SEARCH

SEARCH BY CITATION

Keywords:

  • asthma;
  • biomarkers;
  • endotypes;
  • pathomechanisms;
  • phenotypes

Abstract

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

Asthma phenotypes have been developed to address the complexities of the disease. However, owing to a lack of longitudinal studies, little is known about the onset as well as the stability of phenotypes. Distinguishing phenotypes with regard to the severity or duration of the disease is essential. A phenotype covers the clinically relevant properties of the disease, but does not show the direct relationship to disease etiology and pathophysiology. Different pathogenetic mechanisms might cause similar asthma symptoms and might be operant in a certain phenotype. These putative mechanisms are addressed by the term ‘endotype’. Classification of asthma based on endotypes provides advantages for epidemiological, genetic, and drug-related studies. A successful definition of endotypes should link key pathogenic mechanisms with the asthma phenotype. Thus, the identification of corresponding molecular biomarkers for individual pathogenic mechanism underlying phenotypes or subgroups within a phenotype is important. Whether newly defined asthma endotypes predict the individual course of asthma has to be validated in longitudinal studies. The accurate endotyping reflects natural history of asthma and should help to predict treatment response. Thus, understanding asthma endotypes might be useful in clinical practice.

Asthma is a complex disease or a syndrome that includes several disease variants. A disease ‘phenotype’ describes ‘clinically observable characteristics’ of a disease without direct relationship to an underlying pathophysiology. ‘Endotypes’, however, describe subtypes of a disease defined by an intrinsically ‘distinct pathogenetic mechanism’. In asthma, phenotypes describe clinical and morphologic characteristics as well as unique responses to treatment. Phenotypes are clinically relevant in terms of presentation, triggers, and treatment response but do not necessarily relate to or give insights into the underlying pathological mechanism. Asthma endotypes, however, describe disease subtypes based on cellular and molecular mechanisms, including the reactivity of structural cells. Understanding these events would allow us to understand and classify asthma endotypes by the use of biomarkers from body fluids and/or affected tissues [1]. Endotyping asthma based on disease mechanisms could eventually lead to an individualized management [2, 3].

Disregarding endotypes might lead to two major setbacks: (i) unsuccessful clinical trials because of a bulk patient selection and (ii) unequivocal results because of insufficient preselection of patients based on their phenotype and endotype in large cohorts.

Therefore, endotype-specific classifications of any participant in clinical studies might be of considerable advantage (Fig. 1).

image

Figure 1. Potential advantages of asthma endotyping.

Download figure to PowerPoint

Current concepts in asthma phenotyping

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

‘Phenotype’ has been used extensively in asthma but is still largely indiscriminate. A wide array of ‘clinically observable characteristics’ have been used to describe asthma phenotypes (Fig. 2). Therefore, there are many ‘definitions’ for asthma phenotypes, many of which are related to differences in symptoms and severity rather than to differences in underlying mechanisms. Any meaningful ‘phenotyping’ requires a correct diagnosis. Alternative diagnosis, comorbidities, environmental exposure (allergen, tobacco smoke, other indoor or outdoor pollutants, etc.), poor adherence to treatment or incorrect inhaler technique, profile of the patient, and level of understanding and coping with the disease require assessment. Asthma can be very heterogeneous, ranging from mild to severe or from intermittent to persistent airway obstruction up to ‘brittle asthma’. Separating asthma heterogeneity into well-defined phenotypes is challenging because of the lack of specific and validated markers. Most of the reported asthma phenotyping is based on cross-sectional or retrospective data, and there is little prospective, long-term validation of the applied ‘phenotype label’. Many phenotypes have been based on epidemiological data, symptom pattern, atopic status, bronchial obstruction pattern, etc. These clinical phenotypes frequently overlap and can change over time.

image

Figure 2. Clinically observed characteristics used to describe asthma phenotypes.

Download figure to PowerPoint

Historically, asthma has been divided into two distinct phenotypes termed extrinsic and intrinsic asthma [4]. Both phenotypes are characterized by eosinophilic inflammation [5]. Allergic asthma exacerbates following the inhalation of specific allergens and is generally considered to be more responsive to treatment than intrinsic asthma where no exogenous allergens can be identified.

