Med. Mycol. J. Vol.Med. 59E, Mycol. E 7-E J. 18, Vol. 2018 59 (No. 1) , 2018 E7 ISSN 2185-6486

Original Article Histopathological Diagnosis of Invasive Fungal Infections in Formalin-Fixed and Paraffin-Embedded Tissues in Conjunction with Molecular Methods: Comparison of Reproducibility and Reliability of Histopathological Evaluation, Polymerase Chain Reaction, and In Situ Hybridization

Minoru Shinozaki 1, Naobumi Tochigi 1, 2, Sota Sadamoto1, 2, Somay Yamagata Murayama3, Megumi Wakayama1, 2 and Tetsuo Nemoto2

1 Department of Pathology, Toho University Graduate School of Medicine 2 Department of Surgical Pathology, Toho University School of Medicine 3 Laboratory of Molecular Cell Biology, School of Pharmacy, Nihon University

ABSTRACT

The main objective of this study was to evaluate the relationship between histopathology, polymerase chain reaction (PCR), and in situ hybridization (ISH) for the identification of causative fungi in formalin-fixed and paraffin-embedded (FFPE) tissue specimens. Since pathogenic fungi in tissue specimens can be difficult to identify morphologically, PCRand ISH have been usually employed as auxiliary procedures. However, little comparison has been made on the sensitivity and specificity of PCR and ISH using FFPE specimens. Therefore, to compare and clarify the reproducibility and usefulness of PCRand ISH as auxiliary procedures for histological identification, we performed histopathological review, PCRassays, and ISH to identify pathogenic fungi in 59 FFPE tissue specimens obtained from 49 autopsies. The following are the main findings for this retrospective review: i) even for cases classified as “ not otherwise specified” (MNOS), two cases could be identified as by molecular methods; ii) all cases classified as non-zygomycetes mold (NZM) were Aspergillus species and were not identified by molecular methods as other fungi; iii) all 3 cases classified as zygomycetes mold (ZM) could be identified by molecular methods as ; iv) except for 1 case identified by molecular methods as spp., 5 cases were originally identified as dimorphic (DY). As a measure of nucleic acid integrity, PCRand ISH successfully detected human and fungal nucleic acids in approximately 60% of the specimens. Detection of Aspergillus DNA by nested PCRassay and by ISH against the A. fumigatus ALP gene were similarly sensitive and significant (p < 0.01). Thus, our findings demonstrated the potential risk of error in the classification of fungi based on pathological diagnosis. Combining molecular methods such as ISH and PCRon FFPE specimens with pathological diagnosis should improve diagnostic accuracy of fungal infection. Key words : Histopathological diagnosis, Polymerase chain reaction, In situ hybridization, Invasive fungal infection

mortality1, 2). Indeed, our epidemiological analysis suggests Introduction that the prevalence of IFIs in patient groups with hematologic- al disorders was significantly higher than that for other patient Invasive fungal infections (IFIs) are a rapidly progressive groups and was also detected in an immunocompromised state and major threat that typically affects severely immunocom- caused not only by primary disease but also by treatment with promised patients with hematological and oncological anti-tumor drugs and corticosteroids3). Although early diagno- malignancies, and can result in a high degree of morbidity and sis of IFIs is crucial for favorable outcomes4), conventional

Address for correspondence: Tetsuo Nemoto Department of Surgical Pathology, Toho University School of Medicine, 6-11-1 Omori-Nishi, Ota-Ku, Tokyo 143-8541, Japan Received: 29, October 2017, Accepted: 14, December 2017 E-mail: [email protected] E8 Medical Mycology Journal Volume 59, Number 1, 2018 diagnostic methods such as fungal culture generally are neither Table 1. Underlying diseases of the 49 patients sensitive nor rapid 5, 6), and they often do not detect IFIs until Underlying diseses No. of cases late in the disease course7). Although pathological and Malignancy cytological diagnoses of IFIs can provide definitive diagnoses, Acute myeloid leukemia 8 morphological similarities of causative fungi in histopatholo- Chronic myeloid leukemia 7 gical specimens complicate the rendering of a diagnosis based on pathological and cytological findings alone8).Hence,the Adult T-cell leukemia 6 development of an adjunct system to diagnose fungal infection Chronic myeloid leukemia 1 in histological specimens is necessary9). To assess the Adult T-cell leukemia 1 performance of polymerase chain reaction (PCR) assays and in Multiple myeloma 1 situ hybridization (ISH) for the identification of fungal Solid-organ malignancies 5 pathogens, we evaluated a total of 59 formalin-fixed and Hematological disorder paraffin-embedded (FFPE) tissue specimens from autopsy Aplastic anemia 2 cases with histologically confirmed IFIs. To date, few studies Myelodysplastic syndromes 1 have examined the usefulness of pathological diagnosis of IFIs Others 17 that involved a combination of PCRand ISH in FFPE tissues to detect and identify causative fungi. This novel diagnostic approach could be developed as an adjunct tool for the respectively. These categories were morphologically defined diagnosis of IFIs in FFPE tissues. as follows: NZM: hyphae with septate, acute-angle (45º) or dichotomous branching; ZM: broad, thin-walled, ribbon-like Materials and methods hyphae with sparsely septate and right-angle branching; MNOS: hyphae with irregular branching; MY, round or ovoid Tissue samples organisms only; DY: round or ovoid organisms with A total of 59 formalin-fixed and paraffin-embedded tissues pseudohyphae and/or hyphae occasionally with chained beads were retrieved from the archives of the Department of pattern. Pathology at Toho University Medical Center, Omori Tissue specimen preparation Hospital. The autopsies from which the samples were taken All tissues were fixed in 15% formalin, followed by were performed between 1955 and 2006, and the tissues were dehydration with ethanol and embedding in paraffin. Selected collected from 59 different autopsy specimens (49 cases) taken case slides and paraffin blocks were retrieved from the from cases with histologically proven fungal infections (46 hospital archives. Tissue sections (3 µm) were mounted on lung, 7 heart, 2 intestine, 2 pleurae, 1 kidney, and 1 aortic aminoalkylsilane-coated slide glasses (Matsunami Glass Ind., aneurysm). In this study, mycological culture was not Ltd., Osaka, Japan). Specific paraffin blocks containing performed. The 49 autopsy cases included 25 men and 24 infected lesions were confirmed by histological examination women. Median patient age was 53 years old (range: 1.6-87 using the sections stained with hematoxylin and eosin (H&E) years). Postmortem interval (PMI) was 2:47 hr. (range: 0:47- and Grocott’s stains10). 10:30), but the duration of fixation during tissue processing DNA isolation was unknown. Malignancy was the most common underlying DNA extraction using xylene-ethanol deparaffinization was disease in 29 cases, 24 of which showed hematological carried out as previously described11) with minor modifica- malignancy. The underlying diseases are summarized in Table tions. Briefly, three 5-µm-thick sections were cut from the 1. FFPE tissue blocks with a standard microtome and collected in Review of histopathological findings microcentrifuge tubes. To avoid cross-contamination, we used We undertook a retrospective review of morphological a new, sterile disposable microtome blade for each block. classification using five morphological categories; namely, Deparaffinization was performed by incubation with 1 ml non-zygomycetes mold (NZM), zygomycetes mold (ZM), xylene for 15 min at room temperature (RT). The tubes were mold not otherwise specified (MNOS), monomorphic yeast centrifuged at 12,000 rpm for 5 min, and the supernatant was (MY), and dimorphic yeast (DY) (Fig. 1). Aspergillus species discarded. This step was repeated twice. The samples were and Mucorales were categorized as NZM and ZM, respective- rehydrated twice in 100% ethanol and subsequently air-dried. ly. Filamentous fungi that did not correspond with the typical For DNA extraction, 200 µl lysis buffer composed of 200 mM morphological characteristics of Aspergillus species or Tris-HCl (pH 8. 0), 25 mM EDTA, 0. 5% sodium dodecyl Mucorales were categorized as MNOS. Morphological sulfate (SDS) (Merck KGaA, Darmstadt, Germany), 250 mM features of the yeast were categorized as MY and DY NaCl, and 20 mg/ml proteinase K (Nippon Gene Co., Tokyo, according to monomorphic or dimorphic morphology, Japan) were added to each tube. The suspension was incubated Med. Mycol. J. Vol. 59 (No. 1) , 2018 E9

