A Winogradsky-based culture system shows an association between microbial fermentation and cystic fibrosis exacerbation

Abstract

There is a poor understanding of how the physiology of polymicrobial communities in cystic fibrosis (CF) lungs contributes to pulmonary exacerbations and lung function decline. In this study, a microbial culture system based on the principles of the Winogradsky column (WinCF system) was developed to study the physiology of CF microbes. The system used glass capillary tubes filled with artificial sputum medium to mimic a clogged airway bronchiole. Chemical indicators were added to observe microbial physiology within the tubes. Characterization of sputum samples from seven patients showed variation in pH, respiration, biofilm formation and gas production, indicating that the physiology of CF microbial communities varied among patients. Incubation of homogenized tissues from an explant CF lung mirrored responses of a Pseudomonas aeruginosa pure culture, supporting evidence that end-stage lungs are dominated by this pathogen. Longitudinal sputum samples taken through two exacerbation events in a single patient showed that a two-unit drop in pH and a 30% increase in gas production occurred in the tubes prior to exacerbation, which was reversed with antibiotic treatment. Microbial community profiles obtained through amplification and sequencing of the 16S rRNA gene showed that fermentative anaerobes became more abundant during exacerbation and were then reduced during treatment where P. aeruginosa became the dominant bacterium. Results from the WinCF experiments support the model where two functionally different CF microbial communities exist, the persistent Climax Community and the acute Attack Community. Fermentative anaerobes are hypothesized to be the core members of the Attack Community and production of acidic and gaseous products from fermentation may drive developing exacerbations. Treatment targeting the Attack Community may better resolve exacerbations and resulting lung damage.

Figure 1

(a) Schematic of the WinCF model showing setup and principles of microbially induced physiochemical changes after incubation. (b) Fluorescent ruthenium oxygen optode before and after incubation with a pure culture of P. aeruginosa in the WinCF model showing removal of oxygen. (c) 40 × magnified capillary tube plug. (d) Capillary tube gas bubbles. (e) WinCF model image of a pure culture of P. aeruginosa (left panel) and an exacerbation sputum sample (right panel) showing the different physiochemical responses.

Figure 2

Box plots of the quantitative measures of the triplicate WinCF tubes from the MATLAB output of the capillary tube images. Each tube is measured independently in the MATLAB script, and the distribution of these measures for each sample is shown.

Figure 3

Fluctuations in WinCF model variables through two separate exacerbation events in a single CF patient. Quantitative MATLAB output of the images is plotted through time where each point represents a sputum sample. Arrows in the plot indicate where antibiotics were administered for an exacerbation, the black arrow was intravenous and the striped arrow was oral administration of antibiotics.

Figure 4

Significantly elevated gasses (permutation test) detected from WinCF capillary tube headspace using gas chromatography and mass spectrometry. Gasses were captured after inoculation with a sputum sample from patient CF1 and incubation at 37 °C in sealed glass containers.

Figure 5

(a) Microbial community composition indicated by sequencing of V4 region of 16S rRNA gene from cross-sectional WinCF and sputum samples in both CF and non-CF individuals (genera with at least 5000 total reads from Miseq Illumina sequences analyzed with mothur). Profiles from the same sputum sample are paired in the graph to demonstrate the WinCF community that developed from the same sputum sample. (b) 16S rRNA gene microbiome profile of WinCF communities from longitudinal sputum samples through two exacerbations in one patient. The clinically determined disease state is indicated for each sample along with the days since collection of the first sample.

Figure 6

Multidimensional scaling plot of a Bray–Curtis distance matrix comparing output of quantitative WinCF model variables for all samples in this study. Squares are longitudinal sputum samples where blue squares are exacerbation samples, purple squares are during treatment and orange squares are during stability. Circles are cross-sectional sputum samples from different CF (red) and non-CF (yellow) patients. Crosses are various regions of a homogenized explant lung, each spot representing a different location in the lung. The oval is a pure culture of P. aeruginosa.