Phenotyping of asthma with inflammometry, especially with the use of sputum cytology [6], might offer advantages such as predicting treatment responses [7, 8], highlighting mechanistic pathways involved in disease pathogenesis [9], and predicting future risk [10]. Investigations of granulocyte infiltration in induced sputum suggested eosinophilic, neutrophilic, mixed granulocytic, and paucigranulocytic phenotypes. Whether these phenotypes are stable or subject to change over time is still unclear. A prospective evaluation of inflammatory phenotypes in induced sputum in moderate and severe asthma showed that only one-third maintained a stable inflammatory phenotype [11]. In addition, inflammatory phenotypes appear to differ with age. Eosinophilic inflammation is more prevalent in both acute and stable childhood asthma, while neutrophilic inflammation is more frequent in acute asthma in older adults [12]. In intrinsic asthma, which is frequently of late onset, eosinophils predominate [13].

Recently, novel statistical and mathematical methods such as cluster or factor analysis and principal component techniques have been used to phenotype asthma, suggesting novel phenotypes in both adults and children [14, 15]. Machine-learning approaches have also been applied to cluster phenotypes in childhood asthma [16]. This resulted in a 5-class model where children were clustered into four atopic phenotypes, and only those from the ‘multiple early (sensitization) atopic phenotype' had an increased risk to develop asthma.

From phenotypes to endotypes

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

Pathogenetic mechanisms described in asthma include Th2/allergic and eosinophilic inflammation, airway hyper-responsiveness, neuro-immune interactions, bronchial and possibly lung remodeling as well as changes in airway secretions, and angiogenesis. Steroid resistance and decreased response and/or toxicity prone to beta2-agonists' overuse have also been described as possible phenotypes [17, 18]. Recent data also indicate the involvement of neutrophils, NKT cells, and mechanisms of innate immunity or the resolution of inflammation and tissue repair [9, 19] (Table S1, online repository).

A properly defined asthma endotype should include the complex and fluctuating nature of the disease, as well as its possible progression to fixed airflow obstruction or other outcomes such as aspirin sensitivity. A multidimensional approach combining clinical features and disease physiopathology (inflammation and remodeling), evaluated as complex parameters, has been recommended [20], which would also require to consider age and sex in such definitions.

Whether newly defined asthma endotypes are of any value should be investigated and subsequently validated in three steps: (i) longitudinal replication across different populations should predict meaningful differences among individuals; (ii) observations implicated in such endotypes should reflect the diseases' biology and its natural history and predict response to treatment; and (iii) endotypes should be easily applicable and useful in daily clinical practice and cost-efficient.

Accordingly, any successful definition of an endotype should eventually link the key pathogenic mechanism with a clinical phenotype of asthma. An acceptable starting point to define endotypes would be the identification of corresponding molecular biomarkers for such a pathogenetic mechanism. Several asthma endotypes might be identified following this approach (Fig. 3).

image

Figure 3. Linking essential pathogenic mechanisms with phenotypes of asthma.

Download figure to PowerPoint

Allergic asthma

The phenotype of allergic asthma has been extensively studied. It usually starts in childhood and is often accompanied or preceded by allergic rhinitis and/or atopic eczema. Initial symptoms are typically driven by allergen exposure, leading to increased airway inflammation which in many cases can persist and progress even in the absence of allergen. Histologically, allergic asthma is characterized by mucosal infiltration with eosinophils, CD4+ cells, mast cells, expression of the high-affinity IgE receptors on inflammatory and resident cells and epithelial damage, half-desmosomes, goblet cell hyperplasia, reticular basal membrane thickening, and smooth muscle hypertrophy. Until today, it remains unclear whether these latter findings occur as a consequence or possibly independent of eosinophilic airway inflammation.

Key pathogenic mechanism underlying allergic asthma is the Th2-driven inflammation, which is, however, only pronounced in the presence of or following allergen exposure [21] possibly due to an inefficient response of T-regulatory cells [22]. Established biomarkers measured in sputum, BAL, serum, and bronchial biopsies are IL-4 (difficult to measure after allergen exposure but actually absent without allergen), IL-5 (only present if eosinophils are present), IL-9, IL-13, and periostin [23], and CCR8 with emerging implication of new molecules (TLSP, TARC, IL-25, CRTH2, and DP1 receptors, ICOS/ICOS-L, IL-31, IL-33, and T1/ST-2, IL-19, IL-22, H4 receptors) (Table S1, online repository). Studies that comprehensively analyze these markers and their potential patterns cross-sectionally and longitudinally are lacking, but might be of crucial importance in future attempts to define a specific endotype. NO might be increased in exhaled air and has been suggested as a surrogate for eosinophilic inflammation [24].

The diagnosis of allergic asthma requires the determination of atopic status. Positive skin prick tests, specific IgE antibodies in serum, eosinophilia in the peripheral blood and/or in induced sputum/BAL/bronchial biopsies as well as an association with allergic rhinitis and/or eczema are characteristic for this phenotype.