(a) (b) (c)

(d) (e) Fig. 1. Categories of histopathological review. Based on histopathological findings, the 49 cases were divided into five categories: (a) Non-zygomycetes mold (NZM), (b) Zygomycetes mold (ZM), (c) Mold not otherwise specified (MNOS), (d) Monomorphic yeast (MY), and (e) Dimorphic yeast (DY). These categories were morphologically defined as follows: NZM: hyphae with septate, acute-angle (45º) or dichotomous branching; MNOS: hyphae with irregular branching; ZM: broad, thin- walled, ribbon-like hyphae with sparsely septate and right-angle branching; MY: round or ovoid organisms only; DY: round or ovoid organisms with pseudohyphae and/or hyphae occasionally with chained beads pattern. at 48ºC overnight. After inactivating the proteinase K at 94ºC important fungi were performed to assess the quality of DNA for 10 min, 500 µl of phenol-chloroform-isoamyl alcohol (25: from fungal organisms. Primers U1/U213) and 18SF1/58SR114) 24:1) (Wako Pure Chemical Industries, Ltd., Osaka, Japan) were used to amplify a region of the fungal 28S rRNA gene was added to each microcentrifuge tube. The microcentrifuge and the internal transcribed spacer 1 (ITS1) region, tubes were centrifuged at 12,000 rpm for 15 min, and the respectively. These primer pairs consistently amplified supernatants were transferred to a new microcentrifuge tube. fragments of 260 bp (U1/U2) and approximately 300 bp This step was repeated twice. DNA was precipitated by the (18SF1/58SR1) with species-dependent variable fragment addition of one-tenth the volume of 3 M sodium acetate lengths. The nucleotide sequences of the primers used were 5'- solution (pH 5.2) and twice the volume of cold 100% ethanol, GTG AAA TTG TTG AAA GGG AA (U1), 5'- GAC TCC incubated at -20ºC overnight, and subsequently centrifuged at TTG GTC CGT GTT-3' (U2), 5’-AGG TTT CCG TAG GTG 12,000 rpm for 20 min at 4ºC. Finally, the pellet was washed AAC CT-3’ (18SF1), and 5'-TTC GCT GCG TTC TTC ATC with 70% ethanol and centrifuged at 12,000 rpm for 15 min at GA-3' (58SR1). The annealing temperature of the reactions 4ºC. After air-drying, the DNA was dissolved in 20 µl DEPC- using the U1/U2 and 18SF1/58SR1 primer pairs was 55ºC and treated distilled water and stored at -20ºC until use. 60ºC, respectively. PCR Primers for two human β-globin PCR assays PCR primers for two Aspergillus PCR assays To assess the quality of human DNA preserved in the FFPE Aspergillus -specific PCRassays that amplified tissues, we performed human β-globin PCRassays using two regions of the 18S rRNA gene15) and 5.8S rRNA gene16) were primer pairs12). The primers 5'-ACA CAA CTG TGT TCA performed as a nested PCRwith two sets of primers. For the CTA GC-3' (PC03), 5'-CAA CTT CAT CCA CGT TCA CC- first PCRassay, the primers used for the first step were 5′- 3' (PC04), and 5'-TGA CAA AGC CCC ATA CG CT-3’ CGG CCC TTA AAT AGC CCG-3′ (AFU7S) and 5′-GAC (GH21) consistently amplified 110 bp (PC03/PC04) and 250 CGG GTT TGA CCA ACT TT-3′ (AFU7AS). The primers bp (PC03/GH21) fragments. The annealing temperature of the used in the second step of the first PCRassay were 5′-AGG reactions was 53ºC. GCC AGC GAG TAC ATC ACC TTG-3′ (AFU5S) and 5′- PCR primers for two panfungal PCR assays GGG GGT CGT TGC CAA CCC CCC TGA-3′ (AFU5AS). Two panfungal PCRassays to detect DNA of medically For the second PCRassay, the primers used in the first step E10 Medical Mycology Journal Volume 59, Number 1, 2018 were 5′-TCT TGG TTC CGG CAT CGA T-3′ (ASP nest 1) product was used for the second step of nested PCR. The and 5′-TGA CAA AGC CCC ATA CGC T-3′ (ASP nest 2). products amplified by PCRwere detected using 3% agarose The primers used in the second step of the second PCRassay gel electrophoresis and visualized under UV irradiation. were 5′-GAA GAA CGC AGC GAA ATG C-3′ (ASP nest 3) DNA sequencing of PCR products and 5′-AAC ACA CAA GCC GTG CTT GA-3′ (ASP nest 4). Primers 18SF1/58SR1, Asp Nest3/4, and/or ZM1/3 were The above-mentioned primer pairs amplified fragments of 405 used to perform DNA sequencing analysis. Amplicons of bp (AFU7S / AFU7AS), 236 bp (AFU5S/AFU5AS), 250 bp positive PCRwere purified using QIAquick PCRpurification (ASP nest1/ASP nest2), and 146 bp (ASP nest3/ASP nest4). kit (Qiagen, Tokyo, Japan). Both strands of the amplified The annealing temperature of the reactions using AFU and DNA were sequenced using BigDye Terminator version 3.1 ASP primers was 65ºC and 55ºC, respectively. cycle sequencing kit (Applied Biosystems, Foster City, CA). PCR primers for PCR assays The protocol included a total reaction volume of 20 µl The following PCRprimers were designed to amplify a consisting of 3 µl Big Dye, 2.5 µl 5 × sequencing buffer, region of the 28S rRNA gene of the Fusarium species: 5′- 5 pmol primer, and 1 µl PCRproduct. Sequencing reactions TGG GAG GTA TAT GTC TTC TAA-3′ (FS S) and 5′-GAT were carried out in 28 cycles of 96ºC for 10 s, 50ºC for 5 s, GAT CAA CCA AGC CCA-3′ (FS AS). The above- and 55ºC for 4 min. Sequencing products were purified by mentioned primer pair generated a 170-bp fragment. PCR ethanol precipitation and run on an ABI3130 genetic analyzer reactions were performed at 57ºC annealing temperature. (Applied Biosystems). PCR primers for Mucorales PCR assays In situ hybridization (ISH) To amplify a region of the 18S rRNA gene of Mucorales Two types of probes were used for ISH: double-stranded species, we performed semi-nested PCR17, 18). The primers used DNA (dsDNA) and peptide nucleic acid (PNA). These probes were 5′-ATT ACC ATG AGC AAA TCA GA-3′ (ZM1) and 5′ were conjugated with fluorescein isothiocyanate (FITC). -TCC GTC AAT TCC TTT AAG TTT C-3′ (ZM2) for dsDNA Probe amplification of a 407-408 bp fragment (for the first step of To detect and identify Aspergillus species in FFPE tissues, PCR), and ZM1 and 5′-CAA TCC AAG AAT TTC ACC TCT we used a dsDNA probe. The dsDNA labeling was performed AG-3′ (ZM3) for amplification of a 175-177 bp fragment (for using PCRamplification of the A. fumigatus alkaline the second step of PCR). PCR was carried out at an annealing proteinase gene (ALP)21). The 583-bp dsDNA probe was temperature of 55ºC. synthesized using a PCRfluorescein labeling mix (Roche PCR primers for Scedosporium PCR assays Diagnostics K. K. Tokyo, Japan) and primers for ALP Scedosporium genus-specific PCRassays that amplify the amplification. The primers used were 5′-GCC TAT CCG TGT 5.8S rRNA gene and ITS 2 were performed as a semi-nested ACT TGA TG-3′ (ALPn1) and 5′-AAG GTT CAG CAC GAT PCR19). The primers used were 5'-TTA CTA CGC AGA AGG CAA C-3′ (ALPn2). The dsDNA probes were synthesized by CAA-3' (Pbo1), 5′-AAT CTT TGA ACG CAC ATT-3′ (SP1), PCRusing A. fumigatus (NBRC 6344) DNA as template, and Pbo1 and 5′-TGT CCG AGC GTC ATT TC-3′ (SP2). deoxynucleoside triphosphate mix with FITC-dUTP from the These primers consistently amplified fragments of 197 bp PCRfluorescein labeling mix (RocheDiagnostics K.K.), and (Pbo1/SP1) and 149 bp (Pbo1/SP2). The primer annealing Taq DNA polymerase (Expand High Fidelity PCRSystem: temperature used was 55ºC. Roche Diagnostics K. K.), according to the manufacturer’ s PCR assays recommendations. Each PCRreaction contained 1 × PCR TM Extracted DNA was amplified using a MyCycler buffer containing 4 mM MgCl2, 200 µM PCRfluorescein thermocycler (Bio-Rad Laboratories, Richmond, CA, USA). labeling mix dNTPs, 0.5 µM of each primer, 1 U Taq DNA The primers and PCRconditions used are specified below. polymerase, 100 pg template, and distilled water to a final Conventional PCRwas performed as described by Zhi et al. 20) volume of 100 µl. Thermocycling conditions were initial with some modifications. Nested PCRamplification was denaturation at 94ºC for 4 min, 35 cycles of denaturation at carried out in two steps. Reactions were performed in 50 µl 94ºC for 30 s, annealing at 53ºC for 30 s, extension at 72ºC for total volume containing 1 × PCRbuffer, 0. 2 mM of each 90 s, and final extension at 72ºC for 5 min. PCRproducts were dNTP, 1.5 mM MgCl2, 0.4 µM each of forward and reverse purified using QIAquick PCRpurification kit (Qiagen). primers, 1.25 units of iTaq DNA polymeraseTM (Bio Rad Co. PNA probes Ltd., Richmond, CA, USA), and 1 µl template DNA. Thermal To estimate retention and hybridizability of the target RNA cycling conditions were initial denaturation at 94ºC for 5 min in the samples, we used a panfungal antisense PNA probe (N- followed by 35 cycles of denaturation at 94ºC for 1 min, terminus- TAC TTG TGC GCT ATC GGT)10, 22, 23).Toidentify annealing at these optimal temperatures for 1 min, and C. albicans and Fusarium species, we also used an antisense extension at 72ºC for 1 min followed by a final extension at PNA probe targeting the 26S rRNA of C. albicans (N 72ºC for 7 min. Approximately 1 µl of the first-round PCR terminus-ACA GCA GAA GCC GTG)24) and 28S rRNA of Med. Mycol. J. Vol. 59 (No. 1) , 2018 E11

