Precision Medicine in Asthma: The Role of Biomarkers

Introduction

Asthma is the most common chronic respiratory condition affecting both children and adults, with an estimated prevalence of over 260 million cases globally.¹ Although asthma is now understood to be an extremely heterogeneous condition underpinned by multifactorial immunopathological mechanisms and diverse genetic and environmental factors,² it continues to largely be treated as a singular disease. Conventional management in most treatment guidelines is structured in a stepwise algorithmic approach, where pharmacological treatment is escalated based on symptom control.³

The advent of monoclonal antibody biologic therapies has significantly improved clinical outcomes in patients with severe asthma. However, a considerable subset of patients exhibit suboptimal responses. This heterogeneity is likely attributable to the paucity of robust, specific biomarkers capable of delineating underlying disease mechanisms and thereby facilitating the selection of therapies tailored to individual patient profiles.⁴ Consequently, asthma management frequently follows a “treat-to-failure” paradigm, whereby standard therapies are applied uniformly, and treatment modifications are implemented only after patients experience persistent loss of disease control and recurrent exacerbations.⁵ Delayed targeted intervention results in morbidity and mortality resulting from exacerbations and hospitalisation, chronic airway remodelling and iatrogenic effects of oral corticosteroid exposure.

A treat-to-target approach in asthma, grounded in precision medicine, involves tailoring therapy to a patient’s underlying disease biology early and monitoring treatment response. By integrating biomarkers and endotyping into assessment at disease onset, the goal is to ensure that the right patient receives the right therapy at the right time, optimising outcomes while avoiding unnecessary exposure to ineffective treatments.⁶ This review will summarise the evidence for and critically appraise current biomarkers in asthma and highlight emerging biomarkers that may help to guide personalised management of asthma in the future.

Moving From Asthma Phenotypes to Endotypes

In the past few decades there has been a paradigm shift from asthma being viewed as a single disease entity, to a heterogeneous condition with significant variability in patterns of airway inflammation, severity of disease and response to treatment. Traditionally, the predominant dogma was that patients could be split into two cohorts-allergic (“extrinsic”) and non-allergic (“intrinsic”). Our understanding has evolved over time that this oversimplifies a much more complex picture of interrelated immunopathogenic pathways.

Practice initially shifted toward stratifying patients by phenotype, defined as clusters sharing similar observable clinical characteristics. Analyses from several large multicentre asthma cohorts have been pivotal in defining disease phenotypes, with the most influential being the Severe Asthma Research Program (SARP) in the United States and the Unbiased Biomarkers in Prediction of Respiratory Disease Outcomes (U-BIOPRED) cohort in Europe.^7,8 These studies utilised different clustering techniques and computational approaches incorporating clinical, inflammatory, transcriptomic, proteomic and metabolic data. The clusters identified were similar across studies, which included (1) early-onset, mild allergic asthma with preserved lung function (2) early-onset allergic asthma with fixed airflow limitation, (3) late-onset non-allergic, eosinophilic disease, (4) late-onset, non-allergic, neutrophilic severe asthma with fixed airflow obstruction and (5) late-onset nonatopic asthma in obese female patients.^7,9 These clusters showed that sputum inflammatory profile appeared important in determining stratification.

However, a major limitation of this approach is the significant overlap between clusters and the inability to link distinct molecular mechanisms driving the clinical phenotype. The ability to subtype based on specific disease mechanism, or endotype, has not yet been achieved beyond being able to identify patients with type 2 (T2)-driven inflammation versus T2-low asthma.¹⁰ Several biomarkers of T2 inflammation exist, however, these biomarkers do not discriminate well between endotypes and cannot predict a patient’s response to a specific T2 biologic. There is a paucity currently of validated non-T2 biomarkers. In the next section, we will discuss some of the currently utilised established biomarkers, the evidence for their role in asthma stratification and their limitations (Figure 1).

Figure 1 Biomarker matrices in asthma. Biomarkers in asthma from breath, airway lumen, airway wall, and more distal sites including blood and urine. Created in BioRender. Horn, N. (2025) https://BioRender.com/3byzpyi. Abbreviations: FeNO, Fractional Exhaled Nitric Oxide; VOCs, Volatile Organic Compounds; PexA, particles in exhaled air.

Biomarkers in Asthma

T2 Biomarkers

Sputum Eosinophils

Sputum inflammatory cell analysis is a non-invasive method to enumerate differential inflammatory cell pattern in the airways. A number of studies have confirmed eosinophilic airway inflammation to be the most common phenotype accounting for 50–60% of the asthma population,¹¹ though it has been observed to be higher in other populations.¹²

Sputum eosinophilia has been shown to correspond to worse baseline lung function, greater airway hyperresponsiveness and higher risk of exacerbations.^13,14 It has also been shown to predict response to corticosteroids,¹⁵ and a sputum eosinophil-guided management strategy to titrate inhaled corticosteroid doses significantly reduced exacerbation rates compared to guideline-based care with symptom-based management.^16,17 However, obtaining sputum through induction is not widely used as it is difficult and time-consuming for patients to perform, costly and requires laboratory expertise that is not available in most clinical settings. As a result, other biomarkers that are surrogates of airway eosinophilia are more widely used, such as blood eosinophils and FeNO.

Blood Eosinophils

Blood eosinophils have become a key surrogate biomarker for eosinophilic airway inflammation as full blood count testing is often routinely done in clinical practice. Blood eosinophil count has been shown to be related to significantly higher risk of asthma exacerbation and failure to achieve asthma control,¹⁸ and has also been shown to be associated with airflow obstruction and accelerated decline in lung function.¹⁹ The relationship between eosinophils in blood and airway compartments however has not been consistent, with studies showing varying sensitivity and specificity to predict raised sputum eosinophils.^20–22

Blood eosinophil counts have been demonstrated to exhibit dose response to both inhaled and oral corticosteroids.^23,24 Importantly, blood eosinophils have been valuable as a biomarker to predict response to multiple T2 biologics, including those targeting anti-interleukin-5 (IL-5) (mepolizumab, reslizumab), anti–IL-5 receptor α (anti-IL-5Rα, benralizumab), anti–interleukin-4 receptor α (anti-IL-4Rα, dupilumab) and anti-thymic stromal lymphopoietin (anti-TSLP, tezepelumab).²⁵ Not enriching for underlying phenotypes can increase the risk of unsuccessful clinical trials by enrolling patient populations unlikely to benefit. This was exemplified by the initial trial of mepolizumab, which failed to demonstrate a significant reduction in airway hyperresponsiveness in an unselected cohort and nearly resulted in the discontinuation of its development as a biologic therapy.²⁶

Although blood eosinophils reflect the presence of T2 inflammation, they are not specific to the dominant cytokine pathway driving disease and cannot reliably predict specific biologic response. While IL-5 drives eosinophil maturation, activation and survival, interleukin-14 (IL-14) and interleukin-13 (IL-13) contribute to eosinophilia by driving eotaxin production, enhancing eosinophil trafficking and recruitment into airway tissue.²⁷ This is borne out by biologic switch studies which have found that despite selection for patients with high eosinophil counts, there are patient cohorts which have shown response to dupilumab after suboptimal response to mepolizumab, suggesting additional benefit of IL-14/IL-13 blockade beyond eosinophil suppression.²⁸

Fractional Exhaled Nitric Oxide (FeNO)

Fractional Exhaled Nitric Oxide (FeNO) has in recent years been more widely used as a biomarker in asthma. Nitric oxide is produced by airway epithelial cells as a result of IL-4 and IL-13/Signal Transducer and Activator of Transcription 6 (STAT-6) induced upregulation of inducible nitric oxide synthase (iNOS).²⁹ Increased nitric oxide levels can contribute to bronchial hyperresponsiveness, vascular permeability, mucus hypersecretion and airway inflammation. It can be measured reliably and non-invasively by handheld devices; the American Thoracic Society (ATS) guidelines recommend a level of less than 25 parts per billion (ppb) to be considered as normal in adults and greater than 50ppb to be elevated.³⁰ FeNO levels of greater than 50ppb and concurrent raised blood eosinophil count has been shown to be correlated with greater risk of exacerbation and asthma symptoms.^31,32

A systematic review demonstrated that FeNO had a summary area under the receiver operator curve (AUC) of 0.75 (95% CI 0.72–0.78) for sputum eosinophils of 3% or more.³³ A meta-analysis showed only modest diagnostic accuracy for asthma diagnosis (pooled sensitivity of 0.65).³⁴ Establishing a threshold to diagnose asthma in mixed populations can be challenging as it can be confounded by a number of factors, most significantly inhaled corticosteroid use and smoking, and can vary with age, sex, height and ethnicity.³⁰

FeNO has been shown to decrease rapidly with inhaled corticosteroids. A trial comparing FeNO-guided titration of inhaled corticosteroid versus conventional treatment found that the FeNO-tailored group had significant reductions in final mean daily doses of inhaled corticosteroid.³⁵ The use of FeNO as a biomarker for selection of biologics has also been demonstrated. Subgroup analysis of the EXTRA study which evaluated the effect of anti-IgE monoclonal antibody omalizumab found a significantly greater reduction in exacerbations in the high FeNO group (>19.5ppb) of 53% vs 16% in the low FeNO cohort (<19.5ppb).³⁶ In the LIBERTY ASTHMA QUEST study, the efficacy of dupilumab in reducing annualised exacerbation rate was significantly greater in the subgroup with FeNO≥50ppb.³⁷ FeNO has not been extensively evaluated as a biomarker in other biologic studies, however the DREAM study reported no significant effect of mepolizumab on change in FeNO pre and post treatment compared to placebo,³⁸ and a real-world systematic review found no significant change in FeNO pre and post treatment with benralizumab.³⁹ A higher FeNO level may therefore be an indicator that targeting the IL-4/IL-13 pathway may be more effective than targeting IL-5.⁴⁰

Serum Periostin

Periostin is an extracellular matrix protein produced by bronchial epithelial cells and fibroblasts in response to IL-4 and IL-13. As a potential surrogate marker of T2 inflammation it is unique in being related to airway remodelling. In mouse models, it has been implicated in the process of subepithelial fibrosis having been shown to bind to other extracellular matrix proteins such tenascin, collagen V and fibronectin, and has been significantly associated with fixed airway obstruction.^41,42

Serum periostin has been shown in a randomised controlled trial (RCT) to significantly decrease in response to inhaled corticosteroids, and this decrease was associated with improvements in lung function and decrease in computed tomography (CT) measurements of wall thickness.⁴³ It was shown in the EXTRA study to potentially be able to predict response to omalizumab, where the reduction in exacerbations in the high serum periostin ≥50ng/mL group was 30% vs 3% in the low serum periostin group, though this did not reach statistical significance.³⁶ High serum periostin was also found in a trial of anti-IL-13 biologic lebrikizumab to predict greater improvement in forced expiratory volume in 1 second (FEV1).⁴⁴ However follow-on Phase III studies LAVOLTA I and LAVOLTA II failed to show consistent improvement in exacerbation rates or FEV1 improvement compared to placebo, and selection of patients with high serum periostin did not consistently identify those who would benefit from treatment with lebrikizumab.⁴⁵

The clinical use of serum periostin has so far stalled due to several key limitations. It was shown in a study to have poor accuracy in predicting sputum eosinophilia compared to blood eosinophils, with a ROC AUC of only 55% (p=0.44).⁴⁶ In addition, large intraindividual and interindividual variability, and a lack of validated and standardised enzyme-linked immunosorbent assay (ELISA) limits reproducibility and ability to set normal ranges.