Intrinsic (nonatopic) asthma

Intrinsic (nonatopic) asthma appears to represent about one-third of all adult asthmatic patients [25] and about half in childhood asthma [26]. Intrinsic asthma in adults usually begins in the second half of life; it has a female preponderance and is often more severe than atopic asthma. Chronic sinusitis, nasal polyps, and aspirin sensitivity are more prevalent. In children, nonallergic asthma has an earlier start in their symptoms, and respiratory infections and exposure to environmental tobacco smoke appear to be more frequent. In contrast to adults, children with intrinsic asthma show less severe symptoms [27].

Airway inflammation shares similarities with allergic asthma, with increased numbers of T-helper type 2 cells, mast cell activation, and infiltration of eosinophils, and similar mediators of inflammation are expressed, including Th2 cytokines and eosinophilotactic chemokines [28]; however, in intrinsic asthma, IL-2 and IFN-gamma are increased in BAL fluid, but not IL-4 [29]. BAL T cells in intrinsic asthma express markers of chronic activation, suggesting ongoing T-cell stimulation, possibly by an endogenous antigen [4].

The key drivers of inflammation in intrinsic asthma are unknown. Allergens have no obvious role in driving the inflammatory process in the airways. Local IgE synthesis has been suspected because of endobronchial mRNA for IgE but their pathogenetic relevance remains unclear. House dust mite–specific IgE has been reported in the sputum of patients with intrinsic asthma [30], but this IgE's reactivity did not elicit clinical responses to Der p, and it remains unclear whether these patients were correctly labeled as intrinsic asthma. A second signal that promotes IgE-mediated asthmatic responses through FcεRI might be lacking in intrinsic asthma [30]. Superantigens may contribute to its pathogenesis via class switching of local B cells, resulting in endobronchial polyclonal IgE production and also by generating specific IgE against superantigens that also cause clonal expansion of T cells, resulting in increased Th2 cells and CD8+ cells, while suppressing regulatory T cells. Superantigens could also reduce the responsiveness to corticosteroids, resulting in more severe asthma [31]. Free light chains eliciting immediate hypersensitivity responses via mast cell degranulation have also been suggested as a pathogenic mechanism in intrinsic asthma [32]. Autoreactive immune mechanisms induced either by molecular mimicry or as a consequence of chronic allergic inflammation and the generation of cytotoxic autoantibodies have been described [33]. IgG antibodies directed against epithelial proteins, such as cytokeratin-18, have been detected in adults with intrinsic asthma [33] where dysregulated epithelial repair and impaired resolution of the inflammation might also play a role.

Biomarkers for intrinsic asthma have not been validated. Th1 cells, IL-18, IL-15, IP-10, antibodies directed against epithelial proteins, and CD26 have been described in intrinsic asthma but require validation (Table S1, online repository).

Noneosinophilic asthma

Noneosinophilic asthma has been characterized by the absence of airway eosinophilia. Instead, neutrophilia can be observed. Until today, it is still unclear whether this represents a unique form of asthma or just a different stage of severity or persistent bacterial colonization or infection of the airways. In a study using induced sputum, neutrophilic asthma was identified in 59% of adult patients with symptomatic asthma [34]. The attributable risk of asthma owing to eosinophilic inflammation in the population is about 50%. Conversely, this means that up to 50% of all asthma has been attributed to inflammation dominated by other cell types [35]. However, as longitudinal studies are lacking, it remains unclear whether neutrophilic asthma developed on a background of long-standing eosinophilic asthma or recurrent or persistent bacterial infections superimposed on the background of a previously eosinophilic asthma. Less is known about the incidence of neutrophilic asthma in children, where eosinophilic asthma seems to predominate both in acute and in stable disease [12]. Neutrophilic inflammation appears more frequent in children with nonatopic asthma [36]. In adults, it is more frequent in females [37]. While subepithelial layer thickness appeared normal in adults with noneosinophilic asthma [35], children with noneosinophilic asthma have characteristic features of airway remodeling similar to patients with eosinophilic disease [38]. Children with noneosinophilic as well as eosinophilic asthma had thickened basement membrane, increased epithelial loss, and a higher number of vessels. Moreover, both groups of asthmatics expressed increased IL-4 and IL-5, while TGF-βRII was reduced [38].

Novel mechanisms implicated in the pathogenesis of noneosinophilic asthma involve the activation of innate immune responses with a possible role of bacteria, viruses, and diet, activation of neutrophil elastase, and impaired nuclear recruitment of histone deacetylase (HDAC) 2 [37, 39]. There is also evidence for an increase in factors promoting airway neutrophil viability in severe asthma [40]. Furthermore, the role of Th17 lymphocytes in asthmatic airway inflammation is being investigated [41].