Fusarium species (N terminus-GAT GAT CAA CCA AGC value less than 0.05 was considered statistically significant. CCA)10, 22). The N-terminus of the PNA probes was conjugated All statistical analyses were carried out using Excel Tokei with FITC. 2010 for Microsoft Excel 2010 (SSRI, Tokyo, Japan). Primers and probes Ethics statement The oligonucleotide primers and PNA probes used in this This study protocol was approved by the Institutional study were synthesized by Fasmac Co., Ltd. (Kanagawa, Research Ethics Committee of Faculty of Medicine, Toho Japan). University: A17003_26006_25080. ISH procedure The ISH procedure was performed as described Results previously10, 22, 23). Briefly, the sections were deparaffinized and rehydrated according to standard procedures. To expose target For this study, histopathological review, PCR, and ISH nucleic acids in the formalin-fixed tissue, we treated the were performed to evaluate the diagnosis of fungal infections sections with 1 mM EDTA pH 8.0 in a water bath (Thermo in a total of 49 cases (59 autopsy specimens). These included Fisher Scientific K.K., Yokohama, Kanagawa) for 20 min at PCRassays specific for Aspergillus spp., Mucorales, 98ºC and digested them with 10 µg/ml proteinase K (Nippon Fusarium spp., and Scedosporium spp., and ISH specific for Gene Co., Ltd., Tokyo, Japan) for 10 min at 37ºC. The Aspergillus spp., ,andFusarium spp. standard hybridization medium contained 10 mM Tris-HCl The review of fungal infections and evaluation of fungal (pH 7. 6), 50% formamide (Nacalai Tesque Inc., Kyoto, pathogens were made on the basis of histopathological Japan), 1 × Denhardtʼs solution (Sigma-Aldrich Japan, findings in all cases. Histological sections from 49 cases Tokyo), 10% dextran sulfate, 600 mM NaCl, 0. 25% SDS, reviewed in this study were classified as: Non-zygomycetes 500 µg/ml yeast tRNA (Sigma), and 1 mM EDTA (pH 8.0). mold (NZM), 34 cases; Zygomycetes mold (ZM), 2 cases; The concentrations of the dsDNA and PNA probes were 10 % combined NZM and ZM, 1 case; Mold not otherwise specified and 1 µg/ml, respectively. The hybridization solutions were (MNOS), 5 cases; monomorphic yeast (MY), 1 case; and applied directly to the sections and covered with a cover glass dimorphic yeast (DY), 6 cases (Table 1). Of the 49 cases (30 mm × 30 mm, Matsunami Glass Ind., Ltd.). The applied analyzed in this retrospective review, discrepancies in probes and target nucleic acids in the tissues were denatured in morphological identification using tissues between previous a block thermostatic oven (DTU-2B; TAITEC Co. Saitama, histological diagnosis and the current review were found in 7 Japan) for 5 min at 94ºC, followed by immediate cooling on cases (14%) that included ZM (n = 1, case 4), MNOS (n = 5, ice for 5 min. Hybridization experiments were carried out at cases 1, 5, 8, 24, and 35) and combined NZM and ZM (n = 1, 50ºC for 24 hr and at 56ºC for 90 min for the dsDNA and PNA case 37). Both the 2 cases classified as ZM (including a probe, respectively. After repeated washings with 0. 2 × combined NZM and ZM case) were originally identified as standard saline citrate (SSC) at 50ºC and 56ºC for the dsDNA Aspergillus spp. Of the 5 cases classified as MNOS, all were and PNA probe, respectively, the signals were detected by also originally identified as Aspergillus spp. Of these 7 enzyme immunohistochemistry using an anti-FITC antibody discrepant cases, 4 could be identified by molecular methods. (Roche Diagnostics K.K.) and horseradish peroxidase-labeled Of the 5 cases classified as MNOS, 3 cases could not be polymer solution (Nichirei Biosciences, Inc., Tokyo, Japan). identified by molecular methods. The results of histopatholo- Finally, the sites of peroxidase activity were visualized by 3,3’ gical review and sequence-based identification at the genus -diaminobenzidine tetrahydrochloride (DAB; Dojindo Labor- and species level are shown in Table 2. The results from our atories, Kumamoto, Japan) in the presence of H2O2 as well as molecular methods indicated that 2 cases (case 4 and 37) nickel and cobalt ions. among the 7 discrepant cases involved diagnostic errors in Data analysis differentiating Mucorales from Aspergillus spp. These 2 cases The selected sequences were analyzed using the BLAST were identified by PCRdirect sequence analysis as Rhizopus alignment program of GenBank (National Institutes of spp. (case 4) and Rhizomucor spp. (case 37). One case (case Health). The computer alignment provides a list of matching 1) classified as MNOS was identified as non-fumigatus organisms, ranked in order of similarity between the unknown Aspergillus species by PCRdirect sequence analysis. The sequence and the sequence of the corresponding organism remaining case (case 24) classified as MNOS was identified as from the database. In particular, ≥ 99% identity was required Aspergillus spp. by ISH (Table 3). Although not misclassified, for identification at the species level, and ≥ 97% identity was one Candida spp. classified as DY (case 12) was identified as required for identification at the genus level. Trichosporon spp. by PCRdirect sequence analysis. This Statistics misinterpretation may be derived from morphologic similar- Univariate analysis was performed with Fisher’s exact test ities that exist when classifying yeast, wherein Candida spp. is to compare frequencies among the categories. A calculated p- most commonly confused with Trichosporon spp. E12 Medical Mycology Journal Volume 59, Number 1, 2018

Table 2. Results of histopathological review and sequence-based identification Histology Sequence analysis Case No Sample No Organs Previous Recorded ITS-1 18S or 5.8S rDNA 1 1 Lung Aspergillus MNOS Aspergillus flavus, ect. Aspergillus flavus, ect. 2 Lung Aspergillus MNOS 2 3 Lung Aspergillus NZM 3 4 Lung Aspergillus NZM Aspergillus spp. 4 5 Lung Aspergillus ZM Rhizopus spp. 5 6 Lung Aspergillus MNOS 6 7 Lung Aspergillus NZM Aspergillus fumigatus Aspergillus spp. 8 Lung Aspergillus NZM Aspergillus spp. 9 Heart Aspergillus NZM Aspergillus spp. 10 pleurae Aspergillus NZM Aspergillus spp. 7 11 Lung Candida MY 12 Intestine Candida MY 8 13 Lung Aspergillus MNOS 9 14 Lung Candida DY 15 Heart Candida DY 10 16 Lung Aspergillus NZM 11 17 Lung Aspergillus NZM 12 18 Heart Candida DY Trichosporon ovoides,ect. 19 Intestine Candida DY Trichosporon ovoides,ect. 20 kidney Candida DY Trichosporon ovoides,ect. 13 21 Lung Aspergillus NZM Aspergillus spp. 14 22 Lung Candida DY 15 23 Lung Aspergillus NZM Aspergillus spp. 16 24 Lung Aspergillus NZM 17 25 aortic aneurysm Candida DY 18 26 Lung Aspergillus NZM 19 27 Lung Candida DY 20 28 Lung Aspergillus NZM Aspergillus spp. 21 29 Lung Aspergillus NZM Aspergillus fumigatus Aspergillus spp. 22 30 Lung Aspergillus NZM Aspergillus spp. 23 31 Lung Aspergillus NZM Aspergillus spp. 24 32 Lung Aspergillus MNOS 25 33 Lung Aspergillus NZM 34 Heart Aspergillus NZM 26 35 Lung Candida DY 27 36 Lung Aspergillus NZM Aspergillus fumigatus Aspergillus spp. 28 37 Lung Aspergillus NZM Aspergillus spp. 29 38 Lung Aspergillus NZM Aspergillus spp. 30 39 Heart Aspergillus NZM Aspergillus spp. 31 40 Lung Aspergillus NZM Aspergillus spp. 32 41 Lung Aspergillus NZM Aspergillus spp. 33 42 Lung Mucorales ZM 34 43 Heart Aspergillus NZM Aspergillus spp. 35 44 Lung Aspergillus MNOS 36 45 Heart Aspergillus NZM Aspergillus spp. 371) 46 Lung Aspergillus NZM Aspergillus fumigatus Aspergillus spp. Aspergillus ZM Rhizomucor spp. 38 47 Lung Aspergillus NZM Aspergillus spp. 39 48 pleurae Aspergillus NZM 40 49 Lung Aspergillus NZM Aspergillus spp. 41 50 Lung Aspergillus NZM 42 51 Lung Aspergillus NZM Aspergillus fumigatus Aspergillus spp. 43 52 Lung Aspergillus NZM Aspergillus spp. 44 53 Lung Aspergillus NZM Aspergillus spp. 45 54 Lung Aspergillus NZM Aspergillus spp. 46 55 Lung Aspergillus NZM 56 Lung Aspergillus NZM 47 57 Lung Aspergillus NZM 48 58 Lung Aspergillus NZM 49 59 Lung Aspergillus NZM Aspergillus spp. 1): Case 37 was categorized as combined NZM and ZM. Med. Mycol. J. Vol. 59 (No. 1) , 2018 E13

