Antibiotic failure: Beyond antimicrobial resistance

Abstract

Despite significant progress in antibiotic discovery, millions of lives are lost annually to infections. Surprisingly, the failure of antimicrobial treatments to effectively eliminate pathogens frequently cannot be attributed to genetically-encoded antibiotic resistance. This review aims to shed light on the fundamental mechanisms contributing to clinical scenarios where antimicrobial therapies are ineffective (i.e., antibiotic failure), emphasizing critical factors impacting this under-recognized issue. Explored aspects include biofilm formation and sepsis, as well as the underlying microbiome. Therapeutic strategies beyond antibiotics, are examined to address the dimensions and resolution of antibiotic failure, actively contributing to this persistent but escalating crisis. We discuss the clinical relevance of antibiotic failure beyond resistance, limited availability of therapies, potential of new antibiotics to be ineffective, and the urgent need for novel anti-infectives or host-directed therapies directly addressing antibiotic failure. Particularly noteworthy is multidrug adaptive resistance in biofilms that represent 65 % of infections, due to the lack of approved therapies. Sepsis, responsible for 19.7 % of all deaths (as well as severe COVID-19 deaths), is a further manifestation of this issue, since antibiotics are the primary frontline therapy, and yet 23 % of patients succumb to this condition.

Introduction

Throughout their lives, humans encounter numerous microorganisms that can range from beneficial commensal bacteria to potentially fatal pathogens. The dynamic interplay between the host’s responses, the commensal microbiota, and potential pathogens reflects overall health and the occurrence of diseases. Infections occur when foreign invaders multiply within the body, causing damage to cells and tissues and triggering an immune response (Libertucci and Young, 2018).

The discovery of antibiotics revolutionized medicine, significantly improving the quality of life, extending human lifespans and making many complex medical procedures and therapies possible (Hutchings et al., 2019), while ensuring food safety, and advancing public health. Unfortunately, genetically-encoded antimicrobial resistance/AMR has increased while, especially in the early part of this century, the rate of new discovery declined. Nevertheless, there have always been infectious situations where antibiotics have failed, and these continue to outstrip antibiotic resistance as a cause of morbidity and mortality (Table 1). This continuing problem of antibiotic failure is only recently becoming appropriately recognized and addressed. Here we discuss the underlying causes and potential solutions.

Table 1.

Major causes of antibiotic failure.

Cause	Number of cases annually	Number of deaths	Economic impact
Biofilms (Global Report on Infection Prevention And Control, 2022)	103 million cases of Hospital Acquired Infections (HAI)	No data	$387 Billion (Cámara et al., 2022)
Sepsis	49.5 million	11 million	~$1000 Billion (van den Berg et al., 2022)
Antimicrobial Resistance; AMR (includes some sepsis cases)	32 million HAI (National Infection & Death Estimates for Antimicrobial Resistance, 2021)	1.27 million direct; 4.95 million indirect	~$300–1000 Billion (Dadgostar, 2019)
Primary immunodeficiency	6 million (Primary Immune Deficiency Diseases PIDDs, 2016; Meyts et al., 2021)	~4000 (Fernández Pérez et al., 2019)	~$125 billion (USA)
Secondary Immunodeficiency	Very common includes cancer, AIDS, diabetes/metabolic diseases, malnutrition, immunosuppressive drug-induced, etc.

Antibiotic failure is broadly defined as instances in which antimicrobial treatment is not effective in eliminating bacterial infection leading to persistent or worsening clinical conditions (Haney and Hancock, 2022). While antimicrobial-resistance/AMR is one cause of antibiotic failure that has been discussed in depth, other factors such as pathogen colonization (especially biofilm formation,) and the host’s response (e.g., immune dysfunction or genetic defects) also play major roles in antibiotic failure (Fig. 1; Table 1).

Fig. 1.

Potential factors contributing to antibiotic failure. Antibiotic failure refers to situations where the administration of antibiotics does not lead to improved patient outcomes. While antimicrobial resistance (AMR), due to genetic exchange or mutation is a significant factor, other common causes of antibiotic failure include biofilm-associated infections, infections in immunocompromised patients (e.g., individuals with primary and secondary immunodeficiencies or with dysbiosis at the level of the microbiome), and the severe condition known as sepsis. Particularly, biofilm-grown bacteria resist antimicrobial treatment 1000 times more than planktonic forms, causing persistent infections like abscesses. Primary immunodeficiencies increase infection risk from birth, with most affected individuals succumbing to infections before one year of age, despite antibiotic treatment. Similarly, traumas that lead to secondary immunodeficiencies and compromised host defenses, induce intense inflammation, weakening immunity, and increasing vulnerability to sepsis and organ failure. The microbiome’s diverse microorganisms support digestion, produce vital metabolites, and shape the immune system, benefiting overall health and influencing pathogen susceptibility. Additionally, sepsis, a life-threatening condition due to uncontrolled infection response, is a major global cause of mortality. Early and appropriate antibiotic treatment is critical, as delays can increase mortality rates.

Contributors to antibiotic failure and underlying mechanisms

Biofilms

Bacterial biofilms are colonial aggregations of microbes encased in a polymeric matrix, and usually found on a surface (Davies, 2003). Biofilms, a major contributor to infections (65 %) and antibiotic failure, are implicated in a wide range of infections, including medical devices and implants (Khatoon et al., 2018), chronic infections, lung, bladder, wound, dental, skin, ear, nose and throat, sinusitis, and orthopedic infections (Hancock et al., 2021). Since this topic was recently reviewed in depth (Hancock et al., 2021), it is briefly summarized here.