Established biomarkers that drive neutrophilic inflammation are IL-8, IL-17A, LTB4, and possibly IL-32, PAMS, DAMPS, or SDF (Table S1, online repository). Airway pathophysiology in neutrophilic asthma is characterized by (fixed) airflow limitation, a decrease in airway hyper-responsiveness to mannitol, and low levels of FeNO [42]. The key diagnostic tool to diagnose neutrophilic asthma is induced sputum. Whether this is sufficient to define a separate phenotype remains to be determined.

Aspirin-intolerant asthma

Aspirin-intolerant asthma (AIA) affects approximately 5–10% of all adult asthmatics and is more common in nonatopic asthmatics [43]. It is rare in children. Women are more often affected than men with reports ranging from 5.5 : 1 to 1.32 : 1 [44]. Aspirin-intolerant asthma occurs on the background of an asthma, which is frequently progressive and causes signs and symptoms of asthma even in the absence of aspirin or other nonsteroidal anti-inflammatory drugs. It has been suggested to describe this condition as aspirin-exacerbated respiratory disease (AERD). A typical course starts with symptoms of rhinitis during the third decade, often after a viral respiratory illness. Over a period of months, chronic nasal congestion, anosmia, rhinorrhea, and nasal polyps develop, which are followed by asthma and sensitivity to aspirin [44]. The disease runs a protracted course even if COX-1 inhibitors are avoided, and is often severe. Many patients require systemic corticosteroids to control their sinusitis and asthma. Aspirin is a common precipitant of life-threatening attacks of asthma. In a large survey, 25% of asthmatic patients requiring emergency mechanical ventilation were found to have AIA. Aspirin and NSAIDs should be strictly avoided to prevent life-threatening asthma attacks. Highly specific COX-2 inhibitors are occasionally well tolerated and can be tried under supervision.

Histologically, AERD is characterized by an intense eosinophilic inflammation of nasal and bronchial tissues [44]. In addition to the underlying asthma, the eicosanoid metabolism is uniquely altered and combined with an increased sensitivity to leukotrienes C4, D4, and E4 [45]. Overproduction of cysteinyl leukotrienes (Cys-LTs) has been implicated as major pathogenetic feature. In AERD, free radical–mediated prostaglandin generation has also been demonstrated by measuring the urinary enantiomer PGF [2] alpha [46]. Although both COX-1 and COX-2 are expressed in the airways, COX-1 is the functionally dominant enzyme, which explains some of the clinical observations related to drug specificity in patients with aspirin-sensitive asthma (tolerance of COX-2 inhibitors) [47]. Although COX-2 inhibitors precipitate asthma attacks less likely, some patients do react; though, the subsequent asthma attack following the oral uptake of specific COX-2 inhibitors appears to be less severe. Systematic studies on this clinically important question are still missing. Leukocytes from aspirin-sensitive, but not from aspirin-tolerant, asthmatics generate 15-hydroxyeicosatetraenoic acid (15-HETE) when incubated with aspirin [48]. 15-HETE generation involves the activation of 15-LO and is modulated by prostaglandin EP1-3 receptors [48]. Expression of alternatively spliced variants of COX-1 mRNA is increased and correlates with aspirin-triggered 15-HETE generation [49]. Furthermore, MCP-3 and RANTES are overproduced in AERD [50]. In addition, bronchial challenge with aspirin involves systemic reactions and is associated with the mobilization of leukocyte and eosinophil progenitor cells from the bone marrow [51]. Staphylococcal superantigens may contribute to airway inflammation and the development of airway hyper-responsiveness in AIA [52].

Although the search for biomarkers to differentiate AERD from aspirin-tolerant asthma included eicosanoid pathway metabolites in urine, blood, induced sputum, saliva, or exhaled breath condensate, none of them is able to diagnose AERD with reliable sensitivity and specificity. While in AERD significantly higher leukotriene E4 and lower PGE2 concentrations are present in urine compared to aspirin-tolerant asthma [53], there is still considerable overlap. CysLT and LTB4 concentrations in saliva quantified by enzyme immunoassay (EIA) following purification by high-performance liquid chromatography (HPLC) were also higher in AERD [54]. Using an analytic approach based on mass spectrometry of EBC, subjects with AERD were distinguished by a sharp increase in the level of prostaglandin D2 and E2 metabolites. 5- and 15-HETE levels were also higher compared to aspirin-tolerant asthma [55]. Still, none of these tests is able to reliably diagnose AERD, and the observed quantitative differences might be related to the higher intrinsic inflammation observed in patients with intrinsic asthma and aspirin sensitivity.