Table 3. Results of PCR and ISH

PCR ISH Fusarium Scedosporium Aspergillus Fusarium β globin Panfungal Aspergillus spp. Mucorales Panfungal C. albicans Case Sample spp. spp. spp. spp. No No PC PC03/ 18SF1/ AspNest 1/2 AFU 7S/7AS FSS/ ZM 1/2 Pbo 1/SP1 U1/U2 A. fumigatus 03/04 GH21 58SR1 AspNest 3/4 AFU 5S/5AS FS AS ZM 1/3 Pbo 1/SP2 28S rRNA 26S rRNA 28S rRNA ALP1) gene 110 bp 250 bp 260 bp 300 bp 146 bp 236 bp 170 bp 175 to177 bp 149 bp 11 + - - + + + - - - + + - - 2+---+ + ------23------+ - - 34---- + - - - - - + - - 45------+ - ---- 56------67 + - - + + + - - - + + - - 8+--- + + - - - + + - - 9+--++ + - - - ++- - 10----+ - - - - - + - - 711+------+ - - - 12+------+ - - - 813+------+ - - - 914+------+ - + - 15+------+ - + - 1016------+ + - - 11 17 + + ------+ + - - 12 18 + - + + - - - - - + - - - 19+-++---- -+--- 20+-++- --− -+--- 1321++++ + + - - - + + - - 14 22 + ------+ - + - 1523----+ - - - - + + - - 1624-+------+-- 1725------+ - 1826------+ + - - 1927++------2028----+ - - - - + + - - 21 29 + - - + + + - - - + + - - 22 30 + - - - + - - - - + + - - 2331----+ - - - - + + - - 2432------+ - - 2533------34---- - + - - - + + - - 2635---+------+- 27 36 - - - + + - - - - + + - - 2837----+ - - - - - + - - 29 38 + - - + + + - - - + + - - 30 39 - - + + + - - - - + + - - 31 40 + - - - + - - - - + + - - 32 41 + - - - + - - - - + + - 3342+++--- - + - ---- 3443----+ ------3544------3645+---+--- --+-- 372) 46+-++++-+ -++-- - 38 47 + - - - + - - - - + + - - 3948+------40 49 + - - - + - - - - + + - - 41 50 + ------+ + - - 42 51 + - - + + + - - - - + - - 43 52 + - - - + - - - - + + - - 44 53 + + - - + - - - - + + - - 4554----+ - - - - + + - - 4655+------+-- 56+------+-- 4757+------4858+------+-- 4959--+-+--- --+-- No. of positive 36 6 8 14 29 11 0 3 0 35 38 5 0 % of positive 61 10 14 24 49 19 0 7 0 59 63 8 0 1), ALP: Alkaline Proteinase 2), ALP ISH were separately assessed in case 37 (combined NZM and ZM). E14 Medical Mycology Journal Volume 59, Number 1, 2018

Table 4. Comparison of PCRand ISH for assessing nucleic acid preservation in tissue sections

Human β globin Human β globin Panfungal Panfungal PCR PCR PCR PCR PC03/PC04 C03/ GH21 U1/U2 18SF1/58SR1 (110 bp) (250 bp) (260 bp) (300 bp) +-+-+-+- Panfungal ISH + 26 9 3 32 6 29 12 23 -1014321222222 Chi-square P-value p < 0.05 ns ns ns

Table 5. Correlations among three methods for detecting Aspergillus spp.

Nested PCR Nested PCR AFU7S/7AS Asp Nest1/2 ISH: ALP AFU5S/5AS Asp Nest3/4 (236 bp) (142 bp) ISH: ALP Chi-square P-value ― ns p < 0.01 Nested PCR AFU7S/7AS Chi-square P-value ns ― p < 0.01 AFU5S/5AS (236 bp) Nested PCR Asp Nest1/2 Chi-square P-value p < 0.01 p < 0.01 ― Asp Nest3/4 (142 bp)

To demonstrate the presence of detectable nucleic acids in Mucorales fragment in 7% (3/59), 170-bp Fusarium spp. tissues, we performed PCRtargeting the human β-globin gene fragment in 0% (0/59), and 149-bp Scedosporium spp. and fungal rRNA gene cluster, as well as ISH targeting the fragment in 0% (0/59) of samples. Aspergillus spp., C. 28S rRNA of pathogenic fungi. β -globin PCR with PC03/ albicans and Fusarium spp. were detected by ISH in 63% PC04 primers (110 bp) gave a positive identification rate of (38/60), 8% (5/59), and 0% (0/59), of samples respectively 61% (36/59) (Table 2). On the other hand, positive rates for β- (Table 3). The differences in the above-mentioned results globin PCRwith PC03/GH21 primers (250 bp), panfungal reflect the separate assessment of ALP ISH in case 37 PCRwith U1/U2 primers (260 bp), and panfungal PCRwith (combined NZM and ZM). ISH targeting C. albicans 26S 18SF1/58SR1 primers (300bp) were 10% (6/59), 14% (8/59), rRNA detected yeast rRNA, except for Trichosporon spp., in 5 and 24% (14/59), respectively. The amplification of human β- of 6 DY samples (83%). The best sensitivity results were globin and panfungal genes yielded poor results in terms of obtained for ISH targeting the A. fumigatus ALP gene to detect sensitivity. ISH targeting 28S rRNA of pathogenic fungi Aspergillus spp. The detection of Aspergillus DNA by nested detected 59% (35/59) of the samples (Table 3). Significant PCRassay with Asp Nest 1/2 and Asp Nest 3/4 primers, and correlation was found only between human β-globin by PCR ISH against the A. fumigatus ALP gene was similarly sensitive with PC03/PC04 primers (110 bp) and pathogenic fungi 28S and relevant. The concordance between the two assays was rRNA detected by ISH (p < 0.05) (Table 4). There were no 96% (27 of 28 samples: sequence-based evidence for significant relationships among postmortem interval, duration Aspergillus spp.). For Aspergillus spp. detection, statistically of block storage, and degree of nucleic acid preservation in significant correlations were observed (p < 0.01) (Table 5). tissue sections (data not shown). DNA sequence analysis of PCRproducts was successfully To detect and identify fungi in specimens, we performed performed in 33 of 59 samples (56%) (Table 2). The PCRand ISH analyses. These included PCRassays specific diagnostic value of ISH targeting the A. fumigatus ALP gene for Aspergillus spp., Mucorales, Fusarium spp., and Scedos- for the diagnosis of Aspergillus infection was evaluated based porium spp., and ISH assays specific for Aspergillus spp., on results of sequence analysis, which showed sensitivity of Candida albicans,andFusarium spp. PCRdetected the 146- 96.4%, specificity of 100%, positive predictive value (PPV) of bp Aspergillus spp. fragment in 49% (29/59), 236-bp 100%, and negative predictive value (NPV) of 83.3% (Table Aspergillus spp. fragment in 19% (11/59), 175-177-bp 6). Med. Mycol. J. Vol. 59 (No. 1) , 2018 E15

