2-pentylfuran: biosynthesis by Aspergillus fumigatus in-vitro and investigation of confounders of a breath test in-vivo

Abstract

A. fumigatus is the principle cause of fatal invasive pulmonary aspergillosis (IPA) in severely immunocompromised hosts and rapid treatment is compromised by the limitations in the current diagnostic tests. 2-pentylfuran (2-PF) is produced in-vitro by A. fumigatus and detectable by gas chromatography with tandem mass spectroscopy (GC-MS/MS) in the breath of patients with chronic lung disease who are colonised with this organism (sensitivity 77%, specificity 78%). The 2-PF breath test has the potential of being a rapid and reliable test for the diagnosis of IPA.

The aims of this thesis are: 1) to confirm that 2-PF is produced consistently by A. fumigatus under conditions similar to the microenvironment of lung tissue, and identify a biochemical pathway for 2-PF biosynthesis, 2) to demonstrate that 2-PF is detectable in the breath of patients with IPA, 3) to identify causes of false positive breath tests, and 4) to validate a 2-PF breath test in an animal model. Using GC-MS/MS, A. fumigatus cultured on potato dextrose agar in closed vials produced higher 2-PF levels in the vegetative phase than conidiating phase (P<0.001). Both vegetative and conidiating cultures incubated under 0% and 20% oxygen flow produced 2-PF (P<0.05). To determine whether oxygenases contributed to the synthesis of 2-PF mutant strains were studied. In A. fumigatus, RNAi silenced triple ppo (precocious sexual inducer (psi) producing oxygenase) strain (RNAippoABC) produced 36% less, and the double lipoxygenase (lox) mutant (∆loxAB) produced 43% less 2-PF than the wild-type AF293 strain (p<0.05). In A. nidulans, the triple ppo mutant (∆ppoABC) produced 91% less 2-PF than the wild-type strain (p<0.001). The 2-PF breath test was repeatedly positive in two clinical cases of IPA and became negative with effective antifungal treatment. Subsequently 2-PF was found acontaminant of environmental air and glass breath collection bulbs so that Tedlar® breath collection bags were substituted. The precision and accuracy of this breath analysis method were 90.40% and 94.89%, respectively, and the empirical limit of quantitation for 2-PF was ~1.13 attograms per 2L. Foods containing 2-PF included soy products and vegetable oils. There was no detectable difference in 2-PF levels between morning and afternoon, or fasting and non-fasting breath samples in healthy subjects consuming a 2-PF poor diet. 2-PF was present in breath samples immediately after a soy milk mouth rinse (p<0.001), and intermittently after soy milk ingestion followed by a water mouth rinse. In a pilot study, a small proportion of subjects with pneumonia (consolidation) (5 of 19) or acute exacerbations of COPD (2 of 22) had quantifiable 2-PF in breath samples but this was not significantly different from controls. These findings do not exclude the possibility that 2-PF is produced by an inflammatory response in-vivo. The development of a sheep model of IPA was hindered by the susceptibility of sheep to cyclophosphamide (CPA) toxicity but quantifiable 2-PF was not present in the 115 breath samples of eight normal sheep or in sheep with CPA-induced diffused alveolar damage. To our knowledge, this is the first report providing evidence that 2-PF is an endogenous metabolite of A. fumigatus and is synthesized via oxygenase pathway. A 2-PF breath test may be confounded by consumption of foods containing 2-PF but lung inflammation as a cause of elevations in 2-PF in breath has not been excluded. Development of a breath collection system that can eliminate environmental contamination with 2-PF is a key improvement necessary to optimise the performance of this breath test.