High Diversity of Fusarium Species in Onychomycosis: Clinical Presentations, Molecular Identification, and Antifungal Susceptibility

Fusarium are uncommon but important pathogenic organisms; they cause non-dermatophyte mould (NDM) onychomycosis. Patients typically respond poorly to treatment owing to Fusarium’s native resistance to multiple antifungal drugs. However, epidemiological data for Fusarium onychomycosis are lacking in Taiwan. We retrospectively reviewed the data of 84 patients with positive Fusarium nail sample cultures at Chang Gung Memorial Hospital, Linkou Branch between 2014 and 2020. We aimed to investigate the clinical presentations, microscopic and pathological characteristics, antifungal susceptibility, and species diversity of Fusarium in patients with Fusarium onychomycosis. We enrolled 29 patients using the six-parameter criteria for NDM onychomycosis to determine the clinical significance of Fusarium in these patients. All isolates were subjected to species identification by sequences and molecular phylogeny. A total of 47 Fusarium strains belonging to 13 species in four different Fusarium species complexes (with Fusarium keratoplasticum predominating) were isolated from 29 patients. Six types of histopathology findings were specific to Fusarium onychomycosis, which may be useful for differentiating dermatophytes from NDMs. The results of drug susceptibility testing showed high variation among species complexes, and efinaconazole, lanoconazole, and luliconazole showed excellent in vitro activity for the most part. This study’s primary limitation was its single-centre retrospective design. Our study showed a high diversity of Fusarium species in diseased nails. Fusarium onychomycosis has clinical and pathological features distinct from those of dermatophyte onychomycosis. Thus, careful diagnosis and proper pathogen identification are essential in the management of NDM onychomycosis caused by Fusarium sp.


Introduction
Fusarium is a widely distributed hyaline mould genus with at least 300 phylogenetically different species in 23 species complexes [1]. These species have notable human pathogenicity; however, despite their diversity, only a few, such as F. solani species complex (FSSC), F. oxysporum species complex (FOSC), and F. fujikuroi species complex (FFSC), cause disease in humans [2]. In immunocompetent patients, locally invasive onychomycosis and keratitis are the most frequent manifestations; however, in immunocompromised patients, severe disseminated disease may cause mortality [3]. Owing to late diagnosis, intrinsic resistance to azole antifungals, and the emergence of multidrug resistant strains due to agricultural antifungal overuse, treatment of fusariosis is a major challenge.

Materials and Methods
We retrospectively reviewed data obtained from 84 patients with positive nail cultures of Fusarium at the Chang Gung Memorial Hospital, Linkou branch, between 2014 and 2020. Demographic data, history of soil contact, associated predisposing factors (including diabetes mellitus, malignancy, and an immunocompromised status), treatment (topical or systemic antifungals, surgery, and laser therapy), and prognosis were collected from medical records and by telephone interview. Photographs of the affected nails were taken with patient consent, and this study was reviewed and approved by the Institutional review board (IRB) of the Chang Gung Medical Foundation (approval number 202101575B0). Patient consent was waived by the IRB.
Diagnosis of Fusarium onychomycosis was made based on the six-parameter criteria for NDM onychomycosis proposed by Gupta et al., as follows: (1) identification of NDMs by microscopy using potassium hydroxide (KOH) preparation, (2) culture isolation of NDMs, (3) repeated isolation of the same NDM in culture, (4) failure to isolate a dermatophyte in culture, (5) culture of the same NDM from 5 out of 20 inoculations of nail fragments, and (6) NDM identification using molecular techniques or histological findings [9]. In this study, we made a diagnosis of Fusarium onychomycosis when parameter (4) and at least two to three other parameters were fulfilled.

