The current coronavirus disease (COVID-19) pandemic illustrates the diversity in response on exposure to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Response to infection ranges from asymptomatic to death from organ failure, of which the latter is most commonly observed in the elderly (1). Such differences are associated with aging, but may also be influenced by the genetic constitution of the host.

Diversity in response to SARS-CoV-2 exposure may be related to host factors associated with airway defense. The gel-forming mucin 5B (MUC5B) is part of the mucus that covers the surface of the respiratory epithelium and forms a key barrier defense against respiratory pathogens (2, 3). In vivo studies in mice showed that Muc5B deficiency caused accumulation of materials in the upper and lower airways, leading to chronic infection and inflammation that failed to resolve normally. By contrast, in mice that overexpress Muc5B, macrophage function was improved. Hence, the presence of Muc5B in the lung is essential for controlling infections, maintaining immune homeostasis, and mucociliary clearance (4). Aged mice had significantly reduced Muc5b levels in comparison with young mice (5) and decreased expression of Muc5B in mouse models was associated with reduced mucociliary clearance (4, 5). In both humans (6) and mice (5) decreased mucociliary clearance was shown to be associated with aging.

Constitutive expression levels of MUC5B are associated with a common promoter polymorphism, rs35705950 of the encoding gene MUC5B. The minor rs35705950 T allele is associated with high expression levels of MUC5B and the major G allele is associated with low expression levels (7, 8). The high expressing T-allele is a known risk factor for idiopathic pulmonary fibrosis (IPF) (7), a fatal aging lung disease of unknown cause predominately affecting older males with a history of smoking. IPF is a non-infectious disease of the distal lung caused by damage of the alveolar epithelium followed by progressive fibrogenesis (9).

Recently it was shown that aging lung diseases such as IPF and chronic obstructive pulmonary disease (COPD) share disease loci but have opposite risk alleles (10). Given the fact that the alleles of these loci influence expression levels we proposed a theory of trade-offs in aging lung disease (11). A trade-off exists whenever a benefit in one context entails a cost in another (12). In aging lungs, the high expressing MUC5B T-allele may be important for optimal airway defense against infections while it provides an increased risk for IPF in the alveolar compartment.

Therefore, we examined if MUC5B rs35705950 is associated with severe COVID-19. To investigate this, we performed a retrospective candidate rs35705950 case-control study in a Dutch cohort and included an UK cohort for validation and an Italian and Spanish cohort for replication.

In this study we used a candidate gene case-control approach to examine whether a genetic polymorphism that influences expression of MUC5B is associated with susceptibility to severe COVID-19. We observed a significant association between the MUC5B rs35705950 promoter polymorphism and severe COVID-19 in four white European cohorts. The results of the meta-analyses demonstrate the protective effect of the MUC5B T allele against severe COVID-19. The T-allele frequency and T-carrier frequency was lower in severe COVID-19 patients than in controls.

Beneficial effects of carriage of the T-allele have been reported before. In smoking non-Hispanic white COPD patients with interstitial HRCT features, T-allele carriers experienced less acute respiratory disease and a longer time-to-first event (17). Furthermore, in IPF patients, carriers had a lower bacterial burden than non-carriers (18) and better survival (19).

In the human respiratory system, MUC5B is secreted throughout the lung by submucosal glands and the superficial epithelium of trachea, bronchi, bronchioles and alveoli, and by salivary glands and nasal mucosa (3, 4, 20, 21). The T-allele was shown to increase the MUC5B promoter activity and carriers of the T-allele demonstrated increased RNA expression of MUC5B in lung tissue (7, 8, 22). A recent paper shows that MUC5B rs35705950 resides within a gene enhancer that is subject to epigenetic remodeling (23). In the airway epithelium of an explant lung of a severe ICU admitted COVID-19 patient, dramatically reduced MUC5B protein and mRNA expression was found when compared with control lung (24).