Bacterial biofilms grow in five main phases: (1) attachment to a surface of free-moving (planktonic) cells, (2) essentially-irreversible surface attachment and formation of a protective polymeric extracellular matrix, (3) formation of cell clusters embedded in the biofilm matrix (maturation I stage), (4) proliferation and maturation of microcolonies (maturation II stage), and (5) detachment and dispersal of some bacterial cells enabling new biofilm foci to form (Hancock et al., 2021; Sauer et al., 2022). Throughout these phases, hundreds of genes including dozens of regulators, manage the process, leading to significant alterations in growth state and cellular metabolism. Biofilm cells are arguably a stress adaptation and exhibit distinct transcriptional and proteomic profiles that enable them to survive in harsh environmental conditions, making them resilient against various stressors, including antibiotics (Hancock et al., 2021). Furthermore, biofilms evade host defense mechanisms, contributing to their ability to persist and cause chronic infections. In addition to genetically-encoded resistance borne by particular bacteria, biofilms are adaptively resistant to most antibiotics by about 10–1000 fold compared to their planktonic counterparts. When bacteria disperse from biofilms, they return to normal susceptibility (Davies, 2003; Hancock et al., 2021). Many factors are proposed to contribute to adaptive resistance, but we have proposed that altered gene expression of multiple resistance genes is the primary mechanism (Hancock et al., 2021). Other important factors include restricted antibiotic diffusion into the biofilm matrix, slow growth or quiescence at the base of the biofilm, etc (Hancock et al., 2021).

Abscesses, responsible for 3.2 million emergency room (ER) visits/year in the USA, are localized areas of infection characterized by high bacterial cell density, pus accumulation, inflammation, and granulation tissue formation, analogous to biofilm-associated infections. Adaptive antibiotic resistance exhibited by abscess microbes frequently contributes to infection recurrence (Darvishi et al., 2022; Rowe et al., 2021). Similarly chronic rhinosinusitis is very common (500,000 ER visits annually) and is frequently caused by bacterial biofilms but not effectively treatable with antibiotics.

Adaptive resistance

Biofilms archetypically exhibit adaptive resistance. This process is related to, but somewhat different from “tolerance” and “persistence”, where bacteria survive but fail to grow in the presence of antibiotics. In the case of “tolerance” bacteria can transiently survive exposure to high concentrations of antimicrobials (Kaldalu et al., 2020). In the case of “persistence”, bacteria survive in a quasi-dormant state (Bigger, 1944). In both circumstances bacteria can regrow when relieved from the conditions driving tolerance/persistence.

Adaptive resistance, unlike genetically encoded acquired resistance (Sandoval-Motta and Aldana, 2016), is dependent on the way bacteria grow. It is triggered by external factors such as growth state (biofilms/surface motility), nutrient conditions, various stressors, or subinhibitory antimicrobials, and usually reverses when the inducing conditions are removed (Rizi et al., 2018). Altered gene expression is the primary mechanism (Rowe et al., 2021; Sun et al., 2018), while epigenetic plasticity, as suggested by studies on DNA methylation mechanisms, might also contribute to adaptive resistance (Rizi et al., 2018).

Sepsis

Sepsis is a life-threatening condition characterized by an uncontrolled host response to an infection, with the most-severe cases of sepsis characterized by multi-organ dysfunction. Pre-COVID-19, it was responsible for 49 million cases and 11 million deaths annually (Rudd et al., 2020), a staggering 19.7% of all deaths, making it one of the leading causes of mortality worldwide. Essentially all COVID-19 deaths could also be ascribed to sepsis (Vincent, 2021; Baghela et al., 2023). Despite modest advances in our understanding of sepsis pathogenesis, therapies directed at treating the early hyper-inflammatory “cytokine storm” have universally failed, although new insights into common and endotype-specific mechanisms (Baghela et al., 2022) offer hope for new drugs addressing the aberrant immune response (Baghela et al., 2023). Currently, the most effective interventions involve promptly controlling the presumptively-causative infection (not initially known in the majority of sepsis cases) and providing supportive management (e.g., fluid resuscitation) (Polat et al., 2017). Delayed administration of antibiotics is directly associated with disease progression and increased mortality (María Luisa et al., 2020), such that for every hour’s delay in starting appropriate antibiotic treatment there is a 7.6% increased risk of death (Kumar et al., 2006).

Although antibiotics are commonly used in sepsis treatment, their effectiveness in eliminating the infection is severely limited as evidenced by the high mortality rate (23–35%) and the downstream consequences of not controlling sepsis (with high rates of progression to multi-organ dysfunction and severe post-sepsis conditions amongst survivors) (María Luisa et al., 2020). Prolonged or inappropriate exposure to antimicrobials can lead to complications such as the development of resistant or persistent bacterial infections (Prescott and Iwashyna, 2019). Additionally, the use of potent antibiotics carries the risk of adverse side effects, including drug toxicity (Barnhill et al., 2012). Therefore, while antibiotics offer apparent benefits, their efficacy and potential drawbacks must be carefully considered in sepsis management. Critically we need to better diagnose sepsis such that the best antibiotics are only administered to the individuals who will actually acquire severe sepsis and organ dysfunction.

In sepsis management, physicians often empirically prescribe broad-spectrum antibiotics and/or combination therapies as the initial treatment to avoid delays until the causative pathogen is known, and antimicrobial-resistance concerns often delay physician decisions to administer antibiotics (Pradipta et al., 2013). Additionally, the Covid era has taught us that viral sepsis is very similar to bacterial sepsis (An et al., 2023), and in this case antibiotics are not warranted. Identifying the specific bacteria causing the infection, obtaining blood cultures, and analyzing biomarkers can be time-consuming (24–72 h), although new multiplexed DNA amplification methods (see below) deliver results in ~6 h. Instead, non-specific antimicrobial prescriptions are based on the likely body location of the infection, the most-common bacterial species associated with sepsis, and local patterns of antibiotic resistance.