Key diagnostic tools to diagnose AERD today are still the lysine-aspirin bronchial challenge test [56] or an oral challenge protocol [57]. Recent guidelines [56] recommend a bronchial or a nasal challenge and if the results are negative, an oral challenge to confirm the diagnosis. However, nasal provocation tests are not well validated and lack specificity.

Based on the above-mentioned features, it is difficult to attribute the term ‘endotype’ to AERD because it is an acquired condition on top of an intrinsic or less frequently allergic asthma and thus, despite its peculiar sensitivity to NSAIDs, still has major overlap with these conditions.

Extensive remodeling asthma

Another variant of asthma has been described as extensive remodeling asthma, which is histologically characterized by minimal inflammation, but extensive remodeling that can be characterized into subtypes such as thickened small airways, alveolar detachment and loss of elastin, airway smooth muscle (ASM) hypertrophy, goblet cell hyperplasia and mucus production, angiogenesis, lymph angiogenesis, and reticular basement membrane (RBM) thickening [58-61]. Recent reports indicate that the increased mass of ASM is one of the essential features of airway remodeling. Morphometric analysis of ASM layers in fatal asthma has revealed two subphenotypes of fatal asthma: type I with increased muscle layers only in the larger airways and type II with thickened muscular layers in the entire bronchial tree, from central to peripheral regions [62]. These increased muscle layers surrounding the airways might be the morphologic correlate of hypersensitive airway constriction in these patients. This could contribute to airway narrowing by more forceful and more frequent constriction, and result in more rigid airway walls with increased folding, and secretion of pro-inflammatory cytokines [63]. In the airways of fatal asthma, muscle bundles and fibroconnective tissues disrupted the tissue's lymphatics resulting in impaired airway clearance and increased mucosal edema [64].

The precise cellular defects, the longitudinal course, and in particular the initial events leading to this putative subgroup of asthma remain unclear. It has been suggested that an abnormal epithelial–mesenchymal trophic unit (EMTU) with abnormal activation and inappropriate tissue repair mechanism might be involved in its pathogenesis [65]. An intrinsic abnormality of ASM might co-exist favoring ASM cell proliferation. In severe asthma, a gallopamil-sensitive calcium influx and the activation of calcium-calmodulin kinase IV leading to enhanced mitochondrial biogenesis through the activation of various transcription factors (PGC-1α, NRF-1, and mt-TFA) have been described. The altered expression and function of sarcoplasmic/endoplasmic reticulum Ca2+ pump could play a role in ASM remodeling in moderate to severe asthma [66]. Abnormal expression of tight-junction proteins within the ASM might also contribute to ASM proliferation. Increased claudin-1 expression, which was observed in the nucleus and cytoplasm of ASM in asthma, might play a role in ASM cell proliferation and could promote angiogenesis via VEGF [67]. Th2-cell-derived mediators can mobilize and recruit pro-inflammatory and pro-angiogenic precursors from the bone marrow into the airway wall where they induce angiogenesis [60]. Effector cells (human lung mast cells, basophils, eosinophils, macrophages, etc.) are also sources of angiogenic and lymphangiogenic factors.

Biomarkers investigated include MMPs, TIMP, TGF-β, IL-13, VEGF subsets, ADAM 33, ADAMTS, pro-angiogenic hematopoietic progenitor cells, as well as newly described molecules or tissue structures such as IL-13, ICOS-L, CC, and CXC chemokines, osteopontin, amphiregulin, periostin, fibulin, decorin, oncostatin M, LIGHT, relaxin, endothelin-1, airway basal stem cells, retinoid receptors, tight junctions (Table S1, online repository). None of these, however, is able to explain the observed features satisfactorily, and these markers do not allow a clear separation from other forms of asthma.

Clinically, most patients have long-standing asthma that often has been insufficiently treated. Data in children are scarce, and there are no longitudinal or epidemiological data. Airway physiology varies from less reversible/fixed to highly collapsible airways or a fast decline in pulmonary function [68, 69].

Key diagnostic tools are dynamic evaluation of airway physiology and high-resolution computed tomography [70]. The use of bronchial biopsies in this setting needs to be determined.

Genetics of asthma and endotypes

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

In asthma, genetic associations of highly replicated susceptibility genes lack consistency across populations. It has been postulated that epigenetics and prenatal influences reshape and modulate the genetic background expression toward the expression of an endotype [71]. Stratification of the subjects by endotype in genetic studies of asthma may be necessary to improve the sensitivity and reliability of association results. On the other hand, the candidate genes yielded so far by association or genome-wide association studies are biologically plausible in some but not all cases, because many of them are encoding mediators, receptors, or cytokines, which are considered to be of relevance in pathogenesis of asthma. Data published so far relating a specific asthma endotype to a genetic profile are rather scarce (Table 1).