Eosinophil Peroxidase (EPX)

Eosinophil peroxidase (EPX) is emerging as a promising biomarker as an indicator of eosinophil activation rather than merely reflecting the presence of eosinophilia. EPX is known to have direct toxic effects due to its generation of reactive oxygen species (ROS) and oxidative modification of tyrosines and lipids, leading to downstream T2 inflammatory cascades, airway remodelling and mucus hypersecretion.⁴⁷ It has also been linked to excessive eosinophil extracellular trap formation, which is increasingly recognised to play a key role in chronic airway inflammation and epithelial damage seen in asthma.⁴⁸

It has previously been shown that sputum EPX levels are strongly correlated with sputum eosinophil counts,⁴⁹ and that sputum and nasal EPX are significantly increased in asthma vs healthy controls.⁵⁰ The utility of nasal EPX to predict treatment response has also been demonstrated in a trial of oral eosinophil lowering therapy dexpramipexole in eosinophilic asthma, where it was demonstrated to reduce EPX in a dose-dependent manner.⁵¹ Furthermore, it has been shown from analysis of the SARP cohort that a cohort of patients on mepolizumab had persistently raised sputum EPX despite blood eosinophils normalising, and that persistently raised sputum EPX was associated with greater risk of exacerbations, worse airflow obstruction and higher mucus plug scores.⁵²

Eosinophil Extracellular Traps (EETs)

As part of innate immune defense against external pathogens, neutrophils are activated and degranulate, releasing neutrophil extracellular traps: web-like chromatin structures composed of DNA, proteins and granules which allow the host to bind and kill microbes.⁵³ Recent studies have shown that eosinophils can form extracellular traps (EETs) similar to neutrophils, which are also important for host defence against infection. However, excessive production of EETs (or “EETosis”) can lead to airway inflammation and remodelling. Activated eosinophils have been shown to not only release cytotoxic granule proteins such as EPX and eosinophil-derived neurotoxin (EDN) but have also been shown to release galectin-10 and undergo a transformation to Charcot Leyden crystals in airway tissue. These can activate macrophages and epithelial cells to release proinflammatory cytokines.⁴⁸ In humans, EETs measured in bronchoalveolar lavage fluid were found to be significantly higher in asthmatics compared to healthy controls, positively correlated with IL-4, IL-5 and IL-13, and associated with worse lung function and disease severity.⁵⁴

The feasibility of EETs as a therapeutic target is still under evaluation, with different drugs targeting different phases of EET formation and release, including DNase treatments, peptidyl arginine deiminase 4 (PAD4) inhibitors and NADPH/ROS inhibitors.⁴⁸ At present, there are a lack of easily measurable biomarkers for EETs. Other serum circulating biomarkers have been proposed as surrogates of airway EETs including circulating citrullinated histone H3 (citH3), calmodulin and Galectin-10 (see below), however further studies are required to evaluate these biomarkers in asthma patients.

Galectin-10

Galectin-10 is a member of the lectin family and is an eosinophil-specific cytosolic protein that has shown to be a major component of Charcot Leyden crystals, which have been observed in multiple eosinophilic pathologies. During EETosis, non-apoptotic cell death leads to release of free Galectin-10 after plasma membrane disintegration and crystallises extracellularly into Charcot Leyden crystals.⁵⁵ Galectin-10 can be measured in sputum, nasal lining fluid, serum and tissue. Serum Galectin-10 was showed to be significantly higher in asthma subjects with evidence of fixed airflow limitation and small airways dysfunction.⁵⁶ Sputum Galectin-10 and Charcot Leyden crystals were also shown in another study to be significantly raised in asthmatics compared to controls, and had a significant correlation with sputum eosinophil count.⁵⁷ A small study in 20 patients has shown that mepolizumab significantly decreased levels of serum and sputum Galectin-10.⁵⁸ However further studies are required to evaluate Galectin-10 as a biomarker to stratify response to other treatments.

Proteomics

As the field has moved from disease phenotyping to endotyping, strategies employing “-omics” approaches – large scale studies enabling analysis and clustering of genes, proteins, and metabolites – can potentially identify and refine new endotypes through a more systematic and quantitative characterisation at the molecular level. Proteomics, by generating a fingerprint of the pattern of protein expression, can provide a window into the current inflammatory state of the airway.

U-BIOPRED was the first large study to apply – omics strategies in an unbiased manner to stratify asthma endotypes. Liquid chromatography with high-definition mass spectrometry was performed to profile the proteome of sputum supernatants in 246 participants.⁵⁹ Clustering of patients was based on proteomic data using topological data analysis (TDA), which identified 10 clusters. When sputum granulocyte profile was overlaid, three of the proteotypes associated with a highly eosinophilic subphenotype, two of the clusters had atopy as their main feature, and three clusters were characterised by neutrophilia. Ten proteins were strongly predictive of the eosinophilic phenotype: including histone H4, complement C3, α-1-antitrypsin and galectin-3-binding protein. The neutrophilic phenotype was associated with 14 differentially abundant proteins, including histone H4, annexin A1 and A3 and neutrophil gelatinase associated lipocalin. This study demonstrated the ability of proteomics to identify several candidate biomarkers and potential novel therapeutic targets.

Bronchial Biopsies

Bronchial biopsies are considered the gold standard for understanding structural changes in the airway, and have identified key histopathological features including epithelial cell shedding and desquamation,⁶⁰ eosinophilic and neutrophilic airway inflammation,⁶¹ goblet cell hyperplasia,⁶² subepithelial fibrosis,⁶³ epithelial basement membrane thickening,⁶⁴ airway smooth muscle hypertrophy⁶⁵ and angiogenesis.⁶⁶

The CASCADE trial was an exploratory Phase 2 study assessing Tezepelumab where the primary endpoint was the change in number of airway mucosal inflammatory cells before and after treatment in biopsy samples.⁶⁷ There was a significant reduction in the number of airway submucosal eosinophils in the treatment arm but no significant change in any other inflammatory cells. Other studies have examined the effects of biologics including benralizumab on airway smooth muscle mass,⁶⁸ lebrikizumab on airway subepithelial eosinophils in basement membrane,⁶⁹ and fevipiprant on submucosal eosinophilic count.⁷⁰ Due to the invasive nature of biopsy sampling, there is increasing interest in other methods of assessing the effects of biologics on airway structure including use of quantitative imaging approaches to the airways.

Beyond T2 Biomarkers

Sputum Neutrophils

Type 2-low asthma, or asthma in the absence of T2 inflammation, has thus far lacked any specific validated biomarkers. Studies have associated it with both neutrophilic and paucigranulocytic sputum inflammatory profiles.⁷¹ A previous study has found sputum neutrophilia present in around 20% of asthmatics.⁷² Importantly, sputum neutrophilia has been demonstrated to predict poor response to inhaled corticosteroids, where this subgroup were shown to have significantly less improvement in symptom scores, FEV1 and airway hyperresponsiveness after a trial of inhaled budesonide.⁷³ These patients therefore represent an urgent unmet clinical need as they respond poorly to the mainstay of treatment for asthma and are not candidates for T2-targeting biologics.

Unlike sputum eosinophils, blood neutrophil count does not appear to predict sputum neutrophil count.⁷¹ Several other biomarkers have been associated with sputum neutrophilia, including interleukin-17 (IL-17), CXC chemokine receptor (CXCR2), tumour necrosis factor-α (TNF-α), interleukin-23 (IL-23) and interleukin-1-beta (IL-1β). A number of monoclonal antibodies against these therapeutic targets have been evaluated, which will be discussed in the following sections.