Optimizing antibiotic dosage is challenging in sepsis patients due to physiological changes such as hyperdynamic circulation, alterations in volume distribution, and impaired renal and liver functions. Additionally, they may undergo extracorporeal oxygenation or renal replacement therapy, further complicating antibiotic pharmacokinetics (María Luisa et al., 2020). Lower survival rates occur in certain patient populations, e. g. elderly individuals (≥65 years), newborns, and those with preexisting medical conditions (including compromised immune systems, diabetes, or cancer), in part due to a diminished response to antibiotic therapy in sepsis (Im et al., 2022; What is Sepsis?, 2022). These factors can influence antibiotic effectiveness, contributing to poorer outcomes.

Unfortunately, antibiotic failure in treating sepsis is often linked to delays in determining if a patient actually has sepsis (Baghela et al., 2022; Prescott and Iwashyna, 2019), since early symptoms (termed sudden inflammatory response syndrome) are quite non-specific (Singer et al., 2016), as well as early delays in identifying the causative pathogen and initiating appropriate therapy (Im et al., 2022). Epidemiological studies indicate that the success of antimicrobial treatment in sepsis is also influenced by age, body mass index, and the presence of comorbidities (Prescott and Iwashyna, 2019; Im et al., 2022). Therefore, the overall health, physiological condition, and immune response of the host play a significant role in antibiotic effectiveness.

Compromised immune and host defenses

The host defense system is a complex network of mechanisms that protect against microbes, involving physical barriers (e.g., skin, stomach acid, mucociliary clearance), innate immunity (e.g., phagocytic cells, complement), and adaptive immunity (e.g., T- and B-lymphocytes, antibodies). When any of these components are compromised, susceptibility to pathogens increases, leading to frequent and/or recurrent infections (Dropulic and Lederman, 2016). This can result from genetic mutations or external factors, leading to two categories of immunocompromised patients. In primary immunodeficiencies, immune function is altered due to a single gene disorder. Individuals with secondary immunodeficiencies, can be affected by another illness, condition (e.g. multigenic disorders like cystic fibrosis, asthma, inflammatory bowel disease, diabetes, etc.) or immunosuppressive therapy (such as for cancer and transplants), all of which affect the host immune system (Haney and Hancock, 2022; Morelli et al., 2022). Antibiotic treatment and aggressive therapies are often necessary for immunocompromised patients to prevent the long-term complications associated with chronic infections, while short term aggressive therapy is often warranted for individuals with secondary immunodeficiencies. However, the dysfunctional immune system poses challenges to the effectiveness of antibiotic therapy, increasing the likelihood of treatment failure (Haney and Hancock, 2022; Dropulic and Lederman, 2016) since antibiotics cooperate with host responses to clear infections (Berti et al., 2020). Moreover, specific deficiencies in individual compartments of the immune system contribute differently to the host’s susceptibility to particular pathogens. Clinical data demonstrates correlations between specific immunodeficiencies and distinct microbial infections (Dropulic and Lederman, 2016). The immune profile of the host and the underlying mechanisms influencing vulnerability to specific pathogens thus play a crucial role in determining the success or failure of antibiotic treatment.

Primary immunodeficiencies encompass a diverse range of conditions, with more than 400 recognized types, typically inherited as autosomal recessive or X-linked disorders. These disorders are often categorized based on the specific compartments of the immune system in which functionality is impaired, such as antibodies, cell-mediated immunity, complement, and phagocytes. A fundamental basis for suspicion of a primary immunodeficiency is recurring infections (Dropulic and Lederman, 2016; Berti et al., 2020; Marciano et al., 2015), with an average of 6.38 acute infections/year pre-diagnosis and 1.78 post diagnosis (Primary Immune Deficiency Diseases PIDDs, 2016). The success of antibiotics is at best transient in these patients. For example, individuals with immunodeficiencies associated with the absence or partial absence of humoral and cell-mediated adaptive immunity are vulnerable to a wide range of infections and typically present clinical symptoms within the first month of life (Dropulic and Lederman, 2016). Despite antibiotic treatment, most of these individuals succumb to infections within their first year of life.

Secondary immunodeficiencies are acquired conditions that arise from diverse external factors, including cancer (e.g., leukemia, lymphoma), autoimmune or metabolic diseases (e.g., AIDS, lupus, diabetes (Caputo et al., 1997)) and malnutrition, interventions such as chemotherapy (Dropulic and Lederman, 2016), surgeries (e.g., transplantation), or the natural aging process (e.g., immunosenescence (Szepanowski et al., 2021)). Each of these conditions presents a distinct immune profile and requires intensive antibiotic treatments, which may not always be successful in preventing and treating bacterial infections (Szepanowski et al., 2021). For example, sequelae of chemotherapy, including neutropenia causing chemotactic and phagocytic defects in neutrophils, and lymphocyte dysfunction, significantly increase the host’s susceptibility to certain infections and reduce the effectiveness of antibiotic treatment.

Furthermore, there are various clinical conditions that can compromise the host’s immune defenses, either as a result of other diseases (such as asthma, cystic fibrosis, inflammatory bowel disease, heart and liver diseases) or due to traumas (including surgeries, injuries, or burns). These conditions weaken the immune system, making individuals more susceptible to infections and decreasing the success of therapy (Dropulic and Lederman, 2016; Kiedrowski and Bomberger, 2018; Burns, 2018).