Table 6. Diagnostic accuracy of ALP ISH for Aspergillus infection: Comparison of ALP ISH with PCRdirect sequencing

Sequence analysis Positive Negative Diagnostic Sensitivity Specificity 33/59 (56%) predictive predictive method (%) (%) Aspergillus spp. Trichosporon spp. Mucorales value (%) value (%) +270 0 ISH: ALP 96.4 100 100 83.3 -1 3 2 (p < 0.01) ALP: Alkaline Proteinase

(a) (b) (c)

(d) (e) (f) Fig. 2. Photographs from a representative case analyzed in this study. This case (No 37) had a combined infection with A. fumigatus and Rhizomucor spp. (a) Pathological findings with hematoxylin and eosin stain. Histological examination revealed foci consisting of filamentous growth of organisms. (b) Findings with Grocott’s stain. Grocott’s stain showed hyphae with acute-angle and dichotomous branching within foci of infection. (c) Result of ISH against the A. fumigatus ALP gene. The probe was reactive with the fungal elements of A. fumigatus. (d) Pathological findings with hematoxylin and eosin stain. Histological examination revealed foci consisting of filamentous growth of organisms. (e) Findings with Grocott’s stain. Grocott’s stain showed ribbon-like hyphae with sparsely septate and right-angle branching. (f) Result of ISH against the A. fumigatus ALP gene. The probe was not reactive with the fungal elements of Rhizomucor spp. Original magnification, × 400.

