Rubén Javier López | Freelancer Portfolio Item #447953

ORIGINAL RESEARCH published: 12 October 2021 doi: 10.3389/fcimb- Analysis of Salivary Mycobiome in a Cohort of Oral Squamous Cell Carcinoma Patients From Sudan Identiﬁes Higher Salivary Carriage of Malassezia as an Independent and Favorable Predictor of Overall Survival Edited by: Thuy Do, University of Leeds, United Kingdom Reviewed by: Hubert Low, Chris O’Brien Lifehouse, Australia Maha Abdelkawy, Beni-Suef University, Egypt *Correspondence: Daniela Elena Costea-Specialty section: This article was submitted to Microbiome in Health and Disease, a section of the journal Frontiers in Cellular and Infection Microbiology Received: 27 February 2021 Accepted: 27 August 2021 Published: 12 October 2021 Citation: Mohamed N, Litlekalsøy J, Ahmed IA, Martinsen EMH, Furriol J, JavierLopez R, Elsheikh M, Gaafar NM, Morgado L, Mundra S, Johannessen AC, Osman TA-H, Nginamau ES, Suleiman A and Costea DE (2021) Analysis of Salivary Mycobiome in a Cohort of Oral Squamous Cell Carcinoma Patients From Sudan Identiﬁes Higher Salivary Carriage of Malassezia as an Independent and Favorable Predictor of Overall Survival. Front. Cell. Infect. Microbiol. 11:673465. doi: 10.3389/fcimb- Nazar Mohamed 1,2, Jorunn Litlekalsøy 1, Israa Abdulrahman Ahmed 1,3, Einar Marius Hjellestad Martinsen 4, Jessica Furriol 5, Ruben Javier-Lopez 6, Mariam Elsheikh 2,7, Nuha Mohamed Gaafar 1,2, Luis Morgado 8, Sunil Mundra 8,9, Anne Christine Johannessen 1,10, Tarig Al-Hadi Osman 1, Elisabeth Sivy Nginamau 1,10, Ahmed Suleiman 2,7 and Daniela Elena Costea 1,10* 1 Gade Laboratory for Pathology, Department of Clinical Medicine, and Center for Cancer Biomarkers CCBIO, University of Bergen, Bergen, Norway, 2 Department of Oral and Maxillofacial Surgery/Department of Basic Sciences, University of Khartoum, Khartoum, Sudan, 3 Department of Operative Dentistry, University of Science & Technology, Omdurman, Sudan, 4 Department of Clinical Science, University of Bergen, Bergen, Norway, 5 Department of Nephrology, Haukeland University Hospital, Bergen, Norway, 6 Department of Biological Sciences, The Faculty of Mathematics and Natural Sciences, University of Bergen, Bergen, Norway, 7 Department of Oral & Maxillofacial Surgery, Khartoum Dental Teaching Hospital, Khartoum, Sudan, 8 Section for Genetics and Evolutionary Biology (EvoGene), Department of Biosciences, The Faculty of Mathematics and Natural Sciences, University of Oslo, Oslo, Norway, 9 Department of Biology, College of Science, United Arab Emirates University, Al Ain, Abu Dhabi, United Arab Emirates, 10 Department of Pathology, Laboratory Clinic, Haukeland University Hospital, Bergen, Norway Background: Microbial dysbiosis and microbiome-induced inﬂammation have emerged as important factors in oral squamous cell carcinoma (OSCC) tumorigenesis during the last two decades. However, the “rare biosphere” of the oral microbiome, including fungi, has been sparsely investigated. This study aimed to characterize the salivary mycobiome in a prospective Sudanese cohort of OSCC patients and to explore patterns of diversities associated with overall survival (OS). Materials and Methods: Unstimulated saliva samples (n = 72) were collected from patients diagnosed with OSCC (n = 59) and from non-OSCC control volunteers (n = 13). DNA was extracted using a combined enzymatic–mechanical extraction protocol. The salivary mycobiome was assessed using a next-generation sequencing (NGS)-based methodology by amplifying the ITS2 region. The impact of the abundance of different fungal genera on the survival of OSCC patients was analyzed using Kaplan–Meier and Cox regression survival analyses (SPPS). Results: Sixteen genera were identiﬁed exclusively in the saliva of OSCC patients. Candida, Malassezia, Saccharomyces, Aspergillus, and Cyberlindnera were the most Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 1 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC relatively abundant fungal genera in both groups and showed higher abundance in OSCC patients. Kaplan–Meier survival analysis showed higher salivary carriage of the Candida genus signiﬁcantly associated with poor OS of OSCC patients (Breslow test: p = 0.043). In contrast, the higher salivary carriage of Malassezia showed a signiﬁcant association with favorable OS in OSCC patients (Breslow test: p = 0.039). The Cox proportional hazards multiple regression model was applied to adjust the salivary carriage of both Candida and Malassezia according to age (p = 0.029) and identiﬁed the genus Malassezia as an independent predictor of OS (hazard ratio = 0.383, 95% CI = 0.16–0.93, p = 0.03). Conclusion: The fungal compositional patterns in saliva from OSCC patients were different from those of individuals without OSCC. The fungal genus Malassezia was identiﬁed as a putative prognostic biomarker and therapeutic target for OSCC. Keywords: oral squamous cell carcinoma (OSCC), mycobiome, toombak, biomarker, overall survival (OS), Malassezia, Candida Aureobasidium, Saccharomycetales, Aspergillus, Fusarium, Cryptococcus, and Malassezia (Ghannoum et al., 2010; Dupuy et al., 2014). Evidence is accumulating on the role of fungi in neoplasia (Rindum et al., 1994; McCullough et al., 2002; Barrett et al., 2008; Hebbar et al., 2013; Berkovits et al., 2016; Zhu et al., 2017; Conche and Greten, 2018; Al-Hebshi et al., 2019; Aykut et al., 2019). Some earlier studies have suggested a possible role of Candida in the initiation of carcinogenesis (Field et al., 1989; Krogh, 1990). Candida may have a causal role in oral precancer and cancer, albeit an indirect one, implying that Candida, along with other cofactors, e.g., tobacco consumption, is involved in the initiation and promotion of carcinogenesis (Bakri et al., 2010; Sanjaya et al., 2011). Some C. albicans strains may contribute to oral carcinogenesis by producing endogenous nitrosamine (Krogh et al., 1987). An immune-mediated role in the acceleration of pancreatic ductal adenocarcinoma has also been suggested recently for another genus, namely, Malassezia (Aykut et al., 2019). There are only sparse reports in the literature on the mycobiome in oral squamous cell carcinoma (OSCC). Perera et al. revealed a dysbiotic mycobiome characterized by lower species diversity and increased relative abundance of C. albicans in tissue biopsies of OSCC in a cohort of patients from Sri Lanka (Perera et al., 2017). Berkovits et al. (2016) used cultivation techniques coupled with matrix-assisted laser desorption/ ionization time-of-ﬂight mass spectrometry (MALDI-TOF MS) and identiﬁed a more diverse mycobiome associated with OSCC, mainly consisting of Candida species in addition to Rhodotorula, Saccharomyces, and Kloeckera. The oral microbiota is dynamic and responsive to environmental and biological changes, so discoverable shifts in its composition and/or function might offer new biomarkers useful for the diagnosis of oral cancers (OCs) and oropharyngeal cancers (OPCs) (Aas et al., 2005). While host biomarkers are subject to individual biological variations (Lim et al., 2017), there are indications that the core oral microbiome is consistently conserved among unrelated subjects (Lim et al., 2017; Shaw INTRODUCTION The oral cavity is a habitat for a diverse and ﬂuctuating collection of microorganisms (Aas et al., 2005; Nasidze et al., 2009; Yang et al., 2016). The oral microbiome, which includes, in addition to complex bacterial communities, oral fungi, viruses, and phages (Baker et al., 2017), is one of the most diverse microbial communities in the human body (Dewhirst et al., 2010; Huttenhower et al., 2012), and this is related to its multiple ecosystems (Arweiler et al., 2016). The oral microbiota represents a critical component of health and diseases (Jenkinson and Lamont, 2005; Avila et al., 2009), and balance is maintained by a continuous interplay with the host (Vasquez et al., 2018). Dysbiosis of the oral microbiome has been proposed as a marker, initiator, or modiﬁer of oral diseases (Ghannoum et al., 2010; Hooks and O’Malley, 2017; Iliev and Leonardi, 2017; Rosier et al., 2018). Recent advances in microbial detection techniques allowed the transition from culture-dependent studies of a single species to complex in vitro multispecies community detection and characterization studies (Baker et al., 2017). Large nextgeneration sequencing (NGS)-based projects, such as the Human Microbiome (Huttenhower et al., 2012), the Integrative Human Microbiome Project with a focus on the mechanisms of host–microbiome interactions (Proctor et al., 2019), and the Human Oral Microbiome