Interleukin-17 (IL-17)

Interleukin-17 (IL-17) a cytokine released by T helper 17 (Th17) cells, which is known to play an important role in antimicrobial defence against extracellular pathogens. IL-17 triggers epithelial release of chemoattractants CXCL8 (interleukin-8) and chemokine C-C motif ligand 2 (CCL2) which enhance neutrophil chemotaxis. IL-17 has further pro-inflammatory and airway remodelling effects including myofibroblast differentiation and airway smooth muscle migration, mucin expression and mucus hypersecretion, and induction of proinflammatory cascades through Nuclear Factor kappa-light-chain-enhancer of activated B cells (NF-κβ) and mitogen-activated protein kinase (MAPK) activation.⁷⁴ Sputum IL-17A and interleukin-8 (IL-8) mRNA levels have been found to be significantly elevated in asthma patients and correlated with sputum neutrophil counts.⁷⁵

Trials of anti-IL-17 biologics have yielded disappointing results. Brodalumab, a human anti-IL-17 receptor A antibody, did not result in any significant treatment effects on Asthma Control Questionnaire (ACQ) change, lung function, or symptom scores.⁷⁶ Secukinumab was investigated in a Phase II RCT, however no improvements in ACQ were found and the study was discontinued by the sponsor.⁷⁷

CXC Chemokine Receptor 2 (CXCR2)

CXC chemokine receptor 2 (CXCR2) is a receptor for IL-8 which plays a key role in neutrophil chemotaxis and regulation of neutrophil release from bone marrow.⁷⁸ It has also been shown to promote angiogenesis through recruitment of endothelial progenitor cells independent of the effect of vascular endothelial growth factor (VEGF).⁷⁹ A phase study found that a CXCR2 antagonist, SCH527123, caused a significant reduction in sputum neutrophils compared to placebo.⁸⁰ However a subsequent RCT of a selective CXCR2 antagonist in patients with a blood neutrophilia, failed to show any significant reduction in the frequency of severe exacerbations despite causing a reduction in sputum and blood neutrophils.⁸¹

Tumour Necrosis Factor-Alpha (TNF-α)

Tumour necrosis factor-alpha (TNF-α) has been implicated in the pathophysiology of multiple inflammatory conditions including rheumatoid arthritis and inflammatory bowel disease. It plays a key role in innate immune response and is primarily released by macrophages after activation of pattern recognition receptors by pathogen associated molecular patterns. TNF-α receptor activation leads to phosphorylation of Iκβα and activation of NF-κβ, which leads to downstream transcription of proinflammatory genes including IL-1β, interleukin-6 (IL-6), IL-8 and TNF-α.⁸² It is also known to increase the expression of adhesion molecules by epithelial cells, such as vascular cell adhesion molecule 1 (VCAM-1) and intercellular adhesion molecule 1, which further enhance chemoattraction of inflammatory cells such as eosinophils and neutrophils to the airway.

TNF-α has been observed to be increased in bronchoalveolar lavage from atopic asthmatics.⁸³ Exposure of normal subjects to TNF-α has also been shown to cause airway hyperresponsiveness and airway neutrophilia.⁸⁴ The effect of TNF-α was evaluated in an RCT of etanercept, however no significant differences were observed between etanercept and placebo for change in baseline FEV1, asthma exacerbations, change in ACQ or provocation concentration 20% (PC20).⁸⁵ Another TNF-α monoclonal antibody, golimumab, did not significantly decrease exacerbations or improve FEV1. An unfavourable risk-benefit profile was also observed, with 30% of patients treated with golimumab registering serious adverse events including serious infections and malignancy.⁸⁶

Interleukin-23 (IL-23)

Interleukin-23 (IL-23) is a heterodimeric cytokine composed of an interleukin-12B subunit and an interleukin-23A (IL-23A) subunit. In animal models, IL-23 has been shown to stimulate Th17 CD4 cell proliferation, promote IL-17 production and stimulate neutrophil recruitment and chemotaxis.⁸⁷ They have also been shown to upregulate Th2 cell-mediated eosinophilic airway inflammation independent of the Th17 axis.⁸⁸ Serum IL-23 has been observed to be high in children with asthma and associated with impaired lung function.⁸⁹ A phase 2a RCT of risankizumab, an anti-IL-23p19 monoclonal antibody, did not demonstrate clinical efficacy.⁹⁰ The time to first asthma worsening was shorter and annualised rate of asthma worsening was worse in the risankizumab group.

Markers of Airway Microbial Dysbiosis – 16s rRNA, Sputum PCR

The relationship between the host and its microbiome has been well characterised in diseases such as those affecting the gastrointestinal tract, but its role in airway disorders like asthma is gaining increasing attention.⁹¹ Research is now focusing on how both the airway and gut microbiomes influence immune maturation in early life, as disturbances in these microbial communities are thought to disrupt the balance between protective immunity against pathogens and the development of immune tolerance.⁹²

The development of culture-free methods based on the high-throughput sequencing of genetic material of microbes or identification of antibodies has allowed more sensitive detection of microbial communities rather than a single pathogenic organism.⁹¹ The most common technique is amplicon sequencing of the 16s ribosomal RNA (rRNA) of the microbiome. Studies that have utilised 16S rRNA sequencing have been concordant in finding that Proteobacteria phylum and Haemophilus species are often present in upper and lower airways of asthma patients.^93,94

The potential for treatments to alter lung and gut microbiome in asthma is currently being explored. There have been several studies of the effects of antibiotics in patients with asthma, including the AMAZES trial of azithromycin which found a significant reduction in asthma exacerbations and improved quality of life in the treatment group.⁹⁵ 16S rRNA sequencing and quantitative PCR analysis were performed to assess the effect of azithromycin on sputum microbiology and found that although total bacterial load did not decrease, the load of Haemophilus influenzae was significantly decreased.⁹⁶ Five macrolide resistance genes were also significantly increased highlighting the potential unintended consequence of antibiotics-resistance. The role of probiotics and prebiotics to modify the gut microbiota in the treatment of asthma has also shown some promise in some preliminary studies, but needs further evaluation.^97,98

Biomarkers to Guide Treatment Selection

Randomized controlled trials have provided important insights into the role of biomarkers in guiding treatment selection for asthma. Key evidence from these trials is summarized in Table 1.

Table 1 Use of Biomarker Stratification in Phase II and III Randomised Controlled Trials (RCTs)

Inhaled Corticosteroids and Traditional Controller Therapies

Titration of inhaled corticosteroids in treatment algorithms has conventionally been guided by subjective measures of symptoms and physiological measures such as peak flow. However, there are a significant proportion of patients who experience uncontrolled asthma despite maximal inhaled corticosteroid therapy.

Sputum eosinophils have been recognised as an important biomarker in being able to predict response to corticosteroids. In a landmark RCT, Green et al demonstrated that a sputum eosinophil-guided strategy to adjust corticosteroid treatment was significantly more effective in reducing exacerbations than usual guideline-based management.¹⁶ A systematic review comparing sputum eosinophil-guided management versus symptom-based management also found a significant reduction in exacerbation risk with titration to normalise sputum eosinophils, with a number needed to treat of 6 patients to avoid one exacerbation over 16 months.¹⁷

Due to the practical challenges in obtaining and analysing induced sputum, the use of FeNO in titrating inhaled corticosteroids has also been evaluated. A Cochrane review evaluating the efficacy of FeNO-guided asthma treatment compared to standard care found that across seven studies there was a significantly lower risk of having one or more exacerbations in the FeNO group compared to control (OR 0.60, 95% CI 0.43–0.84).¹¹⁵ A FeNO suppression test has also been to be a useful tool to differentiate nonadherence to inhaled corticosteroid versus patients who are genuinely non-responsive to inhaled corticosteroid.¹¹⁶ Global Initiative for Asthma (GINA) guidelines currently recommend for FeNO to be used in the context of titration of inhaled corticosteroids.³

Biologic Selection Based on Biomarkers: RCT Evidence

Anti-IgE (Omalizumab)

Immunoglobulin E (IgE) has been well recognised to play a role in atopic diseases including asthma, allergic rhinitis and atopic dermatitis. After exposure to allergens in immediate hypersensitivity reactions, Th2 cells are activated which stimulate B cells to class switch and produce allergen-specific IgE. IgE then binds to mast cells and basophils through high-affinity IgE receptors (FcϵRI), and on re-exposure to allergen, the allergen cross-links the bound IgE and triggers degranulation of mast cells to release histamine, prostaglandins, leukotrienes and cytokines.¹¹⁷ This drives smooth muscle contraction, mucosal oedema and mucus hypersecretion. Total serum IgE levels have been shown to correlate with asthma severity, worse lung function, risk of admission and inhaled steroid use.¹¹⁸ Studies have also shown an association between sensitisation in early life and lung function impairment in adulthood in asthmatics.^119,120

Omalizumab, a monoclonal antibody which prevents free IgE from binding to FcϵRI receptors, was the first biologic approved for use for asthma. The INNOVATE study showed a significant reduction in exacerbations (RR 26%, p=0.04), however response was not proportional to IgE level.¹¹¹ Similarly the EXTRA study found a significant treatment effect of omalizumab in reduction in exacerbations, but did not identify serum IgE to be a significant predictor of response.³⁶ It did find however that patients stratified by high FeNO, blood eosinophils and serum periostin experienced more marked exacerbation reduction. Interestingly, a recent one-year open-label trial (SOMOSA study) found that standard biomarkers (blood and sputum eosinophils, FeNO and serum IgE) did not predict response, however, omic biomarker scores comprised of (i) five exhaled volatile organic compounds and (ii) five plasma lipid biomarkers strongly predicted a reduction in exacerbations of more than 50% with a ROC AUC of 0.780 and 0.922, respectively.⁹⁹ Although total serum IgE does not appear to predict response, the pretreatment level of IgE and body weight is used to determine dose and frequency of dosing, as it reflects the amount of circulating IgE that needs to be neutralised effectively.

Anti–IL-5/IL-5R (Mepolizumab, Reslizumab, Benralizumab)

Interleukin-5 (IL-5) exerts a key function in the pathogenesis of eosinophilic asthma through its pleiotropic effects on eosinophil differentiation, growth, activation and recruitment to the airways. It is produced by several cells including Th2 lymphocytes, type 2 innate lymphoid cells (ILC2), mast cells and eosinophils themselves. Release of IL-5 is dependent on stimulation of key transcription factors such as STAT6 and GATA binding protein 3 (GATA3) by IL-4, and from ILC2s through induction by alarmins such as TSLP, interleukin-25 (IL-25) and interleukin-33 (IL-33) in response to inhaled allergens and pathogens. IL-5 binds to IL-5 receptor (IL-5R), driving janus kinase 2 (JAK2) and STAT 1, 3 and 5 which stimulate transcription of genes involved in eosinophil proliferation.¹²¹ In addition, via activation of the Extracellular signal-regulated kinase (ERK) subfamily of MAPK, several genes involved in eosinophil maturation, eosinophil proliferation are stimulated, as well as production of eosinophil chemoattractants such as leukotriene C4.

The first major mepolizumab trial represented a significant turning point in precision medicine and became a landmark case for the importance of biomarker-guided stratification. A first proof-of-concept study evaluating the effect of mepolizumab found that although it reduced sputum and blood eosinophils effectively, it did not improve clinical outcomes including lung function, symptoms or daily beta-agonist use.²⁶ The outcome of this trial almost led to the abandonment in IL-5 as a therapeutic target. However subsequent DREAM and MENSA studies showed significantly reduced annualised exacerbation risk in patients with a raised blood eosinophil count. Higher blood eosinophil counts were associated with greater reductions in exacerbations.¹⁰⁷ This later led to approval from the US Food and Drug Administration and European Medicine Agency as an add-on treatment for patients with severe eosinophilic asthma.