In addition, a representative case we studied is shown in diagnostic tool in mycology26). However, pathological diagno- Fig. 2. sis of IFIs is frequently difficult because the causative fungal pathogen is morphologically similar to other fungi in Discussion pathological specimens8, 27). Therefore, an auxiliary diagnostic method would be valuable for the diagnosis of fungal infection In general, cultivation and subsequent identification are in FFPE tissues. In particular, molecular diagnostic standard approaches for laboratory diagnosis of invasive approaches may be considered for improving diagnostic fungal infections25). If fungal culture is unsuccessful for accuracy of fungal infections. Many recently published studies various reasons, however, FFPE material can be a valuable demonstrated that PCR-based detection studies are highly alternative sample material. In these situations, identification sensitive, efficient, and reliable for identifying causative of fungi based on pathological diagnosis serves as a major fungi28, 29). Applications of ISH in the identification of E16 Medical Mycology Journal Volume 59, Number 1, 2018 causative fungi, however, have not been fully discussed. causative fungi in FFPE tissues using PCR. Well-controlled Furthermore, a molecular diagnostic approach using a formalin fixation times are preferable to overcome these combination of PCRand ISH has not been systematically difficulties. Unfortunately, the duration of fixation during explored, and little is known regarding its epidemiological tissue processing was unknown in this study. We believe that diagnostic approach. outcomes of molecular methods in routine surgical pathology The 49 cases analyzed in this retrospective study were can be improved by the use of standardized formalin fixation divided into five categories based on histopathological protocols. findings. These retrospective classifications were subsequent- Aspergillus spp. have been identified and can be reliably ly confirmed using molecular methods such as PCRand ISH. detected using a nested PCRassay with Asp Nest 1/2 and Asp The results suggested that the morphological similarities of Nest 3/4 primers, and ISH against the A. fumigatus ALP gene, causative fungi in histopathological specimens resulted in which showed excellent concordance. For identification of errors in diagnosis based on pathological findings alone. Aspergillus spp., the relative levels of the positive identifica- Several notable inconsistencies were found in the identifica- tion rate revealed by Aspergillus nested PCRwith Asp Nest tion of the more commonly misdiagnosed fungi. The most 1/2 and Asp Nest 3/4 primers were broadly consistent with common diagnostic inconsistency was observed in dif- those estimated by ISH targeting the A. fumigatus ALP gene. ferentiating Aspergillus spp. from Mucorales. The difficulty of The data showed significant correlation between the Aspergil- differential diagnosis between Aspergillus spp. from Mucor- lus spp. 146-bp fragment detected by nested PCRand ales may also extend to other septate hyphae, such as non- Aspergillus spp. detected by ISH (p < 0.05). Furthermore, our Aspergillus hyalohyphomycetes, as well as non-septate sequence analysis results provided evidence for the diagnostic hyphae, such as Mucorales. Furthermore, although the value of ISH targeting the A. fumigatus ALP gene for the morphologic findings were suggestive of Mucorales, results of diagnosis of Aspergillus infection. Our results demonstrate differential diagnosis included non-fumigatus Aspergillus that these primers and probes would be useful for the detection species, such as A. flavus. Although not a misclassification, and identification of Aspergillus spp. in FFPE tissues. If misinterpretation also occurred in the identification of sequence analysis using panfungal primers is not successful, Trichosporon spp. Histopathological differentiation of Candi- PCRusing specific primer appears to be a useful approach for da spp. from other such as Trichosporon spp. can also accurate diagnosis of pathogenic fungi. be challenging. The same pitfalls associated with morphologic The reliability of ISH targeting the A. fumigatus ALP gene identification of fungal infections in histological specimens to detect Aspergillus spp. was already demonstrated using were noted in a previous study26). Thus, awareness of these infected mouse models21). Our results further provide evidence pitfalls may help in further increasing diagnostic accuracy of for the reliability of these methods for identification of fungal infections. Although the 3 cases classified as MNOS Aspergillus spp. Furthermore, our previous study (data not could not be identified by molecular methods, they may shown here) demonstrated that PNA ISH targeting 26S rRNA include important non-Aspergillus hyalohyphomycetes (e.g., is an accurate means of differentiating Candida spp. from Fusarium, Scedosporium,andPenicillium spp.). As Walsh et Trichosporon spp. in FFPE tissues23). al. noted, infections due to these fungal pathogens are In this report, we assessed two molecular techniques, PCR emerging and are of particular medical importance30). and ISH, for the detection and identification of fungal Generally, PCRamplification was successful when smaller pathogens from FFPE tissue specimens. However, the lack of products were amplified from FFPE tissue because standardization in these techniques had not been addressed formaldehyde-induced DNA-protein cross-linking can inhibit until recently. Nucleic acid preservation must be assessed to amplification of longer fragments31). Our results confirmed this ensure predictive validity of FFPE tissues. We previously observation, and in fact, β-globin PCRwith PC03/PC04 and reported that use of ISH with a panfungal probe was valuable Aspergillus nested PCRwith Asp Nest 1/2 and Asp Nest 3/4 for estimating levels of hybridizable nucleic acids for the primers, with final amplification products of 110 bp and 146 specific detection of human pathogenic fungi in FFPE bp, respectively, showed the best results in terms of tissues10, 23). Indeed, significant correlations were found sensitivity. Results of β -globin PCR with PC03/GH21, between human β-globin PCRwith PC03/PC04 primers (110 panfungal PCRwith U1/U2, panfungal PCRwith bp) and panfungal ISH targeting the 28S rRNA of pathogenic 18SF1/58SR1, and Aspergillus nested PCRassay with fungi (p < 0. 