Phaeohyphomycosis Caused by a Novel Species, Pseudochaetosphaeronema martinelli

Abstract

Among the opportunistic mycoses that are emerging in patients with immunosuppression or severe underlying illness, many isolates lack of characteristic sporulation and until recently could not be identified. Clinical signs are mostly nonspecific and therefore such infections have often been disregarded. In the present paper we describe a novel, nonsporulating fungal species causing subcutaneous phaeohyphomycosis in two patients of different origin. One is a 73-year-old female from Martinique who suffered from rheumatoid arthritis, while the other case concerns a 72-year-old male from Mexico who had a history of type 2 diabetes mellitus. Sequencing of the partial ribosomal operon revealed that in both cases a member of the order Pleosporales was concerned which could not be affiliated to any family within this order. Multilocus analysis revealed that the fungus was related to another, unaffiliated agent of human mycetoma, Pseudochaetosphaeronema larense, and therefore the name Pseudochaetosphaeronema martinelli was introduced.

INTRODUCTION

Melanized opportunistic fungi compose a large and heterogeneous group belonging to different orders (1, 2). They are inhabitants of plant debris, or thrive in extreme environments on rock, on smooth inert surfaces or in hypersaline waters (3, 4). In recent decades the frequency and biodiversity of melanized fungi as a cause of human or animal infection has increased dramatically (5, 6). These infections are particularly seen in immunocompromised patients, although also immunocompetent individuals may be affected subcutaneously, as well as in deep seated levels (7).

Infections caused by melanized fungi are clinically classified as phaeohyphomycoses. In routine diagnostics, they are characterized by the presence of dark hyphae of the causative agent in the infected tissue (5, 7). This feature is seen in a wide range of medical conditions varying from superficial cutaneous to deep or disseminated infections, while also the clinical presentation may change with the patient's immunity status (8). Subcutaneous phaeohyphomycosis is the most commonly reported form of this disease, and a large number of species have been involved (5). The infection mostly occurs as a result of traumatic inoculation of a saprobic fungus into subcutaneous tissue. In immunocompetent individuals the disease progresses, slowly leading to either an encapsulated cyst, to a swelling without ulceration (9), or to local necrosis; when the fungal hyphae are compacted to grains, the infection is classified as a mycetoma (10). Dissemination may be a result of immunosuppression but may also occur in patients without known immune dysfunction (11).

With the use of molecular identification tools for pathogenic fungi, the number of species and genera recognized as causing phaeohyphomycosis has increased tremendously. We also know that the agents are mostly affiliated to a limited number of fungal orders, particularly the Chaetothyriales and the Pleosporales. Members of the Chaetothyriales belong to the black yeasts and relatives, whereas members of the Pleosporales are frequently plant saprobes and remain sterile when recovered from clinical materials (6, 12); some produce coelomycetous fruit bodies after prolonged incubation. Identification of such species in routine diagnostics can be achieved with the use of DNA-based methods, for which the fungal barcode marker ribosomal DNA internal transcribed spacer region (ITS) is particularly versatile (6).

Clinical isolates of species of which the sequence has as yet not been deposited in public databases are frequently encountered. In the present manuscript, we describe two cases of phaeohyphomycosis caused by a sterile fungus of which the ITS sequence did not show identity with any known taxon and therefore a new species is proposed. The species was found to belong to the order Pleosporales clustering close to Pseudochaetosphaeronema larense. This species and the related taxa could not be affiliated with any pleosporalean family.

CASE REPORTS

Case report 1.