For example, in cystic fibrosis (CF) patients there is a dysfunction in the transmembrane conductance regulator (CFTR) anion channel due to one of a host of different mutations in the CFTR gene. This genetic defect leads to mucous dehydration manifesting as thick mucus on the surface of the airway epithelium, impairing mucociliary clearance in the lungs of CF patients. Consequently, microorganisms have a greater opportunity to infect the airways (especially the lungs and sinuses), triggering significant local inflammation characterized by the accumulation of neutrophils and high levels of proinflammatory cytokines in the airways. Despite receiving intensive antibiotic treatments, CF patients often experience chronic infections primarily caused by Pseudomonas aeruginosa, as well as Staphylococcus aureus, Burkholderia spp. and coinfection with other microorganisms. Patients have a much-elevated risk of frequent hospitalization for so-called pulmonary exacerbations. As another example, despite the administration of intensive antimicrobial therapy in patients with severe burns, there is a lack of immune support, and skin infections frequently lead to severe clinical complications and significant mortality.

Microbiome

Recent years have seen significant progress in microbiome research, leading microbiologists to adjust their focus from merely describing the taxonomic composition of the species within the human microbiota to understanding the functional roles of consortia of microbes and their interactions with and influence on the host. This shift has allowed for a deeper exploration of the complex dynamics and relationships between the host and the microbiome (Gao et al., 2023). The massive complexity of the microbiome has created enormous challenges to proving cause and effects, however it is clear that the addition of new relatively-benign microbes, e.g. probiotics, can have a substantial influence on host susceptibility to infections (Tegegne and Kebede, 2022). The administration of intensive and prolonged antibiotic treatments, intended to eliminate pathogens, has a profound impact on the microbiome (Patangia et al., 2022). It can disrupt the microbial composition, compromise beneficial functions of the microbiota, cause immune dysbiosis, and promote the emergence of antibiotic-resistant strains (Patangia et al., 2022).

During or after antibiotic treatments, some patients experience antibiotic-associated diarrhea (AAD), which can lead to the colonization of opportunistic pathogens such as Clostridioides difficile causing inflammation (i.e., colitis), fever and abdominal pain. In healthy individuals, C. difficile is present in low abundance in the gut. However, the disruption of the gut microbiome caused by antibiotics allows for its proliferation and dominance in the intestinal environment (Theriot et al., 2014).

Looking ahead: proposed approaches to counter antibiotic failure

While genetically-encoded antimicrobial-resistance is a prominent and growing cause of antibiotic failure (1.27 million direct deaths (Murray et al., 2022)), it is by no means the only or even the most prominent one (Haney and Hancock, 2022). Indeed, other instances of antibiotic failure including sepsis (11 million deaths (Rudd et al., 2020) and biofilm infections (65 % of all infections (Hancock et al., 2021) have caused problems throughout the antibiotic era, and yet no major advances in dealing with these instances of antibiotic failure have been forthcoming. For example, although adaptively multiple-antibiotic resistant biofilms are by far the most prominent types of infections in human health, not a single treatment for biofilms has been approved to date. We ascribe this rather shocking fact to the general lack of appreciation of the underlying causes of antibiotic failure and the paucity of simple model systems to enable screening and testing of such agents. To address this challenge and mitigate the enormous impact on human health, it is essential to adopt alternative screening methods and strategies that address the underlying basis for failure, especially utilizing non-antibiotic approaches. Such approaches lend themselves to being used in combination with conventional antibiotics (Hancock et al., 2021). In the subsequent sections, we explore diagnostic and therapeutic strategies designed to prevent and treat situations where antibiotics fail (Fig. 2).

Fig. 2.

Potential strategies to address antibiotic failure. Several potential strategies are shown here, including the development of advanced diagnostic tools and specialized therapeutics tailored for situations where antibiotics prove ineffective. These approaches aim to tackle the complexities associated with antibiotic failure and offer innovative solutions to improve patient outcomes.

Rapid diagnostics of factors that contribute to antibiotic failure

First and foremost, it is imperative that we are able to recognize situations that contribute to antibiotic failure. When bacterial infection occurs, initial therapy is often empiric (Kollef et al., 2021; Bassetti et al., 2022), taking into account the region of the body where the infection occurs, the most likely bacteria involved, known resistance patterns, and the need for urgency, etc. Unfortunately, this has the potential downside of overprescription of the most-effective broad-spectrum antibiotics with attendant concerns regarding the loss of these drugs over time due to genetic resistance development. Another enormous issue related to antibiotic stewardship is to avoid the use of antibiotics where they are not warranted such as acute respiratory tract or ear infections which are often viral in nature. For this reason, a critical need is diagnostics that avoid the use of antibiotics when unnecessary, such as PCR-based diagnostics that distinguish between bacterial and viral infections, by differentiating host responses typical of each infectious agent (Tsao et al., 2020), or utilizing combinations of host markers and clinical symptomology (Bhuiyan et al., 2019). The appropriate class of antibiotics can be inferred by rapidly determining the infectious species using multiplex PCR-based assays, and a range of other high-throughput molecular methods (Rentschler et al., 2021). Such methods can also detect the most common classes of transmissible antibiotic resistance, while ultra rapid whole genome sequencing methods that take only a couple of hours and cost less than $100 are now making it possible to rapidly detect specific organisms (Kiedrowski and Bomberger, 2018) and their resistance genes. While these are new-age approaches with a traditional bent, they can provide critical clinical information and can inform therapeutic options considerably more rapidly than culturing bacteria.