Database (Proctor et al., 2019), give deeper insights into the human microbiome. Despite advances in the understanding of the microbiome, majority of the studies have focused on the bacterial part of the microbiome. Little is known about the fungal part of the human microbiome, recently deﬁned as the mycobiome (Ghannoum et al., 2010; Cui et al., 2013; Chandra et al., 2016). The few existing studies have revealed that the diversity of the oral mycobiota is lower when compared to that of the oral bacteriome (Iliev and Leonardi, 2017), and it is dominated by members of the phylum Ascomycota, mainly Candida spp., with Candida albicans as the dominant species. The other commonly identiﬁed fungi in the oral mycobiome are Cladosporium, Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 2 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC was reported in pack-years (PY) (Masters, 2018), with calculations for consumption of the smokeless tobacco “toombak” adjusted according to the average of manually prepared portions in Sudan (Idris et al., 1995). et al., 2017). The incorporation of the oral microbiome panel in other tumor biomarkers may therefore help reduce human biological variations, which prevented, so far, the utilization of molecular diagnosis and stratiﬁcation in OCs and OPCs (Zaura et al., 2014; Lim et al., 2017). Moreover, salivary diagnostics is a rapidly developing ﬁeld, and combined with biomarker identiﬁcation and validation, it may provide a platform for the development of a noninvasive, salivary-based tool for the stratiﬁcation of OSCC patients and for individualized treatments. This study aimed to investigate the salivary mycobiome in a cohort of OSCC patients and in non-OSCC controls from Sudan and its possible impact on clinical variables, including overall survival (OS). We employed the NGS methodology to explore fungal diversities and communities in saliva and describe the salivary fungal compositional patterns in OSCC patients compared to individuals without OSCC. The fungal genus Malassezia was identiﬁed as an independent prognostic biomarker for OS of OSCC patients. Saliva Sample Collection Unstimulated saliva samples were collected. Brieﬂy, the donor was asked not to eat and not to use oral hygiene products 1 h before saliva collection. At least 2 ml of unstimulated saliva was collected on ice and then kept in a portable liquid nitrogen container until further storage at −80°C at the end of the collection day. The sample collection time did not exceed 20 min. Fungal DNA Extraction and Control Sample Setting The recommendation for standardized DNA extraction for microbiome studies was followed (Leigh Greathouse et al., 2019). A combined enzymatic–mechanical extraction method was chosen and modiﬁed, when needed, for fungi (Huseyin et al., 2017; Rosenbaum et al., 2019). Of the saliva, 300 µl was used for DNA extraction. Sputasol® (300 µl, Oxoid Ltd., Basingstoke, UK) was added and incubated, with shaking, at 37°C for 15 min. Following centrifugation, pellets were reconstituted in 250 µl of phosphate-buffered saline (PBS). For enzymatic digestion, an enzyme cocktail of lysostaphin (4,000 U/ml), mutanolysin (25,000 U/ml), and lysozyme (10 mg/ml) was diluted in TE5 buffer (10 mM Tris-HCl and 5 mM EDTA, pH 8.0) (all from Sigma-Aldrich, Saint-Louis, MO, USA). Fifty microliters of the enzyme cocktail was added to each reconstituted pellet and mixed well, then incubated at 37°C with slight shaking at 350 rpm for 1 h. The FastDNA™ Kit (MP Biomedicals, Irvine, CA, USA) was used after enzymatic digestion. The samples were centrifuged and the pellets were lysed with 800 µl CLS-Y buffer (FastDNA™ Kit, MP Biomedicals, Irvine, CA, USA). The beadbased protocol for isolation was followed according to the manufacturer’s instructions. Two biological fungal mock communities (M1 and M2) were included in the study. Both were constituted from environmental fungi: M1 was composed of wood-decomposing polypore fungi (Mycena galopus, Mycena galericulata, Mycena leptocephala, Mycena epipterygia, Serpula lacrymans, and Amanita muscaria), and M2 was constructed from eight fungi isolated from air (Boeremia exigua var. exigua, Cladosporium, Penicillium chloroleucon, Aspergillus fumigatus, Discostroma fuscellum, Paraphaeosphaeria michotii, Mucor hiemalis, and Leptosphaerulina chartarum). Three single-species positive controls were also prepared from three Candida reference strains (C. albicans ATCC 10231, Candida parapsilosis ATCC 22019, and Candida glabrata ATCC MYA-2955). Serially diluted samples of fungal species isolated from a healthy volunteer and grown on Sabouraud dextrose agar (SDA; Sigma-Aldrich, St. Louis, MO, USA) at 37°C for 48 h were also included as controls. Dilutions (from 1:10 up to 1:106) were done in both artiﬁcial saliva (Saliva Orthana®, NycoDent, Asker, Norway) and human saliva from a volunteer that did not MATERIALS AND METHODS Ethical Considerations This is a prospective study involving OSCC patients (n = 59) and healthy non-cancer controls (n = 13) recruited between 2012 and 2015 at Khartoum Dental Teaching Hospital, Sudan. The National Health Research Ethics Committee, Federal Ministry of Health, Sudan, approved the research in Sudan (fmoh/rd/SEC/09). Written informed consent was obtained from both patients and controls. The Regional Ethical Committee in Norway approved the project (REKVest- REKVest-). Study Participants The inclusion criteria were as follows: age older than 18 years, with histologically conﬁrmed primary OSCC, did not receive any previous surgical and chemo- or radiotherapy, and consented to participate in the study. Critically ill patients, patients under medication, and those positive for human immunodeﬁciency virus (HIV) and hepatitis B surface antigen (HBs Ag) were excluded from the study. Human papilloma virus (HPV)positive cases were also excluded from the study. Detailed clinical information (age, gender, tobacco habits, and alcohol use) was obtained through interviews. A routine dental examination was performed on participating individuals, which included registration of the periodontal status, plaque, gingival index, community periodontal index (CPI), simpliﬁed oral hygiene, ﬁllings and missing teeth, and carious teeth by a team of trained and calibrated dentists speciﬁcally for this project. Non-cancer controls were included after informed consent and consecutively recruited from patients attending the outpatient clinic for trauma and benign conditions. The tumor localization, tumor size, TNM stage, comorbid conditions, last date of followup, and survival data were obtained from patients’ hospital records. TNM stage was noted according to the guidelines of the American Joint Committee on Cancer, version 7.0. Information on current smoking habits and history of smoking Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 3 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC grow fungi when cultured on SDA. The experimental setup also included three negative controls, two of which were negative extraction controls and the third one just nuclease-free water added before library normalization. et al., 2019). The ITSxpress QIIME 2 plugin (v.1.3) (Rivers et al., 2018) was used to extract the ITS2 region. The sequences were then passed through the DADA2 pipeline (Callahan et al., 2016) for ﬁltration, dereplication, chimera detection, and the merging of paired-end reads to create the so-called amplicon sequence variants (ASVs). The resultant ASVs were included for further analysis. The UNITE database (version 8) (Nilsson et al., 2019b) was trained to create a naive Bayes classiﬁer in order to classify the sequences obtained from the DADA2-generated ASV table. Postclustering curation using LULU (Frøslev et al., 2017) was performed to avoid diversity overestimation. Unidentiﬁed ASVs in the UNITE database were blasted to NCBI and the taxonomy for each was reassigned (considering an e-value and similarity or coverage ≥99% of the best hit). Various taxonomic levels were used to classify the sequence data. Species with low abundance (20 reads in less than ﬁve samples) were discarded. Three OSCC saliva samples and one non-OSCC control were excluded due to the exclusion criteria for low-abundance samples. ITS Amplicon PCR PCR ampliﬁcation was performed in a 25-µl reaction volume using 12.5 µl of KAPA HiFi HotStart® ReadyMix PCR Master Mix (Kappa Biosystems, Sigma-Aldrich) and 1 µl of the DNA template, in addition to 0.5 µl of each reverse and forward primers and nuclease-free water. The internal transcribed spacer 2 (ITS2) subregion was targeted for ampliﬁcation, as