A 73-year-old, blind female was referred to the Dermatology Department of University Hospital of Martinique in July 2011 for a skin lesion located at the left knee (Fig. 1). She suffered from rheumatoid arthritis, which was treated by methotrexate (15 mg/week) and the corticosteroid cortancyl (9 mg/day; prednisone), and she presented no history of trauma. The lesion appeared 8 months earlier as a small pustule and evolved into painless, erythematous, scaly plaques with subcutaneous nodules. The first evoked diagnosis was bacterial, and the patient was treated with gentalline (gentamicin) at 3 mg/kg given intramuscularly for 10 days without improvement. Mycological examinations of scales and pus were made, and material was inoculated onto Sabouraud glucose agar with chloramphenicol and without cycloheximide (Bio-Rad, Marnes-la-Coquette, France), followed by incubation at 30 and 37°C. Direct examination revealed the presence of numerous septate and branched, melanized hyphae and chlamydospore-like structures (Fig. 2). The culture from pus was positive after 4 days at 30°C and after 11 days at 37°C; the strain was deposited in the CBS collection (CBS 135986). A biopsy was later performed. Histological examination of the skin biopsy specimen demonstrated a scar fibrosis with no specific focal chronic inflammation, and no fungal elements were seen. The patient was treated with voriconazole at 200 mg administered twice daily and then examined every month for pus production and eventual clinical evolution. After 4 months, voriconazole treatment was terminated despite significant improvement because of the appearance of side effects (nausea, abdominal pain, and disturbances in liver function indicated by an increase of gamma-glutamyl transpeptidase [453 IU/liter] and transaminases [aspartate transaminase, 146 IU/liter; alanine aminotransferase, 168 IU/liter]). Three years later, another sample was taken which was found positive upon direct examination, and the same fungus was cultured and identified based on the ITS sequences. The clinical appearance had slightly improved, with healing of some of the nodules and decrease of discharge in remaining nodules. Liver functions had remained within normal limits. Treatment with methotrexate was terminated in order to reintroduce voriconazole, but the patient was afraid of a possible recurrence of painful arthritis, so she refused to stop methotrexate treatment and was reluctant to restart voriconazole.

FIG 1.

Clinical presentation of the case report 1 showing lesions in the left knee (A) First visit, July 2011; (B) follow-up visit, February 2014; (C) follow-up visit, December 2014.

FIG 2.

Direct examination of pus from patient lesions. (A) Hyphae and chlamydospore-like structures (case report 1). (B) Dark brown hyphae with swollen chlamydospores (case report 2).

Case report 2.

A 72-year-old male farmer from Palenque Chiapas, southeast Mexico, presented in the Dermatologic Center of Yucatán with two subcutaneous cystic lesions on the inside of the left foot, which occasionally drained serosanguineous discharge (Fig. 3). The cysts had started to grow 2 years earlier; the patient reported mild pruritus and tenderness. His medical history was relevant for type 2 controlled diabetes mellitus and for repeated trauma on the feet, which was evidenced by scars near the cysts. Magnetic resonance imaging (MRI) was carried out showing two cystic, delimited lesions (Fig. 3), which made a clinical impression of being fungal in nature. Sinuses injuries were pressed for greater serosanguineous discharge, which was collected and used for microbiological examination. Direct examination in 20% KOH showed multiple broad, irregular, branched, and pigmented hyphal elements, some with internal inclusions (Fig. 2). Culture on Sabouraud glucose agar demonstrated growth of a filamentous, melanized fungus that failed to exhibit reproductive structures on any of the media tested. Itraconazole treatment at 300 mg/day per os was administered for 7 months, leading to clinical and mycological cure. After a 4-year follow-up, the patient continues to be healthy, without signs of fungal infection.

FIG 3.

Presentation of patient of case report 2. (A and B) Clinical presentation; (C) MRI view of the affected foot.

MATERIALS AND METHODS

Phenotypic characters.

For phenotypic description, the isolates were grown on plates of malt extract agar (MEA; Oxoid, United Kingdom) and oatmeal agar (OA; home-made at CBS) and then incubated at 30°C for 2 weeks. To induce the sporulation, the fungus was grown on water agar with or without plant stems (lupine), and on OA under near-UV light for 8 weeks. Microscopic mounts were prepared in lactic acid and examined with a microscope (Nikon Eclipse 80i) using differential interference contrast illumination. Micrographs were taken using a digital camera attached to the microscope (Nikon, digital sight, DS-5M). Cardinal temperatures were determined on MEA plates incubated at temperatures ranging from 6 to 36°C at 3°C intervals and including 37 and 40°C. For the detection of enzymes involved in melanin production, laccase was tested using ABTS medium containing 0.03% ABTS [2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt] (13). ABTS plates were inoculated with fresh mycelia from the margin of colonies grown on MEA and incubated for 1 week at room temperature. The presence of a green zone around the colony indicates positive extracellular laccase production.

Genotypic characters.