Sepsis, as mentioned above, is a life-threatening organ dysfunction caused by a dysregulated response to infection. Early sepsis (e.g. in the emergency department/ED), is very difficult to diagnose and yet there are huge penalties to the patient for delayed initiation of treatment (Kumar et al., 2006). Rapid identification of the bacterial infectious agent, combined with patient symptomology, as defined by recent Sepsis 3 definitions (Singer et al., 2016), should assist physicians in making decisions about whether to initiate intensive treatment. However, as a mitigating factor, bacteria are only identified by classical methods in around 50% of patients (Phua et al., 2013), while PCR-based methods are not more effective (Tsalik et al., 2010); thus international Sepsis 3 criteria state “suspicion of infection” as a criterion.

Newer diagnostics aim at characterizing the dysfunctional host response that is a hallmark of sepsis. A variety of gene expression signatures have been proposed (Pena et al., 2014; McHugh et al., 2015; Sweeney et al., 2015; Scicluna et al., 2015) and have reasonable accuracy, sensitivity, and specificity in the Intensive Care Unit (ICU). For example, a gene expression signature based on a type of sepsis immune dysfunction termed cellular reprogramming (where macrophages and monocytes become unable to recognize microbes), has revealed very good balanced accuracy (average of sensitivity plus specificity) in predicting the progression to organ dysfunction within 24 h, both in the ED (69–75 %) (Baghela et al., 2022; Pena et al., 2014), and early ICU (>80 %) (Baghela et al., 2023). Such a diagnostic would increase certainty and likely decrease the potential for inappropriate prescription that drives resistance development (Prescott and Iwashyna, 2019).

Another important factor that hinders therapy is the massive complexity of sepsis. Although there are undoubtedly common factors in this disease (e.g. the signatures described above), there is also substantial heterogeneity in stage of disease progression, presentation and outcomes. This includes responses driven by individual genetic variation, demographic factors, the infection source and agent, appropriateness of therapeutic intervention, comorbidities including pre-existing immune-suppressive conditions, epigenetics, etc (Leligdowicz and Matthay, 2019). Indeed, we now understand that sepsis comprises a set of endotypes, which are distinct subgroupings of disease with discrete mechanisms and pathophysiology. Several attempts to classify sepsis patients into endotypes have been made, but one recent RNA-Seq study identified 5 distinct endotypes in 348 patients that could be already identified and assorted in the ED using expression of just 2 genes/endotype as signatures (Baghela et al., 2022); these endotypes also discriminated groups amongst patients with severe COVID-19 sepsis (Baghela et al., 2023). Critically, two of these endotypes demonstrated substantially increased severity and mortality, and their ~200 unique differentially expressed genes respectively indicated neutrophilic/immune-suppressive (NPS endotype) and inflammatory (INF) phenotypes. Since these are effectively opposite phenotypes, it might explain the failure to date of dozens of therapies aimed at addressing the hyperinflammatory nature of early sepsis (Polat et al., 2017). Drug targets based on the unique genes for each endotype were identified, using two methods (hub identification by network biology and drug-gene interactions) and endotype-specific repurposed drugs predicted (e.g., dexamethasone, a known anti-sepsis agent, for the most lethal NPS endotype) (Baghela et al., 2023). Thus, since a major reason why antibiotics fail in sepsis is likely the deficiency of support from a dysfunctional immune response, diagnosis of endotypes gives hope for new early immune-corrective drugs that can synergize with antibiotics and impact strongly on outcomes.

Similar strategies can be applied for all of those situations in which underlying immunodeficiencies or compromised host defences explain the failure of antibiotics. For example, in cystic fibrosis, CFTR modulators can improve the efficacy of antibiotics (Cigana et al., 2023).

Therapies that selectively target biofilms

Severe biofilm-associated infections are currently managed by debriding the infected tissue and locally delivering antimicrobials to control the infection and prevent the dispersal of biofilm cells to other parts of the body (Koo et al., 2017). Since biofilms are highly resistant, combinations of antibiotics are often utilized, but tend to be poorly effective in preventing recurrence. For example, in the case of chronic rhinosinusitis, although antibiotics are frequently prescribed, there is surprisingly little evidence to support their efficacy (Lux et al., 2020). To address such challenges, specific antibiofilm strategies that target the unique aspects of the biofilm growth state, are being developed, although to date not a single biofilm-specific therapeutic has been approved (Hancock et al., 2021).

One approach involves physical-mechanical methods such as high-velocity spray and jet irrigators, which utilize fluid as a mechanical force to remove the biofilm and deliver antibiotics to the infection site (Fabbri et al., 2016). These methods are moderately effective, but commonly used to treat dental biofilms, exudates, and necrotic tissues.

Medical devices such as catheters, mega-prostheses, endotracheal tubes, and wound dressings can also be coated with antifouling chemicals (Yu et al., 2021), to prevent biofilm attachment and formation (Howlin et al., 2015). Coatings incorporating antimicrobials (including peptides or bio-compatible silver or copper alloys (Howlin et al., 2015), or with slow-release antibiotic properties, have shown promise for killing microbes and preventing biofilm formation (Stærk et al., 2021; Cesaro et al., 2023a, 2022a). The combination of an anti-adhesive surface with anti-biofilm peptides attached using polymer-brush technology showed efficacy in a rat model of catheter-associated biofilm infection of the urinary tract (Yu et al., 2017), avoiding a common issue for antimicrobial surfaces of devices whereby dead bacteria coat the antimicrobials on the surface blocking further killing.