Reslizumab is another monoclonal antibody that targets IL-5, and was also found in two pooled phase III studies in patients with a blood eosinophil count of 400 cells/μL or higher to significantly reduce frequency of asthma exacerbations.¹²² Since reslizumab requires intravenous administration, this has prevented widespread use given the availability of subcutaneous formulations of anti-IL-5/anti-IL-5R biologics.

Benralizumab is an antibody that specifically binds to the alpha subunit of IL-5R expressed on eosinophils and basophils. It also binds the Fc-gamma receptor 3a on natural killer (NK) cells, further causing eosinophil depletion by cell-mediated apoptosis by NK cells. As it targets eosinophils independent of cytokine involvement, it causes almost complete depletion of blood and airway eosinophils compared to anti-IL-5 antibodies. Across Phase 3 trials SIROCCO and CALIMA, benralizumab has been shown to significantly reduce exacerbation rates compared with placebo.^108,109 Although there have been no head to head studies evaluating mepolizumab versus benralizumab in asthma, the MANDARA study in eosinophilic granulomatosis with polyangiitis showed greater depletion of eosinophils with benralizumab compared to mepolizumab, and although it was non-inferior in induction of remission, it demonstrated greater oral corticosteroid sparing effect.¹²³

Anti-IL-4Rα (Dupilumab)

Interleukin-4 (IL-4) receptors are made up of two heterodimeric complexes: type 1 and 2 receptors. Type 1 receptors consist of IL-4Rα and ϒ chain subunits, and type 2 receptors are receptor complexes consisting of IL-4Rα and interleukin-13 Receptor Alpha 1 (IL-13Rα1) subunits. Dupilumab binds to both receptors, thereby blocking downstream signalling of both IL-4 and IL-13. IL-4 promotes Th2 differentiation and proliferation, B cell expansion, class switching to IgE and eosinophil trafficking. IL-13 in addition also stimulates goblet cell hyperplasia and promotes smooth muscle hypertrophy.¹²⁴ Due to the overlapping effects of IL-4 and IL-13, blockade with dupilumab of both pathways is required for complete inhibition of T2 signalling cascades.

The LIBERTY ASTHMA QUEST trial in over 1900 patients showed a significantly higher reduction in annualised rate of severe exacerbations in patients compared to placebo (RR 0.46 vs 0.87).³⁷ When patients were stratified by a baseline blood eosinophil count of 300 cells/μL the relative risk of exacerbations decreased further (RR 0.37, 95% CI 0.29–0.48) versus unselected (RR 0.46, 95% CI 0.39–0.53). A post-hoc analysis showed that FENO was also able to predict response to dupilumab, with the relative risk of severe exacerbations 22.7%, 58.3% and 69.3% lower for dupilumab with FeNO<25ppb, FeNO 25–50ppb, and FeNO≥50ppb subgroups, respectively. The ability of FeNO to stratify risk was found to be independent of blood eosinophil levels.¹²⁵ This may reflect that FeNO is mainly generated by airway epithelial cells through iNOS, which is upregulated by IL-13 via STAT6 signalling, whereas eosinophil differentiation and recruitment are mainly driven by IL-5.

Anti-TSLP (Tezepelumab)

Thymic stromal lymphopoietin (TSLP) is part of the alarmin family, a group of epithelial cytokines which are released by airway epithelial cells in response to exposure to exogenous triggers such as pathogens, allergens and chemical irritants. TSLP drives innate and adaptive inflammatory cascades by activating multiple cell lineages including dendritic cells, T cells, ILC2 cells, eosinophils and mast cells. TSLP results in dendritic cell maturation and upregulation of OX40-ligand, which facilitates dendritic cell priming of naïve CD4+ve T cells to differentiate to Th2 cells.¹²⁶ ILC2s are also recognised to be directly activated by TSLP to produce type 2 cytokines that drive eosinophil differentiation and trafficking, such as IL-5 and IL-13.¹²⁷ Therefore, it was hypothesised that blockade of TSLP could simultaneously inhibit multiple allergic cascades upstream of type 2 cytokines and could therefore be an attractive therapeutic target in asthma.

Tezepelumab is an IgG2 monoclonal antibody that prevents binding of TSLP to its receptor. Three large phase III RCTs have demonstrated the clinical efficacy of Tezepelumab in both T2-high and T2-low patient groups. The PATHWAY trial in 550 patients showed Tezepelumab significant reduced annualised exacerbation rates compared to placebo by 71% (p<0.0001).¹²⁸ Exploratory analyses showed significant reduction of T2 biomarkers including blood eosinophils, FeNO, total IgE, periostin, serum TSLP, serum IL-5 and serum IL-13, though exacerbations were reduced irrespective of baseline biomarkers.¹¹⁴ The NAVIGATOR trial also found a significant reduction in annualised exacerbation rates across all baseline blood eosinophil levels.¹¹⁰

Mechanistic studies investigating the effect of tezepelumab on airway remodelling has also found reduction in mucus plug scores and airway eosinophilic infiltration.^67,105 Although biomarkers such as blood eosinophils can indicate a greater magnitude of response, currently there are no biomarkers that can predict TSLP response. TSLP appears to be effective in both T2-high and T2-low patient cohorts by acting upstream of both inflammatory cascades.

Emerging Therapeutic Targets and Associated Biomarkers

Emerging therapeutic targets and their associated biomarkers are expanding the landscape of precision medicine in asthma, with the potential to address unmet needs in both T2- and non–T2-driven disease. These advances highlight novel avenues for individualised therapy, as summarised in Figure 2 (T2 targets and therapies) and Figure 3 (non–T2 targets and therapies).

Figure 2 Current and emerging T2 targets for biologic therapy. Created in BioRender. Horn, N. (2025) https://BioRender.com/cuol3fr. Abbreviations: APC, antigen presenting cell; BTK, Bruton’s tyrosine kinase; IL, interleukin; IL-5Rα, interleukin-5 receptor alpha subunit; ST2, suppression of tumorigenicity 2 receptor; TSLP, thymic stromal lymphopoietin; TSLPR, thymic stromal lymphopoietin receptor, ILC2, type 2 innate lymphoid cells.

Figure 3 Current and emerging beyond T2 targets for biologic therapy. Created in BioRender. Horn, N. (2025) https://BioRender.com/2qaj2ar. Abbreviations: APC, antigen presenting cell; BTK, Bruton’s tyrosine kinase; FLAP, 5-lipoxygenase-activating protein; IL, interleukin; JAK, Janus kinase; KIT, KIT proto-oncogene receptor tyrosine kinase; LIGHT, lymphotoxin-like, exhibits inducible expression, and competes with herpesvirus glycoprotein D for herpesvirus entry mediator; OM-85 BV, Broncho-vaxom, bacterial strain lysate; OX-40L, OX40 ligand; ST2, suppression of tumorigenicity 2 receptor; TH1, T-helper type 1 cell; Th0, naïve T-cell; TKI, tyrosine kinase inhibitor; TSLP, thymic stromal lymphopoietin; TSLPR, thymic stromal lymphopoietin receptor.

5.1 – Emerging T2 Targets

Anti-Interleukin-33 (Anti-IL-33)

Interleukin-33 (IL-33) is an IL-1 family member which binds to a heterodimeric receptor composed of ST2 receptor (IL-1 receptor-like 1) and its coreceptor IL-1 receptor accessory protein. IL-33, similarly to TSLP, acts as an alarm signal in response to environmental insults and plays a key role in regulating innate and adaptive immune responses. IL-33 recruits myeloid differentiation primary response gene 88 (Myd88) into the receptor complex, which activates downstream signalling via interleukin-1 receptor-associated kinase 4 (IRAK4) including c-Jun N-terminal kinases (JNK), NF-κβ, and Janus kinase pathway 2 (JAK2).¹²⁹ Activation of ILC2 cells leads to promotion of pro-inflammatory cytokines including IL-4, IL-5 and IL-13 which polarise T cells to Th2 differentiation. IL-33 also affects structural cells, including proliferation of airway smooth muscle cells, inducing collagen deposition by airway fibroblasts and stimulating goblet cell hyperplasia.

Several biologics targeting IL-33 are currently in development, including astegolimab (an ST2-inhibitor), itepekimab and etokimab (anti-IL-33 antibodies). The ZENYATTA study found astegolimab was associated with a reduction in annualised exacerbation rates and was effective across a broad population including eosinophil low strata.¹³⁰ Currently astegolimab has not progressed into phase III studies in asthma with the main focus of RCTs now within COPD.

A phase II proof of concept RCT in 296 patients of itepekimab showed that endpoints of improvement in loss of asthma control and significant reduction in blood eosinophils were met compared to placebo, however when combined with dupilumab there was no increased treatment benefit compared to dupilumab monotherapy.¹³¹

Anti-Interleukin-25 (Anti-IL-25)

Interleukin-25 (IL-25), otherwise known as IL-17E, is another alarmin which is part of the IL-17 family. After binding to IL-25 receptor, IL-25 forms a complex with CD3 and CD28 and induces recruitment of Act1, leading to upregulation of NF-κβ, MAPK, GATA3 and STAT5 and the transcription of multiple cytokines including T2 cytokines, CXCL10 and IL-6.¹³²

A phase II RCT was undertaken to evaluate brodalumab, an anti-IL-17RA monoclonal antibody, however this did not demonstrate any clinical benefit in ACQ scores, lung function improvement, or symptom frequency other than in a subgroup of patients with high bronchodilator reversibility.⁷⁶ Other anti-IL-25 therapies are still in preclinical or early clinical phases. As the majority of evidence for IL-25 comes from animal models and in vitro studies, and due to the overlap with other biologics such as anti-TSLP which has already demonstrated broad anti-inflammatory effects and greater clinical impact, IL-25 currently remains an interesting but unproven target in asthma.