Table 1. Genetics of asthma and endotypes
Asthma endotypeProposed genetic linkageLevel of evidence
  1. For references, see online repository.

Allergic asthmaSNP located inside the coding sequence of C5 Other loci regulating immune- or inflammation-mediated mechanisms and airway smooth muscle contraction [1]Data from pooled genome-wide association studies in pediatric allergic asthma describing genes specific for asthma among all allergic subjects
Several Th2 cytokine SNPs Filaggrin ORMDL3 [2]Association observed mostly in adult population. Might be only an association and not a true risk for asthma [overall OR for filaggrin of 1.3 if not associated with eczema [3]]
TSLP promoter gene polymorphism [4]Associated with disease susceptibility in both children and adults Correlated with pulmonary function
Nonatopic asthmaHomozygosity for MMP-9 variants [5]Increased risk but not limited to intrinsic asthma
Aspirin-intolerant asthmaSNPs of CEP68 gene [6]Related to the magnitude of FEV1 decline following aspirin bronchial provocation test
Galectin-10 gene [7]mRNA increased in aspirin-intolerant asthma
The human EMI domain–containing protein 2 (EMID2) gene [8]Related to extracellular matrix deposition and epithelial–mesenchymal transition
Solute carrier family 22 member 2 (SLC22A2), DTD1, STK 10, ADAM 33, HLA-DPBI*0301, genes encoding the leucotriene C4 synthase, ALOX5, CYSLT, PGE2, TBXA2R, and TBX21 [9-12] 
Extensive remodeling asthmaADAM 33 and DPP10 polymorphisms [13, 14]Unclear if related to asthma or rather suggest a generally increased vulnerability of the lungs
Hedgehog interacting protein (HHIP) gene Family with sequence similarity 13, member A (FAM13A), Patched homolog 1 (PTCH1) [15]Subset of normal lung function genes Together predict lung function abnormalities as a measure of severity in non-Hispanic White and African American subjects with asthma

Whole-genome gene expression profiling could provide information relating to asthma heterogeneity, classifications, and mechanisms. This approach has the potential to define novel pathways as well as to identify the phenotypes that relate to endogenous mechanisms [72]. The study by Baines et al. characterized three asthma phenotypes on a transcriptional level by using unsupervised hierarchical clustering of induced sputum gene expression profiles. These three distinct phenotypes were related to both the clinical asthma status and the type and degree of airway inflammation. Two of the described phenotypes were inflammation predominant, one eosinophilic and the other neutrophilic, while the third phenotype was paucigranulocytic. Both inflammation-predominant phenotypes had an up-regulation of genes related to immune defense, inflammatory responses, and responses to stimuli, wounding, and stress. In the neutrophilic phenotype, the up-regulated genes were predominantly related to cell chemotaxis, indicating active cellular recruitment, and also to IL-1 and TNF-α/NF-κB pathways [72].

Treatment response and asthma endotypes

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

Response to beta2-agonists

The response to beta2-adrenergic receptor (beta2AR) agonists (full or partial) varies individually and has been associated with beta2AR polymorphisms. Some beta2AR haplotypes are associated with increased cell-surface beta2AR protein expression (expression phenotypes), and thus, the initial bronchodilator response from intermittent administration of beta-agonist is beta2AR haplotype dependent. Other haplotypes showed increased agonist-promoted down-regulation of beta2AR protein expression (down-regulating phenotypes), which might be relevant in chronic use of beta2-agonists. Finally, some haplotypes showed the ideal clinical combination of high expression and less down-regulation [73].

For pediatric asthma, a slower responder phenotype to i.v. beta2-agonists was described, requiring significantly higher total doses of IV terbutaline, higher maximum administration rates, and longer ICU and hospital length of stay [74].

Impaired β-agonist-induced ASM relaxation in asthmatics might also be important in extensive remodeling asthma with prominent ASM hypertrophy. Decreased β-agonist-induced cAMP in ASM from asthmatics results from enhanced degradation because of increased PDE4D expression. This defect is directly related to ASM because it requires no inflammatory environment [75].

