The Impact of CFTR Modulator Triple Therapy on Type 2 Inflammatory Response in Patients with Cystic Fibrosis

Background Treatment of cystic fibrosis (CF) has been revolutionized by the use of cystic fibrosis transmembrane conductance regulator (CFTR) protein modulators such as elexacaftor/tezacaftor/ivacaftor (ETI) triple therapy. Prior studies support a role for type 2 (T2) inflammation in many people with CF (PwCF) and CF-asthma overlap syndrome (CFAOS) is considered a separate clinical entity. It is unknown whether initiation of ETI therapy impacts T2 inflammation in PwCF. We hypothesized that ETI initiation decreases T2 inflammation in PwCF. Methods A single center retrospective chart review was conducted for adult PwCF. As markers of T2 inflammation, absolute eosinophil count (AEC) and total immunoglobulin E (IgE) data were collected longitudinally 12 months prior to ETI therapy initiation and 12 months following therapy initiation. Multivariable analyses adjusted for the age, gender, CFTR mutation, disease severity, inhaled steroid use, and microbiological colonization. Results There was a statistically significant reduction (20.10%, p < 0.001) in 12-month mean IgE following ETI initiation; this change remained statistically significant in the multivariate model. The longitudinal analysis demonstrated no change in AEC following therapy initiation. Conclusion This study shows reduction in IgE but no change in AEC after ETI therapy initiation. We think that the lack of influence on AEC argues against an impact on previously established T2 inflammation and that the reduction in IgE is likely related to antigen load reduction post ETI. Further studies are warranted to determine the underlying mechanism of ETI impact on T2 inflammation and possible role for asthma immunomodulator therapy post ETI initiation in CFAOS.


Introduction
Cystic brosis (CF) is one of the most common respiratory genetic diseases, affecting an estimated 37,000 people in North America and ~162,000 people worldwide 1 . It is characterized by mutations in the CFTR gene which encodes the cystic brosis transmembrane conductance regulator (CFTR) protein, most commonly a deleted phenylalanine in the 508 th position (F508del). CFTR is a chloride channel that regulates transport of sodium, chloride, and bicarbonate. 2 Classic pulmonary symptoms of CF include thick, mucous secretions in the airway causing chronic pulmonary infection, increase bacterial load and fungal colonization, and obstructive lung symptoms, which progress to airway edema, mucus plugging with airway collapse, and bronchiectasis. 2 Many studies have shown that CF patients commonly display evidence of type 2 (T2) in ammation 3 characterized by evidence of airway and peripheral eosinophilia, as well as high levels of total and allergen-speci c immunoglobulin E (IgE). 3,4 T2 in ammation is classically implicated in the pathogenesis of allergic as well as non-allergic eosinophilic subtypes of asthma. 3,5 but can also develop in other lung in ammatory disorders including, for example, being highly prevalent in eosinophilic chronic obstruction pulmonary disease. Similar epithelial differentiation can be conjectured as often evolving in the in amed airway of CF patients. In these scenarios, T2 in ammation represents, in large part, the response of the airway epithelial cells to airspace in ammation with the consequent production of numerous cytokines (including the alarmins thymic stromal lymphopoietin (TSLP), interleukin (IL)-25, and IL-33) and chemokines (including CCL5, CCL11, CCL24, CCL26) that are associated with promoting the T2 in ammation signature. These mediators both directly drive the generation of the classic T2 cytokines IL-4, IL-5, and IL-13 from innate lymphoid 2 cells (ILC2s), mast cells, and others but are also responsible for the differentiation of the antigen-speci c Th2 effector cells that also produce these cytokines. Loss of epithelial barriers allow for allergens and microbes to access stromal tissue and further promote the activation of Th2 effector lymphocytes and B cells. 5 Together these cytokines underlie the presence of both eosinophilic in ammation (IL-5) and elevated concentrations of IgE (IL-4 and IL-13).
Several studies support this role for T2 in ammation in many patients with CF. Siedlecki, et. al. found a prevalence of peripheral eosinophilia of 45% in CF patients hospitalized for exacerbation, and eosinophilia was associated with signi cantly longer hospital stays. 6 Our group previously found a correlation between T2 in ammation, frequent exacerbations and lung function decline, based on peripheral eosinophilia and IgE levels and we have reported on the ability of eosinophil-targeting therapeutics to produce bene cial outcomes in CF patients. 7,8 Interestingly, according to the 2021 CF Foundation Patient Registry Annual Data Report, the prevalence of asthma in CF patients is 30.8 %. 9 Experts in the CF community have acknowledged the association between asthma and CF, now considering CF-asthma overlap syndrome (CFAOS) a separate clinical entity. 10 Previous studies support this concept that elevations in type 2 in ammatory markers likely re ect an intrinsic eosinophilic phenotype of CF. 4,5 Elexacaftor/tezacaftor/ivacaftor (ETI) modulator therapy has recently been approved by Food and Drug Administration and has become the main therapy prescribed for most people with CF. ETI has been shown to improve CFTR function, decrease sweat chloride concentrations, and improve both lung function and functional status as demonstrated in the PROMISE clinical trial 11 . While addressing the defect in CFTR function, it is unknown whether these agents impact the presence of a previously established T2 in ammatory state that developed as a complication of CF in ammation.
The aim of this study was to investigate the impact of CFTR modulator triple therapy on Absolute Eosinophil Count (AEC) and IgE as markers of the type 2 in ammatory response in patients with CF.