Siglec-8

Siglecs (sialic acid-binding immunoglobulin-like lectins), cell surface proteins that primarily bind to sialic acid, are known to play an inhibitory function in regulation of immune responses. Siglec-8 cross-linking has been shown to trigger apoptosis of eosinophils, through production of reactive oxygen species, mitochondrial dysfunction and cleavage of caspase-3, 8 and 9, leading to cell death.¹³³ Engagement of Siglec-8 with anti-Siglec-8 monoclonal antibody in a mouse model has also been shown to inhibit high-affinity Fc receptor for immunoglobulin E (FcεRI)-dependent mast cell degranulation and IL-33 dependent mast cell activation and neutrophil influx and cytokine production.¹³⁴ Gene expression profiling in asthmatics also found that Siglec-8 was highly expressed on eosinophils and mast cells from sputum. A Siglec-8 monoclonal antibody significantly depleted sputum eosinophils and inhibited FcεRI-activated mast cells in lung tissue.¹³⁵

Lirentelimab (AK002) is a humanised anti-Siglec-8 monoclonal antibody which binds with high affinity to CD16a, an Fc-receptor that mediates the antibody-dependent cell-mediate toxicity of Natural Killer cells against eosinophils. Lirentelimab has yet to be evaluated in human studies in asthma, but has been trialled in other eosinophil and mast cell-driven conditions such as chronic urticaria, allergic conjunctivitis, and eosinophilic gastritis.¹³⁶

OX40 Ligand (OX40L)

OX40 Ligand (OX40L), also known as CD252, is a member of the tumour necrosis factor (TNF) receptor family. The OX40 receptor and OX40L are primarily expressed by antigen-presenting cells following activation by TSLP. OX40L acts as a costimulatory molecule that promotes Th2 polarisation of naïve CD4+ T cells to produce IL-4, IL-5, IL-13 and TNF-α. In mouse models, ovoalbumin-challenged OX40 knockout mice demonstrated significantly abrogated T2 response, with decreased eosinophilia, goblet cell hyperplasia, mucus secretion and attenuated airway hyperresponsiveness compared to wild-type mice.¹³⁷ OX40 and OX40L were also shown to be more highly expressed in the bronchial mucosa in mild asthmatics compared to healthy controls.¹³⁸

A phase II RCT of a humanised anti-OX40L monoclonal antibody in 28 mild, atopic asthmatics using an allergen challenge model found that despite significant reduction in total IgE and sputum eosinophils, there was no effect on airway hyperresponsiveness.¹³⁹ Amlitelimab is another anti-OX40L monoclonal antibody that is currently in a phase II RCT in moderate-severe asthmatics (TIDE-Asthma).

Combined Biologics

Although treatment with monoclonal antibodies targeting single pathways have demonstrated efficacy, there is still a significant proportion of severe asthmatics who have suboptimal response.¹⁴⁰ Several mechanisms have been proposed, including the impact of dose and route of administration on drug pharmacokinetics, inadequate patient endotyping, the development of anti-drug antibodies, and the presence of overlapping inflammatory pathways. In the latter scenario, the use of combined biologic therapy has been suggested as a potential strategy.

Beyond a combined itepekimab/dupilumab phase II study which did not find additional benefit combining the two versus singular treatment with dupilumab,¹³¹ there have been no other RCTs evaluating two approved biologics in combination. Most evidence is from clinical case series from real-world use of pairing biologics which have generally shown improved asthma control and few serious adverse effects.¹⁴¹ Another approach that is being evaluated is bispecific monoclonal antibodies, which are molecules designed to target two epitopes. The first combined biologic to be trialled in asthma is lunsekimig, a bispecific nanobody targeting both IL-13 and TSLP. A Phase 1b study in 36 participants with mild-moderate asthma that were enriched for FeNO≥25ppb, showed that lunsekimig potently suppressed FeNO and reduced blood eosinophils, serum IL-5, eotaxin and serum IgE.¹⁴² Lunsekimig is currently being evaluated in a phase IIb dose ranging study.¹⁴³

Oral Anti-T2 Agents

Dexpramipexole is an oral agent which was previously developed for the treatment of amyotrophic lateral sclerosis (ALS). It is an enantiomer of an antiparkinsonian drug pramipexole, however it differs in that at clinically used doses, it has virtually no anti-dopaminergic activity. Unexpectedly during trials for ALS, it was found to significantly lower blood eosinophil counts.¹⁴⁴ The mechanism of action is incompletely understood, however it is thought to inhibit bone marrow maturation and release of eosinophils.¹⁴⁵ In a phase 2 trial investigating the effects of dexpramipexole in asthma patients enriched for blood eosinophilia, dexpramipexole resulted in a significantly reduced blood eosinophil count at both low and high dose strata. It was also found to significantly reduce an exploratory biomarker, nasal eosinophil peroxidase.⁵¹ Dexpramipexole is currently in phase III clinical trials.

Bruton’s tyrosine kinase (BTK) inhibitors are another class of oral medications currently under investigation in asthma and other allergic diseases. BTK is primarily expressed by B cells and plays a key role in B cell development, activation and survival, dysregulation of which can lead to B-cell malignancies.¹⁴⁶ BTK inhibitors were first developed for the treatment of B cell malignancies. However, BTK is also a key enzyme that is essential for signalling through the FcεRI pathway in human mast cells and basophils following antigen binding to IgE.¹⁴⁷ In vitro studies from BTK knockout mice bone marrow-derived mast cells have shown that stem cell factor mediated degranulation was significantly reduced, as well as production of pro-inflammatory cytokines including IL-13, IL-6 and TNFα.¹⁴⁸

The application of BTK inhibitors in treating allergic diseases is therefore currently being investigated. A phase 2 proof-of-concept trial of rilzabrutinib found a significant reduction in loss of asthma control events, mainly driven by reduction in SABA use.¹⁴⁹ In exploratory RNA sequencing analysis of nasal brushings from participants administered rilzabrutinib, there was significant downregulation of gene expression of B-cell activating factor, IL-33, IL-5 and GCSF.¹⁵⁰

Lastly, inhibition of signal transducers and activators of transcription (STATs) including Signal Transducer and Activator of Transcription 6 (STAT6) are another potential target for oral therapy in asthma. Binding of IL-4 and IL-13 to IL-4Rα leads to phosphorylation of its tyrosine residues by janus kinase 1 (JAK1) and STAT6 activation. This leads to Th2 polarisation via GATA3 and prostaglandin D2 receptor 2 (CRTH2) signalling. STAT6 signalling can induce other airway epithelial genes including mucin 5AC (MUC5AC) which stimulate goblet cell hyperplasia, and eotaxins which stimulate eosinophil chemotaxis.¹⁵¹ Ovalbumin allergen challenge in STAT6-/- mice did not induce eosinophilia, airway hyperresponsiveness (AHR) and mucus hypersecretion compared to wild-type BALB/c mice.¹⁵² Evidence for STAT6 degraders as a potential therapeutic remain in pre-clinical stages currently. KT0621 has been reported in vitro in human primary immune and tissue cells, and in vivo in intranasal house dust mite (HDM) allergen challenge in mice to potently degrade STAT6 and block downstream IL-4/IL-13 as measured by BAL inflammatory infiltrate and cytokine levels, with efficacy comparable to dupilumab.¹⁵³

Ultra Long-Acting Biologics

Ultra long-acting biologics represent a new area of treatment in asthma that could potentially provide several advantages over current biologics. Existing biologics require administration every four to eight weeks; ultra long-acting biologics have a reduced dosing frequency of 8–24 weeks which could improve convenience for patients, improve adherence, and lower healthcare utilisation.¹⁵⁴ Depemokimab is an anti-IL-5 monoclonal antibody with half the clearance rate of mepolizumab and roughly 29-fold greater potency, allowing 6-monthly dosing. Recent phase IIIa RCTs in 792 patients with severe asthma and an eosinophilic phenotype showed a significant reduction in annualised rate of exacerbation at 1 year, with a rate ratio (RR) of 0.46 with depemokimab versus placebo (95% CI 0.36–0.59).¹⁵⁵ There was no clear relationship between the baseline blood eosinophil count and efficacy of depemokimab, unlike mepolizumab. Other long-acting biologics are currently in clinical development, including verekitug, a novel long-acting anti-TSLP monoclonal antibody.¹⁵⁶

5.2 – Emerging Beyond T2 Targets

Janus Kinase Inhibitors (JAK Inhibitors)

The Janus kinase–signal transducer and activator of transcription (JAK-STAT) pathway is understood to play a critical role in both T2 high and T2 low inflammatory cascades. JAK signalling is critical in differentiation of CD4+ T cells towards a Th1 or Th2 phenotype through IL-12 and IL-4, respectively. Th2 polarisation is stimulated through activation of STAT5 and STAT6 by interleukin-2 (IL-2) and IL-4 binding to receptors coupled to JAK1 and janus kinase 3 (JAK3). Conversely, signalling through an IL-12 receptor leads to Th1 polarisation through activation of JAK2 and STAT4, inducing transcription of IFN-ϒ.¹⁵⁷ Furthermore, JAK-STAT signalling is also recognised to be important in IgE regulation. Binding of IL-4 to ligand-specific IL-4Rα initiates downstream activation of JAK1 and JAK3, which allows docking of phosphorylated STAT6. STAT6 activates transcription of genes involved in B cell differentiation and class switching to produce IgE.¹⁵⁸

Inhaled formulations are under investigation to minimise risks of systemic exposure such as serious infections, venous thromboembolism, cardiovascular events and malignancy, while maximising target organ deposition. A phase 1 study of AZD4604, an inhaled JAK inhibitor, showed approximately 50% reduction in FeNO, but only transient minimal suppression of systemic target engagement as measured by IL-4 induced STAT6 phosphorylation in peripheral CD3+ T cells.¹⁵⁹ Phase II clinical trials are currently underway for AZD4604.

Tyrosine Kinase Inhibitors (TKI)

Originally used in cancer therapy, tyrosine kinase inhibitors (TKI) such as imatinib have also shown potential in treating asthma. Tyrosine kinase signalling is understood to be important in the pathogenesis of airway inflammation and remodelling. Epidermal growth factor receptors (EGFR) and platelet derived growth factor receptors (PDGFR) are transmembrane receptor tyrosine kinases that upon activation, recruit and activate downstream molecules such as src homology 2 (SH2) and phosphoinositide 3-kinase (PI3K) that result in airway smooth muscle proliferation and mucus secretion.¹⁶⁰ Tyrosine kinase is also involved in inducing mast cell degranulation through the FcεRI pathway. Non-receptor tyrosine kinases such as the Janus kinases (JAK1, JAK2 and JAK3) lead to tyrosine phosphorylation of STAT transcription factors including STAT6, which is critical in Th2 polarisation and Th2 cytokine expression.