Response to inhaled corticosteroids

Eosinophilic inflammation in asthma is a marker of steroid responsiveness in both allergic asthma [10, 76] and AERD (Table 2). Allergic asthma usually responds well to inhaled corticosteroids (ICS), which have become the mainstay of therapy also for AERD, which, however, frequently also requires systemic steroids. TSLP gene expression also proved responsive to both ICS and LABA. The induction of TSLP mRNA and protein expression was synergistically impaired by a corticosteroid and salmeterol [77]. A functional glucocorticoid-induced transcript 1 gene (GLCCI1) variant was associated with reduced lung function in response to ICS. The genotype accounted for 6.6% of overall ICS response variability [78]. SERPINE 1 gene variant was also linked to short-term steroid responsiveness [79] (Table 2). Although some inflammatory patterns are similar to atopic asthma, the response to inhaled steroids is less pronounced in intrinsic asthma. Patients with noneosinophilic asthma have a poorer short-term response to ICS [80]. It is not known to what extent airway remodeling might be sensitive to corticosteroids [81].

Table 2. Asthma endotypes and response to inhaled corticosteroids
BiomarkersSurrogate marker predicting responseSubphenotypePredictionStrength of association
Eosinophilic inflammation [10, 76]Sputum eosinophils ?Exhaled NO Pediatric and adult allergic asthmaGood response+++
Aspirin-intolerant asthmaGood response+/−
SERPINE1 gene [79]?Fast FEV1 decline?Good response+
TSLP gene [77]?Allergic asthmaGood response+
GLCCL1 variant [78]?Poor response+
Neutrophilic inflammation [37, 39, 80]Sputum neutrophilsNeutrophilic asthmaPoor response++
Airway remodeling [81]HRCT; lack of inflammationSevere asthma with extensive remodelingPoor response?

Response to antibiotics/antioxidants

In intrinsic asthma, antibiotic therapy to eradicate the relevant superantigen-producing microorganisms so far has not been convincingly shown to challenge the efficacy of corticosteroids. Noneosinophilic asthma might respond to antibiotics or antioxidants possibly due to some immune modulator effect of certain antibiotics, such as macrolides [82]. By modulating the innate immune response in the lung, including the effects on neutrophils and the signaling pathway and chemokine release, macrolides might be of beneficial particularly in noneosinophilic asthma [83]. Unfortunately, the trials evaluating macrolides in asthma focused mainly on the infectious status of the patients and not on the endotype and thus failed to provide a substantial benefit [84].

Response to targeted treatment

Response to a targeted intervention in asthma may be improved if a better selection of patients based on an exact phenotyping or even on endotypes can be used (Tables 3 and 4). Response to anti-IL-5- and anti-IL-4-targeted treatment was improved when patients were preselected based on biomarkers related to their presumed endotype (sputum eosinophils for anti-IL-5 and serum periostin for anti-IL-4) [23, 85, 86].

Table 3. Asthma endotypes and response to targeted treatment
Targeted treatmentSurrogate marker predicting better responseSubphenotypeStrength of association
  1. AIA, aspirin-intolerant asthma; ICS, inhaled corticosteroids.

  2. For references, see online repository.

Antileukotriene agentsYounger age, shorter disease duration, and increased LTE4/FENO ratio [1, 2]Pediatric asthma+
Comorbid allergic rhinitis, younger age, shorter duration of asthma, and treatment with only ICS and not ICS + LABA [3]Adult asthma+
Less severe aspirin-induced bronchospasm Lack of history of aspirin hypersensitivity or sinusitis [4]AIA+/−
Allergen-specific immunotherapyMonosensitized and/or in whom a single allergen is predominantly driving asthma symptoms [5, 6]Mild, well-controlled allergic asthma+
Aspirin desensitizationRecurrent nasal polyposis Overdependence on systemic corticosteroids [7, 8]AIA+
Restoration of HDAC 2 nuclear recruitment with theophylline [9]?Neutrophilic asthma? [10, 11]+/−
?Smokers with asthma [12]+/−
Table 4. Asthma endotypes and response to targeted treatment with biologicals
Targeted treatmentSurrogate marker predicting better responseSubphenotypeStrength of association
  1. For references, see online repository.

