Predictors of response to <scp>anti‐IL‐5</scp> biologics

Eosinophils have long been associated with asthma. However, it is not known if they are the dominant effector cells that cause symptoms and contribute to severity in all patients with asthma. Interleukin (IL)-5 is the cytokine that has the most pronounced effect in modulating eosinophil biology. Therefore, blocking IL-5 signalling (either the ligand with mepolizumab or reslizumab, or the receptor with benralizumab) is a sensible approach to treat patients whose asthma is associated with eosinophils. Targeting the IL-5 cytokine is logically likely to have maximum efficacy only in those patients with asthma whose severity is predominantly caused by eosinophils, rather than just be associated with eosinophils (which probably explains why anti-IL5 biologics have only modest beneficial effects in patients with very mild asthma). This is not easy to figure out. These are likely to be older patients, with non-atopic late-onset severe asthma, who have associated rhinosinusitis (with or without nasal polyps) and have significant and persistently increased numbers of eosinophils in blood and in the airway. A favourable clinical response at 4 months of treatment seems to predict longer term response.1 All the pivotal clinical trials of these biologics selected patients based on their absolute blood eosinophil count and reported clinical efficacies (largely annualized reduction in exacerbation rates) of between 35% and 60%. Treatment effects increase with increasing thresholds of blood eosinophil counts, with reduction in exacerbation of up to 70% with mepolizumab when blood eosinophil counts are >500/μL.1 It is ironic that this is hailed as 'precision medicine', when it does not provide any insight into mechanisms of benefit, and when a substantial proportion of patients have only a modest response or no response at all. There could be number of reasons for this. These include patient characterization, mechanisms of eosinophil recruitment and factors related to strategies to target IL-5. An important factor is using low thresholds of blood eosinophil counts to define 'eosinophilic asthma' when raised eosinophil numbers in blood maybe non-specific and do not reflect luminal eosinophil numbers. After all, circulating eosinophils are rarely activated, and they do not cause luminal obstruction and asthma symptoms unless they track their way into the airway and can be quantified in sputum (intact cells, granules or extracellular DNA). This discordance is mostly pronounced in those patients with asthma who are on very high doses of inhaled or oral glucocorticosteroids.2 This is partly contributed by the persistence of eosinophils >3% in sputum (in situ eosinophilopoiesis) driven by locally derived IL-5 from innate lymphoid cell-2 (ILC2) cells.3 Greater the number of circulating eosinophils, the greater the chance that this would be associated with sputum eosinophils, but the converse is not true. The assumption that circulating eosinophil numbers are a better predictor than sputum eosinophils of the treatment effect of anti-IL-5 monoclonal antibody (Mab) is based on an erroneous and under-powered interpretation of the mepolizumab phase 2 data of response to one of the three intravenous doses of mepolizumab.4 As there are no direct head–head comparisons of selecting patients based on sputum eosinophils versus blood eosinophils, one cannot make a definitive conclusion, but a number of clinical trials5 and observational cohort studies6 demonstrate ongoing symptoms when sputum eosinophils persist despite normalization of blood eosinophil numbers. Thus, while raised blood eosinophil numbers (>300–400 cells/μL) may be a reasonable choice to select patients for a biologic therapy, they have limitations to monitor response to therapies. Sub-optimal response to anti-IL-5 Mab is also plausible if IL-5 is not the dominant cytokine that drives eosinophilia. Sputum transcriptome data of severe eosinophilic asthmatics from the U-BIOPRED consortium would suggest that although IL-5 may be the dominant cytokine in most patients, other cytokines related to ILC2 biology may also play a role.7 Recent observations suggest that galectin-10, coded by the CLC gene, may also regulate airway eosinophilia and this is independent of IL-5.8 In our cohort study of over 250 severe asthmatic patients treated with mepolizumab or reslizumab,6 the majority of sub-optimal responders (with persistent sputum eosinophils but normal blood eosinophils) had elevated IL-5 in their sputum and not the other potential target proteins, suggesting that sub-optimal response was indeed due to inadequate neutralization of IL-5 in the airway. This leads to perhaps the most important factor which is the strategy (choice of the anti-IL-5 biologic, dose and route of administration of the drug) to control eosinophilia. As far as I know, there are no pharmacokinetic data of drug levels in the airway relating to optimal neutralization of airway IL-5 (which is the best predictor of response to an anti-IL-5 Mab5). Under-dosing with an anti-IL-5 Mab not only may lead to sub-optimal response, but may also potentially worsen both airflow and eosinophilia in 10–15% of patients in whom the treatment is indicated.6 This tends to happen in patients who already have endogenous immunoglobulin G (IgG)-type antibodies in their sputum to various cell products such as eosinophil peroxidase. Anti-IL-5 forms heterocomplexes with IL-5 and the endogenous IgG autoantibodies, and could potentially activate complement as well.6 These immune complexes could trigger the release of tumor necrosis factor-like cytokine-1 (TL-1) from monocytes and macrophages, which would act as 'danger signals' and stimulate ILC2 cells (through death receptor 3) to produce more IL-5,9 thus perpetuating sputum eosinophilia. The patients most likely to have this phenomenon tend to be those on high doses of oral prednisone (usually >10–12.5 mg/day), eosinophils and granules in sputum, history of recurrent respiratory bacterial infections (leading to airway lymphoid follicle formation) and lymphopenia (associated with loss of immune tolerance).10 The optimum strategy to treat these patients would have been higher doses of anti-IL-5 Mab than what is currently approved by regulatory agencies. These immune complex-mediated worsening are distinct from reduced efficacy of the drug due to circulating anti-drug antibodies, which are exceedingly uncommon in the clinical trials. We currently do not know the optimal strategy to treat this phenomenon. Benralizumab, by its dual action of preventing IL-5 by binding to its receptor, and antibody-dependent cytotoxic killing of eosinophils, progenitor cells and ILC2 cells,11 effectively depletes eosinophils in blood and sputum (although there has not been any direct head–head comparison studies with mepolizumab or reslizumab). Consistent with this, majority of sub-optimal response to benralizumab is not associated with eosinophilia, rather with intense neutrophilic bronchitis due to airway infections.12 History of respiratory infections in preceding years was a predictor of more airway infections. The reason for this infrequent phenomenon is currently unknown, but it does not appear to be related to eosinophil suppression. A clinical algorithm to choose from the three anti-IL-5 strategies is illustrated in Figure 1. This needs to be prospectively validated in a clinical trial. In summary, in clinical practice, sub-optimal response to anti-IL-5 biologics ranges from 25% to 50%. The reasons may relate to poor endotyping of 'eosinophilic asthma' limited to low thresholds of blood eosinophil counts and under-dosing of the currently available anti-IL5 Mab. We have very effective drugs at our disposal. We need to get better at administering them. P.N. is supported by the Frederick E. Hargreave Teva Innovation Chair in Airway Diseases. P.N.'s university has received grants from AstraZeneca, Teva, Sanofi and Boehringer Ingelheim. P.N. has received honoraria for lectures and scientific advisory boards from Astra Zeneca, Sanofi, Teva, GSK, Methapharm, Novartis, Merck and Equillium.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.