Imatinib is a KIT/PDGFR inhibitor that has been evaluated in a 24-week RCT in 62 severe asthmatics. At 6 months, there was a significant decrease in airway hyperresponsiveness.¹⁶¹ Potential biomarkers for efficacy were also assessed including serum tryptase and airway mast cell counts which were both significantly reduced by imatinib. Masitinib is another tyrosine kinase inhibitor that targets the c-Kit receptor, and this was shown in a 419 participant RCT to decrease exacerbations by 35% relative to placebo.¹⁶² Both imatinib and masitinib demonstrated acceptable safety profiles, with no evidence for increased risk of infection compared to placebo.

5-Lipoxygenase-Activating Protein (FLAP) Inhibitors

5-lipoxygenase-activating protein (FLAP) inhibitors are another category of novel oral medications being assessed in clinical trials, which work by inhibiting 5-lipoxygenase-activating protein (FLAP). Unlike montelukast which blocks the CysLT1 receptor, FLAP is crucial in the synthesis of leukotrienes through facilitating the interaction of 5-lipoxygenase with arachidonic acid, enabling conversion of leukotriene A₄ (LTA₄) to leukotriene B₄ (LTB₄) or to leukotriene E₄ (LTE₄) by the action of different peptidases.¹⁶³ LTB₄ is a potent chemoattractant for neutrophils.¹⁶⁴ LTE₄ is a potent bronchoconstrictor, mediates migration of inflammatory cells into the airway and increases vascular permeability. In addition, FLAP inhibitors have a potential non-invasive biomarker in the form of urinary LTE₄.¹⁶⁵

FLAP inhibitor GSK2190915 was shown in an inhaled allergen challenge model to attenuate both early and late asthmatic FEV1 decrease compared to placebo, and there was a significant reduction in sputum eosinophil count.¹⁶⁶ There was no effect on methacholine challenge response. The effect of the same FLAP inhibitor was later evaluated in patients with elevated sputum neutrophils. This however showed that GSK2190915 had no short term effect on sputum neutrophil count or clinical endpoints including FEV1 or ACQ after 13–16 days treatment, though sputum LTB₄ and urinary LTE₄ levels were potently suppressed.¹⁶⁷ A phase IIa RCT is currently ongoing for Atuliflapon, an orally administered flap inhibitor, in moderate to severe asthma.¹⁶⁸

Interleukin-1-Beta (IL-1β)

There is growing evidence to implicate NLRP3 (nucleotide-binding domain, leucine-rich-containing family, pyrin domain-containing-3) inflammasome complexes and IL-1β signalling in the pathogenesis of neutrophilic, steroid resistant asthma.¹⁶⁹ NLRP3 is an intracellular pattern recognition receptor (PRR) that recognises pathogen and damage associated molecular patterns (PAMPs and DAMPs). Priming by PAMPs and DAMPs results in increased levels of inactive NLRP3 and pro-IL-1β secondary to upregulated NF-κβ transcription. On further exposure, ROS release then results in activation of NLRP3 inflammasome, which produces caspase-1 that cleaves pro-IL-1β to its active pro-inflammatory form, IL-1β. IL-1β has been demonstrated to mediate lung neutrophilia, IL-33 and MUC5AC expression in mouse models of viral-induced asthma.¹⁷⁰ IL-1β has also been shown to exert effects on airway smooth muscle by activating NF-κβ and Cyclooxygenase-2 (COX-2), preventing relaxation of smooth muscle and promoting bronchoconstriction.¹⁷¹ IL-1β and TNF-α/NF-κβ genes have been found to be overexpressed in patients with a neutrophilic phenotype, and higher levels of sputum NLRP3 and caspase-1 have also been found in patients with neutrophilic asthma.^172,173

Therapeutic strategies targeting IL-1β are largely still in pre-clinical or early clinical phase. Anakinra, an IL-1 receptor antagonist with indications for treatment of rheumatoid arthritis, familial Mediterranean fever and Still’s disease, was investigated in an inhaled endotoxin (LPS) human challenge model. Anakinra pre-treatment significantly decreased airway neutrophilia versus placebo in 17 healthy volunteers, and significantly attenuated airway IL-6, IL-8 and IL-1β levels.¹⁷⁴ NLRP3 inhibitors are currently under evaluation in a number of clinical trials in other inflammation-driven pathologies including COVID-19, cardiovascular disease, rheumatoid arthritis and malignancy,¹⁷⁵ however currently there are no active human clinical trials in asthma.

LIGHT (TNFSF14) Pathway

Members of the tumour necrosis factor (TNF) family have been extensively studied in the pathogenesis of asthma, most prominently TNF-α, which has been shown to promote airway remodelling.¹⁷⁶ LIGHT (TNFSF14) is a TNF family cytokine that binds to herpesvirus entry mediator (TNFRSF14) and lymphotoxin-β receptor (LTβR). LTβR engagement initiates NFκβ-dependent gene transcription of pro-inflammatory cytokines.¹⁷⁷ LIGHT deficient mice have been shown to have impaired fibrosis and smooth muscle proliferation, as well as reduced IL-13 and transforming growth factor-beta (TGF-β) expression,¹⁷⁸ and LIGHT blockade was also demonstrated to decrease the differentiation of CD4+ T cells into Th1, Th2 and Th17 cells and reduce expression of NF-κβ in a mouse OVA-challenge model of chronic allergic asthma.¹⁷⁹ Higher sputum levels of LIGHT have also been found to be associated with persistent airflow limitation in asthma patients.¹⁸⁰

Quisovalimab (AVTX-002), an anti-LIGHT monoclonal antibody, was evaluated in phase 2 trial in patients with poorly controlled non-eosinophilic asthma. Although it did not meet its primary endpoint of a reduction in asthma-related events, it demonstrated a trend towards reduced events in a group with raised baseline serum LIGHT levels.¹⁸¹ This suggests that selection of patient strata using LIGHT as a biomarker might identify those that might benefit from treatment targeting this pathway.

Bacterial Lysates

Bacterial lysates, inactivated mechanically fractionated particulates of bacterial extracts, have historically been used for the prevention of current respiratory tract infections in children. The rationale for their use has been to stimulate the innate immune system response via antigen presentation to mucosal dendritic cells, which then generate antigen-specific T cell responses that modulate B cell isotype switching to plasma cell and IgA production.¹⁸² The hygiene hypothesis suggests that reduced exposure to infections in early childhood can lead to the developmental of allergic sensitisation,¹⁸³ therefore there has been increasing interest in the application of bacterial strain lysates as potential immunomodulators in asthma.

In a murine OVA-challenge model, orally administrated Broncho-Vaxom, a commercially available bacterial lysate, suppressed airway inflammation through IL-10 and increased the number of FoxP3+ T regulatory cells in the trachea, inhibiting AHR, mucus hypersecretion and airway eosinophilia, and production of antigen-specific IgE.¹⁸⁴ Bacterial lysates have also been shown to suppress Th2 inflammation with significant abrogation in BAL eosinophilia, expression of IL-13 and IL-5 mRNA, and ILC2-driven IL-33 expression with direct local administration of bacterial lysate, OM-85, in mice sensitised and challenged with ovoalbumin and Alternaria.¹⁸⁵ A systematic review and meta-analysis showed a significant decrease in both wheezing episodes and asthma exacerbations in children.¹⁸⁶ Bacterial lysates are potentially promising as adjunctive immunomodulatory agents in early asthma by enhancing mucosal immunity and suppressing Th2 dominance.

Triglyceride-Lowering Therapy

The link between obesity and asthma is well established. Obese patients have a higher risk of asthma, and obese asthmatics have been shown to have a higher burden of symptoms, exacerbations and reduced response to steroids.¹⁸⁷ Epidemiological studies have shown an association of metabolic dysfunction with asthma risk, including with high serum triglycerides and LDL cholesterol,^188,189 and furthermore that serum triglycerides are linearly related to FeNO,¹⁹⁰ and that metabolic syndrome is associated with lung function impairment.¹⁹¹ The proposed mechanistic pathways include direct activation of PRRs by free fatty acids and downstream activation of NF-κβ and transcription of pro-inflammatory cytokines such as TNF-α and IL-6, and of NLRP-3 inflammasome-mediated inflammation.¹⁹² Leptin expression in the airways may also modulate immune responses by biasing to a Th1 phenotype, as well as the induction of insulin-like growth factor 1 (IGF-1) on airway smooth muscle hyperplasia and subepithelial fibrosis.

