Background
Chronic obstructive pulmonary disease (COPD) is a lung disease characterised by airflow-limiting inflammation and mucus production. Short periods of worsening symptoms, called exacerbations, progressively aggravate the disease [
1]. COPD is the most common chronic respiratory illness in older adults and the third leading cause of mortality worldwide, responsible for about 6% of all global deaths in 2019 [
2]. COPD substantially impacts quality of life and imposes a heavy socioeconomic burden [
3]. The total annual cost of COPD in Europe was €141.4 billion in 2011 [
3].
Acute Exacerbation of COPD (AECOPD) is a major cause of COPD-related morbidity and mortality and accounts for a large proportion of COPD’s economic burden [
3‐
7]. AECOPD is triggered by various causes, such as pneumonia, pulmonary embolism, or inhalation of irritants [
1], but is mostly associated with viral and bacterial infections. Bacterial infections alone are associated with about 50% of exacerbations, and non-typeable
Haemophilus influenzae (NTHi),
Streptococcus pneumoniae, and
Moraxella catarrhalis (Mcat) are the bacteria most often identified [
8].
One of the strategies to prevent exacerbations would be vaccinating COPD patients against the most common pathogens causing AECOPD. Vaccines against Streptococcus pneumoniae already exist as they were developed to prevent streptococcal pneumonia and meningitis. No vaccine exists against NTHi and Mcat. Since NTHi lacks a polysaccharide capsule, vaccine development efforts have been concentrated on identifying NTHi surface proteins that are immunogenic and highly conserved within the species.
A candidate vaccine against NTHi and Mcat composed of three NTHi antigens (type IV pilin protein [PilA], protein E [PE], and protein D [PD]) and one Mcat antigen (ubiquitous surface protein A2 [UspA2]) has been developed and tested in a Phase 2 efficacy trial [
9]. PilA is a pilus subunit involved in biofilm formation and mobility whereas PE and PD are involved in serum resistance and adherence. UspaA2 mediates both bacterial adherence and complement-mediated killing resistance through the binding of extracellular matrix proteins, including vitronectin, collagen, and laminin. To improve our understanding of the role of infections in AECOPD, the prospective longitudinal epidemiological study Acute Exacerbation and Respiratory InfectionS in COPD (AERIS) was conducted [
10]. With the current study, we aimed to build upon the findings from AERIS by assessing the expression of NTHi-Mcat vaccine candidate genes in sputum of COPD patients who participated in the AERIS study. Sputum samples were processed to limit their possible alterations and to allow RNA quantification in conditions mirroring in-vivo situation, and referred to as “ex-vivo” in this manuscript. Additionally, we determined whether expression of such genes differed between samples at stable routine visits (ST) and exacerbation visits (EX), in relation to COPD severity and AECOPD frequency.
Discussion
In a prospective, observational cohort study we analysed expression of NTHi-Mcat vaccine candidate genes in 174 sputum samples from 69 COPD patients positive for the presence of NTHi and/or Mcat bacteria. We selected patients that had sputum samples positive for the bacterial pathogens during at least one ST and one EX. We showed for the first time that the expression of almost every NTHi-Mcat vaccine candidate gene was detectable in sputum samples and that expression was detectable during both ST and EX. In addition, we found that there was no difference in expression of these genes between ST and EX. Finally, we observed a slight association of pd and pilA gene expression with the number of AECOPD events before enrolment that is completely undetectable when antigen expression is normalized to the housekeeping genes expression level.
Normalization of gene expression data against the expression of multiple household genes has been recommended to limit errors and increase the accuracy of results essential for a good comparison of gene expression from sputum samples [
15]. Therefore, in addition to normalization of NTHi vaccine genes against
ompP6 and
gapA individually
, we also performed normalization to the average combination of these two household genes using the method proposed by Riedel et al. [
14]. The results were not different from the data normalized against the single household genes.
NTHi and Mcat are often present during exacerbation states and these pathogens are therefore thought to play a role in triggering exacerbations [
8]. In this AERIS study, the presence of NTHi and Mcat, but not of other bacteria, was associated with a heightened exacerbation risk [
12], therefore we wanted to investigate whether this association is perhaps due to a modulation of expression of NTHi and Mcat virulence factors during exacerbations compared to stable state.
In a similar study in Spain that was performed from October 2009 to October 2010,
Pseudomonas aeruginosa and
Streptococcus pneumoniae were isolated most frequently from AECOPD sputum while NTHi and Mcat were isolated as well [
16], indicating pathogens may vary by year and/or geographic location.
It was previously concluded, based on the presence of the pathogens in the sputum, that NTHi and Mcat detection was associated with AECOPD occurrence [
12]. The current study confirms this based on the continued expression of key virulence genes during the ST and EX. In addition, because the expression of the housekeeping genes was not different between ST and EX, it appears that the pathogens are not increasing or modulating the quantity of the transcripts of these genes during the exacerbations.
In the AERIS study, analysing the lung microbiome in AECOPD, the relative abundance of
Moraxella was found to be increased in samples taken during exacerbations compared to those taken during ST [
17]. Indeed, in a previous study, the lung microbiome was found to be dynamic, microbiome changes were associated with exacerbations, and the relative abundance of
Moraxella was overall increased, although there was some heterogeneity, as in some patients
Moraxella abundance was decreased at exacerbation [
18]. In that same study, it was found that treatments potentially alter the lung microbiome. In particular, a reduction in microbial diversity and an increased Proteobacteria:Firmicutes ratio were observed in patients treated with steroids alone, whereas the trend was reversed in patients treated with antibiotics [
18]. In this previous study, the technique used to determine the relative abundance of bacteria (16s rRNA gene sequencing) did not allow identification at the species level, only at the genus level [
17,
18], so the changes they found may not be relevant for specific individual species. For the current analyses, we selected ST and EX samples from the same patients who were Mcat or NTHi positive, which is different from the two previous studies, where any samples from ST and EX were analysed regardless of whether they contained the pathogen or not [
17,
18]. Relative abundance of
Moraxella might have been similar in those studies as well if ST and EX samples had been used that were matched for being Mcat positive.
In two small, earlier studies of the lung microbiome, overall microbial diversity was less in severe or very severe patients than in mild-to-moderate or moderate-to-severe COPD patients [
19,
20]. We found that absolute expression of the Mcat housekeeping gene
polA, and absolute and relative expressions of the Mcat gene
uspA2, were not different between ST and EX. These results may indicate that if the total lung microbiome diversity during exacerbations was decreased, while Mcat remained present in the same amount, then we have an increase of the relative amount of Mcat.
One limitation of this study was the limited sample size analysed in addition to the non-consideration of the impact of treatment on the NTHi and Mcat population. Another limitation is that we were not able to analyse gene expression in patients with AECOPD for which no sputum samples were available at stable visits.
A strength of the study is the application of RT-ddPCR, which is an ultrasensitive method for RNA quantification. This method allows for monitoring targets in complex backgrounds such as detecting mRNA from individual bacterial genes in samples that also contain human material and material from various other microbes.
Acknowledgements
The authors thank Nathalie Devos (GSK) for her critical review of this manuscript.
The authors also thank the members of the AERIS Study Group for their contribution to the study: J. Alnajar (University of Southampton, Southampton, UK), R. Anderson (University of Southampton, Southampton, UK), E. Aris (GSK, Wavre, Belgium), W.R. Ballou (GSK, Rockville, MD, USA), A. Barton (University Hospital Southampton NHS Foundation Trust, Southampton, UK), S. Bourne (University of Southampton, Southampton, and Portsmouth Hospitals NHS Trust, Portsmouth, UK), M. Caubet (GSK, Wavre, Belgium), S. C. Clarke (University of Southampton, Southampton, UK), D. W. Cleary (University of Southampton, Southampton, UK), C. Cohet (GSK, Wavre, Belgium), N. A. Coombs (University of Southampton, Southampton, UK), K. Osman (University of Southampton, Southampton, UK), J-M. Devaster (GSK, Rixensart, Belgium), V. Devine (University of Southampton, Southampton, UK), N. Devos (GSK, Rixensart, Belgium), E. Dineen (University Hospital Southampton NHS Foundation Trust, Southampton, UK), T. Elliott (University of Southampton, Southampton, UK), R. Gladstone (Wellcome Sanger Institute, Hinxton, UK), S. Harden (University Hospital Southampton NHS Foundation Trust, Southampton, UK), J. Jefferies (University of Southampton, Southampton, UK), V. L. Kim (University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK), P. Moris (GSK, Rixensart, Belgium), K. K. Ostridge (University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK), T. G. Pascal (GSK, Wavre, Belgium), M. Peeters (GSK, Rixensart, Belgium), S. Schoonbroodt (GSK, Rixensart, Belgium), B. Sente (GSK, Rixensart, Belgium), K. J. Staples (University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK), A. C. Tuck (University of Southampton, Southampton, UK), L. Welch (Solent NHS trust, Southampton, UK), V. Weynants (GSK, Wavre, Belgium), T. M. A. Wilkinson (University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK), A.P. Williams (University of Southampton, Southampton, UK), N. P. Williams (University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK), C. Woelk (University of Southampton, Southampton, UK), M. Wojtas (University of Southampton, Southampton, UK), and S. A. Wootton (University Hospital Southampton NHS Foundation Trust, Southampton, UK).
The authors thank Business & Decision Life Sciences platform for editorial assistance and manuscript coordination, on behalf of GSK. Esther van de Vosse (Business & Decision Life Sciences, on behalf of GSK) provided medical writing support.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.