Patient Population
A single center retrospective chart review was conducted for adult cystic brosis patients (n = 108) seen at the University of Virginia Adult Cystic Fibrosis Clinic between 2018 and 2021. Inclusion criteria were a diagnosis of CF and initiation of ETI therapy. Exclusion criteria were a history of transplant, IL-5 (benralizumab, mepolizumab) or IgE (omalizumab)-targeting biologics, and insu cient AEC or IgE data. Following exclusion, there were 54 patients with adequate AEC data and 80 patients with adequate IgE data for a total of 85 patients. This study was approved by our institution's Institutional Review Board Human Subjects Research committee and followed procedures with ethical standards. (IRB-HSR 23337).

Data Collection
AEC and IgE data were collected for each patient 12 months prior to initiation of ETI therapy and 12 months following initiation of therapy. AEC measurements from the center-associated lab were collected from the electronic health record (EHR). Given that infection results in activation of in ammatory responses, we excluded data collected during pulmonary exacerbation and/or use of IV antibiotics. AEC measurements were excluded if they were obtained within 7 days prior to determination of an exacerbation and/or initiation of IV antibiotic therapy or if they were within 14 days following exacerbation and/or completion of IV antibiotic therapy.
We were unable to exclude IgE measurements taken during an exacerbation as IgE measurements were obtained on a less frequent basis, often only once annually. As such, IgE data were collected from the EHR including both data from the center associated lab and outside facilities, where available. Additional data collected included age, sex, disease severity, infection/colonization history, CFTR mutation information, and inhaled steroid use. Disease severity was de ned based on percent predicted Forced Expiratory Volume in one second (ppFEV 1 ) and was categorized as normal (ppFEV 1 > 90%), mild (70% <ppFEV 1 ≤ 90%), moderate (40% <ppFEV 1 ≤ 70%), and severe (ppFEV 1 ≤ 40%). With regards to bacterial infection/colonization history, data regarding prior Pseudomonas infection, MRSA infection, and other bacterial colonization were collected. Other bacteria were de ned as Stenotrophomonas spp., Achromobacter spp., Acinetobacter spp., or Burkholderia spp. With regards to fungal colonization history, data regarding prior Aspergillus spp., Exophiala spp., and Rasamsonia spp. colonization were collected. Inhaled steroid use was coded as yes or no, de ned as use prior to ETI initiation.

Statistical Analysis
All statistical analyses were performed in R (version 4.2.2). A two-sided signi cance level of 0.05 was selected as the null hypothesis rejection criterion. Linear mixed-effect models were t using the lme4 12 and lmerTest 13 packages. In multivariable models, all clinically relevant covariates were chosen a priori. Before performing inferences on any regression model, residual diagnostics were checked to verify assumptions of normality and homoskedasticity. The responses, AEC and mean IgE, were natural log transformed to induce normality and correct heteroskedasticity. Due to the log transformation, coe cients are interpreted as percent changes rather than absolute changes in the response. Inferences on model coe cients and contrasts were conducted with t-test using the emmeans package 14 and the Ftest.
To assess how log-transformed 12-month mean IgE changed after ETI initiation, we t a multivariable linear mixed-effect model with a random intercept of patient, regressing over an indicator variable for ETI initiation and adjusting for current age, gender, CFTR mutation, initial disease severity level, Pseudomonas spp. colonization, MRSA colonization, other bacterial colonization, fungal colonization, and inhaled steroid use. A linear mixed-effect models was constructed to further investigate the mediating effect of initial disease severity on the ETI treatment effect. The model treated initial disease severity as a factor with four levels.
The relationships between the change in mean IgE after ETI and prior fungal colonization, Pseudomonas colonization, or MRSA colonization were each examined individually with two linear mixed-effect models (a univariate and multivariable model with covariates age, gender, CFTR mutation, initial disease severity level, and inhaled steroid use) that had a random intercept for patient. All models included an interaction between the presence of the pathogen and ETI initiation.
To assess how log-transformed AEC changed after ETI initiation, we t two linear mixed-effect models, a univariate and multivariable model, with a random intercept for patient. While AEC was collected longitudinally, it was still regressed over an indicator variable for ETI initiation because time since initiation was not meaningful. The multivariable model adjusted for age, gender, CFTR mutation, initial disease severity level, Pseudomonas spp. colonization, MRSA colonization, other bacterial colonization, fungal colonization, and inhaled steroid use.