The increased MUC5B production in T-allele carriers may protect carriers from adverse events related to airway defense. This may be of particular importance in aging, because mucus production and mucociliary clearance have been described to decrease with aging (5, 6). Furthermore, decreased mucociliary clearance may underlie the observed age-related increase in the incidence of severe community-acquired pneumonia in the elderly (25). Similar to previous reports on COVID-19 (1) we also observed that severe COVID-19 and death from COVID-19 is predominantly found in the elderly.

Severe COVID-19 is primarily associated with a respiratory system pathology. Autopsy of patients who died from COVID-19 showed presence of diffuse chronic and tracheobronchial inflammation and alveolar type 2 (AT2) cell hyperplasia in the majority of biopsies (26). SARS-CoV-2 virus was detected in both airway epithelium and AT2 cells and the authors concluded that histology suggests progressive disease that begins in the airway and extends to alveolar zones (26). High constitutive levels of MUC5B in the elderly may protect the airway from SARS-CoV-2 viral infection.

Death rates of the Dutch and UK COVID-19 cases are high. The UK biobank positive COVID-19 cases even had a death rate of 27% which is considerably higher than the maximum 15.2% (27) evolving case fatality rate in the UK from mid-March to end May 2020. After community testing was rolled out, the case fatality rate in the UK decreased below 4% (27). The high death rate of affected UK biobank cases may be due to the older age of biobank participants with a median of 69 years old. With increasing age, the lung changes to the extent that alleles which in younger people confer non-essential divergent expression, may influence the risk of disease in aged tissue. In aging lung diseases such as IPF, COPD and lung cancer, a pattern is emerging of shared disease loci. Although loci are shared, it is of considerable interest that the respective diseases associate with opposite risk alleles and with opposing expression levels (10, 11, 28). Previously we summarized findings and presented a theory in which trade-offs in the aging respiratory system exist (11). The present study shows that the MUC5B rs35705950 polymorphism may be added to this list of shared loci with opposite risk alleles. The MUC5B T-allele, which appears beneficial in this study, is best known as a major risk allele for IPF (7). IPF is a rare non-infectious pulmonary aging disease of unknown cause characterized by insidious onset of disease in patients without a history of pulmonary health problems. Subsequent studies showed that the MUC5B T-allele not only predisposes to IPF but to a variety of chronic progressive forms of pulmonary fibrosis (29–32).

Because severe COVID-19 is associated with substantial pneumonitis and shares multiple risk factors with IPF, Fadista et al. recently investigated whether a genetic correlation between IPF and severe COVID-19 exists using a Mendelian randomization approach (33). They found that genetically increased risk of IPF indeed associated with increased COVID-19 severity, except for the MUC5B allele. The MUC5B risk allele had a different effect compared with other IPF predisposing alleles and protected against COVID-19 hospitalization in the elderly. Because the MUC5B results contradicted their hypotheses the authors were concerned that the analysis might have been influenced by possible selection bias: 1) due to the rs35705950 T allele carriers undertaking strict self-isolation, and 2) due to survival bias of the rs35705950 non-IPF T allele carriers (33). With the unique data of the UK biobank cohort, we were able to address these questions. First, the MUC5B T allele is only known to be associated with progressive fibrosing ILD. These patients may have been isolating due to clinical vulnerability. However, instead of underrepresentation, we were able to show the significant overrepresentation of patients with ILD in the biobank COVID-19 cases vs. non-cases. These data include 5 IPF cases and 1,014 IPF controls, bias introduced by effective shielding of these patients is therefore not present. Second, it is suggested that MUC5B T-allele carriers may have increased survival in the population. This would result in increasing T-allele frequencies with age. However, we found a very slight decrease in the T-allele frequency with age. When we delete ILD subjects from the cohort, the association remained, meaning that there is no survival bias of T-allele carriers in the non-ILD population.

