Objectives:
Chronic pulmonary aspergillosis (CPA) and allergic bronchopulmonary aspergillosis (ABPA) are progressive, chronic pulmonary diseases causing significant morbidity and mortality in the UK. Many CPA and ABPA patients harbour Aspergillus fumigatus as a result of inhaling the abundant airborne conidia produced by this opportunistic fungus. Our high volume culture (HVC) procedure improves culture positivity, allowing identification and EUCAST susceptibility testing; moreover, Aspergillus-specific qPCR allows detection when culture fails, 25-49% of the time. Consequently, long-term antifungal treatment is complicated by lack of susceptibility data and the rising frequency of triazole resistance in A. fumigatus isolated in patients at the National Aspergillosis Centre (NAC) in Manchester.
As a result of the limited coverage of polymorphisms in the target site, cyp51A, by commercial qPCR kits and the insensitivity conferred by Sanger sequencing, we developed a pyrosequencing assay to elucidate whether we could detect polymorphisms earlier in the treatment of CPA and ABPA patients.

Methods:
Patients were perceived as potentially failing therapy and as candidates for pyrosequencing analysis if they maintained therapeutic antifungal drug levels but produced respiratory samples positive for Aspergillus spp. by qPCR but negative by HVC. Their sputum or bronchopulmonary lavage (BAL) extracts underwent PCR amplification of cyp51A to yield biotinylated products for pyrosequencing using the Qiagen Advanced PyroMark system. Positions G54, L98, Y121, M220 and T289 within cyp51A were assayed as well as possible insertions in the promoter region and results were confirmed by Sanger sequencing.

Results:
Ninety-one patient samples were analysed from March-November 2017: 40 (44.0%) had no polymorphisms, 23 (25.3%) had cyp51A polymorphisms, 15 (16.5%) were processed by Sanger sequencing only and 13 (14.3%) failed to be amplified. We detected only four samples with TR34 and/or L98H, potentially conferring pan-azole resistance; 5 with G54 (E, R or V), potentially conferring itraconazole and posaconazole resistance, and 14 with Y121F/T289A polymorphisms, potentially conferring voriconazole resistance, although the accompanying TR46 insertion was not found. Pyrosequencing results agreed with susceptibilities derived by HVC only half the time, possibly due to amplification insensitivity or PCR inhibition and definitively when species other than A. fumigatus were cultured or resistance was due to non-cyp51A polymorphisms, confirmed by Sanger sequencing.

Conclusion:
Pyrosequencing, although technically challenging for some patient samples, identified A. fumigatus cyp51A polymorphisms when culture failed and susceptibility testing was not possible and Aspergillus qPCR indicated escalation. Moreover, clearly discernible mixed wild-type and polymorphic A. fumigatus cyp51A templates were used as an early indicator of infection with resistant isolates. The impact of resistance profiling by pyrosequencing on therapeutic management incorporating analysis of patient benefit, effect on antifungal drug stewardship and antifungal drug cost-effectiveness is being investigated.

Pyrosequencing of Aspergillus fumigatus cyp51A reveals triazole resistance when susceptibility testing cannot