One problem with biofilms is the immense variety of adhesins, matrix components, and specific elements including regulators that guide the development of biofilms. The demonstration of common broad-spectrum action of a class of host defense peptides termed anti-biofilm peptides (Hancock et al., 2021), a subset of host defense peptides, revealed the likelihood that there is at least one common targetable function, the stringent stress response that encourages biofilm growth. These peptides effectively hinder mature biofilms and eradicate diverse species within the biofilm matrix, including the drug-resistant ESKAPEE (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, Enterobacter species and Escherichia coli) pathogens. Their broad spectrum activity is due in part to their targeting a common stress signal, guanosine tetraphosphate (ppGpp) that is required for biofilm formation in diverse species. Additionally, they demonstrate strong synergy with multiple antibiotics (e.g., ciprofloxacin, ceftazidime, imipenem, and tobramycin) resulting in the complete eradication of preformed biofilms (Pletzer and Hancock, 2016). Antibiofilm peptides exhibit effects even against mature biofilms, at concentrations lower than their MICs against planktonic cells (de la Fuente-Núñez et al., 2014; Cesaro et al., 2022b), work in preclinical mouse models and human organoid models of biofilm-associated infections (Wu et al., 2021; Alford et al., 2021; Pfalzgraff et al., 2018) including sinusitis (Alford, 2021; Rentschler et al., 2021), wound (Pletzer et al., 2018) and dental (Gonçalves da Costa Sousa et al., 2022) models, and show synergy with conventional antibiotics (Pletzer et al., 2018).

Another potentially common target is promoting dispersal since this would render biofilms susceptible to conventional antibiotics. One promising method is to use nitric oxide (NO) or mimics to encourage dispersal (Ren et al., 2016), since NO is already used in human medicine for respiratory complications. NO disperses biofilms and enhances antibiotic activity by stimulating c-di-GMP-degrading phosphodiesterase, switching bacterial growth from the biofilm to planktonic mode (Boon et al., 2013). Nanoparticles including NO-releasing versions have shown promising efficacy (Pelgrift and Friedman, 2013).

Other methods are organism-selective, e.g. disrupting the extracellular polymeric substance (EPS) of the biofilm matrix using enzymes such as deoxyribonuclease (DNase) I, Dispersin B, and α-amylase. Targeted alternatives to antibiotics including antibodies (e.g. targeting EPS adhesins (Totsika et al., 2013) or matrix components (DiGiandomenico et al., 2012)) and bacteriophages (Sharma et al., 2019) (killing through phage-encoded peptidoglycan hydrolases) have also shown efficacy in reducing bacterial adhesion and biofilm initiation. Targeting the signaling cascade of quorum sensing (QS), that regulates biofilm structure/formation, is another organism-specific strategy (Jiang et al., 2019). For example, CRISPR interference (CRISPRi) technology has also been explored to target E. coli biofilms by knockdown of specific QS signaling (luxS) or adhesin (fimH) genes (Alshammari et al., 2023).

Combination antibiotic therapies

Combining antibiotics has proven to be a good strategy for overcoming antibiotic failure. Application of two or more antimicrobials has been used for different purposes in medicine such as achieving synergy (e.g., sulfamethoxazole-trimethoprim), limiting resistance (through combinations like β-lactams and β-lactamase inhibitors, or antitubercular regimens), and broadening the spectrum of action in anti-infective treatments (for example, in empiric treatments for sepsis) (Cacace, 2023; Pletzer et al., 2018; Reffuveille et al., 2014). With a repertoire of more than 300 antibiotics available today, yielding ~44,850 potential drug pairs, and considering variable factors such as dosage, metabolic conditions, and bacterial strains, there exists significant potential to enhance the effectiveness of individual antibiotics (Lv et al., 2022). In the backdrop of antimicrobial failure, which poses a pervasive global threat across all facets of modern medicine, the use of drug combinations to effectively combat infections by countering resistance, to attack multiple bacterial weaknesses simultaneously, and to provide a swift and comprehensive response in critical clinical scenarios like sepsis has emerged as a promising approach for untreatable infections. However a limitation is that organisms can become resistant to one of the antibiotics, thus muting the synergistic impact of combinations. While this strategy targets failure due to antibiotic resistance, other types of combination therapies are discussed below (e.g. using a combination of a host defence peptides and an antibiotic).

Anti-virulence therapies

Advanced understanding of bacterial pathogenesis and intracellular communication have revealed alternative therapeutic strategies for bacterial-mediated diseases. To successfully cause disease, invading bacteria must overcome various obstacles including initial physical barriers, the immune response, the resident microbiome, and prevailing growth conditions/limitations. This is promoted by the activation, under host conditions, of virulence factors that aid in colonization, evasion and/or suppression of the immune response, and nutrient acquisition; host cell damage is often a consequence of deployment of virulence determinants. By inhibiting virulence factors, host susceptibility to bacterial colonization can be reduced and the effectiveness of host immune response can be enhanced. This topic has been well covered by recent reviews (Sharma et al., 2019; Jiang et al., 2019). Such anti-virulence agents tend to be species-specific (requiring effective rapid diagnostics) and unlike traditional antibiotics are hard to pre-screen in vitro (since bacteria generally do not die). Examples include targeting of toxins/secretion-systems (Sharma et al., 2020), adhesins (Dehbanipour and Ghalavand, 2022), two-component systems (Stephenson et al., 2000; Zheng et al., 2020), bacterial communication elements (Alshammari et al., 2023), and non-coding RNAs (Dehbanipour and Ghalavand, 2022).