recommended (Knot et al., 2009; Nilsson et al., 2019a). ITS2 universal primer 5, 8S ITS2-F GTGAATCATCGARTCT TTGAA, and 28S1 ITS2-R TATGCTTAAGTTCAGCGGGTA (TIB, MOLBIOL, Berlin, Germany) were used to amplify the region of interest. The Veriti Thermal Cycler® (Applied Biosystems, Foster City, CA, USA) was used for ampliﬁcation. Thermal cycling was done as follows: 3 min at 95°C, initial denaturation followed by 45 cycles of 30 s at 95°C: denaturation, 60 s at 58°C as annealing, 30 s at 72°C for the extension, and a ﬁnal extension at 72°C for 5 min. The PCR products were examined by electrophoresis in a 1% (w/v) agarose gel in 1× TAE buffer. Statistical Analyses Differences in the composition of the mycobiome between the OSCC and healthy control groups, and within samples, were tested for signiﬁcance using relevant statistical tests in MicrobiomeR (Lahti et al., 2017), Phyloseq (McMurdie and Holmes, 2013), and MicrobiomeAnalystR (Dhariwal et al., 2017). Alpha diversity was calculated and plotted in Phyloseq, R version 4.0.3. QIIME2 ANCOM parameters (Bandara et al., 2019) and ALDEx2 (Fernandes et al., 2013) plugins were used for the analysis of the composition of microbiomes. The Kaplan– Meier survival estimator and Cox proportional hazards models (with “enter” method) were used for survival analysis, with OS of 2 years after diagnosis as the end point; all patients who were alive or lost to follow-up at the end of data collection were censored. Survival analysis was performed using Statistical Package for Social Sciences (SPSS), version 25 (IBM, Armonk, NY, USA). For all analyses, p-values ≤0.05 were considered to be signiﬁcant. PCR Clean-up and Library Preparation Two rounds of clean-up, one after amplicon PCR and the other after index PCR, were performed using a bead-based method (Agencourt AMPure XP, Beckman Coulter, Brea, CA, USA). After the ﬁrst round, 5 µl from each cleaned up sample was transferred to a 96-well PCR plate for indexing. The indices were arranged according to the manufacturer’s protocols. Index PCR and Library Normalization and Denaturation Nextera XT index primers (Illumina, San Diego, CA, USA) were used for indexing. Index PCR was carried out on the Veriti Thermal Cycler ® (Applied Biosystems) with parameters recommended by Illumina (San Diego, CA, USA). One microliter of a 1:50 dilution of each sample was used for library validation using a Bioanalyzer® DNA 1000 Chip (DNA LabChip® using 2100 Bioanalyzer, Agilent Technologies, Santa Clara, CA, USA). The DNA concentrations of the index PCR products were measured with the Qubit 3.0 Fluorometer® (Invitrogen, Carlsbad, CA, USA), and the DNA concentration was calculated in nanomolars based on the size of the DNA amplicons determined using Bioanalyzer®. The normalized library was combined with HT1 and PhiX, as recommended by Illumina. The MiSeq Reagent Kit v.3 (600 cycles; Illumina, San Diego, CA, USA) was used for library denaturation and MiSeq sample loading. Sequencing was performed on the Illumina MiSeq platform using a 2 × 300-bp paired-end protocol. RESULTS Cohort Description The prospective cohort included 59 patients (age range = 25–87 years, mean = 50.6 years, median = 60 years) with histologically proven OSCC and 13 non-OSCC controls (age range = 30–70 years, mean = 46.5 years, median = 45 years). Patients in the OSCC group presented more tobacco consumption (expressed in pack-years for both smoking and smokeless tobacco taken together) than did the controls, although the difference was not statistically signiﬁcant (p = 0.06) (Table 1). The average number of decayed teeth (DT) was similar to the general Sudanese population, as previously evaluated (Khalifa et al., 2012), except for the age groups 25–44 and >65 years in our cohort, which showed a higher number of decayed teeth compared to the general Sudanese population. The same was found for missing teeth (MT) (Table 1). The mean plaque index of OSCC patients Bioinformatics Processing Demultiplexed Illumina-generated paired-end sequences were processed using QIIME 2 (version QIIME2-2020.8) (Bolyen Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 4 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC TABLE 1 | Demographics, oral health ﬁndings, and clinicopathological ﬁndings of the cohort. Cohort demographics No. of individuals Age (years), mean Males Females No. of users Pack-years (PY), mean (p = 0.06**) Non-OSCC controls 13 (7 males, 6 females) 45.4 - (39–70) 2 (15%, all males) 4.5 Patients 59 (42 males, 17 females) 60 - (40–80) 33 (56%, all males) 51.1 Non-OSCC controls- ± 0.67 1.19 ± 0.31 OSCC patients- ± 0.64 1.67 ± 0.56 Oral ﬁndings DT (p = 0.630) MT (p = 0.287) Community periodontal index, mean ± SD (CPI: p = 0.013*) Gingival index, mean ± SD (p = 0.014*) Missing teeth and decayed teeth in the OSCC cohorta Age groups (years-+ DT OSCC- Non-OSCC 0 1.4 3.2 5 – – General population*- Non-OSCC- – MT OSCC- General population*- Tobacco and alcohol consumption History Yes—current user No Past user Unknown Toombak, N (%) Patients Non-OSCC controls 5 (8.5) 0 (0) 30 (50.8) 11 (84.6) 22 (37.3) 1 (7.7) 2 (3.4) 1 (7.7) Smoking, N (%) Patients Non-OSCC controls 7 (11.9) 2 - (77) 12 0 (0) 2 (3.4) 1 (7.7) Alcohol, N (%) Patients Non-OSCC controls 1 (1.7) 0 (0) 39 (66.1) 11 (86.6) 13 (22) 1 (1.7) 6 (10.2) 1 (1.7) OSCC patients: clinical ﬁndings Tumor location Buccolabial–sulcus N (%) 32 (54.2) Tongue Retromolar–palatal–alveolar 5 (8.5) 15 (25.4) Tumor stage N (%) I II III IV 0 (0) 2 (3.4) 13 (22) 37 (62.7) T1 T2 T3 T4 Tx Missing, N (%) T stage N (%) 2 (3.4) 11 (18. 6) 15 (25.4) 20 (33.9) 4 (6.8) 7 (11.9) N stage N (%) N0 4 (6.8) N1 19 (32.2) N2 27 (45.8) N3 1 (1.7) Nx 1 (1.7) M stage N (%) M0 39 (66.1) M1 1 (1.7) Mx 12 (20.3) OSCC, oral squamous cell carcinoma; DT, mean number of decayed teeth; MT, mean number of missing teeth. Signiﬁcantly different at p < 0.05 (*Kruskal–Wallis and **Mann–Whitney U test). a Based on Khalifa et al. (2012). was comparable to that of the control group (p = 0.59), while the gingival index (mean ± SD = 1.67 ± 0.56) was signiﬁcantly higher for the OSCC group (p = 0.014) than that for the control group (mean ± SD = 1.19 ± 0.31). The localization of OSCC lesions was predominantly lower buccal or labial (40.4%); only ﬁve cases (6.9%) were localized on the tongue. Of all OSCC patients, 47 (79.6%) presented with locoregional lymph node metastases at the time of diagnosis. Nearly all OSCC patients (96.8%) presented at a late stage (Table 1). The extracted ITS2 region was merged and temporarily clustered into 6,699,920 amplicon reads. After DADA2 ﬁltration, dereplication, chimera detection, and merging of paired-end reads, a total of 3,514,250 reads were retained for further analysis. The quality-ﬁltered, denoised, chimera-removed sequence reads were clustered into 514 ASVs. Post-clustering curation using LULU (Frøslev et al., 2017) and removal of contaminants using the Decontam algorithm (Davis et al., 2018) retained 340 ASVs. The rarefaction curves curves are presented in Supplementary Figure 1. The positive and negative controls showed the expected reference strains (for positive ones) and negative outputs (negative ones). The M2 mock community showed good distribution, while M1 showed a generally quite good coverage, Method Performance A total of 21,698,808 Illumina-generated demultiplexed fungal ITS raw paired-end sequences were imported into QIIME2. Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 5 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC Agaricus, Alternaria, Cladosporium, Clavispora, Naganishia, Nakaseomyces, Penicillium, Rhizopus, Vishniacozyma, and Sarocladium were the second most commonly identiﬁed fungal genera (Figure 1B). Candida was found to have a higher relative abundance in the saliva of females than that of males and accounted for more than half of the genera present in females (Figure 2A). There was no difference in the diversity of the salivary mycobiome between females and males (Figure 2B). Eight genera were detected exclusively in the saliva of tobacco users (when analyzing together toombak dippers and smokers), of which seven were shared by smokers and users of toombak (Figures 2C–E): Macrophomina, Schizophyllum, Cinereomyces, Leucosporidium, Rhodosporidiobolus, Cutaneotrichosporon, and an unidentiﬁed one belonging to the family Ustilaginaceae. Lodderomyces was detected only in the saliva of smokers. Phlebiopsis and Filobasidium were detected only in the saliva of non-tobacco users. No statistically signiﬁcant differences in the overall oral mycobiome diversity were observed between non-tobacco users and smokers or toombak users, even when considering only the OSCC cases (Figures 2F–H), although a although taxonomic assignment was obtained correctly only to the class level (Amanita and Mycena in M1 were identiﬁed at the order level, i.e., Agaricomycetes). The composition of our mock communities is reﬂected in the analysis results, indicating minimal cross-contamination and tag switching. The distribution pattern of the total reads followed the serial dilutions we made (Supplementary Figure 2). Abundance analysis of the serially diluted samples showed a pattern corresponding with the inputs of the diluted samples (Supplementary Figure 3). Candida, Saccharomyces, Malassezia, Aspergillus, and Cyberlindnera Were Identiﬁed to Be the Most Common Fungi Present in the Salivary Mycobiome Processed, quality-ﬁltered ASVs were assigned to 36 different fungal genera. Relative abundance analysis showed that the salivary mycobiome was dominated by ﬁve genera, namely, Candida, Saccharomyces, Malassezia, Aspergillus, and Cyberlindnera (Figure 1A and Supplementary Figure 4). A B FIGURE 1 | (A) Relative abundance of the top ﬁve genera in the saliva of the individuals investigated in our cohort. (B) Heat map showing the relative abundance of the top 20 genera (X-axis sorted non-OSCC controls to the left and OSCC to the right) in each of the investigated sample. OSCC, oral squamous cell carcinoma. Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 6 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC A B C D E F G H FIGURE 2 | (A, B) Relative abundance of the top ﬁve genera (A) and diversity of the salivary mycobiome (B) of the individuals investigated in our cohort grouped by gender. (C) Relative abundance of the top ﬁve genera in tobacco users versus non-smokers. (D) Relative abundance of the top ﬁve genera in smokers, toombak users, and non-smokers. (E) Venn diagram showing the distribution of genera in smokers, toombak users, and non-smokers. (F) Relative abundance of the top ﬁve genera in tobacco users versus non-smokers in the oral squamous cell carcinoma (OSCC) group. (G) Relative abundance of the top ﬁve genera in smokers, toombak users, and non-smokers in the OSCC group. (H) Venn diagram showing the distribution of genera in smokers, toombak users, and non-smokers in the OSCC group. needed complex periodontal treatments such as root planing or periodontal surgical procedures (CPI higher than 3) showed a higher relative abundance of Aspergillus and a lower relative abundance of Malassezia than did those in the other two groups (Supplementary Figure 7B). Individuals with intermediate CPI (1.1–3), who needed to undergo plaque control procedures, showed the lowest diversity of fungi compared to other subjects with clinically higher or lower CPIs. Individuals with poor oral hygiene, quantiﬁed by the use of a simpliﬁed oral trend toward somehow restricted diversities in smokers or toombak users was observed (Supplementary Figure 6). Individuals aged 55–64 years showed the least relative abundance of Candida and the highest abundance of Aspergillus in their saliva (Supplementary Figure 5A). Individuals with severe gingivitis showed a predominance of species other than the identiﬁed top ﬁve genera in their saliva (Supplementary Figure 7A), along with a gradually reduced diversity, compared to the other two groups. Individuals who Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 7 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC cases), while Malassezia arunalokei was the second most predominant species in the saliva of non-OSCC controls (66.70% of controls, n = 12; 64.3% of OSCC patients, n = 56). Additionally, different species of Malassezia were identiﬁed in the saliva of OSCC patients and non-OSCC controls. Malassezia globosa (64.3%), Malassezia restricta (16%), Malassezia dermatis (5.4%), Malassezia furfur (3.5%), and Malassezia sloofﬁae (1.8%) were identiﬁed in the saliva of OSCC patients. In the saliva of non-OSCC controls, only M. restricta (33.3%) and M. globosa (58.3%) were identiﬁed. Cyberlindnera had lower abundance in the saliva of OSCC cases than that of non-OSCC controls (relative abundance and log-transformed count in each case shown in Figures 3B and 4E, respectively). Cyberlindnera jadinii (synonym: Pichia jadinii) was detected in the saliva of 50% of the OSCC patients, while it was present in 58.3% of the non-OSCC controls, showing an inverse relation to C. albicans in more than half of the whole group (56% of the whole cohort), although the bivariate correlation was statistically not signiﬁcant (correlation = −0.257, p = 0.1). hygiene index, showed higher relative abundance of Candida, Aspergillus, and Saccharomyces and a trend toward a lower diversity of fungi (Supplementary Figure 7C). Individuals with the number of decayed teeth higher than that of the mean value for the Sudanese population had a lower relative abundance of Candida but a higher relative abundance of Saccharomyces than the rest of the participants (Supplementary Figure 7D). The opposite was observed for individuals with the number of missing teeth higher than that of the mean value for the Sudanese population (Supplementary Figure 7E). The alpha diversity median was also higher, although statistically not signiﬁcant, for the salivary mycobiome of individuals with more decayed and missing teeth. Sixteen Genera Were Identiﬁed Exclusively in the Salivary Mycobiome of OSCC Patients The extracted DNA content in the samples from OSCC patients was signiﬁcantly higher than that in the samples from nonOSCC controls, as evaluated using two different approaches (Qubit® and Bioanalyzer®) (p < 0.05). Twenty genera were found in the saliva of both the OSCC and non-OSCC groups. Sixteen genera were found exclusively in the saliva of OSCC patients (Figure 3A). Univariate statistical comparison of the relative abundance of the top ﬁve genera showed no statistically signiﬁcant differences between the two groups; the same top ﬁve most abundant genera were found in both groups (Figures 3B and 4). Alpha diversity analysis, considering richness and evenness, did not show statistically signiﬁcant differences between OSCC patients and non-OSCC controls (Figure 3C). Non-metric multidimensional scaling (NMDS) and analysis of similarities (ANOSIM)/permutational multivariate analysis of variance (PERMANOVA) were applied in order to test for dissimilarities in the mycobiome composition between OSCC patients and non-OSCC controls. There was no shift observed between the study groups (NMDS stress > 0.2); statistical signiﬁcance was marginal with ANOSIM (pANOSIM = 0.056) and non-signiﬁcant with PERMANOVA (p = 0.265). Differential abundance analysis using ANCOM and ALDex2 did not show any differentially abundant genera when comparing the OSCC group and the non-OSCC control group. Although not statistically signiﬁcant, the salivary carriage of Candida was higher in the saliva of OSCC cases than that in nonOSCC controls (relative abundance and log-transformed count in each case shown in Figures 3B and 4A, respectively). The Candida species identiﬁed in the saliva of those in the OSCC group were C. albicans (78.8% of all OSCC cases), Candida tropicalis (32.1%), C. parapsilosis (37.5%), C. glabrata (16.1%), Candida orthopsilosis (3.6%), and Candida sake (9%). C. orthopsilosis and C. sake were among the fungi identiﬁed exclusively in the saliva of OSCC patients. In the saliva of OSCC cases, Saccharomyces also had a higher abundance than that in the saliva of non-OSCC controls (relative abundance and log-transformed count in each case shown in Figures 3B and 4C, respectively). Saccharomyces cerevisiae was second to C. albicans in the saliva of OSCC cases (76.8% of OSCC Frontiers in Cellular and Infection Microbiology | www.frontiersin.org Malassezia Was Identiﬁed as an Independent Predictor of OS for OSCC Patients The saliva of OSCC patients with tumors located in labial, buccal, or alveolar areas (toombak dipping areas) showed a lower relative abundance of Candida but a higher relative abundance of Cyberlindnera compared to patients with OSCC located in other sites (Supplementary Figure 5B). OSCC patients with locoregional lymph node involvement showed higher relative abundance of Candida and Aspergillus and a lower relative abundance of Malassezia compared to the group with no lymph node involvement (Supplementary Figure 5C). The same trend was observed for the OSCC patients who died during the follow-up period compared to those still alive at the end of the study (relative abundance in and log-transformed count in each case shown in Figures 3D and 4F–J, respectively). Alpha diversity analysis revealed that lower diversity index values were more commonly found in OSCC patients with locoregional lymph node involvement and those with poorer survival (Figure 3E), although not statistically signiﬁcant. A trend toward a lower relative abundance of Saccharomyces and a higher relative abundance of Aspergillus with stage has also been observed (Supplementary Figure 5D). Alpha diversity analysis showed no statistically signiﬁcant differences between stages. Kaplan–Meier analysis revealed that a high relative abundance of Candida was associated with poor OS in OSCC patients (Breslow test: p = 0.043) (Figure 5A). On the contrary, a high relative abundance of Malassezia showed association with favorable survival in OSCC patients (Breslow test: p = 0.039) (Figure 5B). The Cox proportional hazards multiple regression model was applied to adjust the salivary carriage of both Candida and Malassezia for age (p = 0.029) and identiﬁed Malassezia as an independent predictor of OS (hazard ratio = 0.383, 95% CI = 0.16–0.89, p = 0.03). 8 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC A B C D E FIGURE 3 | (A) Venn diagram showing the number of genera identiﬁed in the oral squamous cell carcinoma (OSCC) and non-OSCC groups. Those that were found exclusively in the OSCC group were: Macrophomina, Ramularia, Aureobasidium, Alternaria, Ulocladium, Lodderomyces, Meyerozyma, Schizophyllum, Cinereomyces, Phlebiopsis, Rhodosporidiobolus, Rhodotorula glutinis, Filobasidium, Cutaneotrichosporon, unidentiﬁed1, and unidentiﬁed2. (B) Relative abundance in the OSCC and non-OSCC groups showing the top ﬁve most predominant genera. (C) Alpha rarefaction curve showing the observed features (richness) at different sequencing depths. (D) Relative abundance of individuals (alive and dead) in the OSCC groups showing the top ﬁve most predominant genera. (E) Alpha rarefaction curve showing the observed features (richness) at different sequencing depths for OSCC patients stratiﬁed by overall survival (OS). Mukherjee et al., 2014; Chandra et al., 2016), studies on the mycobiome in disease and health are scarce, and the actual contribution of the mycobiota in carcinogenesis has only recently been explored (Perera et al., 2017; Al-Hebshi et al., 2019; DISCUSSION AND CONCLUSION Although the baseline mycobiome proﬁles utilizing NGS have been established for some time (Ghannoum et al., 2010; Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 9 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC A B C D E F G H I J FIGURE 4 | (A–E) Box plots for the abundance of the top ﬁve genera in the the oral squamous cell carcinoma (OSCC) group versus non-OSCC controls. (F–J) Box plots for the relative abundance of the top ﬁve genera in OSCC patients who were dead or alive at the end of the study period. Classical univariate statistical comparison of the relative abundance showed no statistically signiﬁcant differences. Aykut et al., 2019). This study is one of the ﬁrst characterizing the salivary mycobiome in OSCC and provides signiﬁcant information despite the fact that it has been run on a relatively smaller number of cases, particularly of non-OSCC controls. Another limitation of this study is that data on antibiotic use were missing. In Sudan, despite instructions, the misuse of antibiotics is a common problem, and the use of antibiotics is known to affect the results of mycobiome analysis (Awad et al., 2007; Oleim et al., 2019). Most mycobiome studies focused on either the ITS1 or the ITS2 subregion of typically 250–400 bases. Targeting the ITS2 subregion has the additional advantage of including lower length variations and more universal primer sites, resulting in less taxonomic bias than when targeting ITS1 (Nilsson et al., 2019a). In our study, by utilizing 2 × 300-bp sequencing and by merging paired reads, we obtained better taxonomic resolution since the full ITS2 length was covered. We used a relative abundance cutoff of 1%, as used by other studies (Ghannoum et al., 2010; Perera et al., 2017). The inclusion of negative controls (no saliva sample), positive controls (known species most likely to be found in the samples), and of mock communities was done as a standard for proper assessment and quantiﬁcation of tag switching, chimera formation, ASV inference stringency, and abundance shifts (Bakker, 2018; Nilsson et al., 2019a). After evaluating the controls, the overall methods used here for DNA extraction, Frontiers in Cellular and Infection Microbiology | www.frontiersin.org sequencing, and bioinformatics analysis were considered to be sensitive for salivary mycobiome identiﬁcation, under the conditions and aims of our study. Although the concept of a healthy core oral mycobiome (Ghannoum et al., 2010) was redeﬁned (Dupuy et al., 2014), with 14 core genera detected in healthy individuals, the overall abundance and diversity of fungal taxa may also be somewhat individualized (Witherden et al., 2017). It is considered that the vast majority of the mycobiome consists of a few genera, with C. albicans and C. parapsilosis as the major species of the human oral mycobiome (Naglik et al., 2013). The most abundant genera found in our study are in line with these previous baseline ﬁndings and with other OSCC-associated salivary mycobiomes reported in previous studies (Ghannoum et al., 2010; Dupuy et al., 2014; Mukherjee et al., 2017; Perera et al., 2017). Previous studies on the dynamics of the oral bacterial community showed enrichment in both abundance and function with OSCC staging (Yang et al., 2018). We found an enriched but somehow less diverse fungal mycobiome in the most advanced stage group. This might be related to the limited number of cases in the early stages in our cohort. Late tumor stage presentation is typical for OSCC in Sudan (Osman et al., 2010), and as mentioned, this is limiting the conclusions we could draw on the differences between stages in the salivary mycobiome of this cohort. The salivary microbiome was previously found to be related to dental ﬁndings. Gazdeck et al. found a lower bacterial diversity 10 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC to enhance the invasion of OSCC cells by producing speciﬁc proteinases capable of degrading the basement membrane and the extracellular matrix (Bakri et al., 2010). The inﬂammatory response to C. albicans is mediated by NF-kB (Müller et al., 2007), which is frequently involved in carcinogenesis where cancer-related inﬂammation is a feature (Mantovani et al., 2008). Taking this into consideration, our ﬁnding of the association between Candida and poor prognosis might rely on a biological explanation. On the other hand, we found the salivary carriage of Malassezia as an independent predictor of better prognosis. Malassezia has a unique pattern of interaction with pattern recognition receptors compared to C. albicans (Goyal et al., 2018). Additionally, Malassezia has large intraspecies diversity. The exact composition of different Malassezia species at a time point may contribute to different outcomes in the interaction between the fungus and the host (Sparber and LeibundGut-Landmann, 2017). Malassezia might have been overrepresented in our study, although it was described as part of the redeﬁned core oral mycobiome in humans (Dupuy et al., 2014). Since Malassezia is a normal skin commensal with population densities peaking between 20 and 45 years (Ashbee, 2007), the sample collection method we used might have included some contamination from the lips, in addition to the age-related differences in Malassezia enrichment. However, the interest in discovering microbiome proﬁles associated with survival is growing (Plantinga et al., 2017; Koh et al., 2018). Different methods are used to associate the microbiota at the community level and censored survival time, such as MiRKAT-S (Plantinga et al., 2017) and its follower OMiSA (Koh et al., 2018). We chose to test the categorization of microbial proportions into high and low and run conventional survival analysis. This seems an attractive way to incorporate microbiome signatures in clinically applicable diagnostic tools. Of interest is that we did not identify Hannaella and Gibberella, which were found enriched in a cohort of OSCC patients in a recent study and considered to be contaminants (Perera et al., 2017). This might be an indication of the effects of dietary habits, among others, and population-related differences in the mycobiome proﬁles. These two species are known plant fungi. Nevertheless, the contribution of environment-related fungi to the carcinogenic process cannot be ignored. The sample type, the method of collection, and, very importantly, the methods for DNA extraction and bioinformatics processing could explain the observed differences between species reported in different studies (Brooks et al., 2015). Curation of the databases used for taxonomical assignment could also affect the ﬁndings (Seed, 2015), in addition to the more classical factors such as ethnic differences and diet (Deschasaux et al., 2018). Ethnic differences could be related to different single nucleotide polymorphisms associated with susceptibility to fungi (Romani, 2011). The role of genetic host susceptibility should not be ignored when considering the diversity changes or the associations of the mycobiota with diseases. In any case, a further, more thorough investigation of mycobiome meta-transcriptomes and metaproteomes is needed to answer such questions related to the epidemiological patterns of mycobiomes (Huttenhower et al., 2012). A B FIGURE 5 | Kaplan–Meier survival curves showing the impact of salivary Candida (A) and Malassezia (B) on the overall survival of oral squamous cell carcinoma (OSCC) patients. in edentulous patients (Gazdeck et al., 2019). We report here higher diversity median values associated with more missing and decayed teeth. This might indicate relevant fungal–bacterial interactions (Deveau et al., 2018) that need further longitudinal studies for ﬁnal elucidation. For a long time, our understanding of periodontal disease has been based on its bacterial origin (Hajishengallis and Lamont, 2012). However, the crosstalk between fungi and bacteria seems to result in different outcomes for the host. This relationship ranges from synergism to antagonism for different speciﬁc microbial interactions (Krüger et al., 2019). Peters et al. showed Candida species to be more represented in subjects with periodontal disease and more missing teeth count (Peters et al., 2017), and, in accordance with this, C. albicans was shown to enhance Porphyromonas gingivalis invasion in vitro (Tamai et al., 2011). We observed the same trend for individuals with higher number of missing teeth. When it comes to its role in carcinogenesis, in addition to the direct role of Candida by producing nitrosamines, it was shown that it also affects the metabolism of procarcinogens and inﬂuences other bacteria, which may play a role in carcinogenesis (Hooper et al., 2009). The combinatorial effect of carcinogens and C. albicans was shown to promote OC in a murine model (Dwivedi et al., 2009). C. albicans was also shown Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 11 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC Also worth mentioning is the fact that the cohort analyzed in this study included patients consuming a special type of smokeless tobacco, the “toombak” (the local form of smokeless tobacco used in Sudan). Regarding our ﬁndings on tobacco consumption (both smoking and the smokeless tobacco toombak), it is worth noting that there may be bias related to self-reporting. However, the predominant site for tumor localization in our OSCC cohort was lower buccal or labial and sulcular, consistent with a toombakrelated OSCC etiology, and this also correlated with the selfreported habit of packing toombak in the mouth in our cohort. Self-reporting of alcohol consumption should be considered with caution as well since it is illegal in Sudan and may carry a social stigma (Gadelkarim Ahmed and Ahmed, 2013). Previous studies have shown that tobacco exposure was associated with a shift of the oral bacteriome at the population level (Beghini et al., 2019). Our study showed the same trend for the oral mycobiome. Some of the genera were identiﬁed exclusively in tobacco users, including toombak users, since many consume toombak besides smoking, and some of the genera, such as Schizophyllum, are known plant pathogens; thus, they may be related to the processed tobacco product. In conclusion, the present study reveals that Candida, Malassezia, Saccharomyces, Aspergillus, and Cyberlindnera are the most relatively abundant fungal genera in the salivary microbiome of this cohort of Sudanese individuals. Candida and Malassezia were shown to have an impact on the survival of OSCC patients: a higher salivary carriage of the genus Candida was found to be associated with poor prognosis, while Malassezia was enriched in patients with favorable prognosis, although only the salivary carriage of Malassezia emerged as an independent prognostic biomarker for the survival of OSCC patients. This can serve as groundwork for performing mycobiome-based biomarker studies in larger cohorts of OSCC patients. RJ-L, LM, SM, TO, and EN did the formal analysis. NM, JL, IA, ME, TO, EN, AS, AJ, and DC contributed to the investigation. NM prepared the original draft. All co-authors reviewed and edited the manuscript. NM, EM, JF, RJ-L, SM, TO, and DC contributed to visualization. AJ, TO, EN, AS, and DC supervised the study. EN, AS, AJ, and DC administered the project. AJ and DC helped with funding acquisition. All authors contributed to the article and approved the submitted version. FUNDING This work was supported by the Research Council of Norway through its Centers of Excellence funding scheme (grant no. 22325), Helse Vest (grant no. 911902/2013 and 912260/2019), and the Norwegian Centre for International Cooperation in Education (project no. NORPART-2018/10277). ACKNOWLEDGMENTS The authors would like to thank to Prof. Audun Nerland, Dr. Øyvind Kommedal, Dr. Christine Drengnes, Mrs. Tharmini Kalananthan, and Mrs. Sonja Ljostveit for help with optimization of DNA extraction and sequencing. SUPPLEMENTARY MATERIAL The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb-/full#supplementary-material Supplementary Figure 1 | Sequencing depth curves for OSCC and nonOSCC controls. DATA AVAILABILITY STATEMENT Supplementary Figure 2 | Pie charts depicting expected and detected distribution of mock communities. Stars refers to identical genera detected. The raw data represented in this study has been deposited and is publicly available from SRA (PRJNA722859); NCBI PRJNA722859. Supplementary Figure 3 | Bar-graph representing the number of reads according to the serial dilution of samples. Y axis represents log scale of proportions. AS - sample diluted in Artiﬁcial Saliva, HS - sample diluted in Human Saliva that did not grow fungi when cultured on Sabraud’s Dextrose agar medium. ETHICS STATEMENT Supplementary Figure 4 | Bar plots showing the relative abundance in saliva of OSCC cases and non-OSCC controls. Five most dominant genera were shown. The National Health Research Ethics Committee, Federal Ministry of Health, Sudan, approved the research in Sudan (fmoh/rd/SEC/09). Also, the Regional Ethical Committee in Norway approved this project (REKVest- REKVest-). The patients/participants provided their written informed consent to participate in this study. Supplementary Figure 5 | (A) Distribution and differences in abundance of 5 topmost salivary fungi according to age. (B) Distribution and differences in abundance of 5 topmost salivary fungi according to tumour localization. (C) Distribution and differences in abundance of 5 topmost salivary fungi according to lymph node metastasis status. (D) Distribution and differences in abundance of 5 topmost salivary fungi according to tumour stage. Supplementary Figure 6 | Diversity of the overall oral mycobiome of the individuals of our cohort grouped by tobacco use (no tobacco users, smokers and Toombak users). AUTHOR CONTRIBUTIONS Supplementary Figure 7 | (A) Distribution and differences in abundance of 5 topmost salivary fungi according to Gingival Index. (B) Distribution and differences in abundance of 5 topmost salivary fungi according to Community Periodontal Index. NM and DC conceptualized the study. NM, JL, IA, EM, JF, RJ-L, LM, SM, TO, and EN helped with the methodology. NM, EM, JF, Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 12 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC (C) Distribution and differences in abundance of 5 topmost salivary fungi according to Oral Hygiene Index (Simpliﬁed). (D) Distribution and differences in abundance of 5 topmost salivary fungi according to mean number of decayed teeth. (E) Distribution and differences in abundance of 5 topmost salivary fungi according to mean number of missing teeth. (F) Distribution and differences in abundance of 5 topmost salivary fungi according to diabetes status. REFERENCES Davis, N. M., Di, M., Holmes, S. P., Relman, D. A., and Callahan, B. J. (2018). Simple Statistical Identiﬁcation and Removal of Contaminant Sequences in Marker-Gene and Metagenomics Data. Microbiome 6, 226. doi: 10.1186/ s- Deschasaux, M., Bouter, K. E., Prodan, A., Levin, E., Groen, A. K., Herrema, H., et al. (2018). Depicting the Composition of Gut Microbiota in a Population With Varied Ethnic Origins But Shared Geography. Nat. Med. 24,-. doi: 10.1038/s- Deveau, A., Bonito, G., Uehling, J., Paoletti, M., Becker, M., Bindschedler, S., et al. (2018). Bacterial-Fungal Interactions: Ecology, Mechanisms and Challenges. FEMS Microbiol. Rev. 42, 335–352. doi: 10.1093/femsre/fuy008 Dewhirst, F. E., Chen, T., Izard, J., Paster, B. J., Tanner, A. C. R., Yu, W. H., et al. (2010). The Human Oral Microbiome. J. Bacteriol. 192,-. doi: 10.1128/JB- Dhariwal, A., Chong, J., Habib, S., King, I. L., Agellon, L. B., and Xia, J. (2017). MicrobiomeAnalyst: A Web-Based Tool for Comprehensive Statistical, Visual and Meta-Analysis of Microbiome Data. Nucleic Acids Res. 45, W180–W188. doi: 10.1093/nar/gkx295 Dupuy, A. K., David, M. S., Li, L., Heider, T. N., Peterson, J. D., Montano, E. A., et al. (2014). Redeﬁning the Human Oral Mycobiome With Improved Practices in Amplicon-Based Taxonomy: Discovery of Malassezia as a Prominent Commensal. PloS One 9 (3), e90899. doi: 10.1371/journal.pone- Dwivedi, P. P., Mallya, S., and Dongari-Bagtzoglou, A. (2009). A Novel Immunocompetent Murine Model for Candida Albicans-Promoted Oral Epithelial Dysplasia. Med. Mycol. 47, 157–167. doi: 10.1080/- Fernandes, A. D., Macklaim, J. M., Linn, T. G., Reid, G., and Gloor, G. B. (2013). ANOVA-Like Differential Expression (ALDEx) Analysis for Mixed Population RNA-Seq. PLoS One 8:67019. doi: 10.1371/journal.pone- Field, E. A., Field, J. K., and Martin, M. V. (1989). Does Candida Have a Role in Oral Epithelial Neoplasia? Med. Mycol. 27, 277–294. doi: 10.1080/- Frøslev, T. G., Kjøller, R., Bruun, H. H., Ejrnæs, R., Brunbjerg, A. K., Pietroni, C., et al. (2017). Algorithm for Post-Clustering Curation of DNA Amplicon Data Yields Reliable Biodiversity Estimates. Nat. Commun. 8, 1–11. doi: 10.1038/ s--x Gadelkarim Ahmed, H., and Ahmed, H. G. (2013). Aetiology of Oral Cancer in the Sudan. J. Oral. Maxillofac. Res. 4, e3. doi: 10.5037/jomr- Gazdeck, R. K., Fruscione, S. R., Adami, G. R., Zhou, Y., Cooper, L. F., and Schwartz, J. L. (2019). Diversity of the Oral Microbiome Between Dentate and Edentulous Individuals. Oral. Dis. 25, 911–918. doi: 10.1111/odi.13039 Ghannoum, M. A., Jurevic, R. J., Mukherjee, P. K., Cui, F., Sikaroodi, M., Naqvi, A., et al. (2010). Characterization of the Oral Fungal Microbiome (Mycobiome) in Healthy Individuals. PLoS Pathog. 6,-. doi: 10.1371/journal. ppat- Goyal, S., Castrilló n-Betancur, J. C., Klaile, E., and Slevogt, H. (2018). The Interaction of Human Pathogenic Fungi With C-Type Lectin Receptors. Front. Immunol. 9, 1261. doi: 10.3389/ﬁmmu- Hajishengallis, G., and Lamont, R. J. (2012). Beyond the Red Complex and Into More Complexity: The Polymicrobial Synergy and Dysbiosis (PSD) Model of Periodontal Disease Etiology. Mol. Oral. Microbiol. 27, 409–419. doi: 10.1111/ j-.x Hebbar, P. B., Pai, A., and Sujatha, D. (2013). Mycological and Histological Associations of Candida in Oral Mucosal Lesions. J. Oral. Sci. 55, 157–160. doi: 10.2334/josnusd.55.157 Hooks, K. B., and O’Malley, M. A. (2017). Dysbiosis and Its Discontents. MBio 8, e- doi: 10.1128/mBio- Hooper, S. J., Wilson, M. J., and Crean, S. J. (2009). Exploring the Link Between Microorganisms and Oral Cancer: A Systematic Review of the Literature. Head Neck 31,-. doi: 10.1002/hed.21140 Huseyin, C. E., Rubio, R. C., O’Sullivan, O., Cotter, P. D., and Scanlan, P. D. (2017). The Fungal Frontier: A Comparative Analysis of Methods Used in the Study of the Human Gut Mycobiome. Front. Microbiol. 8, 1432. doi: 10.3389/ fmicb- Aas, J. A., Paster, B. J., Stokes, L. N., Olsen, I., and Dewhirst, F. E. (2005). Deﬁning the Normal Bacterial Flora of the Oral Cavity. J. Clin. Microbiol. 43,-. doi: 10.1128/JCM- Al-Hebshi, N. N., Borgnakke, W. S., and Johnson, N. W. (2019). The Microbiome of Oral Squamous Cell Carcinomas: A Functional Perspective. Curr. Oral. Heal. Rep. 6, 145–160. doi: 10.1007/s- Arweiler, N. B., Netuschil, L., Arweiler, N. B., and Netuschil, L. (2016). “The Oral Microbiota,” in Microbiota of the Human Body Implications in Health and Disease. Ed. A. W. Walker (Cham: Springer), 45–60. doi: 10.1007/-_4 Ashbee, H. R. (2007). Update on the Genus Malassezia. Med. Mycol. 45, 287–303. doi: 10.1080/- Avila, M., Ojcius, D. M., and Yilmaz, Ö. (2009). The Oral Microbiota: Living With a Permanent Guest. DNA Cell Biol. 28, 405–411. doi: 10.1089/dna- Awad, A. I., Ball, D. E., and Eltayeb, I. B. (2007). Improving Rational Drug Use in Africa: The Example of Sudan. East. Mediterr. Heal. J. 13,-. doi:-/- Aykut, B., Pushalkar, S., Chen, R., Li, Q., Abengozar, R., Kim, J. I., et al. (2019). The Fungal Mycobiome Promotes Pancreatic Oncogenesis via Activation of MBL. Nature 74 (7777), 264–267. doi: 10.1038/s- Baker, J. L., Bor, B., Agnello, M., Shi, W., and He, X. (2017). Ecology of the Oral Microbiome: Beyond Bacteria. Trends Microbiol. 25, 362–374. doi: 10.1016/ j.tim- Bakker, M. G. (2018). A Fungal Mock Community Control for Amplicon Sequencing Experiments. Mol. Ecol. Resour. 18, 541–556. doi: 10.1111/- Bakri, M. M., Hussaini, H. M., Holmes, A., Cannon, R. D., and Rich, A. M. (2010). Revisiting the Association Between Candidal Infection and Carcinoma, Particularly Oral Squamous Cell Carcinoma. J. Oral. Microbiol. 2 (1), 5780. doi: 10.3402/jom.v2i0.5780 Bandara, H. M. H. N., Panduwawala, C. P., and Samaranayake, L. P. (2019). Biodiversity of the Human Oral Mycobiome in Health and Disease. Oral. Dis. 25, 363–371. doi: 10.1111/odi.12899 Barrett, A., Kingsmill, V., and Speight, P. (2008). The Frequency of Fungal Infection in Biopsies of Oral Mucosal Lesions. Oral. Dis. 4, 26–31. doi: 10.1111/j-.tb00251.x Beghini, F., Renson, A., Zolnik, C. P., Geistlinger, L., Usyk, M., Moody, T. U., et al. (2019). Tobacco Exposure Associated With Oral Microbiota Oxygen Utilization in the New York City Health and Nutrition Examination Study. Ann. Epidemiol. 34, 18–25.e3. doi: 10.1016/j.annepidem- Berkovits, C., Tó th, A., Szenzenstein, J., Deá k, T., Urbá n, E., Gá cser, A., et al. (2016). Analysis of Oral Yeast Microﬂora in Patients With Oral Squamous Cell Carcinoma. Springerplus 5, 1257. doi: 10.1186/s- Bolyen, E., Rideout, J. R., Dillon, M. R., Bokulich, N. A., Chase, J., Cope, E. K., et al. (2019). Reproducible, Interactive, Scalable and Extensible Microbiome Data Science Using QIIME 2. Nat. Biotechnol. 37, 852–857. doi: 10.1038/s- Brooks, J. P., Edwards, D. J., Harwich, M. D.Jr, Rivera, M. C., Fettweis, J. M., Serrano, M. G., et al. (2015). The Truth About Metagenomics: Quantifying and Counteracting Bias in 16S rRNA Studies. BMC Microbiol. 15, 66. doi: 10.1186/ s- Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J. A., and Holmes, S. P. (2016). DADA2: High-Resolution Sample Inference From Illumina Amplicon Data. Nat. Methods 13, 581–583. doi: 10.1038/nmeth.3869 Chandra, J., Retuerto, M., Mukherjee, P. K., and Ghannoum, M. (2016). “The Fungal Biome of the Oral Cavity”, in Candida Species. Methods Mol. Biol. Eds. R. Calderone and R. Cihlar (New York, NY: Humana Press), 1356. doi: 10.1007/-_9 Conche, C., and Greten, F. R. (2018). Fungi Enter the Stage of Colon Carcinogenesis. Immunity 49, 384–386. doi: 10.1016/j.immuni- Cui, L., Morris, A., and Ghedin, E. (2013). The Human Mycobiome in Health and Disease. Genome Med. 5, 63. doi: 10.1186/gm467 Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 13 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC and Identiﬁcation of Fungi. Nat. Rev. Microbiol. 17, 95–109. doi: 10.1038/ s--y Nilsson, R. H., Larsson, K. H., Taylor, A. F. S., Bengtsson-Palme, J., Jeppesen, T. S., Schigel, D., et al. (2019b). The UNITE Database for Molecular Identiﬁcation of Fungi: Handling Dark Taxa and Parallel Taxonomic Classiﬁcations. Nucleic Acids Res. 47, D259–D264. doi: 10.1093/nar/gky1022 Oleim, S. H., Noor, S. K., Bushara, S. O., Ahmed, M. H., and Elmadhoun, W. (2019). The Irrational Use of Antibiotics Among Doctors, Pharmacists and the Public in River Nile State, Sudan. Sudan. J. Med. Sci. 14 (4), 276–288. doi:-/sjms.v14i4.5909 Osman, T. A., Satti, A. A., Bøe, O. E., Yang, Y. H., Ibrahim, S. O., and Suleiman, A. M. (2010). Pattern of Malignant Tumors Registered at a Referral Oral and Maxillofacial Hospital in Sudan During 2006 and 2007. J. Cancer Res. Ther. 6, 473–477. doi: 10.4103/- Perera, M., Al-hebshi, N. N., Perera, I., Ipe, D., Ulett, G. C., Speicher, D. J., et al. (2017). A Dysbiotic Mycobiome Dominated by Candida Albicans Is Identiﬁed Within Oral Squamous-Cell Carcinomas. J. Oral. Microbiol. 