Genomic DNA was extracted from cultures grown on MEA plates using a cetyltrimethylammonium bromide (CTAB) method as described by Möller et al. (14). Amplification and sequencing of the partial ribosomal operon was performed using the primers ITS1 and ITS4 for the ITS (15), NS1, and NS24, with the additional sequencing primers NS2, NS3, NS6, and NS7, for the partial small ribosomal subunit (SSU) (15, 16) and LRoR and LR7 for the partial large ribosomal subunit (LSU) (17). The partial RNA polymerase second largest subunit (rPB2) was sequenced with primers fRPB2-5F and fRPB2-7cR (18), and the partial translation elongation factor 1-alpha (TEF1) with EF-983F and EF-2218R (http://www.aftol.org/pdfs/EF1primer.pdf).

Consensus sequences were computed and edited with SeqMan Lasergene package (DNAStar, Madison, WI). Representative taxa from the Pleosporales were retrieved from GenBank and included in phylogenetic analyses. Sequences were aligned with Mafft v7 (http://mafft.cbrc.jp/alignment/server/) and handled in BioEdit v7.1.3 (19). Alignments for each gene were constructed and a final matrix of combined LSU, SSU, rPB2, and TEF1 was prepared using DataConvert 1.0 (http://www.bigelow.org/research/srs/david_a_mcclellan/david_mcclellan_laboratory/dataconvert/). In order to determine the phylogenetic position of strains, Bayesian analysis and maximum likelihood were used for the final matrix. Bayesian analysis was done in MrBayes v3.1.2 in the CIPRES web server (http://www.phylo.org/) with two parallel runs of four MCMC (Markov Chain Monte Carlo) chains each for a 30,000,000 generations. Maximum-likelihood analysis was constructed with RAxML v7.2.8 from the CIPRES web server. Trees derived from each analysis were edited by MEGA v5.05 and were checked for congruence (20). Ribosomal ITS sequences were aligned with muscle (http://www.ebi.ac.uk/Tools/msa/muscle/) and phylogenetic analysis was performed with maximum likelihood using MEGA v5.05 (20).

Nucleotide sequence accession numbers.

Sequences were deposited in the GenBank database under the following accession numbers: for CBS 135986, KR909314 (ITS), KR909315 (LSU), KR909317 (SSU), KR909321 (rPB2), and KR909320 (TEF1); and for dH 21381, KR909313 (ITS).

RESULTS

Isolates recovered from both cases remained sterile despite various attempts to induce sporulation and therefore could not be identified by morphology. The ITS sequences of the strains from the cases showed 99.6% sequence similarity, indicating that they represented a single taxon. In public databases, ITS sequences did not show identity with any deposition, but the closest hits were all members of the order Pleosporales. Sequences of LSU, SSU, rPB2, and TEF1 were used to determine the phylogenetic position of our isolates within the order. In the combined alignment (3,663 total characters) with representative taxa from the Pleosporales, our sequence was found to belong to the suborder Massarineae. However, the fungus could not be assigned to any known taxon, not even at the family level (Fig. 4; clade 1, 1.0 posterior probability [PP]/100% maximum likelihood bootstrap support [ML-BS]. The closest relation found was the human-pathogenic genus Pseudochaetosphaeronema, which contains a single species, P. larense, based on the type strain CBS 640.73. The second closest taxa, although at a large phylogenetic distance, were two isolates identified as Splanchnonema platani, but this group did not contain a type, and thus the identification was unconfirmed.

FIG 4.

Phylogram resulting from the analysis of concatenated LSU, SSU, rPB2, and TEF1, constructed using maximum-likelihood and Bayesian approaches. Values at the nodes are posterior probability (PP)/maximum-likelihood bootstrap support (ML-BS). Type strains are indicated by a “T” in parentheses.

ITS phylogeny was then performed (609 total alignment characters) focusing on the unidentified isolates and their closest species in Pleosporales. In concordance with the combined data analysis, the ITS confirmed the relation of the isolates to the genus Pseudochaetosphaeronema (ML-BS; 91%), but the phylogenetic profile was distinct from that of P. larense. In addition, the two strains identified as S. platani formed a clade paraphyletic to Pseudochaetosphaeronema (Fig. 5). Since the nearest neighbor of the fungus investigated is Pseudochaetosphaeronema larense, known as an agent of human subcutaneous infection, we describe the new species in the genus Pseudochaetosphaeronema, as follows.

FIG 5.

Phylogenetic tree resulting from analysis of ribosomal ITS using maximum likelihood. Values at the nodes show the percentage of bootstrap support. Type strains are indicated by a “T” in parentheses.

Taxonomy.

Pseudochaetosphaeronema martinelli S.A. Ahmed, Desbois, Miossec, Atoche, Bonifaz, and de Hoog, sp. nov. (Fig. 6), MycoBank MB 812258.

FIG 6.

Pseudochaetosphaeronema martinelli (CBS 135986). (A to E) Growth on MEA (A, obverse; B, reverse) and on OA (C, obverse; D, reverse) after 2 weeks of incubation at 25°C. (E) Hyphae with wart-like incrustation. (F) Swollen hyphae. (G and H) Coiled hyphae. Scale bars, 10 μm.

Etymology.

The martinellum is a collection of pregothic writings honoring Martinus and held by the Utrecht University Library, Utrecht, The Netherlands. Martinus, living in the fourth century, was bishop of Tours in France and the founder of Catholicism in Gallia. He is the patron of the city of Utrecht. The type strain was isolated from subcutaneous nodule of a patient from Martinique, Fort-de-France. The island of Martinique, one of the tropical provinces of France, was named by Christopher Columbus after St. Martin of Tours, whose name is also carried by the Caribbean island of Sint Maarten (affiliated with the Netherlands).

Morphology and physiology.

Colonies on MEA at 25°C attain 1.3 mm in diameter after 2 weeks of incubation, floccose, grayish green, becoming darker with age; reverse black. Colonies on OA are felty and dark gray at the center, with a gray-white periphery; reverse black. Hyphae are verruculose, dark brown to black, branched, regularly septate, and coiled in some parts. Dark brown, swollen hyphae were observed with dark inflations resembling chlamydospores and measuring 5 to 8 μm in diameter; conidiation remained absent. Laccase was positive. Sexual morph was not observed. The minimum growth temperature was below 9°C, the optimum was 30°C, and the maximum was 37°C.

Holotype.

Dried culture CBS H22201, type strain CBS 135986, from subcutaneous nodule of human patient, Fort-de-France, Martinique (N. Desbois). The Second strain analyzed, dH 21381, was from the subcutaneous cystic of a patient from Palenque Chiapas, southeast Mexico (C. Atoche).

DISCUSSION

We report here a melanized fungus causing phaeohyphomycosis in two patients. Case 1 was a patient with rheumatoid arthritis who had undergone prolonged anti-inflammatory and immunosuppressive therapy. This patient developed infection in her left knee, as evidenced by the presence of multiple lesions on the skin. Due to the nonspecific clinical presentation and the absence of a history of trauma, the initial diagnosis was a bacterial infection by cutaneous bacteria, possibly staphylococci. Case report 2 was a diabetic patient who developed two subcutaneous cystic lesions in his foot after repeated trauma. The presumable scenario of acquisition of infection in both patients could be a minor trauma that led to implantation of the fungus into the skin or subcutaneous tissue. This occurs in many cases of subcutaneous mycoses, such as phaeohyphomycosis and mycetoma (5, 21, 22). The ability of the implanted species to produce melanin, which protects from phagocytosis and macrophage killing, as well as the ability to grow at the human body temperature, are among the main pathogenic factors (23, 24). On the host side, many risk factors may also contribute to the infection, including the use of broad-spectrum antibiotics, immunosuppression, and severe underlying diseases (5, 25).

The patients reported here were suffering from the underlying diseases rheumatoid arthritis and diabetes mellitus. Moreover, the patient in the first case had been receiving immunosuppressive drug therapy for some time. A similar presentation was reported by Sutton et al. (26) for a patient with phaeohyphomycosis who had diabetes and rheumatoid arthritis. The fungus isolated by these authors was coelomycetes and belonged to the genus Phomopsis, another member of the order Pleosporales. Sutton et al. (26) also reported the production of conidia from well-characterized conidiomata by the isolated fungus, whereas in both cases reported in the present study the isolates remained sterile without any sporulation (Fig. 6).

