneumonia; AE = adverse event; ig = immunoglobulin; PCR = polymerase chain reaction; Fc = Fragment crystallizable.

1.4. mAbs against P. aeruginosa

The first mAb tested for prevention of P. aeruginosa VAP was KB001 [8], targeting PcrV, a type III secretion protein able to inject cytotoxins inside neighboring cells, which, in the pulmonary environment, facilitates tissue invasion and immune evasion. Based on promising pre-clinical and phase I results, the molecule was tested in a phase II RCT that included 41 mechanically ventilated patients, who were culture positive for P. aeruginosa in the lower respiratory tract (LRT). Although KB001 had a good safety profile and decreased the incidence of P. aeruginosa pneumonia (32% vs. 60% for the placebo arm, p=0.13), clinical development was not continued.

Panobacumab (AR-101; Aerumab^TM) [9], a mAb targeting specifically P. aeruginosa serotype O11 was evaluated in a phase I/II trial as adjunctive therapy to antibiotics in patients diagnosed with nosocomial pneumonia due to P. aeruginosa O11. The use of panobacumab showed a reduced time to clinical resolution and a good safety profile, but again further studies were not pursued. This could partly be associated with the limited serotype coverage of panobacumab. Overall, P. aeruginosa is classified into 20 different serotypes, O11 being one of the most frequent serotypes isolated in pneumonia. Nevertheless, P. aeruginosa is only associated with 14–20% of all VAP, and among these only 26% are of subtype O11 [1,10]. Consequently, the clinical applicability would be restricted, and would require intensive screening of patients at risk of VAP.

More recently, gremubamab (MEDI3902) [11], a modified immunoglobulin G1 (igG1) targeting both PcrV and PsI (a biofilm exopolysaccharide) has been developed. This bi-specific mAb decreased P. aeruginosa cytotoxicity and adherence to epithelial cells in vitro and it prevented mortality in vivo in pneumonia and bacteremia murine models. A large phase II RCT included 160 mechanically ventilated patients with PCR-confirmed colonization by P. aeruginosa in the LRT. Gremubamab had a good safety profile, but the trial failed to demonstrate efficacy to prevent P. aeruginosa pneumonia.

Finally, aerubumab (AR-105; Aerucin®) was developed to target alginate, a component of the mannuronic acid located in the P. aeruginosa capsule. Similarly to gremubamab, aerubumab showed in vitro capacity to mediate opsono-phagocytic killing of a large range of P. aeruginosa serotypes, decreased mortality in an animal model of P. aeruginosa pneumonia, and had a good safety profile in a phase I study. However, the phase II trial (NCT03027609) testing aerubumab’s efficacy as adjunctive therapy to antibiotics in P. aeruginosa VAP, failed to demonstrate an improvement in the proportion of patients reaching clinical cure at days 7, 14 and 21 after VAP diagnosis, when compared to antibiotics alone.

1.5. mAbs against S. aureus

To date, three molecules have been clinically tested for the prevention (ASN100, suvratoxumab) or treatment (tosatoxumab) of S. aureus VAP.

ASN100 [12] is a cocktail of two mAbs (ASN1 and ASN2) targeting the S. aureus toxins Hla, HlgAB, HlgCB, LukED, LukSF and LukAB. The efficacy of ASN100 to prevent S. aureus pneumonia was evaluated in a phase II study in subjects under mechanical ventilation and with a confirmed S. aureus LRT colonization. This trial included 155 patients before being stopped for futility at the interim analysis. Pharmacokinetic analysis led to the hypothesis that the lack of signal was related to the decreased half-life of the cocktail in ICU patients (7.5 days), when compared to phase I data (>24 days in healthy subjects).

Suvratoxumab [13] (MEDI4893), targeting solely S. aureus alpha toxin (Hla) is an IgG mAb modified to reach an extended half-life of at least 100 days. In vivo, suvratoxumab prevented serious disease in pneumonia, dermonecrosis and sepsis animal models, allowing phase I and II trials. In the SAATELLITE phase II trial, in patients with a PCR confirmed S. aureus LRT colonization, suvratoxumab prevented the development of S. aureus pneumonia with a relative risk reduction of 32%. Considering these promising results, a phase III trial was started in 2021 and is currently enrolling patients.