There has been growing evidence for the targeting of the metabolic-asthma phenotype, including from real-world studies which have found a surprising link between statin treatment and a lower risk of asthma exacerbations.¹⁹³ Supplementation of omega-3-fatty acids (n3PUFA) have also been proposed as a potential asthma therapy. However, a multi-centre RCT in 98 overweight asthmatics of supplementation with fish oil did not significantly improve ACQ or risk of exacerbation after 6 months.¹⁹⁴ A meta-analysis on N-3 fatty acid supplementation found significant effects on decline in FeNO and exercise-induced bronchoconstriction, but impact on lung function and bronchodilator use were inconclusive.¹⁹⁵

C-C Chemokine Receptor Type 4 (CCR4)

Asthma is characterised by influx of Th2 lymphocytes into the airway, and one of the mechanisms by which this occurs is by chemokine-directed trafficking of cells. Immunofluorescent analysis of bronchial biopsies from allergen-challenged asthmatics demonstrates predominant expression of T cells of IL-4 and C-C chemokine receptor type 4 (CCR4), as well as CCR4-specific ligands monocyte-derived chemokine (MDC, otherwise known as CCL22ple) and thymus and activation-regulated chemokine (TARC, otherwise known as CCL17).¹⁹⁶ In a mouse HDM challenge model of asthma where mice received peripheral blood mononuclear cells from asthmatic patients, administration of CCR4 blocking antibody resulted in abolition of airway eosinophilia, goblet cell hyperplasia, AHR and IgE synthesis.¹⁹⁷

Human trials targeting CCR4 have been disappointing thus far. A phase 1 trial evaluating GSK2239633 showed low exposure and engagement inhibiting TARC from binding CCR4.¹⁹⁸ Phase II trials of another CCR4 antagonist zelnecirnon have been halted in atopic dermatitis and asthma after a serious adverse event of liver failure was reported.¹⁹⁹

Biomarkers and Airway Remodelling

Structural Changes in Chronic Asthma

Traditionally, therapeutic strategies in asthma have centred on modulation of airway inflammation; however, there is increasing acknowledgement that structural airway remodelling represents a key determinant of disease pathophysiology, severity, and progression. Airway remodelling refers to structural changes that occur due to aberrant inflammatory and repair processes that lead to recurrent inflammation, damage to the airway epithelium, and eventual fixed obstruction. Hallmark pathological changes include thickened reticular basement membrane from collagen deposition, airway smooth muscle (ASM) hyperplasia, goblet cell hyperplasia and mucus hypersecretion, extracellular matrix remodelling and angiogenesis.²⁰⁰ These changes can lead to airway narrowing, persistent airflow limitation and accelerated lung function decline, resistance to bronchodilators and steroids and can contribute to increased symptom burden, poor control and increased risk of exacerbation.²⁰¹

Epithelial tissue injury leads to release of alarmins (IL-33, IL-25 and TSLP) which have diverse effects on inflammatory and mesenchymal structural cells. Alarmins activate lung fibroblasts which increase extracellular matrix deposition (ECM) such as collagen-1, and activate macrophages which release matrix metalloproteinases that can alter ECM structure and turnover.²⁰² Alarmins can also drive epithelial to mesenchymal transition through TGF-β signalling with migration to the lamina propria and synthesis of extracellular matrix.²⁰³ IL-4 and IL-13 enhance subepithelial fibrosis and mucus production via goblet cell proliferation,²⁰⁴ while IL-5 drives eosinophil migration and activation and release profibrotic mediators including TGF-β and matrix metalloproteinase-9 (MMP-9), stimulate reticular basement thickening, and release VEGF contributing to increased vascularity of the airway wall.²⁰⁵ Airway remodelling can also be initiated by mechanical stress alone, as evidenced by bronchial biopsy studies in preschool wheezers which found that airway remodelling was present in the absence of significant airway inflammatory wall infiltrates.²⁰⁶

Targeting Airway Remodelling – Early and Late Effects of Biologics

Biologics have been shown to have both early and late effects on airway remodelling (Table 2). Early effects of biologics have been demonstrated within days to early weeks of starting treatment,^{110,207–209} and these effects are largely thought to be mediated by modulating cytokine drivers of inflammation and mucus production. Mepolizumab has been shown to improve small airway function as measured by multiple breath nitrogen washout after the first dose,²¹⁰ and was found after only 3 administrations to reduce IL-5 mediated eosinophil infiltration, deposition of TGF-β1⁺ eosinophils, ASM proliferation and deposition of ECM proteins in the reticular basement membrane.²¹¹ Dupilumab is thought to reduce ASM proliferation by decreasing actin and collagen deposition and attenuating TGF-β release by blocking IL-4, and reduces mucus secretion and goblet cell metaplasia by blocking IL-13, as demonstrated by early decrease in CT mucus plugging within the VESTIGE trial.^103,212 Omalizumab, which targets IgE, has been shown to reduce reticular basement membrane (RBM) thickening and fibronectin deposits in bronchial mucosa,²¹³ and decrease expression of airway mucus plugs.²¹⁴ Tezepelumab has been shown in animal models to reduce airway remodelling through reduction of MMP and connective tissue growth factor.²¹⁵

Table 2 Trials Evaluating Effect of Biologics on Airway Remodelling Biomarkers

Evidence for durable effects of biologics on remodelling at 12 months onwards, points to modulation of structural components of the airway wall. Omalizumab was shown at 36 months to reduce RBM thickness, and that levels of proteins related to fibrosis, most prominently galectin-3, were significantly reduced.²²⁰ The MESILICO study explored the effects of mepolizumab on airway remodelling markers in bronchial biopsies at the 12 month mark and found significant reduction in RBM thickness, smooth muscle layer thickness and number of submucosal eosinophils.²²¹ Dupilumab was shown to reduce ASM mass and reduce mucosal eosinophil infiltration,²¹² while open-label extension studies have seen sustained improvement in lung function at over 148 weeks.²²²

Biomarkers of Airway Remodelling

Biopsy histology gives the most direct insight into structural changes of remodelling, however due to their invasiveness, other biomarkers for airway remodelling are needed. A number of surrogate markers can potentially capture remodelling, including airway physiology measures of both large and small airways dysfunction, quantitative imaging and biochemical markers.

The most widely utilised measures include physiological indices of airflow limitation, most commonly spirometry to detect reduced bronchodilator reversibility and decline in FEV1 trajectory over time as a measure of accelerated airway remodelling.²²³ Oscillometry biomarkers including resistance at 5Hz minus 20Hz (R5-R20) have also been shown to directly correlate with anatomical airway wall narrowing in computational models generated from CT scans.²²⁴ Spirometry and oscillometry indices may therefore be salient markers to measure the effect of biologics on functional airway limitation including obstruction, burden of small airway disease and air trapping.²²⁵

Quantitative respiratory imaging represents another way to assess the effects of biologics. Computed tomography (CT) imaging can be used to measure airway wall thickness, airway luminal narrowing, and air trapping. Magnetic resonance imaging can visualise ventilation defects with hyperpolarised gas, defects in gas exchange using Xenon 129 (129Xe) dissolved-phase MRI and vascular abnormalities through functional MRI with contrast.²²⁶ The VESTIGE study showed that dupilumab significantly reduced mucus plugging as measured by quantitative CT,¹⁰³ and in a separate study dupilumab was shown to reduce ventilation defect percentage measured by MRI and CT mucus plugging.¹⁰⁴ Tezepelumab has been shown to significantly reduce CT-detected occlusive mucus plugs¹⁰⁵ and increase lumen area across airway generations.⁶⁷ Benralizumab has been shown in a small study to significantly decrease CT mucus plug scores and ventilation defect percentage on 129Xe MRI at both early (4 months) and late phase (2.5 years).^216,217

More exploratory biomarkers include soluble remodelling-associated proteins which can be measured in various matrices including sputum and serum. Matrix degradation enzymes have been measured in induced sputum including matrix metalloproteinase-2 (MMP-2), matrix metalloproteinase-2 (MMP-9) and tissue inhibitor of metalloproteinases-1 (TIMP-1). MMP-9 and TIMP1 have also been measured in the serum of asthmatics, and a low serum MMP-9/TIMP-1 ratio was shown to predict reduced reversibility of FEV1 to oral steroids in asthma.²²⁷ Serum periostin has also been linked to airway remodelling and has been shown to localise to reticular basement membrane and be involved in collagen deposition.²²⁸

Asthma Remission and Biomarkers

Defining Asthma Remission

Goals of asthma treatment have traditionally been symptom control and prevention of exacerbations, however, with the advent of biologics, the concept of “on-treatment” remission as a realistic treatment aim has been gaining traction. The definition of remission has not been universally established, though consensus panels have generally defined this as ≥ 12 months of (1) absence of significant symptoms, (2) no exacerbations and (3) no use of systemic corticosteroids, with some definitions also including stabilisation of lung function.²²⁹

Evidence of Achieving Remission with Biologics

A recent systematic review and meta-analysis of 25 studies looked at attainment of clinical remission using three-component (no use of systemic corticosteroids, no exacerbations and absence of significant symptoms) and four-component (additional inclusion of lung function) definitions. Clinical remission was attained in a minority of patients: 38% for three-component definition and 30% for four-component definition.²³⁰ In another review examining post hoc analyses of clinical trials (RCTs and real-world studies lasting 6–12 months), between 15% and 37% of patients achieved remission, though how remission was defined was heterogeneous with lung function criteria being used variably.²³¹

The parallel use of the term “super-responder”, patients who experience significant and sustained benefits from biologics, has also gained traction. A number of studies have reported super-responder rates on biologics ranging from 14% to 50%, reflecting variations in criteria used, patient cohorts and biological agent.²³²

Given the substantial cost of biologics, the questions remains whether remission can be achieved off-treatment. Evidence so far suggests that cessation of biologic leads to a loss of control. In the COMET study, patients who had been on mepolizumab for 3 years or more and then stopped had a 61% higher risk of a significant exacerbation versus those who continued, and a significant increase in blood eosinophils.²³³ The XPORT study evaluated persistence of response in subjects continuing or withdrawing omalizumab treatment and found 47.7% had an exacerbation in the continuation group vs 67% in the cessation group.²³⁴ Extended follow up examining effect of discontinuation of tezepelumab after 2 years treatment in the DESTINATION study found a gradual increase in blood eosinophils, FeNO and ACQ-6 scores after stopping, indicating a loss of treatment effect off-drug.²³⁵

There has been greater efficacy with tapering of ICS therapy while on biologics as evidenced by the SHAMAL trial which examined whether patients were able to taper formoterol-ICS therapy while on Benralizumab. About 92% of patients were able to reduce their ICS-formoterol dose, and of these, 91% had no exacerbations during tapering.²³⁶ For now, the evidence suggests biologics maintain asthma control and enable tapering of inhaled steroids to reduce steroid-related toxicity but cannot be discontinued without increased risk of disease relapse.

Biomarkers Predicting Remission or Failure to Respond

Understanding the characteristics that predict super-response to biologic may help to improve the chances of successful biologic selection. The main biomarkers which have consistently been found to predict higher odds of remission on treatment have been T2 biomarkers. In the meta-analysis by Shackleford et al, blood eosinophils and FeNO were the strongest predictors of remission (blood eosinophils OR of remission 1.91 [1.42–2.56], FeNO OR 1.95 [1.09–3.46]).²³⁰

The main factors contributing to lower chance of achieving remission have been found to be worse baseline FEV1, worse asthma symptoms, longer asthma duration, use of maintenance oral corticosteroids, emergency department visits in the last year and the presence of comorbidities including depression and obesity.²³⁰ It is well established that longer asthma duration has been linked to persistent airflow limitation,²³⁷ and this is likely to be related to cumulative airway damage over a longer time period leading to airway remodelling.