05). We speculated that host defense and AFU7S/7AS and AFU5S/5AS primers, which amplified antifungal agents may induce degradation of fungal nucleic longer fragments (250 bp, 260 bp, 300 bp, and 236 bp, acids and thereby lead to negative results for molecular respectively) showed lower sensitivity. Similarly, a scheme methods to detect fungal nucleic acids. Therefore, we involving smaller amplification products would be desirable. separately performed a comparative study using human β - Together, these results illustrate the challenges in identifying globin gene and -specific PCRanalysis to detect fungal Med. Mycol. J. Vol. 59 (No. 1) , 2018 E17 nucleic acids and the specimen DNA preserved in the FFPE gy and Oncology (DGHO). Ann Hematol 88: 97-110, 2009. tissues. 3) Shimodaira K, Okubo Y, Nakayama H, Wakayama M, Whereas false-positive results can be caused by PCR Shinozaki M, Ishiwatari T, Sasai D, Nemoto T, Takahashi K, Ishii T, Saji T, Shibuya K: Trends in the prevalence of invasive contamination32), the mechanism of detection in ISH excludes fungal infections from an analysis of annual records of autopsy the risk of cross contamination in detection of causative fungi. cases of Toho University. Mycoses 55: 435-443, 2012. Special preventive measures were therefore employed to avoid 4) Vennewald I, Henker M, Klemm E, Seebacher C: Fungal fungal spore contaminants. Prolonged storage of blocks colonization of the paranasal sinuses. Mycoses 42 Suppl 2: 33- covered with paraffin was done. In particular, disposable 36, 1999. blades provided the greatest protection from block-to-block 5) Thaler M, Pastakia B, Shawker TH, O'Leary T, Pizzo PA: Hepatic in cancer patients: the evolving picture of contamination. Furthermore, it is essential to change gloves the syndrome. Ann Intern Med 108: 88-100, 1988. between the cleaning of the microtome and the sectioning of 6) Kahn FW, Jones JM, England DM: The role of bronchoalveo- each new block. Histological techniques, therefore, such as lar lavage in the diagnosis of invasive pulmonary . ISH, provide useful alternatives if mycological culture is not Am J Clin Pathol 86: 518-523, 1986. successful and contamination is present. In addition, PCR- 7) Groll AH, Shah PM, Mentzel C, Schneider M, Just-Nuebling based molecular techniques can be complicated by inadequate G, Huebner K: Trends in the postmortem epidemiology of invasive fungal infections at a university hospital. J Infect 33: release of DNA during DNA extraction due to the rigidity of 23-32, 1996. 33) fungal cell walls . Altogether, a molecular diagnostic 8) Shinozaki M, Tochigi N, Sadamoto S, Yamagata Murayama S, approach using a combination of PCRand ISH can be used to Wakayama M, Nemoto T, Shibuya K: Technical aspects and overcome the drawbacks of PCR. applications for developing in situ hybridization procedures for In conclusion, this study demonstrated the potential risk of formalin-fixed and paraffin-embedded (FFPE) tissues for misdiagnosis in the classification of fungi based on patholo- diagnosis of fungal infections. Med Mycol J 58: E33-E37, 2017. gical diagnosis alone, indicating that enhanced awareness of 9) Shinozaki M, Okubo Y, Sasai D, Nakayama H, Ishiwatari T, these pitfalls should prove beneficial for increasing diagnostic Murayama S, Tochigi N, Wakayama M, Nemoto T, Shibuya K: accuracy. Additionally, pathological diagnosis in combination Development and evaluation of nucleic acid-based techniques with molecular methods such as ISH and PCRin FFPE for an auxiliary diagnosis of invasive fungal infections in samples should improve the accuracy of the pathological formalin-fixed and paraffin-embedded (FFPE) tissues. Med diagnosis of fungal infections. PCRand ISH could become a Mycol J 53: 241-245, 2012. 10) Shinozaki M, Okubo Y, Sasai D, Nakayama H, Murayama SY, reliable alternative to standard microbiological procedures. Ide T, Wakayama M, Hiruta N, Shibuya K: Identification of PCRand ISH should be used in the diagnosis of IFIs, Fusarium species in formalin-fixed and paraffin-embedded especially when conventional cultures are negative and/or sections by in situ hybridization using peptide nucleic acid culture materials could not be obtained from patients. probes. J Clin Microbiol 49: 808-813, 2011. 11) Sengüven B, Baris E, Oygur T, Berktas M: Comparison of Acknowledgments methods for the extraction of DNA from formalin-fixed, paraffin-embedded archival tissues. Int J Med Sci 11: 494- 499, 2014. This research is partially supported by the Research 12) Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn Program on Emerging and Re-emerging Infectious Diseases GT, Mullis KB, Erlich HA: Primer-directed enzymatic from Japan Agency for Medical Research and development, amplification of DNA with a thermostable DNA polymerase. AMED (17fk0108118h0901 and 17fk0108208h0002). Science 239: 487-491, 1988. 13) Willinger B, Obradovic A, Selitsch B, Beck-Mannagetta J, Buzina W, Braun H, Apfalter P, Hirschl AM, Makristathis A, Conflict of Interest Rotter M: Detection and identification of fungi from fungus balls of the maxillary sinus by molecular techniques. J Clin The authors declare that they have no conflict of interest. Microbiol 41: 581-585, 2003. 14) Makimura K, Tamura Y, Mochizuki T, Hasegawa A, Tajiri Y, References Hanazawa R, Uchida K, Saito H, Yamaguchi H: Phylogenetic classification and species identification of strains based on DNA sequences of nuclear ribosomal internal 1) Pagano L, Caira M, Candoni A, et al: The epidemiology of transcribed spacer 1 regions. J Clin Microbiol 37: 920-924, fungal infections in patients with hematologic malignancies: 1999. the SEIFEM-2004 study. Haematologica 91: 1068-1075, 2006. 15) Skladny H, Buchheidt D, Baust C, Krieg-Schneider F, Seifarth 2) Infectious Diseases Working Party (AGIHO) of the German W, Leib-Mösch C, Hehlmann R: Specific detection of Society of Hematology and Oncology (DGHO), Böhme A, Aspergillus species in blood and bronchoalveolar lavage Ruhnke M, et al: Treatment of invasive fungal infections in samples of immunocompromised patients by two-step PCR. J cancer patients—recommendations of the Infectious Diseases Clin Microbiol 37: 3865-3871, 1999. Working Party (AGIHO) of the German Society of Hematolo- E18 Medical Mycology Journal Volume 59, Number 1, 2018