Finally, tosatoxumab [14], an IgG1 also targeting S. aureus’ Hla was discovered by screening the B cell repertoire for the mAb with the highest neutralizing activity against Hla in a S. aureus pneumonia patient. This molecule showed to be active for prevention and treatment of S. aureus pneumonia in mouse models. A phase I/IIa trial confirmed tosatoxumab’s safety, its promising half-life (25 days), and indicated a trend toward faster microbiologic eradication and shorter duration under mechanical ventilation [14]. Recently, Tosatoxumab completed phase III evaluation as adjunctive treatment of VAP, in addition to antibiotic therapy. Preliminary results indicate that a higher proportion of patients reached clinical cure at day 21 (a composite endpoint including all-cause mortality, need for mechanical ventilation and signs and symptoms of pneumonia) when compared to antibiotics alone, 69% vs. 58%, and more striking effects in the sub-population of patients over 65 years old (64% vs. 30.4%) [15]. In this context, a second phase III trial is currently planned to apply for approval as adjunctive treatment in S. aureus VAP.

1.6. Challenges in approving mAbs for the prevention and treatment of VAP

ICU patients are, by definition, critically ill, meaning they have severe, life-threatening underlying diseases, and often have multiple comorbidities and concomitant treatments. Timing is key in the efficacy of clinical interventions, and proper management requires 24/7 surveillance and rapid decision making. These specificities translate in a high complexity to design and run RCTs, especially in VAP. Enrolling ICU patients in an RCT requires asking for consent from distressed families for whom research might not be the priority, and setting up 24/7 screening and follow-up, often resulting in low inclusion rates, and high costs.

Methodologically, standardized outcomes, such as all-cause ICU mortality, are not very sensitive, since baseline mortality rates in this population are high, and infection-attributable mortality relatively modest. Other traditional outcomes, such as days under mechanical ventilation or ICU discharge, are usually impacted by the underlying disease responsible for ICU admission. Therefore, the efficacy of mAbs is often evaluated by composite endpoints, such as clinical cure (alive and cured), which is a statistically more sensitive outcome. However, clinical cure is difficult to establish in severely ill patients, which raises reliability and reproducibility issues. In addition, existing antibiotic treatment strategies still have a high clinical effectiveness, which makes demonstrating superiority for novel treatment strategies, including mAbs, very challenging.

Prevention of infections is a more relevant intervention and could provide a stronger efficacy signal. However, there is no gold standard for VAP diagnosis. LRT samples of mechanically ventilated patients may be found positive for specific bacteria in the absence of clinical signs of pneumonia (colonization). Conversely, clinical signs of pneumonia (fever, leukocytosis, tracheal secretions, respiratory distress) have a low specificity for VAP in critically ill patients under mechanical ventilation. Many different criteria exist (FDA, EMA, ECDC, CPIS, or PULMIVAP criteria), but their performance to clearly identify patients with VAP from those colonized without infection is still debated. Consequently, VAP diagnosis in RCTs is often confirmed by an adjudication committee, which is a costly and time-consuming process.

On top of the difficulties linked to endpoint selection, the high specificity of mAbs further complicate their applicability. As mAbs are designed to target a single organism, the subpopulation of patients that could benefit from their use need to have confirmed presence of the targeted pathogen in the LRT. This means that for prevention purposes, patients need to be screened frequently, while for adjunctive treatment trials, timely microbiological confirmation is key. This would require the use of molecular techniques that can rapidly identify the pathogen, in combination with short communication links between microbiologists and intensivists. Since only a limited selection of patients will be eligible for pathogen-specific prevention or treatment trials, long periods of patient recruitment can be expected, and experienced ICUs with sufficient clinical and laboratory capacity are required. This could result in long and expensive RCTs, which means that selected mAbs need to be very promising to counterbalance these high costs and risks.