Host-directed therapies

Host defense mechanisms play a crucial role in protecting against infections, but when these mechanisms are impaired due to immunodeficiencies, major injuries or burns, or immunosuppressive therapy, host susceptibility to infections increases and there is a higher risk of antibiotic treatment failure. Host-directed therapies offer a potential solution by compensating for immune deficiencies and host defects, boosting innate or adaptive immune responses, or attenuating infection pathways. By acting as boosters, host-directed therapies have the potential to limit infection progression and mitigate antibiotic failure (Haney and Hancock, 2022). It should be mentioned that host-directed therapy is not new, and the use of exogenously-supplied immunoglobulins is a well-established method for treating ailments, including infections and attendant inflammation, with numerous trials underway (Motley et al., 2019).

As already mentioned above when an immunodeficiency is known, addressing the cause (e.g. CFTR inhibitors in cystic fibrosis (Cigana et al., 2023)), increases the efficacy of antibiotic treatment. Another approach is allogeneic hematopoietic stem cell transplantation (Pereira et al., 2019), which is a highly-effective method for treating immunocompromised patients. However, the success of this approach depends on finding a compatible donor, which is challenging, and requires immunosuppressive drugs which increase the potential for infection.

It can be anticipated that other types of corrective therapy will decrease susceptibility to infections including approaches such as corrective gene therapy that provides an effective copy of the mutated gene (Booth et al., 2019; Kohn and Kohn, 2021). While various viral vectors have been utilized in gene therapy, retroviral vectors are among the most effective tools, demonstrating high transduction efficiency and long-term expression of the introduced gene (Booth et al., 2019). More recently, liposomal nanoparticles are achieving considerable attention as a delivery method for nucleic acids such as inhibitory interfering-RNAs or complementing mRNAs (Kristen et al., 2019). This approach would represent a treatment to be applied in conjunction with antibiotics to suppress infection.

With allogeneic stem cell transplantation and autologous gene therapy facing notable obstacles, the advancement of genome editing technologies has emerged as a promising alternative for addressing primary immunodeficiency diseases (Rai et al., 2021). For example, severe combined immunodeficiency (SCID) is being addressed with CRISPR corrective strategies employing stem cells. Innovative platforms, employing nucleases such as zinc finger nucleases (ZFNs), transcription activator effector nucleases (TALENs), and Cas9, offer the potential for more accurate correction of disease-causing genes, paving the way for new possibilities in the treatment of immunocompromised patients (Rai et al., 2021).

However, gene therapy is expensive and faces a significant limitation in its application when there is a large population of patients. Additionally, individuals with secondary immunodeficiency, whose immune systems are impacted by multiple genetic factors to varying degrees, cannot rely solely on genetic therapies for comprehensive management of their immune disorders. It is thus crucial to understand the mechanisms underlying the pathogenesis of specific microbes, including the host factors involved, to guide strategies to mitigate microbial infections.

Recently, a genome-scale CRISPR/Cas9 knockout library in THP-1 human macrophages was employed to search for loss-of-function mutations that prevent Salmonella Typhimurium uptake (Yeung et al., 2019). The screen revealed 186 genes involved in various pathways, including receptor signaling, cytoskeletal dynamics, calcium transport, and the metabolism of glycosaminoglycan and cholesterol, especially highlighting the role of the NHLRC2 gene in macrophage differentiation and pathogen invasion (Yeung et al., 2019). Such studies could inform the development of host-directed therapies as alternatives to conventional anti-infective treatments, and/or as adjuncts to antibiotics.

Both host uptake of intracellular pathogens and immune responses to those pathogens may involve signal transduction pathways for which many drugs already exist. As mentioned above, the knowledge of differentially-expressed genes in sepsis, and computational methods provided lists of potential repurposed (already-approved) drugs for intervention in the two most deadly endotypes of sepsis (Baghela et al., 2023). Similarly, phosphoinositide-3-kinase inhibitors have been adapted as adjuvant therapeutics for bacterial infections (Fleeman, 2023).

Immunomodulatory therapies offer an approach to boosting host immunity against infections (Del-Rio-Navarro et al., 2006; Hancock et al., 2012). This includes the use of agonists targeting pattern recognition receptors such as Toll-like (TLR) or NOD-like receptors (Hancock et al., 2012), as well as selective immune boosters like host defense peptides (Hancock et al., 2016). While such therapies are less commonly used for bacterial than viral infections, mixtures of killed bacteria are employed in Latin America as therapies for lower respiratory tract infections (Del-Rio-Navarro et al., 2006). Immunomodulators can also be administered alongside conventional antimicrobial agents as adjunctive treatments to enhance activity (Hancock et al., 2012).

TLRs play crucial roles in recognizing microbial signature molecules (e.g., lipopolysaccharide, lipoteichoic acid, DNA and RNA) and triggering immune responses (cytokines, chemokines, and type I interferons against infections) (Hancock et al., 2012). By using molecules that mimic natural ligands, responses mediated by TLR and other pattern recognition receptors, can be modulated. Agonists can enhance the inflammatory response, acting as adjuvants to innate immunity, while antagonists can suppress immune pathways and mitigate inflammation (Hancock et al., 2012). An example of an immunomodulator is DNA-based Cobitolimod (DIMS0150) that targets TLR9 on immune cells and is undergoing Phase III trials for the treatment of ulcerative colitis (Atreya et al., 2018).