Sample Collection, Culture, and Microscopic/Histopathological Examination
Diseased portions of subungual nail debris or plates were collected using a nail clipper or scalpel. Debris was pre-treated with 20% KOH and examined by microscopy to detect fungal elements. Nail plates were sent for histopathological examination and stained with haematoxylin and eosin and Periodic acid-Schiff stains. For fungal culturing, nail debris was inoculated on both inhibitory mould (CMP ® , Creative Life Sciences, Taipei, Taiwan) and Mycosel agar (BD Difco™, BD, Franklin Lakes, NJ, USA) plates and incubated at 25 • C. The fungus grown was purified by subculture on Sabouraud's dextrose agar plates (BD Difco™) for morphological identification and molecular study.

Molecular Identification
All Fusarium isolates were subjected to sequence-based molecular identification. The fungal genomic DNA was obtained using the Smart LabAssist (TANBead, TANBead, Taoyuon City, Taiwan) automatic DNA extraction machine. The internal transcribed spacers (ITS) of ribosomal DNA were amplified with primers-ITS1 (TCCGTAGGTGAAC-CTGCGG) and ITS4 (TCCTCCGCTTATTGATATGC); the partial transcription elongation factor-1α (TEF-1α) gene was amplified with primers EF1 (ATGGGTAAGGARGACAAGAC) and EF2 (GGARGTACCAGTSATCATG). Polymerase chain amplification (PCR) products were confirmed by electrophoresis, purified, and sequenced using an ABI Prism 3730 xl DNA analyser (Applied Biosystems, Foster City, California, USA). Sequences generated in this study were deposited at the DNA Data Bank of Japan (DDBJ) [10]. Preliminary identification performed by comparing the sequences of each Fusarium isolate with sequences deposited in the Fusarium MLST (Multilocus Sequence Typing) database at the Mycobank website (https://fusarium.mycobank.org/ (accessed on 15 February 2022)) and the Fusarium Database (http://isolate.fusariumdb.org (accessed on 15 February 2022)). Identification was confirmed by phylogenetic analysis.
Based on preliminary identification results from the Fusarium MLST database, sequences of Fusarium species similar to the strains used in this study were downloaded from GenBank (https://www.ncbi.nlm.nih.gov/genbank/ (accessed on 15 June 2022)). Atractium crassum was selected as the outgroup for subsequent analysis, and sequences were first aligned by multiple alignment using fast Fourier transform (online version; https://mafft.cbrc.jp/alignment/server/ (accessed on 15 June 2022)). During manual inspection, any poorly aligned regions were removed using Gblocks [11]. Finally, the TEF-1α and ITS regions were concatenated for subsequent analyses. A maximum-likelihood tree was generated using the workflow in IQ-TREE 2.1.3 [12], and DNA models were automatically selected by the built-in ModelFinder algorithm [13]. The support value of the nodes was calculated from 1000 repeated slow standard nonparametric bootstrap. A Bayesian inference tree was obtained by analysing the same dataset with MrBayes v3.2.6 [14]. The analysis started with two MCMC chains of 1,000,000 generations, and one tree was kept every 1000 generations. The last three quarters of the 1000 trees obtained were used to compute the final consensus tree, and trees were visualized using MEGA 7 [15]. All analyses were performed using a Linux Mint 20.3 (64-bit) operating system, and to ensure reproducibility, random seeds were explicitly set to 56 wherever necessary.

Results
Fusarium was isolated from the nail samples of 84 patients, 55 of whom did not fulfil the criteria for diagnosis of NDM onychomycosis and were excluded. Finally, 29 patients were enrolled for further analysis.

