Performance evaluation of multiplex PCR including Aspergillus—not so simple!

Medical Mycology, 2017, 55, 56–62 doi: 10.1093/mmy/myw080 Advance Access Publication Date: 23 September 2016 Review Article Review Article Performance evaluation of multiplex PCR including Aspergillus—not so simple! 1 Paris-Diderot, Sorbonne Paris Cité University, 2 Parasitology-Mycology Laboratory, Saint-Louis Hospital, Assistance Publique-Hôpitaux de Paris (AP-HP) and 3 Institut Pasteur, Molecular Mycology Unit, CNRS URA3012, National Reference Center of Invasive Mycoses and Antifungals, Paris, France ∗ To whom correspondence should be addressed. Stéphane Bretagne, Laboratoire de Parasitologie-Mycologie, Hôpital Saint Louis, 1 rue Claude Vellefaux, 75010, Paris, France. Tel: +33 1 42 49 95 03; Fax: +33 1 42 49 48 03; E-mail: Received 25 April 2016; Revised 4 July 2016; Accepted 8 July 2016 Abstract Multiplex PCRs have been designed for including species other than Aspergillus fumigatus for the diagnosis of invasive aspergillosis, such as microarrays, liquid-phase array, and electrospray-ionization mass spectrometry (PCR/ESI MS). These methods are based on the selection of multiple primers to amplify different species with the specificity checked by hybridization to a probe or by base composition of the amplicon for the PCR/ESI MS. When testing complex samples such as respiratory specimens, some clinically relevant species can be missed. Indeed, it is impossible to design primers able to amplify all the known fungal species with the same efficiency. Therefore, the best amplified species may not be the most clinically relevant. Multiplex assays have also been proposed to detect A. fumigatus DNA and azole resistance. Since the gene responsible for azole resistance is single copy and the gene used for detection is multicopy, only the high fungal loads can be evaluated. Thus, although interesting for investigating mycobiome, the multiplex assays should be used with cautious for the diagnosis of IA or the detection of resistance. For the diagnosis of invasive aspergillosis, validated quantitative PCRs specifically targeting A. fumigatus or a limited set of species to increase sensitivity is a safer option. Key words: Aspergillus, multiplex PCR, microarray, multiplexed PCR and liquid-phase array, electrospray-ionization mass spectrometry, azole resistance detection. Introduction The microbiological diagnosis of invasive aspergillosisis (IA), mainly due to Aspergillus fumigatus,1 includes pathology, direct examination, culture, and three surrogate biomarkers: the antigens galactomannan and (1→3)-β-Dglucan for which commercial assays are available, as well 56 as fungal nucleic acids for which commercial and in-house polymerase chain reaction (PCR) assays have been developed. Given the difficulty to obtain biopsies in immunocompromised patients, the low sensitivity of classical microbiology,2 and the rate of false positivity of galactomannan and the lack of specificity for IA of (1→3)-β-D-glucan,3 PCR  C The Author 2016. Published by Oxford University Press on behalf of The International Society for Human and Animal Mycology. All rights reserved. For permissions, please e-mail: Alexandre Alanio1,2,3 and Stéphane Bretagne1,2,3,∗ Alanio and Bretagne Designing primers for multiplex use The difficulties in developing primers able to amplify every species with the same efficiency are exemplified with the mycobiome studies. The goal is to amplify with one primer pair all the fungi present in the sample, to obtain sequences corresponding to OTU (operational taxonomic unit), and to infer from this result the number of fungi present in the sample tested. The objective of the PCR assay is to have the best representation of the species panel of one sample, if possible with respect of the relative quantification of each fungal species. The first step for developing such assays is an in silico analysis of the sequences available for the fungi in databases. The in silico studies show that biases cannot be avoided, with some species more efficiently amplified than other, for instance, a better amplification of ascomycetes compared to basidiomycetes.14 Due to the high polymorphism of the fungi, it is impossible to get primers able to amplify all the known fungal species without any mismatches with the targeted sequences. Therefore, a better amplification is expected with the species without any mismatches in the primer sequences, which can be neither the most abundant species nor the most clinically relevant species. The result does not represent the real image of the mixture of species in the initial sample but the image of the species that are the most easily amplified. To complete in silico studies for the specific point of detection of the real mixtures, some authors have tested mock communities.15 The number of fungal species and the relative abundance of the species are then perfectly controlled in the samples tested. It is therefore possible to compare the yield of primer pairs designed after in silico analyses. The observation is that no primer pair could reconstruct the known taxonomic distribution with perfect accuracy. Moreover, computational analysis of primer coverage and specificity cannot adequately predict behaviour under mixed biological conditions.15 Of note the worst primers are the ITS1 primers, the most frequently used in mycobiome studies.15 The main conclusion is then that biases are inevitable. Another way to evidence the biases due to primer choice is the difference observed between culture and high throughput sequencing methods. When performing whole genome sequencing (metagenomic) without any specific primers, the correlation with culture is satisfactory whereas using ITS primers (metaxonomic) the correlation is low.16 Impact of primer design for IA diagnosis The mixture of several primer sets raises concern each time multiple sets of primers or primers designed for amplifying several species are used. The impossibility to design primers with no mismatches with all the fungal species must be taken into account when performing multiplex PCR assays for the diagnosis of IA. In serum or blood samples, the presence of contaminant non-fumigatus DNA in the sampling tubes is always a risk.17 This undesirable DNA can come from environment, reagents, enzymes, antibiotics, other perfused products, or the tube itself.18,19 Panfungal primers increase the probability of amplifying environmental non-fumigatus DNA because nontargeted mould DNAs can then compete with A. fumigatus DNA during the amplification reaction. Since the A. fumigatus DNA has always been shown to be in a small amount in patients,6,7 a little non-fumigatus DNA can prevent A. fumigatus amplification. The risk is then to has been extensively studied for more than 20 years in the hope to overcome the limitations of the other methods. Recently, consensual technical requirement has been reached for performing diagnostic quantitative PCR on serum.4,5 One of the reasons why PCR is not in (...truncated)