Both natural and synthetic host defense peptides (HDPs) offer alternative immunomodulatory options for host therapies. HDPs possess the ability to selectively affect various immunomodulatory responses in the host (Hancock et al., 2016), enhancing elements of protective immunity (e.g., chemokine induction, etc.), while inhibiting bacterial-stimulated pro-inflammatory responses. Synthetic HDPs termed innate defense regulators (IDRs) are optimized for immunomodulatory functions ¹⁰¹. IDRs can lack direct antimicrobial effects, but still exhibit anti-infective properties in numerous animal infection models (Hancock et al., 2012; Mookherjee et al., 2020; Madera and Hancock, 2012; Etayash et al., 2020), or combine e.g. anti-biofilm and anti-inflammatory activities (Rowe et al., 2021). The protective effects of IDRs are often attributed to their ability to attract monocytes to the site of infection, thereby enhancing the innate antibacterial immunity of the host (Madera and Hancock, 2012) and they are designed to be used in conjunction with conventional antibiotics.

Other alternatives to antibiotics

There are various unconventional antimicrobial strategies for the prevention and treatment of bacterial infections, including several mentioned above. Probiotics, for example, have shown great potential in protecting the host and reducing the risk of infections (Li et al., 2022). These living therapeutics exert their beneficial effects through multiple mechanisms, including modulating the host’s immune responses to pathogens, fortifying the epithelial barrier, inhibiting pathogen adhesion, releasing antibacterial substances, and neutralizing toxins (Li et al., 2022). For example, efficacy was achieved in a clinical study of 10,000 premature very-low birthweight babies, in which a simple daily dual-probiotic supplementation significantly reduced sepsis mortality and enterocolitis (Denkel et al., 2016; Zheng et al., 2021). Similarly, a meta-analysis revealed that probiotic supplementation enhanced standard therapy resulting in ~13 % increase in eradication of Helicobacter pylori and significantly reduced (~41 %) adverse events (Zhang, 2015). Probiotics also have value in preventing antibiotic-associated diarrhea (Goodman et al., 2021). Similarly, prebiotic substances, have demonstrated significant effects on both host and microbiome health (Xu et al., 2022). They can modify the composition of the gut microbiome, promote the secretion of specific metabolites, and facilitate the proliferation of beneficial bacteria (Mookherjee et al., 2020), and offer potential as adjuncts to improve health and microbiome balance.

Phage therapy, employing viruses that infect bacteria, was one of the earliest antimicrobial strategies employed against infections, but was overwhelmed by the advent of antibiotics. However, it has regained attention due to its ability to specifically target and kill particular strains of bacteria, including biofilm cells, while preserving the host microbiome and exhibiting minimal toxicity (Zrelovs et al., 2021). It also has the strong potential to work in synergy with antibiotics against resistant bacteria (Diallo and Dublanchet, 2022) and boost immunity (Federici et al., 2021). For example, a phase II/III clinical trial was initiated to assess the efficacy of an engineered phage biotherapeutic in treating acute uncomplicated urinary-tract infections caused by multidrug-resistant E. coli (Madeline, 2022).

Apart from phage therapy, lysins have emerged as another promising alternative to antibiotics. Lysins are enzymes synthesized by bacteriophages that disrupt bacterial cell walls (Zrelovs et al., 2021). They offer several potential advantages including selective targeting, low risk of resistance development, rapid action, and minimal toxicity (Zrelovs et al., 2021).

There is also the possibility of reversing the potential antagonistic effects of the infectious microenvironment, for example by reversing anaerobicity at infectious sites. Thus, hyperbaric oxygen (Jensen et al., 2019) has been proposed to reoxygenate tissues and restore antibiotic treatment of infectious biofilms to benefit patients.

Conclusions and outlook

Antibiotic failure is massively important issue in terms of mortality, morbidity, and long-term consequences to human health. Addressing the issues of antibiotic failure requires a comprehensive approach that goes well beyond the discovery of new antibiotics to combat drug-resistant bacteria. While developing new antibiotics is important in combating antimicrobial-resistance, it has severe limitations in addressing the most prominent causes of antibiotic failure. Groups interested in dealing with the threats afforded by bacterial (and other) infections need to consider other strategies as discussed above, including those based on the most exciting, recent developments in science such as cutting-edge AI approaches that can predict new antibiotics and peptides (Cesaro et al., 2022b, 2023b; Maasch et al., 2023; Torres et al., 2022; Wan et al., 2022; Wong et al., 2023; Cesaro et al., 2022; Fangping and Marcelo, 2023; Dias Santos-Júnior et al., 2023; Torres et al., 2023). To effectively address antibiotic failure, it is essential to have a comprehensive understanding of the underlying mechanisms and various factors contributing to failure, and a more holistic approach is necessary. This holistic approach includes implementing stewardship programs to optimize antibiotic use, exploring combination therapies that target multiple aspects of bacterial survival, and investigating alternative antimicrobial, anti-infective and adjuvant approaches. By combining these strategies, the effectiveness of treatment can be enhanced while reducing antibiotic failure. Since such strategies require new drug development paradigms, we must grow beyond our dependence on killing bacteria as a screening tool. This will require the development of model systems to study the infectious situations that contribute to antibiotic failure and enable high throughput screening. For example, new methods for in vitro screening of anti-biofilm agents (Haney et al., 2021) and new simple animal (Pletzer et al., 2018) and organoid (Wu et al., 2021) models are being instituted for biofilm infections.

Integrating rapid diagnostic tests that facilitate streamlined diagnosis, introducing therapeutic tools that target the host or host plus pathogen, and defining appropriate drug regimens are vital steps in minimizing the incidence of antibiotic failure. These measures will provide clinicians with a comprehensive array of resources to effectively manage all types of infections and tackle the challenges associated with antibiotic failure.