Demographic Data and Clinical Manifestations
After rigorous clinical, histological, and mycological confirmation, 29 patients were diagnosed with Fusarium onychomycosis (Table 1). There were 13 men (44.8%) and 16 women (55.2%), with the mean age of 55 (3-87) years. Average disease duration was 20 (1-108) months. Six (20.7%) patients had a personal history of gardening or soil contact, and seven (24.1%) had diabetes (n = 2), immunocompromise (n = 2), and underlying cancer (n = 3) as predisposing factors. The most commonly involved nails were toenails (n = 16), especially the first toe (n = 15); however, there was one case in bilateral thumbs and six in fingernails. Six patients had both fingernail and toenail onychomycosis; however, none of the six had predisposing factors of diabetes, immunocompromise, or cancer. Clinical manifestations of Fusarium onychomycosis differed from those of classic dermatophyte onychomycosis; they included yellow to greenish discoloration, onycholysis, paronychia (mild or severe), and proximal subungual onychomycosis (PSO) (Figure 1).
The most commonly involved nails were toenails (n = 16), especially the first to 15); however, there was one case in bilateral thumbs and six in fingernails. Six pa had both fingernail and toenail onychomycosis; however, none of the six had predi ing factors of diabetes, immunocompromise, or cancer. Clinical manifestation Fusarium onychomycosis differed from those of classic dermatophyte onychomy they included yellow to greenish discoloration, onycholysis, paronychia (mild or sev and proximal subungual onychomycosis (PSO) (Figure 1).

Antifungal Susceptibility Testing
Antifungal susceptibility results are presented in Table 3. Obvious susceptibility differences between and within species complex were noted. In general, all isolates had very

Antifungal Susceptibility Testing
Antifungal susceptibility results are presented in Table 3. Obvious susceptibility differences between and within species complex were noted. In general, all isolates had very high MICs to FLC and ITC and low MICs to EFC, LNC, and LLC. FSSC had higher TRB MICs than non-FSSC, although FIESC had only one strain in the group, and the rough data are not so precise. The range of MICs are similar in VRC and NAT, and lower in AMB. Three F. keratoplasticum isolates were resistant to all antifungals, and one of them had a lower MIC to LNC.

Treatment Response and Prognosis
Among 13 patients who received topical antifungal agents alone (Table 4), more than half (53.8%) had poor response, while two had good response. Nine patients received combination therapy (TRB and topical antifungal agents); however, more than half (55%) of them still had a poor response. Two patients received combination therapy (ITC and topical antifungal agents); one had a good response while the other had a poor response. Only three patients received TRB/ITC treatment with continuous topical antifungal agents, contributing to the good response observed in them.

Presentation of Two Special Cases
Case 1. A 60-year-old woman presented with yellow to greenish discoloration on her right 2nd finger, left thumb, and 2nd and 4th fingers for 2-3 months (Figure 2a,b). She had systemic lupus erythematosus (SLE) and was under treatment for oral azathioprine. She denied gardening history or contact with soil. Direct microscopic examination of the nail specimen demonstrated irregularly segmented hyphae with frequent branching, adventitious sporulation, chlamydospore-like swelling, and thin hyphae, which was different from that of dermatophytes (Figure 2e-g). Histopathology of the nail revealed septated hyphae and chlamydospores (Figure 2h). Fungal cultures from diseased nails all grew Fusarium, and molecular identification showed that her right thumb was infected by F. keratoplascum (CGMHD 0974), and her left thumb and 2nd and 4th fingers were infected by F. solani (CGMHD 0975, CGMHD 0976, CGMHD 0977). Repeated culture 3 months later also revealed the same results. The patient responded poorly to oral itraconazole (200 mg/day) for 3 months and was later treated with oral terbinafine (250 mg/day) combined with nail debridement and topical sulconazole solutions. A growth of new nails was noted three months later (Figure 2c,d).
Case 2. A 55-year-old female patient had yellow to greyish discoloration on the bilateral big toes for several years (Figure 5a). She was a case of HBV chronic hepatitis and goiter of the thyroid. She denied gardening habit or contact to soil or other underlying disease, such as diabetes, malignancy, or under immunosuppressive treatment. Histopathology of the diseased nail demonstrated septated hyphae with a beaded appearance which invaded the nail plate (Figure 5b). Direct microscopic examination revealed septated hyphae and adventitious sporulation (Figure 5c). Six Fusarium isolates were cultured from the diseased nails during the two years of follow-ups. Molecular identification proved that all of them were F. keratoplasticum, but of three different genotypes based on the TEF-1α sequences (Figure 5d) The patient initially received oral griseofulvin 500 mg/day and topical antifungals for 21 days, but in vain. The patient received intermittent nail debridement and treatment with topical sulconazole solution in the following 3-4 years. New healthy nails finally grew with negative culture results. No recurrence was noted.

