Chronic obstructive pulmonary disease (COPD) remains a major cause of respiratory illness worldwide, and despite modern inhaled treatments, many patients suffer from troublesome symptoms such as chronic cough, increased sputum, and increased frequency of exacerbations, potentially leading to life-threatening crisis 1. Smoking cessation is the most effective intervention to slow disease progression 2. However, in daily practice, we often see a frustrating reality: airway symptoms—and the underlying epithelial programs—do not always fully “reset” after smoking stops 3, 4. This is striking because the airway epithelium is thought to renew over months; yet cough, sputum, and exacerbation susceptibility can persist for years, implying a durable shift in the regenerative program rather than a simple persistence of “old” epithelial cells 5. Such a long-lasting shift in epithelial state could plausibly present as ongoing mucus hypersecretion and impaired barrier function, thereby sustaining symptoms and increasing the risk of exacerbation. These findings raise an important question—what maintains pathological epithelial remodelling even after the initiating exposure has been removed? Using patient-derived airway epithelial cultures grown at the air–liquid interface (ALI)—a standard system in which basal cells differentiate into ciliated and secretory lineages 6—Carlier and colleagues reported that airway epithelium from smokers and COPD retains an “injury memory” in ALI cultures, including impaired barrier properties and a mucus-prone phenotype, even after cessation 4. However, underlying molecular mechanisms for such persistent alterations remain undefined. In a recent publication in Respirology, Shiota and colleagues combine patient-derived ALI cultures with single-cell transcriptomic profiling to investigate airway differentiation across never-smokers, current smokers, and former smokers, and to map molecular circuits linked to persistent epithelial remodelling in COPD (Figure 1) 7. A central contribution is the framing of sustained goblet cell differentiation through a coherent transcriptional axis centred on NK2 homeobox 1 (NKX2-1) and SAM pointed domain-containing ETS transcription factor (SPDEF). SPDEF is widely recognised as a driver of secretory and goblet cell programs 8, whereas NKX2-1 (TTF-1)—familiar to clinicians as a diagnostic immunohistochemical marker for primary lung adenocarcinoma and a key lineage transcription factor—supports epithelial homeostasis and, in experimental models, can restrain mucous cell metaplasia, in part by opposing SPDEF-linked programs 9. This antagonistic balance offers a clinician-friendly explanation: when NKX2-1–associated homeostatic control weakens and SPDEF-driven secretory programs dominate, epithelial repair repeatedly “chooses” a mucus-producing fate—potentially sustaining cough/sputum and vulnerability to infection-associated exacerbations. Consistent with this, chronic mucus hypersecretion has been associated with an increased risk of COPD exacerbations across observational studies 10; clinically, excess sputum production may serve as a useful marker of an exacerbation-prone phenotype. At the same time, although the study links a mucus-producing program with reduced expression of junction-associated molecules and barrier readouts, the causal relationship between goblet cell hyperplasia and barrier dysfunction remains to be established. The authors further propose that epigenetic dysregulation helps stabilise this abnormal state. Chromatin-level changes can act as a molecular “scar,” altering how basal cells interpret repair cues long after smoking has ceased. Inflammatory cytokines and environmental factors can trigger secretory programs acutely, but stable regulatory changes may be needed to explain why mucus phenotypes can persist for years and why inter-individual variation remains substantial among patients with similar exposures. Additionally, because chromatin accessibility captures a regulatory “snapshot,” future work analysing DNA methylation at NKX2-1 regulatory elements will help determine whether NKX2-1 suppression reflects transient state shifts or more stable epigenetic locking. Notably, the concept that tobacco smoke leaves a persistent airway “field of injury”—detectable years after cessation—was supported by early transcriptomic work showing smoking-induced gene-expression changes in cytologically normal bronchial epithelium, with a subset persisting after quitting 3. In parallel, lung cancer–focused studies demonstrated that apparently normal bronchial epithelium can also harbour molecular damage at the genetic level, including genome-wide loss of heterozygosity (LOH) alterations in current and former smokers 11. In this context, while the present study emphasizes epigenetic imbalance of the NKX2-1–SPDEF axis, the “field of injury” framework raises the possibility that durable genetic lesions in airway basal/progenitor cell clones may also contribute to persistent epithelial remodelling after cessation. Indeed, deep sequencing has revealed extensive smoking-associated somatic mutations and clonally expanded driver events in normal human bronchial epithelium 12. In practice, genetic and epigenetic mechanisms are not competing explanations—they may cooperate: genetic lesions can establish lasting clonal differences, while chromatin-level remodelling can shape how basal/progenitor cells interpret repair cues and inflammatory signals over time. However, the finding of persistent mucus hypersecretion in long-term ex-smokers with COPD in this study should be interpreted in the context of prior clinical and ALI work suggesting that mucus-associated epithelial features often track more closely with recent cigarette smoke exposure than with COPD status per se. In the ALI immunostaining cohort, the median cessation period among former smokers was 15.5 years (IQR 10–27.5) 7. Notably, the single-cell (scRNA-seq/scATAC-seq) analyses were performed in a subset of donors, including two former smokers who quit 21 and 38 years ago 7. These cessation intervals are well beyond the timeframe in which bronchial biopsy studies have reported attenuation of goblet cell hyperplasia after sustained smoking cessation (e.g., ≥ 3.5 years) 13. This apparent difference may reflect variation in how “mucus” biology is defined and measured (cell-type proportions versus gene programs versus secreted mucin), differences in sampling location and participant characteristics, or both. It is also plausible that COPD-associated inflammatory and remodelling cues can sustain—or reawaken—a mucus-prone program even after long-term cessation. Further studies that directly compare cohorts across cessation intervals using matched ALI protocols and harmonised endpoints will help clarify which components of the muco-secretory phenotype are smoke-dependent, disease-dependent, or jointly maintained. From a disease perspective, the NKX2-1–SPDEF imbalance offers a plausible mechanistic framework for the mucus-predominant phenotype—helping to explain why some patients remain troubled by chronic cough and sputum despite optimised bronchodilation, with downstream consequences that may contribute to exacerbation risk. Beyond mechanistic insight, this framework also points to tractable intervention targets and testable therapeutic hypotheses. The work highlights therapeutic opportunities: dampening SPDEF-linked secretory programs, restoring NKX2-1–associated homeostatic differentiation, or targeting upstream regulators that maintain these states could complement current therapies. Reversing stable epigenetic remodelling is an appealing but challenging longer-term goal, because the therapeutic window may be narrow—reducing pathologic hypersecretion while preserving protective mucociliary clearance. Key questions remain regarding reversibility and airway compartmentalisation, and how these epithelial states interact with microbes and viral triggers in real-world COPD. Overall, these findings support a testable therapeutic concept: persistent cough and sputum may reflect durable and potentially targetable shifts in epithelial differentiation and phenotype, which could be modified as long as we preserve effective mucociliary clearance. The author has nothing to report. During the preparation of this work, the author used ChatGPT5.2 plus to improve language, grammar, and readability. After using this service, he reviewed and edited the content as needed and took full responsibility for the publication's content. This work was supported by Japan Society for the Promotion of Science, grant in aid for scientific research (c)/25K1145. Dr. Mitsuo Sato is an Editorial Board member of Respirology. He was excluded from all editorial decision-making related to the acceptance of this article for publication.
Mitsuo Sato (Tue,) studied this question.