Studies of phaeohyphomycosis in a renal transplant patient from Brazil revealed that, during eight episodes of subcutaneous infections, only sterile isolates were recovered, most being related to Pleosporales (6). Recently, Tsang et al. (27) proposed a new genus Hongkongmyces pedis in Pleosporales for a sterile fungus recovered from a case of subcutaneous phaeohyphomycosis as well. In another case, reported by Ahmed et al. (28), of subcutaneous phaeohyphomycosis in a diabetic patient caused by the novel pleosporalean taxon Roussoella percutanea, the isolated fungus also remained sterile. Interestingly, the isolate recovered by Ahmed et al. (28) had shown identity with another sterile strain, viz. an isolate from a renal transplant patient (28, 29). These authors were unable to detect any sporulation from the strain of the renal transplant patient; however, after being in a collection for 13 years, the strain showed characteristic conidiomata and spores (28, 29). This might indicate that such fungi reversibly lose or gain the ability to sporulate when infecting a human host. Some researchers attributed this to use of drugs such as calcineurin inhibitors which might impair conidium formation (6, 30). In our cases, neither patient received such drugs, and therefore this phenomenon of being sterile remained obscure.

Identification of strains recovered from both patients reported in the present study was hampered by the absence of morphological features. Identification could only be achieved by sequencing of fungal barcode marker ITS. However, the ITS sequences could not be matched with any known taxon in public databases, such as GenBank, and therefore a new species was introduced. The phylogenetic analysis of four loci revealed that the new species belonged to the order Pleosporales. To date, there are more than 20 recognized families in this order (31), and 5 of them were affiliated with the recently introduced suborder Massarineae. Families of Massarineae mostly contain saprobes of aquatic or terrestrial environments and include Montagnulaceae, Massarinaceae, Lentitheciaceae, Morosphaeriaceae, and Trematosphaeriaceae (31). Although our new species was nested within this suborder, phylogenetically it did not belong to any of the families described above. The novel species clustered in a separate clade that contained species of different genera, i.e., Pseudochaetosphaeronema larense and Splanchnonema platani (Fig. 4). Pseudochaetosphaeronema is a monotypic genus characterized by pycnidial fruit bodies produced reluctantly (12). Other neighbors are the two strains of S. platani from the United States previously identified and deposited in the CBS collection as Massaria platani (http://www.cbs.knaw.nl/). Recent phylogenies showed that these strains were poorly affiliated to Lentitheciaceae, and they were reidentified as Splanchnonema platani (31, 32). The taxonomic position of the type species of the genus Splanchnonema, which is S. pustulatum, is still uncertain because of the lack of molecular information, and it has been temporarily affiliated with Pleomassariaceae (31). Consequently, the identity of the two S. platani strains is not confirmed, and the new species could not be classified in the genus Splanchnonema.

The present fungus was added to the genus Pseudochaetosphaeronema as P. martinelli. Both P. martinelli and P. larense caused draining sinuses upon subcutaneous infection, but the new Pseudochaetosphaeronema species produced irregularly swollen hyphal elements in tissue, whereas P. larense formed compact grains and was clinically classified as mycetoma (33). The presence of both species in Massarineae adds another recognizable opportunist to this suborder, in addition to the family Trematosphaeriaceae, which was recently found to harbor known agents of human mycetoma (12).

The two patients described here were from southeast Mexico and Martinique, whereas P. larense was reported from Venezuela (33). It appears that the habitat of Pseudochaetosphaeronema is South and Central America. Species might have similar ecological preferences, and the traveling of humans and animals might have contributed to dispersal (33, 34). Medical mycologists need to be aware of such sterile opportunists.

The two patients infected with P. martinelli were treated with different courses of antifungals and differed in their responses. The case 1 patient was treated for 4 months with voriconazole; therapy was terminated because of its side effects. Despite clinical improvement of the patient's condition, no complete cure was achieved within 4 months, and the infection remained chronic for 3 years. In contrast, the case 2 patient was treated with itraconazole for a period of 7 months. A complete cure was achieved for this patient, which was confirmed after 4 years of follow-up. The sterile nature of P. martinelli frustrates the use of available standard protocols of antifungal susceptibility testing; therefore, no in vitro susceptibility data are available for this species. Judging from the successful outcome in the second case, we recommend itraconazole treatment for prolonged periods.