Discussion
Diagnosis of Fusarium onychomycosis is challenging because NDMs are common contaminants of nails. Published diagnostic criteria vary, and there is no consensus [9]. Approximately 42.8% false negative cases of NDM onychomycosis may be misdiagnosed when only negative dermatophyte microscopic examination and repeated culture are performed [17]. Gupta et al. proposed using three of their six clinical guideline criteria (KOH identification, isolation in culture, repeated isolation, inoculum counting [18], dermatophyte exclusion, and histological proof) to rule out dermatophyte contamination, and this remains the most widely used diagnostic method [8]. Although this method cannot perfectly prevent misdiagnosis of false negative and contaminants, it is straightforward and useful in aiding clinicians in clinical practice. For example, regarding inoculum counting, Gupta et al. had pointed out the low predictive value of inoculum counting as 23.2% of the time [19]. Similar histology finding may also be found in dermatophyte histology, but special appearance of dermatophytoma, irregulated septated hyphae, and terminal swelling are seldom seen in dermatophyte. The systemic review and comparison of histology difference in NDM and dermatophyte is important but still lacking, except in our clinical observations. When only using DME positive and negative dermatophyte culture for NDM diagnosis, there are only 53.6% sensitivity and 70.3% specificity [20]. Classical criteria include positive DME and repeated culture with 92.7% accuracy without the possibility of contaminants, but this method is difficult to be used in clinical practice [20]. Therefore, we follow the criteria of Gupta et al. and furthermore, 21 out of 29 patients in our research had histopathological evidence of fungal invasion with signs of NDM histology