16) Hofman V, Dhouibi A, Butori C, Padovani B, Gari-Toussaint H: Differentiation of Candida albicans and Candida dublinien- M, Garcia-Hermoso D, Baumann M, Vénissac N, Cathomas G, sis by fluorescent in situ hybridization with peptide nucleic Hofman P: Usefulness of molecular biology performed with acid probes. J Clin Microbiol 39: 4138-4141, 2001. formaldehyde-fixed paraffin embedded tissue for the diagnosis 25) Yeo SF, Wong B: Current status of nonculture methods for of combined pulmonary invasive and aspergil- diagnosis of invasive fungal infections. Clin Microbiol Rev 15: losis in an immunocompromised patient. Diagn Pathol 5:1, 465-484, 2002. 2010. 26) Sangoi AR, Rogers WM, Longacre TA, Montoya JG, Baron 17) Bialek R, Konrad F, Kern J, Aepinus C, Cecenas L, Gonzalez EJ, Banaei N: Challenges and pitfalls of morphologic GM, Just-Nübling G, Willinger B, Presterl E, Lass-Flörl C, identification of fungal infections in histologic and cytologic Rickerts V: PCR based identification and discrimination of specimens: a ten-year retrospective review at a single agents of mucormycosis and aspergillosis in paraffin wax institution. Am J Clin Pathol 131: 364-375, 2009. embedded tissue. J Clin Pathol 58: 1180-1184, 2005. 27) Shinozaki M, Okubo Y, Nakayama H, Mitsuda A, Ide T, 18) Hammond SP, Bialek R, Milner DA, Petschnigg EM, Baden Yamagata Murayama S, Shibuya K: Application of in situ LR, Marty FM: Molecular methods to improve diagnosis and hybridization to tissue sections for identification of identification of mucormycosis. J Clin Microbiol 49: 2151- causing invasive fungal infection. Nihon Ishinkin Gakkai 2153, 2011. Zasshi 50: 75-83, 2009. 19) Hagari Y, Ishioka S, Ohyama F, Mihara M: Cutaneous 28) Lau A, Chen S, Sorrell T, Carter D, Malik R, Martin P, infection showing sporotrichoid spread caused by Pseudalles- Halliday C: Development and clinical application of a cheria boydii (Scedosporium apiospermum): successful detec- panfungal PCRassay to detect and identify fungal DNA in tion of fungal DNA in formalin-fixed, paraffin-embedded tissue specimens. J Clin Microbiol 45: 380-385, 2007. sections by seminested PCR. Arch Dermatol 138: 271-272, 29) Moncada PA, Budvytiene I, Ho DY, Deresinski SC, Montoya 2002. JG, Banaei N: Utility of DNA sequencing for direct 20) Zhi Y, Sasai D, Okubo Y, Shinozaki M, Nakayama H, identification of invasive fungi from fresh and formalin-fixed Yamagata Murayama S, Wakayama M, Ide T, Zhang Z, specimens. Am J Clin Pathol 140: 203-208, 2013. Shibuya K: Comparison between the effectiveness of 30) Walsh TJ, Groll A, Hiemenz J, Fleming R, Roilides E, polymerase chain reaction and in situ hybridization in detecting Anaissie E: Infections due to emerging and uncommon the presence of pathogenic fungi by using the preserved DNA medically important fungal pathogens. Clin Microbiol Infect in formalin-fixed and paraffin-embedded tissues. Jpn J Infect 10: 48-66, 2004. Dis 66: 173-179, 2013. 31) Dedhia P, Tarale S, Dhongde G, Khadapkar R, Das B: 21) Hanazawa R, Murayama SY, Yamaguchi H: In-situ detection Evaluation of DNA extraction methods and real time PCR of Aspergillus fumigatus. J Med Microbiol 49: 285-290, 2000. optimization on formalin-fixed paraffin-embedded tissues. 22) Okubo Y, Shinozaki M, Wakayama M, Nakayama H, Sasai D, Asian Pac J Cancer Prev 8: 55-59, 2007. Ishiwatari T, Nemoto T, Tochigi N, Shibuya K: Applied gene 32) Frickmann H, Loderstaedt U, Racz P, Tenner-Racz K, Eggert histopathology: identification of Fusarium species in FFPE P, Haeupler A, Bialek R, Hagen RM: Detection of tropical tissue sections by in situ hybridization. Methods Mol Biol 968: fungi in formalin-fixed, paraffin-embedded tissue: still an 141-147, 2013. indication for microscopy in times of sequence-based 23) Shinozaki M, Okubo Y, Sasai D, Nakayama H, Murayama SY, diagnosis? Biomed Res Int 2015. doi: 10.1155/2015/938721. Ide T, Wakayama M, Ishiwatari T, Tochigi N, Nemoto T, [Epub ahead of print] Shibuya K: Development of a peptide nucleic acid probe to 33) Muñoz-Cadavid C, Rudd S, Zaki SR, Patel M, Moser SA, Trichosporon species and identification of by Brandt ME, Gómez BL: Improving molecular detection of use of in situ hybridization in formalin-fixed and paraffin- fungal DNA in formalin-fixed paraffin-embedded tissues: embedded (FFPE) sections. J Clin Microbiol 51: 295-298, comparison of five tissue DNA extraction methods using 2013. panfungal PCR. J Clin Microbiol 48: 2147-2153, 2010. 24) Oliveira K, Haase G, Kurtzman C, Hyldig-Nielsen JJ, Stender