2. Expert opinion

Seven mAbs have been tested for the prevention or treatment of VAP due to Pseudomonas aeruginosa or Staphylococcus aureus, two of which have reached phase III evaluation. All these trials have successfully proven the strong safety profile and long half-life (up to 100 days) of mAbs, and none of them described any impact on patients’ microbiome composition, or development of antimicrobial resistance.

Despite the promising results, none of the trials has reached statistical significance for clinical efficacy, raising the need for additional studies before approval. Current regulatory pathways testing one mAb at a time require strict patient selection, targeting single pathogen VAP. This requires high throughput techniques of microbiological identification to select the treatment population, but this severely restricts the patient population and adds to the technical complexities of running mAbs trials in VAP.

While waiting for approval through standard RCT evaluation, the potential advantages of mAbs could be proposed to patients through adaptive pathways, such as the ‘compassionate use.’ This mode of usage has allowed to show the efficacy of phage therapy in difficult-to-treat bacterial infections and could be a viable pathway for mAbs in the short term.

Although we still have treatment options available for most bacterial infections among critically ill patients, due to increasing resistance rates, we will soon face a higher frequency of difficult-to-treat infections, and we cannot wait until the number of untreatable infections is overtaking to develop alternative treatment options. Moving beyond antibiotic treatment is an emergency and new drugs with different mechanisms of action are urgently needed. One way to promote research and development of mAbs is to show the possible impact a novel treatment strategy could have on the burden of antimicrobial resistance. The PriMAVeRa (Predicting the Impact of Monoclonal Antibodies & Vaccines on Antimicrobial Resistance) project aims to provide insight in the specific pathogens and the specific patient populations for which novel mAbs, or vaccines, could be the most promising strategy, through application of mathematical models parametrized by empirical data. The results of this project could guide prioritization of pharmaceutical pipelines and provide incentives for future mAb development.

Further areas of exploration of mAbs for VAP management relate to better understanding of the physiological state of the healthy lung. The relevance of the lung as a non-sterile organ, and the importance of the microbiome richness in keeping its healthy state should be evaluated in more details. A better knowledge of the normal lung microbiome, and the effect of mAbs on its composition should be part of evaluations in future mAb trials. Additionally, considering that VAP is frequently caused by multiple pathogens, and to avoid creating selection pressure, interventions should go beyond the single-pathogen interventions and address multiple targets through the use of mAb cocktails. Finally, current mAbs are IgG or IgM, whereas mucosal immune response relies essentially on IgA. The evaluation of new mAbs with this backbone could be of interest for future development.

Current trends indicate that we will be moving from a syndromic approach to individualized, pathogen-directed medicine. Individualized medicine has been the cornerstone for the development of successful therapeutic strategies in oncology. Likewise, infectious disease treatments are moving toward the use of rapid molecular identification of relevant therapeutic targets, to apply timely, specific treatments that improve survival. This has been illustrated during the COVID-19 pandemic, when successful treatments relied on early PCR testing for the identification of SARS-CoV-2 up to the variant level for the timely administration of mAbs.

We can imagine that in the near future, molecular diagnostics become generalized to rapidly identify the causative agent and virulence expression, resulting in personalized mAb cocktails, improving patient outcomes, and reducing collateral damage. However, converging efforts are required to reach this point. First, the key pathogens and virulence factors that should be prioritized need to be better understood, especially in the light of advancing insights in VAP pathogenesis. In parallel, molecular testing to rapidly identify the causative pathogen responsible for VAP should be better developed and become routine practice to select appropriate mAbs in a timely fashion. Finally, regulatory pathways should be adapted to accept alternative endpoints, to be focused on future needs, and to enable more rapid clinical development of cocktail treatments.

List of abbreviations

ICU	=	Intensive care unit
LRT	=	Lower respiratory tract
mAb	=	Monoclonal antibody
VAP	=	Ventilator-associated pneumonia

Declaration of interest

B Francois reports consulting fees from Aridis and AstraZeneca. M de Kraker has received support from the IMI2/EU/EFPIA Joint Undertaking PrIMAVeRa, grant no 101034420. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded