Introduction

Cystic fibrosis (CF) is a genetic disease due to CF transmembrane conductance regulator (cftr) gene mutations (mostly p.F508del mutation). From another point of view, CF can also be considered as an infectious disease, as lung colonisation and infection aggravate the clinical condition and are the main causes of morbidity-mortality in people with CF (PWCF). Key bacterial pathogens (Staphylococcus aureus, Haemophilus influenzae, Pseudomonas aeruginosa and Burkholderia cepacia complex) are known to chronically colonise CF lungs and cause pulmonary exacerbation1. However, CF pulmonary microbiota harbours a wider range of bacterial communities that include certain species thought to play an important role in respiratory disease2. An accurate description of the lung bacterial species, many of which have yet to be identified3,4, could lead to a better understanding of the lung disease2. Strict anaerobes play an important role among these as yet unknown species, and are undeniably important residents of the CF lung5,6, that have been detected in respiratory samples at all ages4. Anaerobes’ growth is stimulated by anoxic conditions, provided partially by mucus hyperviscosity and oxygen consumption by neutrophil afflux or bacterial proliferation5,7. The most frequently detected genera in CF lungs are Prevotella, Veillonella, Fusobacterium, Atopobium, Peptostreptococcus and Porphyromonas4,5,8. However, in spite of their wide diversity and abundance, the role of airway anaerobic bacteria is still controversial. On the one hand, anaerobes exhibit their own virulence factors such as proteases9,10, can enhance others pathogens virulence11,12,13,14 or can promote antibiotic resistance15,16. On the other hand, they can be associated with less lung inflammation6,17 and better lung function4,6,18,19,20. Hence, it remains under debate whether the presence of anaerobes is beneficial and whether targeted antibiotic therapy against anaerobic bacteria is needed or should rather be avoided in case of CF pulmonary exacerbation4,21. In the light of these contradictory observations, we conducted an observational study to go deeper in deciphering anaerobe diversity at species level within the CF lung microbiota. We implemented a dual approach, combining extended-culture and targeted molecular techniques for optimally exhaustive and accurate identification and quantification of anaerobic communities.

Results

Subjects and samples

From March to October 2018, 150 sputa from 101 PWCF were collected in the study: one sample for 57 patients, and two or three samples for 54 patients. Based on Murray-Washington criteria22,23 (Table S1), salivary contamination was low in 58.0% of sputa (87/150), intermediate in 34.0% (51/150), and high in 6.0% (9/150); 2.0% of sputa (3/150) were undetermined because of poor quantity (Fig. 1). The nine samples with high salivary contamination were excluded from further analysis (five patients) and 29 sputa underwent molecular analyses but not extended-culture due to insufficient volume (16 patients) (Fig. 1). After exclusion, a total of 96 PWCF were enrolled in the study; subject characteristics are shown in Table 1 (and in Table S2 for the 80 PWCF whose samples were analysed only by extended-culture approach). According to the Leeds definition for P. aeruginosa colonisation24, the majority of the cohort were categorised as chronic or intermittent (72.9%; 70/96). Concerning lung function, 80.3% of patients (77/96) presented Forced Expiratory Volume in one second (FEV1) inferior or equal to 70%.

Figure 1
figure 1

Flow chart of the patients and samples included in the study.

Full size image
Table 1 Study cohort (96 people with cystic fibrosis): demographic and clinical data.
Full size table

Extended-culture approach for anaerobic bacteria isolation and identification at species level

On the extended-culture approach, 692 strict anaerobic strains were isolated from the 112 sputa analysed: 84.8% were identified by MALDI-TOF MS (587/692), and 15.2% by 16S rRNA gene sequencing (105/692). A total of 91.1% of sputum samples (102/112) were positive for at least one anaerobic species. After the exclusion of duplicate species per patient (Table S3), an average of six anaerobic bacteria were isolated per patient (range 0–16). Thirty-one strict anaerobic genera were cultured from the 112 sputum samples, comprising 69 species (Fig. 2a; Table S4). The 10 most frequent genera were: Prevotella (43.1%; 298/692), Veillonella (15.0%;104/692), Atopobium (6.1%; 42/692), Fusobacterium (5.2%; 36/692), Parvimonas (2.9%; 20/692), Peptostreptococcus (2.7%; 19/692), Solobacterium (2.7%; 19/692), Eubacterium (2.6%; 18/692), Megasphaera (2.0%; 14/692), and Leptotrichia (1.9%; 13/692) (Fig. 2a). Anaerobic genera composition within the different age- and lung function- groups is detailed in Fig. 2b/2c. Two anaerobic genera, poorly represented but both isolated from high-quality sputum (Murray-Washington score at 5), were first detected in the CF lung: Pyramidobacter (0.1%; 1/692; species Pyramidobacter piscolens), and Varibaculum (0.1%; 1/692; species Varibaculum anthropi). At a finer taxonomic level, a wide diversity was found: 17 Prevotella species (P. baroniae, P. buccae, P. denticola, P. histicola, P. intermedia, P. loescheii, P. maculosa, P. marshii, P. melaninogenica, P. nanceiensis, P. nigrescens, P. oralis, P. oulorum, P. pallens, P. salivae, P. timonensis, and P. veroralis) five Veillonella species (V. atypica, V. denticariosi, V. dispar, V. parvula, and Veillonella sp.), and five Fusobacterium species (F. canifelinum, F. naviforme, F. nucleatum, F. periodonticum, and Fusobacterium sp.) (Table S4).

Figure 2
figure 2

Extended-culture approach results. (a) Description of strict anaerobic genera identified in 112 sputum samples: Gram-positive genera are shown in purple, Gram-negative genera in pink, and number of species in brackets. (b) Prevalence of strict anaerobic genera identified according to patient age group (80 patients). (c) Prevalence of strict anaerobic genera identified according to patient lung function evaluated by Forced Expiratory Volume in one second (FEV1) (80 patients).

Full size image

Molecular approach for detection and quantification of Prevotella, Veillonella and Fusobacterium genera

qPCR targeting these three anaerobic genera in the lung confirmed that Prevotella was the most abundant genus in sputa (median of quantification: 1.34 × 106 CFU/mL), compared to Veillonella (median of quantification: 2.71 × 104 CFU/mL) and Fusobacterium (median of quantification: 2.55 × 102 CFU/mL). The molecular approach was significantly more positive than culture for the detection of the less abundant genera, Veillonella and Fusobacterium (p < 0.01 respectively, McNemar test) (Table S5).

Relationship between lung function and Prevotella, Veillonella and Fusobacterium quantification

Lung function was evaluated by FEV1, with a 70% clinical threshold25. Statistical analysis (univariate analysis) of results for the 96 patients (141 samples) showed that a one-log increase of Prevotella, Veillonella and Fusobacterium quantification (CFU/mL) was significantly associated with a FEV1 superior to 70% (OR = 1.41, p = 0.045; OR = 1.92, p < 0.001; OR = 1.45, p = 0.032 respectively). Multivariate analysis confirmed that a one-log increase of the quantity of the genus Veillonella was associated with a better lung function (FEV1 > 70%) (p = 0.02, logistic regression) (Table 2). Except for patient age, no other significant associations were found between lung function and the clinical parameters (gender, body-mass index, oral antibiotic therapy, CFTR modulator treatment, or P. aeruginosa positive sputum culture) (Table 2). Moreover, all 13 samples in which Veillonella was not detected by qPCR came from patients with FEV1 < 70% (10 patients, median age: 36 years). Focusing on the three most identified Veillonella species by culture approach (V. atypica, n = 34; V. dispar, n = 21; V. parvula, n = 47), V. dispar was the most associated with a better lung function (FEV1 > 70%) but this association was not significant (OR = 3.50, p = 0.08).

Table 2 Multivariate analysis: logistic regression modelling the probability of Forced Expiratory Volume in one second (FEV1) > 70%.
Full size table

Discussion

CF lungs harbour diverse communities of bacteria, the abundance and variety of which fluctuate depending on multiple factors: clinical status, treatment and environment3,26,27,28. Within this rich and diversified bacterial microbiota, anaerobes are able to grow in anoxic areas in CF lungs and have been detected in higher quantities in PWCF than in healthy individuals5. These oxygen-sensitive bacteria come from diverse reservoirs (mainly oral and digestive)29 and can be detected in equal or higher proportions than some CF pathogens in lung4. Moreover, numerous anaerobic genera have been identified as part of the CF core lung microbiota: Prevotella, Veillonella, Porphyromonas, Fusobacterium, Catonella and Peptostreptococcus20,28,30. The recent literature emphasised the importance of anaerobic bacteria in lung, but results and hypotheses concerning their role in the pathophysiology and progression of CF disease are contradictory4,6,9,10,11,12,13,14,15,16,17,18,19,20,21,31,32. Anaerobes’ impact is mainly evaluated according to genus rather than species, which may explain these conflicting results, as some species within a given genus may generate contrasting effects33. Thus, anaerobic microbiota description should be exhaustive, as accurate as possible and frequently implemented.

To this end, culture approaches seem to be appropriate to provide descriptive information of anaerobe diversity at species level34. However, due to the fastidiousness of implementing a strictly anaerobic atmosphere, lower sensitivity than molecular approaches, the effect of previous antibiotic therapy, and growth limited to “culturable” bacteria, most descriptive studies of the lung “anaerobiome” are currently provided by 16S-targeted metagenomics. However, the targeting of 16S variable regions with most of the sequencing platforms used did not allow the optimal taxonomic resolution to be achieved35. We decided to conduct an observational study, based on a dual approach combining extended-culture and targeted molecular techniques to go deeper in deciphering anaerobe diversity within the CF lung microbiota. To evaluate sputum quality, we used the Murray-Washington cytological score, which showed that 94.0% of the sputum samples (141/150) were properly sampled and informative. Our findings supplement previous culture data on anaerobe diversity4,5,6,8. Thanks to the innovative collecting device and the culture-extensive protocol (21 days of incubation) coupled with both MALDI-TOF MS and 16S rRNA gene sequencing, 31 strict anaerobic genera and 69 species were described in 112 CF sputa. Considering the seven studies that described strict anaerobes in the CF lung microbiota based on an extensive-culture approach (Table 3), our study provided the greatest diversity ever described so far. Mirkovic et al.13 found 52.3% positive respiratory samples with 8 anaerobic genera detected from 109 samples, Muhlebach et al.4 found 67.0% positive sputa with 18 anaerobic genera detected from 200 samples; O’Neill et al.6 found 12 genera from 41 sputum, Paganin et al.36 found five genera from 78 sputum, Sherrard et al.37 found 23 anaerobic genera from 199 sputa; Sibley et al.38 found 15 genera from 246 sputum, and Tunney et al.5 found 64.0% positive sputa with seven genera from 66 samples. Focusing on the most recent study which reported 18 anaerobic genera on culture in CF lung4, all of them were also detected in the present study, supplemented by description of 13 additional genera. To our knowledge, two anaerobic genera (Pyramidobacter and Varibaculum) were described here in CF lung for the first time in sputa with low salivary contamination. These culture results were contributive as they demonstrate the viability of anaerobes in CF lung, which might not be clear using molecular techniques, and they shed light on anaerobic communities in CF lung according to species. On the molecular side of the study, qPCR improved detection compared with culture for two anaerobic genera (Veillonella and Fusobacterium) which were represented at lower rates in sputum samples. These results showed that the combination of both approaches is essential for optimally exhaustive anaerobic microbiota description.

Table 3 Comparison of studies based on an extended-culture approach for the CF “anaerobiome” description.
Full size table

As well as being descriptive, this study demonstrated a significant positive association between Veillonella quantification and lung function (FEV1 > 70%). This was already seen in previous studies, for the genus Veillonella6 and also for the genus Prevotella18. Moreover, although the observation was based on a small number of samples (n = 13), the genus Veillonella was absent only in sputa from PWCF with poor respiratory function (FEV1 < 70%). Consequently, the genus Veillonella and associated species detected on culture (V. parvula, V. atypica, V. dispar) could in future be used as biomarkers to predict lung function, as previously shown for the anaerobic genus Porphyromonas which could be used to predict earlier P. aeruginosa infection39. In our study, there is a tendency of association of the species V. dispar with a better lung function (non-significant association). The lung role of this species in PWCF is as yet unknown but a recent study in healthy people has shown that lung bacteria are the same as oral or nasal bacteria, but that some, like V. dispar, are enriched in the healthy lung40. This finding suggests that V. dispar may play an important role in lung health40, which needs further investigation. These results also pave the way for discussion of the protective role of anaerobic bacteria. However, little is known about the pathophysiology and mechanisms involved in this potential positive impact, although significant positive correlations were demonstrated between anaerobes and better lung function or/and less airway inflammation, especially for the genera Prevotella, Veillonella and Porphyromonas1,4,6,18,20.

The present study and the methods used here are subject to some limitations. Firstly, the cohort was composed of relatively old PWCF (median age, 27 years; only 12 patients under 18 years of age), which might constitute a recruitment bias for the description of anaerobic bacteria in CF lung, as diversity may be impacted by patient age17,41. Secondly, as the design was single-centre, geographical background was not taken into account, although it may have an impact on the richness and diversity of the anaerobic microbiota. Thirdly, molecular studies were performed only on three anaerobic genera (Prevotella, Veillonella and Fusobacterium). The anaerobe panel needs to be extended, and should target species rather than genera to accurately evaluate the impact of anaerobes on lung function. Fourthly, the Murray-Washington criteria were used to determine sputum quality and to exclude samples suspected of high salivary contamination. However this score does not make it possible to assert the only pulmonary location of the anaerobic bacteria detected. Fifthly, FEV1 may not be the most relevant lung parameter to evaluate lung function. Lung clearance index (LCI) was reported to show greater sensitivity than FEV1 and better correlation with clinical outcome. However, LCI is now indicated mainly for trials in young PWCF or with early or mild lung disease42, which were not the main conditions in our cohort. All in all, new longitudinal studies are required to better understand anaerobic microbiota dynamics and impact on respiratory function, depending on the patient’s clinical status, age (focusing on paediatric patients) and antimicrobial treatments, with inclusion in several CF centres.

In conclusion, this cohort study revealed the largest anaerobe diversity ever reported in CF lungs, thanks to preservation of anaerobic atmosphere in sputa and the powerful performance of the extended-culture approach. Moreover, the study highlighted that a greater quantity of the genus Veillonella was significantly positively associated with better lung function. Consequently, anaerobes should not be underestimated within the lung microbiota. Subsequent longitudinal studies are needed to determine and understand anaerobe impact and pathophysiology according to species and to evaluate anaerobe antibiotic sensitivity under antibiotic pressure.

Methods

People with cystic fibrosis (PWCF) and sample processing

This observational study was conducted between March and October 2018. Samples were collected in the Western Brittany CF centre (Roscoff, France) during follow-up consultations. As samples consisted of sputa, only PWCF able to expectorate were enrolled. Pulmonary transplant patients were excluded. The following sociodemographic and clinical parameters were recorded: age, gender, cftr gene mutation, pancreas status, body-mass index, lung function, diabetes, and treatments (oral antibiotic therapy one month before sampling, chronic antibiotic therapy, corticosteroids, CFTR modulators). A collecting device was specifically designed and patented for the study to ensure preservation of an anaerobic atmosphere for the sputa (EP 20305133.9) (Fig. S1). Between collection and transfer to the anaerobic chamber, samples were conserved at + 4 °C.

Extended-culture approach for bacterial isolation and identification

Sputa quality and salivary contamination were evaluated on Murray-Washington score22,23 (Table S1). Each sputum sample was mixed with an equal volume of dithiothreitol (Digest-EUR, Eurobio, Courtaboeuf, France), and incubated at 37 °C for 30 min to one hour in accordance with the manufacturer’s instructions. One-thousand-fold dilution was prepared in physiological saline supplemented with 0.05% (wt/vol) L-cystein (Sigma-Aldrich, Dorset, UK). One hundred microliters of liquefied diluted sputum was plated onto three different media: anaerobe basal agar with sheep blood (ABA-SB) (ThermoFisher, Waltham, USA), kanamycin-vancomycin ABA-SB, and colistin-nalidixic acid ABA-SB. Plates were incubated in an anaerobic chamber (90% N2, 5% H2, 5% CO2) (Bactron, Sheldon Manufacturing, Cornelius, USA) for 21 days at 37 °C and readings were made every 2–5 days. Isolates of each distinct colony type were subcultured onto ABA-SB, and identified by matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI-TOF MS Biotyper MBT) (Bruker, Billerica, USA) (Supplementary Information). All strict anaerobes isolates were stored at − 80 °C (Microbank, Pro-Lab Diagnostics, Ontario, Canada). In case of mass spectrometry failure (database or technical problems), a 16S rRNA gene (rrs) sequencing method was used for isolate identification. The following primers were used for amplification of the rrs gene (996 bp): sense 5′-CCAGCAGCCGCCGTAATACG-3′; antisense 5′-TACGGYTACCTTGTTACGACT-3′. The PCR reaction was conducted using the Eppendorf Mastercycler Gradient instrument (Eppendorf, Hamburg, Germany). PCR sequencing used the Big Dye Terminator Cycle Sequencing Kit Ready Reaction version 3.1 (Applied Biosystems, Courtaboeuf, France), according to the manufacturer’s recommendations, on an Applied Biosytems 3130xl instrument (ThermoFisher, Waltham, USA). Sequences were then aligned and corrected using the software BioEdit version 7.0.5.3 (http://www.mbio.ncsu.edu/bioedit/bioedit.html). Comparison with the NCBI BLAST database (Basic Local Alignment Search Tool, http://blast.ncbi.nlm.nih.gov/Blast.cgi) enabled bacterial identification, with ≥ 98% similarity determining homology of identity. In parallel to the anaerobic extended-culture approach, all samples were cultured in aerobic conditions to detect usual pathogens such as P. aeruginosa.

Detection and quantification of Prevotella, Veillonella and Fusobacterium genera by quantitative PCR (qPCR)

qPCR was performed for detection and quantification of three anaerobic genera (Prevotella, Veillonella, Fusobacterium) in each sputum sample. 200 µL aliquots of liquefied sputum were stored at − 80 °C. Samples were treated by five minutes-sonication using a bath sonicator (Elmasonic S10, Elma Schmidbauer GmbH, Singen, Germany). After 10 min-centrifugation (5000 g) and prior lysis by proteinase K, total DNA was extracted using the QIAamp DNA Minikit (Qiagen, Courtaboeuf, France) according to the manufacturer’s instructions, with an elution volume of 100 µL. The qPCR reaction was conducted using an ABI Prism 7500 Fast Real-Time PCR System instrument (ThermoFisher, Waltham, USA). Primers (and probe) used for qPCR are detailed in Table S643,44,45. Specificity and efficiency were evaluated for each primer and probe pairs (Supplementary Information).

Statistical analysis

McNemar’s test was used to compare the percentage of positive samples according to culture and molecular approaches. A logistic regression was used for binary outcome (FEV1 > 70%) to estimate odds-ratios relative to a one-log increase of the quantification of the anaerobic genera and to the absence/presence of Veillonella species (V. atypica, V. dispar, V. parvula) in univariate analysis. Only the most associated genus (Veillonella) and species (V. dispar) were considered in the multivariate model, adjusting for potential confounders (age, gender, body-mass index, CFTR modulators, oral antibiotic treatment and P. aeruginosa positive culture). Intra-patient correlation due to repeated samples was handled thanks to an exchangeable correlation structure, using the SAS software GENMOD procedure (SAS Institute, Cary, USA) with binomial distribution and logit link function.

Ethics

The study was approved by the French Ethical Research Committee in March 2018 (2018-A00624-51). All experiments were performed in accordance with relevant guidelines and regulations. Informed consent was obtained from all participants and/or their legal guardians.

Conference presentation

The study was presented in part at the 39th “Réunion Interdisciplinaire de Chimiothérapie Anti-Infectieuse-RICAI”, 16–17 December 2019 in Paris, France.