Baseline Demographics and Characteristics
Summaries of baseline demographics and patient characteristics are presented in Table 1. 108 adult Cystic Fibrosis patients seen at the University of Virginia Adult Cystic Fibrosis Clinic were screened for eligibility, but only 85 unique patients with adequate AEC data, IgE data, or both after exclusion were included in the nal analysis. All but ve patients with complete AEC data also had complete IgE data.
The characteristics of the sample of patients in the IgE analysis were highly similar to the characteristics of the sample of patients in the AEC analysis.

IgE and ETI Therapy
IgE data were log-transformed to induce normality prior to analysis. The estimated change in IgE after ETI initiation is plotted in Fig. 1

Post-ETI Initiation Change in IgE and Lung Function
After determining that ETI initiation was associated with a change in 12-month mean IgE, we investigated whether initial lung function impacted the magnitude of the change in IgE after ETI. In this mixed model, the reduction of IgE following ETI initiation remained statistically signi cant (t = 4.060, P < 0.001). However, IgE at baseline did not signi cantly differ by initial disease severity (F = 1.559, P = 0.206), and the reduction in IgE following ETI therapy did not signi cantly differ by initial disease severity (F = 0.208, P = 0.890). These results show that in our patients, initial lung function is not associated with the magnitude of change in mean IgE after ETI initiation.

Post-ETI Change in IgE and Microbiological Colonization
We also investigated whether fungus colonization, Pseudomonas colonization, or MRSA colonization alone modi ed the magnitude of the change in IgE after ETI using subgroup analysis through univariate and multivariable models. Univariate and multivariable models produced similar results, so all values reported below are adjusted estimates.
In summary, these results indicate that without considering the effects of the presence of other microorganisms, a reduction in mean IgE was only observed in the presence and not absence of a positive fungal culture and the magnitude of reduction in mean IgE after ETI initiation was only impacted by prior MRSA colonization.

AEC and ETI
AEC data were log-transformed to induce normality prior to analysis. The estimated change in log AEC after ETI initiation is plotted in Fig. 3 with institution of biologics that target the T2 mediator IL-5. 8 T2 in ammation re ects in large part the differentiation of airway epithelial cells as characterized by their production of cytokines -especially the alarmins TSLP, IL-25, and IL-33 -that are responsible for immune deviation of Th2 lymphocytes as well as production of chemokines responsible for eosinophil recruitment and activation. 15,16 In addition to Th2 lymphocytes, the alarmins drive secretion of the de ning T2 cytokines IL-4, IL-5, and IL-33 from numerous other cell types including ILC2s and mast cells. The treatment of patients with CF has been revolutionized by the e cacy of CFTR modulator therapy. The mechanism behind the improvement in patient outcomes is a continued topic of investigation as our understanding of the role of CFTR function evolves. While developing presumably as a complication of the in ammatory state associated with CF, it is less clear whether CFTR dysfunction is directly responsible for T2 in ammation, such as through an impact on the expression of these channels on T and B lymphocytes. And, as such, it is unclear whether institution of CFTR modulator therapy will impact a previously established T2 in ammatory state. To contribute to the growing investigation of the pathogenesis and treatment of CF from an immunologic perspective, this study investigated the impact of CFTR modulation via ETI triple therapy on AEC and IgE as components of the T2 in ammatory signature.
To our surprise, multivariable analysis of AEC data after excluding measurements obtained during exacerbation and/or IV antibiotic use, demonstrated no change following therapy initiation. Eosinophilia is the quintessential de ning feature of T2 in ammation. 15,16 Blood eosinophila, in particular, represents a hematopoietic response to the expression of IL-5. IL-5 is produced by a fairly modest number of cell types including primarily Th2 effector lymphocytes, ILC2s, and mast cells. As such the absence of a change in AEC signi es that initiation of CFTR modulator therapy had no impact on these central de ning components of T2 in ammation. This clinically would lead to no impact on the CFAOS presentation and CF exacerbations related to T2 in ammation in PwC and an eosinophilic phenotype.
In contrast to AEC, however, our data demonstrated a ~ 20% reduction in 12-month mean total IgE following ETI triple therapy initiation after controlling for multiple factors including demographic data and inhaled steroid use (Fig. 1). In the absence of an apparent impact on the T2 signature, this impact on total IgE begs an alternative explanation. CFTR dysfunction results in impaired mucus production and clearance, thus promoting microbial colonization in the airway. These microbes and, in particular, fungi, can function as allergens, especially in the presence of a T2 high state and it is well recognized that CF patients frequently display aspergillus-speci c IgE and, indeed, this disease is often complicated by allergic bronchopulmonary aspergillosis [17][18][19] . The CF airway is also frequently contaminated with Staphylococci. While not an allergen, staph are copious producers of numerous superantigens that nonspeci cally (that is, in a T cell receptor-independent fashion) can broadly activate numerous families of T lymphocytes. Indeed, CF has been associated with high expression of staph-derived superantigens in the airway 20 . Insofar as many of these T lymphocytes are Th2-like, staph superantigens can drive a "storm" of Th2 cytokines, including especially, IL-4 and IL-13, cytokines responsible for IgE production. We would propose that while not directly impacting the T2 high state, by improving respiratory function, reducing mucus production, and improving mucociliary clearance CFTR therapy would reduce fungal and staph airway colonization. And, as such, through this indirect mechanism, CFTR therapy could reduce IgE concentration. This hypothesis is supported by our data demonstrating a greater reduction in IgE in patients with prior MRSA (Fig. 2). Moreover, our results also demonstrate that IgE was reduced in patients with a history of fungal colonization but was not reduced in patients with no prior history of fungal colonization. IgE. Thus, reduction in antigen (fungal) and superantigen (Staph) load secondary to ETI therapy initiation could explain the reduced IgE production in the absence of ETI therapy having an effect on ameliorating the underlying T2 high state.
This analysis serves as an important starting point to continue the evaluation of CFTR modulator therapy from an immunologic perspective. These results are intriguing and warrant further studies to elucidate the exact effect of ETI on T2 in ammation. Based on our current ndings and previous studies that show a correlation between T2 in ammation based on AEC and IgE levels and frequency of exacerbations 7 , we suspect the ETI may not in uence the rate of exacerbation in a subgroup of people with CFAOS and an eosinophilic phenotype. Further studies to elucidate the impacts of possible confounding variables in a larger patient population are warranted. Moreover, studies investigating additional markers of the T2 in ammatory response, especially those utilizing samples obtained from airway will prove valuable.
There are important considerations and limitations regarding this study. There are inherent limitations in performing a retrospective analysis due to limited data measurement frequency, varying frequency of clinical assessment of exacerbation status, and possible unknown microbiological colonization history and allergen sensitization prior to subject care establishment at our center. T2 in ammatory markers are impacted by numerous confounders including exposure to other allergens, systemic and topical corticosteroid use, and others.