Discussion
Diagnosis of Fusarium onychomycosis is challenging because NDMs are common contaminants of nails. Published diagnostic criteria vary, and there is no consensus [9]. Approximately 42.8% false negative cases of NDM onychomycosis may be misdiagnosed when only negative dermatophyte microscopic examination and repeated culture are performed [17]. Gupta et al. proposed using three of their six clinical guideline criteria (KOH identification, isolation in culture, repeated isolation, inoculum counting [18], dermatophyte exclusion, and histological proof) to rule out dermatophyte contamination, and this remains the most widely used diagnostic method [8]. Although this method cannot perfectly prevent misdiagnosis of false negative and contaminants, it is straightforward and useful in aiding clinicians in clinical practice. For example, regarding inoculum counting, Gupta et al. had pointed out the low predictive value of inoculum counting as 23.2% of the time [19]. Similar histology finding may also be found in dermatophyte histology, but special appearance of dermatophytoma, irregulated septated hyphae, and terminal swelling are seldom seen in dermatophyte. The systemic review and comparison of histology difference in NDM and dermatophyte is important but still lacking, except in our clinical observations. When only using DME positive and negative dermatophyte culture for NDM diagnosis, there are only 53.6% sensitivity and 70.3% specificity [20]. Classical criteria include positive DME and repeated culture with 92.7% accuracy without the possibility of contaminants, but this method is difficult to be used in clinical practice [20]. Therefore, we follow the criteria of Gupta et al. and furthermore, 21 out of 29 patients in our research had histopathological evidence of fungal invasion with signs of NDM histology features, which can decrease rates of contaminants. New diagnostic methods, including molecular methods and techniques involving PCR, are seeing increasing application and importance. The commercialization of PCR kits may improve fungal diagnosis in the future [8].
In the clinical presentation of Fusarium onychomycosis, only 27.5% of patients had predisposing factors such as diabetes, immunosuppression, and cancers. The majority of Fusarium subtypes vary across the research, with proximal subungual onychomycosis (PSO), total dystrophic onychomycosis (TDO), and paronychia previously regarded as the most common clinical phenotypes [7,21]. However, distal lateral subungual onychomycosis (DLSO) and onycholysis are reportedly the more predominant subtypes [5,[22][23][24][25]. FOSC is predominant in DLSO cases, and FSSC is more commonly involved in PSO, TDO, and SWO phenotypes according to Uemura et al. [25], but research from the north of Iran demonstrated diversity species among different subtypes of onychomycosis, with F. proliferatum, F. keratoplasticum, and F. falciforme predominated in DLSO, and variable appearance in PSO, TDO and endonyx onychomycosis [26]. The clinical differences between Fusarium and dermatophytes onychomycoses are not clear; however, some clues are available. Fusarium onychomycosis is most implicated in (1) periungual inflammation of the nail matrix and purulent discharge [4,9,27], (2) resistance to empirical antifungal treatment [25], (3) trauma history or nail dystrophy and absence of tinea pedis [28], and (4) involvement of the big toes (fingernails are only occasionally involved as combination symptoms) [23,24,29]. In the present study, there were six DLSO, two PSO (one overlapping paronychia), one WSO, and one DLSO phenotype with onychodystrophy from 10 cases.
The histological presentation of Fusarium onychomycosis is only mentioned in the case of reports and is rarely systemically reviewed [30]. Lavorato et al. compared the performance of mycology and histology for dermatophyte and NDM onychomycoses and revealed that direct microscopy was more sensitive for NDM and that nail clippings for histopathology were better for dermatophyte onychomycosis [31]. However, this research only collected DLSO pattern onychomycosis, and only 28.5% cases were Fusarium onychomycosis. Although we cannot differentiate dermatophyte onychomycosis from NDM onychomycosis simply by histology, there may be additional clues. Among the six recognized patterns, the most frequently seen patterns in our study were frequently branching irregularly septated hyphae, moniliform hyphae, and hyphae with terminal swelling. Direct microscopic examination with the findings of chlamydospores with beaded appearance hyphae, terminal enlarging, and adventitious sporulation are also helpful for differentiation.
The pathogenesis of Fusarium's invasion of human nails has been previously elucidated [32]. In vitro, Fusarium species can destroy the stratum corneum through keratolysis without additional nutrients [32]. Further, marked protease activity has been detected in FSSC [33]. Flavia et al. demonstrated that Fusarium oxysporum invades nail plates, resulting in the ex-vitro formation of biofilm composed of hyphae, conidia, and extra matrix. The nail unit is a site of immune privilege with low expression of major histocompatibility antigens, dysfunction of antigen presenting cells, and inhibition of natural killer cell activity [34]. However, studies comparing NDM and dermatophyte onychomycoses in terms of levels of myotoxins, keratinase, and proteases along with components of fungal biofilms remain scarce.
Prior to this study, most of our patients received combination (topical and systemic antifungals-TRB and ITC) or destructive (laser and surgery) therapy. Among them, 53.8% showed poor response to all treatment. In review articles, few treatment methods are listed, with 26.7% clinical and 13.9% mycological cure rates reported [8,25] Gupta et al. proposed a treatment algorithm for NDM onychomycosis using a combination of topical (EFC, Tavaborole, and LLC) and systemic (ITC and TRB) therapies [8], with ITC used as daily or pulse therapy (400 mg/pulse once a week for 3 weeks); both showed mild-to-moderate evidence of Fusarium clearance [35,36]. TRB is commonly used for onychomycosis, but the drug resistance rate is relatively high and requires combination with topical antifungals or keratolytics [37]. Verrier et al. reported that oral TRB and ITC are not effective in the treatment of Fusarium onychomycosis [38]. If treatment fails, an antifungal susceptibility test is indicated, and alternatives should be considered. There is one report of treating recalcitrant Fusarium falciforme with posaconazole pulse therapy (800 mg/pulse one week for each month, with total four months) in the literature; this reportedly achieved clinical and mycological improvement [39]. Combination therapy using topical EFC, oral ITC, and oral fosravuconazole are also approved for the treatment of onychomycosis in Japan [24]. Further, topical treatment with AMB for a year has shown a reduction in recalcitrant cases [40]. Other ablative treatment procedures such as two sessions of Qs Nd-YAG laser therapy (532 nm and 1064 nm) one month apart for patients with FSSC onychomycosis showed good response [41], while another study used 1340 nm laser monotherapy, resulting in persistent onychomycosis (91%) under mycological tests for one year [42]. Methylene blue-mediated photodynamic therapy is another choice, which may be superior to 5% amorolfine nail lacquer for NDM onychomycosis [43].
Currently, there are no clinical breakpoints for antifungal drugs against different Fusarium species. In this study, TRB, FLC, and ITC all showed poor improvement in treatment. However, AMB, EFC, LNC, LLC, VRC, and NAT showed better results and should be considered for clinical applications according to MIC results (Table 3). In the literature, The MIC levels from the north of Iran are compatible with our findings, as LLC and LNC was in the range of 1-0.001 µg/mL [26]. Based on Uemura et al., MIC in Fusarium onychomycosis, ITC, FLC, and 5-Fluorocytosine showed high resistance tendency; TRB showed variable resistance tendency [25]; and VRC and AMB showed low resistance tendency. The recently developed antifungal EFC has shown good treatment response in intractable cases [24,44], and olorofim has also shown promise as a candidate [45] after showing in vitro activity against FSSC and FOSC.
Identification of Fusarium to the species level is challenging for clinical laboratories as the morphological characteristics required for identification are few and require experience with the genus. Furthermore, the recent Fusarium taxonomy is based on molecular phylogeny, making identification by morphology alone unreliable. Among existing reviews and case studies on Fusarium onychomycosis, five identified the pathogen to the genus level and 11 to the Fusarium specie level; only five studies performed molecular identification to the species complex level [4,5,7,[22][23][24]26,29,30,32,[46][47][48][49][50][51][52][53][54]. Molecular identification is typically performed using phylogenetic analysis of sequences of ITS and TEF-1α, with RNA polymerase II's second largest subunit (RPB2) genes sometimes used for better differentiation between species. Molecular identification can provide clues to the source and process of the infection, and the importance of repeated and accurate fungal culture reminds patients to pay attention to pathogen control in public health to identify and avoid the source of Fusarium colonization and invasion. Differences between clinical manifestation and antifungal susceptibility testing highlighted the importance of accurate molecular classification.