Anti-IgEBoth sensitized and exposed [1]Severe allergic asthma [2]+
?Uncontrolled asthma [2, 3]+/−
See Table 2Exacerbation-prone asthma [1]+/−
IP-10?Viral-induced asthma exacerbations (prevention) [1]+/−
Anti-IL-5Sputum eosinophils > 3%Refractory eosinophilic asthma [4, 5]+
Anti-IL-13Serum periostin [6]FeNO? Allergic asthma with dominant IL-13 activation+
Anti-IL-4/IL-13IL-4 receptor α polymorphism [7]Allergic asthma with dominant IL-4/IL-13 activation+/−
Anti-TNF-αIncreased TNF-α in BAL, bronchial biopsies, PBMC [8-10]Severe corticosteroid refractory asthma with up-regulation of TNF-α axis+/−

Conclusion

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

Today, there is increasing evidence for different pathogenetic mechanisms operant within the syndrome of asthma, which appear to require different treatment approaches such as aspirin deactivation in AERD. Available data suggest that specific, pathogenesis-directed treatment approaches such as anti-IL-5 therapy may be beneficial in certain subsets or clusters of patients in which, for example, IL-5-mediated eosinophilia persists despite moderate doses of corticosteroids. For this purpose, identification and validation of phenotype-specific biomarkers, as has been shown in recent studies for periostin, which indicates IL-13 activity, will be essential. Further direct as well as indirect assessments of mediators of asthmatic inflammation will be crucial to determine and/or predict the clinical success of pathogenesis- and phenotype-specific interventions that are currently being developed. Furthermore, more precise phenotyping based on clinical as well as immunological assessments of specific inflammatory constellations leading to the clinical syndrome of asthma may be helpful to tailor individual treatment approaches with more specific, molecular interventions. Thus, specific phenotypes and possibly endotypes may be in the future defined by their response to target-specific interventions. Unfortunately today, there is still a lack of longitudinal studies in asthma with little knowledge about the onset as well as the longitudinal stability of currently described phenotypes, and this is also a severe obstacle to further develop the concept of endotypes. This may be improved by a more specific selection of asthma patients for longitudinal, epidemiological as well as interventional studies based on the expression of specific (inflammatory) biomarkers.

Nevertheless, current data provide a basis to search for new endotypes relevant to genetic associations as well as treatment response.

Further investigations into the underlying mechanisms of asthma will therefore require a careful clinical selection in the presence (or absence) of specific inflammatory markers. In addition, understanding the conundrum of the syndrome of asthma implies a meticulous differentiation between phenotypes on the one hand and any confounding by severity and/or duration of disease/inflammation on the other hand. Apart from documenting biomarkers for new pathogenic mechanisms (innate immunity, neutrophilic or neurogenic inflammation, resolution and repair mechanisms, immune regulation), novel biomarkers for already existing asthma mechanisms and phenotypes (eosinophilic asthma, allergic asthma, severe asthma) or for newly described clinical phenotypes of asthma (‘frequent exacerbators’, postmenopausal asthma, etc.) are needed. Identifying the pathogenetic mechanisms in depth for already predefined endotypes by measuring a complex array of biomarkers (cytokines, chemokines, growth factors, molecules related to the resolution of inflammation, etc.) concomitantly in sputum, EBC, serum, BAL, bronchial biopsies, PBMC and relating these biomarkers to each of the visible properties of the endotype could prove a useful approach. This may also lead to the recognition that some clusters of patients currently grouped within the syndrome of asthma based on their symptoms might not have asthma at all. There is hope for the future that more specific, individually tailored treatment approaches will be possible based on more comprehensive approaches to understand underlying the pathogenesis of asthma. In addition, such an approach would certainly help to exclude patients with asthma-like symptoms because of other diseases and could therefore be helpful in withholding unnecessary treatment to patients who do not need it.

Authors' contribution

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

In this review article, all the authors were involved in the literature review, conception, and design of the manuscript, drafting of the article, or revising it critically for important intellectual content as well the final approval of the submitted version.

Conflict of interest

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

Ioana Agache reports the following conflicts of interest: Relevant financial activities outside the submitted work: consultancy for Data Monitoring Committee for the Review of safety data emerging from VECTURA's clinical studies. Payment for lectures including service on speakers bureaus from AstraZeneca, Merck, Schering-Plough, Boehringer Ingelheim, Nycomed. Cezmi Akdis reports the following conflicts of interest: Relevant financial activities outside the submitted work: consultancy for Actellion, Aventis, Stallergenes, Allergopharma and grants received from Novartis, European Commission's Seventh Framework program 260895 and 261357, Swiss National Science Foundation, Global Allergy and Asthma European Network, Christine Kühne Center for Allergy Research and Education. Marek Jutel reports no conflicts of interest. J. Christian Virchow reports no conflict of interest in relation to the content of this manuscript.

References

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Current concepts in asthma phenotyping
  4. From phenotypes to endotypes
  5. Genetics of asthma and endotypes
  6. Treatment response and asthma endotypes
  7. Conclusion
  8. Authors' contribution
  9. Conflict of interest
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
all2832-sup-TableS1.docWord document474KTable S1. Future directions for endotyping of asthma and references for tables 1, 3, 4.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.