9,-. doi: 10.1080/- Peters, B. A., Wu, J., Hayes, R. B., and Ahn, J. (2017). The Oral Fungal Mycobiome: Characteristics and Relation to Periodontitis in a Pilot Study. BMC Microbiol. 17, 157. doi: 10.1186/s- Plantinga, A., Zhan, X., Zhao, N., Chen, J., Jenq, R. R., and Wu, M. C. (2017). MiRKAT-S: A Community-Level Test of Association Between the Microbiota and Survival Times. Microbiome 5, 17. doi: 10.1186/s- Proctor, L. M., Creasy, H. H., Fettweis, J. M., Lloyd-Price, J., Mahurkar, A., Zhou, W., et al. (2019). The Integrative Human Microbiome Project. Nature 569, 641–648. doi: 10.1038/s- Rindum, J. L., Stenderup, A., and Holmstrup, P. (1994). Identiﬁcation of Candida Albicans Types Related to Healthy and Pathological Oral Mucosa. J. Oral. Pathol. Med. 23, 406–412. doi: 10.1111/j-.tb00086.x Rivers, A. R., Weber, K. C., Gardner, T. G., Liu, S., and Armstrong, S. D. (2018). ITSxpress: Software to Rapidly Trim Internally Transcribed Spacer Sequences With Quality Scores for Marker Gene Analysis. F1000Research 7, 1418. doi:-/f1000research.15704.1 Romani, L. (2011). Immunity to Fungal Infections. Nat. Rev. Immunol. 11, 275– 288. doi: 10.1038/nri2939 Rosenbaum, J., Usyk, M., Chen, Z., Zolnik, C. P., Jones, H. E., Waldron, L., et al. (2019). Evaluation of Oral Cavity DNA Extraction Methods on Bacterial and Fungal Microbiota. Sci. Rep. 9, 1531. doi: 10.1038/s- Rosier, B. T., Marsh, P. D., and Mira, A. (2018). Resilience of the Oral Microbiota in Health: Mechanisms That Prevent Dysbiosis. J. Dent. Res. 97, 371–380. doi: 10.1177/- Sanjaya, P. R., Gokul, S., Gururaj Patil, B., and Raju, R. (2011). Candida in Oral Pre-Cancer and Oral Cancer. Med. Hypotheses 77,-. doi: 10.1016/ j.mehy- Seed, P. C. (2015). The Human Mycobiome. Cold Spring Harb. Perspect. Med. 5: a019810. doi: 10.1101/cshperspect.a019810 Shaw, L., Ribeiro, A. L. R., Levine, A. P., Pontikos, N., Balloux, F., Segal, A. W., et al. (2017). The Human Salivary Microbiome Is Shaped by Shared Environment Rather Than Genetics: Evidence From a Large Family of Closely Related Individuals. MBio 8, e-. doi: 10.1128/mBio- Sparber, F., and LeibundGut-Landmann, S. (2017). Host Responses to Malassezia Spp. In the Mammalian Skin. Front. Immunol. 8, 1614. doi: 10.3389/ﬁmmu- Tamai, R., Sugamata, M., and Kiyoura, Y. (2011). Candida Albicans Enhances Invasion of Human Gingival Epithelial Cells and Gingival Fibroblasts by Porphyromonas Gingivalis. Microb. Pathog. 51, 250–254. doi: 10.1016/ j.micpath- Vasquez, A. A., Ram, J. L., Qazazi, M. S., Sun, J., and Kato, I. (2018). “Oral Microbiome: Potential Link to Systemic Diseases and Oral Cancer,” in Mechanisms Underlying Host-Microbiome Interactions in Pathophysiology of Human Diseases (Boston, MA: Springer US), 195–246. doi: 10.1007/-_9 Witherden, E. A., Shoaie, S., Hall, R. A., and Moyes, D. L. (2017). The Human Mucosal Mycobiome and Fungal Community Interactions. J. Fungi. 3, 56. doi: 10.3390/jof- Yang, F., Ning, K., Zeng, X., Zhou, Q., Su, X., and Yuan, X. (2016). Characterization of Saliva Microbiota’s Functional Feature Based on Metagenomic Sequencing. Springerplus 5, 2098. doi: 10.1186/s- Huttenhower, C., Gevers, D., Knight, R., Abubucker, S., Badger, J. H., Chinwalla, A. T., et al. (2012). Structure, Function and Diversity of the Healthy Human Microbiome. Nature 486, 207–214. doi: 10.1038/nature11234 Idris, A. M., Ahmed, H. M., Mukhtar, B. I., Gadir, A. F., and El-Beshir, E. I. (1995). Descriptive Epidemiology of Oral Neoplasms in Sudan- and the Role of Toombak. Int. J. Cancer 61, 155–158. doi: 10.1002/ijc- Iliev, I. D., and Leonardi, I. (2017). Fungal Dysbiosis: Immunity and Interactions at Mucosal Barriers. Nat. Rev. Immunol. 17, 635–646. doi: 10.1038/nri.2017.55 Jenkinson, H. F., and Lamont, R. J. (2005). Oral Microbial Communities in Sickness and in Health. Trends Microbiol. 13, 589–595. doi: 10.1016/ j.tim- Khalifa, N., Allen, P. F., Abu-bakr, N. H., Abdel-Rahman, M. E., and Abdelghafar, K. O. (2012). A Survey of Oral Health in a Sudanese Population. BMC Oral. Health 12, 5. doi: 10.1186/- Knot, P. D., Ko, D. L., and Fredricks, D. N. (2009). Sequencing and Analysis of Fungal rRNA Operons for Development of Broad-Range Fungal PCR Assays. Appl. Environ. Microbiol. 75,-. doi: 10.1128/AEM- Koh, H., Livanos, A. E., Blaser, M. J., and Li, H. (2018). A Highly Adaptive Microbiome-Based Association Test for Survival Traits. BMC Genomics 19, 210. doi: 10.1186/s- Krogh, P. (1990). The Role of Yeasts in Oral Cancer by Means of Endogenous Nitrosation. Acta Odontol. Scand. 48, 85–88. doi: 10.3109/- Krogh, P., Hald, B., and Holmstrup, P. (1987). Possible Mycological Etiology of Oral Mucosal Cancer: Catalytic Potential of Infecting Candida Aibicans and Other Yeasts in Production of N-Nitrosobenzylmethylamine. Carcinogenesis 8,-. doi: 10.1093/carcin/- Krüger, W., Vielreicher, S., Kapitan, M., Jacobsen, I. D., and Niemiec, M. J. (2019). Fungal-Bacterial Interactions in Health and Disease. Pathogens 8, 70. doi: 10.3390/pathogens- Lahti, L., Shetty, S., and Al, E. (2017) Tools for Microbiome Analysis in R. Microbiome Package Version 1.13.6. Available at: http://microbiome.github. com/microbiome. Leigh Greathouse, K., Sinha, R., and Vogtmann, E. (2019). DNA Extraction for Human Microbiome Studies: The Issue of Standardization. Genome Biol. 20, 212. doi: 10.1186/s- Lim, Y., Totsika, M., Morrison, M., and Punyadeera, C. (2017). Oral Microbiome: A New Biomarker Reservoir for Oral and Oropharyngeal Cancers. Theranostics 7,-. doi: 10.7150/thno.21804 Mantovani, A., Allavena, P., Sica, A., and Balkwill, F. (2008). Cancer-Related Inﬂammation. Nature 454, 436–444. doi: 10.1038/nature07205 Masters, N. (2018). Home | Smoking Pack Years. Available at: https://www. smokingpackyears.com (Accessed March 6, 2019). McCullough, M., Jaber, M., Barrett, A. W., Bain, L., Speight, P. M., and Porter, S. R. (2002). Oral Yeast Carriage Correlates With Presence of Oral Epithelial Dysplasia. Oral. Oncol. 38, 391–393. doi: 10.1016/S- - McMurdie, P. J., and Holmes, S. (2013). Phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. PLoS One 8, e61217. doi: 10.1371/journal.pone- Mukherjee, P. K., Chandra, J., Retuerto, M., Sikaroodi, M., Brown, R. E., Jurevic, R., et al. (2014). Oral Mycobiome Analysis of HIV-Infected Patients: Identiﬁcation of Pichia as an Antagonist of Opportunistic Fungi. PLoS Pathog. 10,-. doi: 10.1371/journal.ppat- Mukherjee, P. K., Wang, H., Retuerto, M., Zhang, H., Burkey, B., Ghannoum, M. A., et al. (2017). Bacteriome and Mycobiome Associations in Oral Tongue Cancer. Oncotarget 8,-. doi:-/oncotarget.21921 Müller, V., Viemann, D., Schmidt, M., Endres, N., Ludwig, S., Leverkus, M., et al. (2007). Candida Albicans Triggers Activation of Distinct Signaling Pathways to Establish a Proinﬂammatory Gene Expression Program in Primary Human E ndothelial Cells. J. Immunol. 1 79,-. doi: 10.4049/ jimmunol- Naglik, J. R., Tang, S. X., and Moyes, D. L. (2013). Oral Colonization of Fungi. Curr. Fungal Infect. Rep. 7, 152–159. doi: 10.1007/s--y Nasidze, I., Li, J., Quinque, D., Tang, K., and Stoneking, M. (2009). Global Diversity in the Human Salivary Microbiome. Genome Res. 19, 636–643. doi: 10.1101/gr- Nilsson, R. H., Anslan, S., Bahram, M., Wurzbacher, C., Baldrian, P., and Tedersoo, L. (2019a). Mycobiome Diversity: High-Throughput Sequencing Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 14 October 2021 | Volume 11 | Article 673465 Mohamed et al. Salivary Mycobiome in OSCC Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Yang, C. Y., Yeh, Y. M., Yu, H. Y., Chin, C. Y., Hsu, C. W., Liu, H., et al. (2018). Oral Microbiota Community Dynamics Associated With Oral Squamous Cell Carcinoma Staging. Front. Microbiol. 9, 862. doi: 10.3389/fmicb- Zaura, E., Nicu, E. A., Krom, B. P., and Keijser, B. J. F. (2014). Acquiring and Maintaining a Normal Oral Microbiome: Current Perspective. Front. Cell. Infect. Microbiol. 4, 85. doi: 10.3389/fcimb- Zhu, F., Willette-Brown, J., Song, N. Y., Lomada, D., Song, Y., Xue, L., et al. (2017). Autoreactive T Cells and Chronic Fungal Infection Drive Esophageal Carcinogenesis. Cell Host Microbe 21, 478–493.e7. doi: 10.1016/ j.chom- Copyright © 2021 Mohamed, Litlekalsøy, Ahmed, Martinsen, Furriol, Javier-Lopez, Elsheikh, Gaafar, Morgado, Mundra, Johannessen, Osman, Nginamau, Suleiman and Costea. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Frontiers in Cellular and Infection Microbiology | www.frontiersin.org 15 October 2021 | Volume 11 | Article 673465