Conclusion
This study demonstrates that there is a statistically signi cant percent reduction (20.10%, p < 0.001) in mean IgE but no change in AEC following ETI initiation. The lack of change in AEC argues against an impact of ETI therapy on previously established T2 disease. In contrast, we suspect that ETI leads to decreased antigen and superantigen load in the airway as a result of improved mucociliary clearance and it is these changes that drive the decline in IgE. Further studies are warranted to determine the underlying mechanism of ETI impact on T2 in ammation and possible role for asthma immunomodulator therapy post ETI initiation in CFAOS. Availability of data and materials: The datasets generated and/or analysed during the current study are not publicly available due to possibly leading to subject identi cation, (single center, and rare disease subjects) but the statistical analysis is available from the corresponding author on reasonable request.

Abbreviations
Competing interests: None of the authors have any competing interests to declare Funding Larry Borish receives salary support from NIH: UO1 AI123337, R21 AI151496, and R56 AI158519.
Authors' contributions Ajay Mehta Substantial contributions to the conception, design of the work, the acquisition, analysis, and interpretation of data for the work; drafting and revising the manuscript critically for important intellectual content; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Irene Lee Substantial contributions to the conception, design of the work, the acquisition, analysis, and interpretation of data for the work; drafting and revising the manuscript critically for important intellectual content; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Galvin Li Substantial contributions to analysis, and interpretation of data for the work; drafting and revising the manuscript critically for important intellectual content; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Marieke Jones Substantial contributions to analysis, and interpretation of data for the work; drafting and revising the manuscript critically for important intellectual content; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Lydia Hanson Substantial contributions to data collection; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Kevin Lonabaugh Substantial contributions to data collection; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Rhonda List Substantial contributions to data collection; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Larry Borish Substantial contributions to the conception, design of the work, the acquisition, analysis, and interpretation of data for the work; drafting and revising the manuscript critically for important intellectual content; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All authors read and approved the nal manuscript.
Dana Albon A Substantial contributions to the conception, design of the work, the acquisition, analysis, and interpretation of data for the work; drafting and revising the manuscript critically for important intellectual content; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. represents the percent change in 12-month mean IgE of a single patient after starting ETI. presence of Pseudomonas, with 95% con dence intervals from the multivariable mixed-effect model. Each point represents the percent change in 12-month mean IgE of a single patient after starting ETI. Pvalues above each group denote the signi cance of the percent change for that group from 0, and Pvalues connecting groups denote the signi cance of the difference in percent change across groups.