The current treatment paradigm follows a “treat to failure” approach whereby treatment is escalated through a one size fits all stepwise approach, only responding to deterioration (worsening symptoms, incidence of exacerbation) before intervening. This strategy leads to progressive airway damage and remodelling and lung function decline over time, as well as accumulation of dose-related adverse events related to corticosteroid exposure. A treat-to-target strategy, in which high-risk patients are identified early and biomarkers are used to select targeted therapy, has the potential to alter the natural history of asthma by enabling the initiation of more aggressive and individualized therapy. This approach has been adopted in other chronic inflammatory diseases such as rheumatoid arthritis, where early treatment with disease modifying agents is commenced shortly after diagnosis with the aim of preventing irreversible joint damage.²³⁸ Whether proactive treatment with biologics may prevent accumulation of airway damage and potentially enable sustained clinical remission needs to be investigated (Figure 4).

Figure 4 Potential effects of biologics on airway remodelling. Schematic to show the potential effects of early targeted intervention with biologics on airway remodelling, and lung function as a proxy. Further research is needed to assess the early, late and potential preventative effects of biologics. Created in BioRender. Horn, N. (2025) https://BioRender.com/fhhf9m4.

Emerging Asthma Biomarkers and Challenges

Transcriptomics and Machine Learning Approaches for Multi-Omic Integration

The dawn of high-throughput gene sequencing has facilitated the ability to identify the expression of thousands of genes in samples such as bronchial and nasal biopsies and brushings, bronchoalveolar lavage cells, sputum and peripheral blood. The generation of gene signatures combined with multi-scale bioinformatic approaches which can be applied to large omics datasets allows a systems-level approach to uncover underlying molecular endotypes. Targeting the mechanisms driving these endotypes may allow a more personalised therapeutic strategy than a clinical phenotyping approach.

In U-BIOPRED, differentially expressed gene (DEG) analysis was employed on sputum samples to compare gene expression between eosinophilic, non-eosinophilic and healthy volunteers.²³⁹ Unsupervised clustering and a supervised learning algorithm were applied to the transcriptome associated clusters (TACs) to determine the predictive gene signatures for each cluster. The first cluster, TAC1, was characterised by the highest blood eosinophils, FeNO and serum periostin, and clinical features included high systemic steroid use, acute exacerbations and severe airflow obstruction. Twenty genes were defined for TAC1 including those related to cytokine signalling such as IL-33R, CCR3 and TSLPR genes. TAC2 was characterised by high sputum neutrophil counts and C-reactive protein (CRP) levels, and moderate airflow obstruction. Genes identified included those associated with the IFN and TNF superfamilies, neutrophil chemotaxis (CXCR1, CXCR2) and inflammasomes (CASP4, MEFV, NAIP). The last cluster, TAC3, had normal to moderate sputum eosinophils, preserved lung function, less severe asthma and lower oral corticosteroid burden. The TAC3 gene signature comprised of genes related to mitochondrial function (MRPL57, PDCD2), ubiquitination (ZYG11B) and lysosomal function and transport (SCARB2).

Beyond U-BIOPRED, transcriptomics are now also being applied to under-recognised clinical phenotypes beyond T2 such as obesity-associated T2-low asthma, revealing novel immune-metabolic signatures. For example, differential gene expression analysis of plasma extracellular vesicle microRNAs (miRNAs) in obese T2-low patients compared to non-obese T2-low patients identified two distinct clusters enriched for pathways related to cytokine signalling (including IL-6, TGF-β and interferons) as well as metabolic pathways such as leptin and insulin signalling.²⁴⁰ Differential gene expression of CD4⁺ T-cell RNA signalling between these two groups also identified upregulated interferon-stimulated genes in obesity-associated T2-low asthma, and pathways related to viral response, gap junctions and G protein-coupled receptor (GPCR)-ligand binding in both groups compared to healthy controls.²⁴¹

In U-BIOPRED, further stratification has also been achieved through multiomics approaches incorporating proteomic signatures and matched transcriptomic data.⁵⁹ In the eosinophilic clusters, there was upregulation of pathways relating to T2 inflammation including IL-13 axis, IL-4 and TSLP. The neutrophilic subphenotypes had gene expression profiles predicted to result from downregulation of downregulation of IL-13 and COL18A1, a gene associated with atopy, activation of virally induced transcription factors and Th1 cytokines.

Limitations remain in that use of such an approach requires obtaining biological samples that are more invasive to obtain (eg bronchoscopic biopsies or induced sputum). Bulk RNA sequencing may also mask cell-specific signals as samples contain mixed populations of airway cells. Newer approaches include single cell transcriptomics to resolve heterogeneity in samples by profiling gene expression in individual cells, and spatial transcriptomics to localise where inflammatory gene expression is taking place and to link remodelling to molecular signals. Integration with multi-omics platforms may help to find multi-dimensional biomarkers that have clearer mechanistic value. The challenge remains being able to distil these complex signatures into clinically measurable biomarker tools that can be translated to the clinic.

Spatial Tissue Biology

Although sampling of bronchoalveolar lavage fluid or airway luminal brushings allows bulk transcriptomics or single cell RNA sequencing, the methods used means that the spatial context of the cells within tissue, and how these cells interact with each other and the stroma is lost. Spatial tissue biology is a new approach that adds the missing dimension of location, elucidating how cells and pathways operate in situ within the airway wall through next-generation molecular gene expression profiling within intact tissue sections.

Single cell spatial transcriptomics have been applied on endobronchial biopsies from patients with mild to severe asthma and healthy volunteers to map the spatial transcriptional landscape of the airway.²⁴² This found that the lung wall was characterised by localised pro-inflammatory “niches” – distinct hubs within the epithelial-subepithelial area and mucous gland area, enriched with a high level of chemokine and alarmin expression. Mast cells and macrophages were among the highest producers of IL-33 and TSLP, while goblet cells exhibited the highest signal for neutrophil chemoattractants CXCL6 and CXCL8.

The impact of biologic treatment was further investigated by examining the effect of imatinib treatment. Imatinib led to a dramatic reduction in the expression of cytokines and chemokines, including IL-33 and CCL19, extracellular matrix genes, regulatory genes including ACKR1, and goblet cell genes including MUC5B. The spatial relationship between cells was also disrupted, with reduced frequency of mast cells near basal cells, and reduced fibroblasts and endothelial cell 2 numbers. This study was an important example of the potential for spatial tissue biology to move beyond understanding which pathways are active, to exactly where and how in the tissue these inflammatory pathways are sustained, garnering further information for therapeutic development on specific drug-target interactions within tissue.

Breath Biomarker Platforms

Breath analysis has the ability to directly mirror metabolic activity taking place in the lungs. It therefore represents a non-invasive patient-friendly diagnostic matrix and potentially a rich platform for novel biomarker discovery. The constitutive components of breath that may provide these biomarkers include the gas phase, where trace volatile organic compounds (VOCs) can be measured, water vapour, which can be trapped as exhaled breath condensate and contains both volatile and non-volatile water-soluble molecules, and larger, non-volatile particles in exhaled breath (particles in exhaled air, or PExA). VOCs can originate from both endogenous (internal metabolic processes) and exogenous sources.

Breath collection can be direct, through “online” sampling with point of care analysis, or indirect, with breath stored in adjuncts such as collections bags or sorbent tubes for lab-based analysis, or “offline” sampling. The most commonly used offline technique is gas chromatography-mass spectrometry (GC-MS) which enables quantitative and robust identification of a wide range of compounds through separation based on retention time and mass spectra matching. This was demonstrated in the UK EMBER trial, a large breath biomarker study where two-dimensional GCxGC-MS was used to identify several novel signatures that were able to distinguish between different acute cardiorespiratory presentations.²⁴³ Specifically, a group of VOCs called reactive aldehyde species (RASPs) were found to be dysregulated in asthma, compounds that have been found to be dysregulated in asthma in other studies.²⁴⁴

Online technologies offer significant advantages in terms of portability and short processing times, allowing bedside clinical application. The main examples include GC-ion mobility spectrometry (IMS) and electronic nose (eNose) technologies. GC-IMS is a technique that separates ionised molecules in the gas phase based on their mobility in a carrier buffer gas, without necessitating pre-concentration. eNoses are made up of array sensors that convert chemicals in breath to electrical signals, and compare them to pre-programmed patterns. eNose analysis in one clinical study was able to discriminate between eosinophilic, neutrophilic and paucigranulocytic inflammatory phenotypes.²⁴⁵

Despite advances made in the field of breathomics, widespread clinical translation beyond FeNO has been limited due to a number of technical challenges. The wide variety of analytical platforms, complex processing and analytical needs, and a lack of standardised protocols for breath collection, analysis and reporting has limited the use of breath analysis in large multi-centre clinical studies. The potential for breath biomarkers to predict response to biologics has been demonstrated by the SOMOSA study.⁹⁹ Going forward, harmonisation of collection and analysis protocols are needed across multicentre trials, with close collaboration between clinicians, analytical chemists, bioinformatics and engineering specialties to translate potential breath signatures to point-of-care sensors that can be used in the clinic.

Conclusion

Current therapeutic approaches in asthma largely follow a one size fits all strategy that fails to consider disease heterogeneity. Despite significant advances made in the treatment of severe asthma with the introduction of biologics, current clinically available biomarkers do not map uniquely to a particular pathway and thus are unable to predict response to precisely targeted treatments. Failing to identify disease endotypes early fosters a treat-to-failure approach, which promotes airway remodelling and contributes to irreversible loss of lung function.

Advances in molecular phenotyping, such as multi-omics integration with machine learning, spatial and single-cell technologies, and minimally invasive sampling, offer the potential to identify novel biomarkers. However, clinical translation will depend on addressing key challenges, including assay standardisation, robust biological and clinical validation in large prospective biomarker-stratified trials, and health economic evaluation to ensure tests are simple, affordable, and widely accessible. Together, these efforts will be crucial to move the field toward a truly precision-based approach, where biomarker-guided therapies can be implemented early to prevent disease progression and fundamentally alter the natural course of asthma.