The finding by Fadista et al. is in line with our current findings and we found no evidence for stratification bias driving these results. We performed a candidate allele study because we hypothesized that the IPF predisposing allele would protect against COVID-19 and confirmed this hypothesis. The current finding of protection against severe COVID-19 combined with the established increased risk for pulmonary fibrosis in T-allele carriers may represent a trade-off that becomes apparent with aging. During the first decades of life the effect of both alleles may be neutral while at an older age differences in constitutive expression levels may predispose to disease. The pleiotropic property of the gene polymorphism is expressed only in the older individuals. This idea complements the well-established principle of pleiotropic antagonism, the theory of aging where one gene is involved in multiple traits (pleiotropy) with a beneficial fitness enhancing effect in early life and a detrimental fitness diminishing effect in later life (34).

A limitation of the study is the focus on white European populations. Minor allele frequencies for MUC5B rs35705950 are known to differ between populations. The allele frequencies of the control cohorts are congruent with previous reports (32, 35–37). The protective T-allele is known to have the highest frequency in populations of European ancestry, but is less frequent to non-existent in non-European populations. It is tempting to speculate that the increased risk for infection with SARS-CoV-2 and the worse clinical outcome in black, Asian and minority ethnic populations in western societies (38) may be associated with low carriage of the protective MUC5B T-allele. Replication cohort I and II are part of the study population used by Fadista et al. (33). In their study, a 89% white patient population and a 99% white control cohort was used, therefor they adjusted the analysis for genetic ancestry principal components (33). We used a white Dutch and UK population, so in order to replicate our findings and allow comparison of the results we included replication cohorts I and II and performed the analysis on white subjects only, which produced similar results. However, future studies aimed at improving understanding of COVID-19 risk in populations worldwide should include genetics of different ethnic groups. Another limitation is the small sample size of the Dutch cohort, yielding a significant result but with a wide confidence interval. Furthermore, specific information on disease severity such as organ involvement, CO-RAD and CT-severity scores are missing. However, all patients in the Dutch cohort were hospitalized for confirmed COVID-19 and had a SpO₂ <94%. Hospital triage during the study period was restricted because the wards were overcrowded. Furthermore, 22% of patients were admitted to the intensive care unit and 24% of patients died. In addition, the death rates of the UK COVID-19 cases was high (27%), all suggestive of case cohorts with severe COVID-19. However, further studies are needed to investigate if the MUC5B polymorphism will associate with specific COVID-19 severity scores.

A strength of our study is the inclusion of the UK Biobank cohort, with cases and controls having been recruited as one cohort 9–13 years prior to the COVID pandemic. This procedure is most ideal to avoid recruitment bias in case-control studies and the cohort yields a highly significant result.

In conclusion, we found that carriage of the T-allele of MUC5B rs35705950 confers protection from development of severe COVID-19. Because the T-allele is a known risk allele for pulmonary fibrosis, this study provides further evidence for the existence of trade-offs between optimal expression levels in the aging lung.

The datasets presented in this article are not readily available because they contain potentially identifying or sensitive patient information. Requests to access the datasets can be directed to the corresponding author.

The ILD biobank and data registry of the St. Antonius Hospital was reviewed and approved by the Medical research Ethics Committees United (MEC-U) of St. Antonius Hospital. The UK Biobank study was reviewed and approved by the National Research Ethics Service Committee North West-Haydock. All participants provided their written informed consent to participate in this study.

The study was conceived and designed by CM and JV. CB, CS, AD, DE, and MQ contributed to the conduct of this study. Data were acquired by MQ, JV, AD, and DE and analyzed by AD and JV. CM reviewed the medical literature, oversaw the conduct of the study, participated in the interpretation of data, drafted, and wrote the manuscript. All authors reviewed and contributed to the manuscript during its development and approved it for publication.

This study was funded by ZonMW TopZorg St Antonius Care grant nr 842002001; ZonMW Topspecialistische Zorg en Onderzoek grant no 10070012010004; Nederlandse Vereniging van Artsen voor Longziekten en Tuberculose COVID-19 grant; GW4 BioMed Medical Research Council Doctoral Training Partnership. Funders were not involved in the study design, collection, analysis and interpretation of the data.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.