Conclusions
Although Fusarium onychomycosis accounts for 1-6% of cases of onychomycosis, its frequent resistance to treatment highlights its importance [5]. Cutaneous fusarium infections can serve as the origin of disseminated and invasive infection poor response to empirical antifungals. Therefore, accurate diagnosis through histological and molecular identification is required. Positive culture results for Fusarium species from nail samples are not a proof of infection; further histological proof or positive repeated culture results for the same species of Fusarium are required. Molecular identification (ITS + TEF-1α) and phylogenetic analysis can be applied for pathogen species confirmation. If a positive culture of Fusarium is simply due to colonization, then destruction of the colony and hygiene to prevent colonization is enough. However, if there is a true infection, then combination therapy (using topical and oral antifungal agents) and even surgical debridement are required. Dermatologists will do well to partner with mycologists specialized in Fusarium identification and application. In this article, we highlighted the clinicopathological features of Fusarium onychomycosis and provided six histopathological hints for differentiating Fusarium onychomycosis from dermatophyte onychomycosis. However, much work is needed to provide a standard effective treatment protocol for Fusarium onychomycosis.
This study has some limitations. The first is its design as a single-centre retrospective study. Additionally, owing to a lack of follow up with clinical photos, we could not determine and classify the percentage of accurate onychomycosis subtypes in this study. Last but not least, although current diagnostic criteria for Fusarium onychomycosis is not perfect, it is crucial for helping clinicians in diagnosis and treatment application.