Microsphaeropsis arundinis: an emerging cause of phaeohyphomycosis in cats and people

Microsphaeropsis arundinis is an anamorphic dematiaceous fungus ubiquitous in soil and fresh water^1–4. It typically inhabits terrestrial plant hosts^1–4 and has a well-known association with Aruno donax, a garden escape weed known as ‘giant reed’ or ‘elephant grass’. M. arundinis (fungi imperfecti) is a coelomycete, which encompasses an emerging group of pathogens capable of causing soft tissue infections, mostly in immunocompromised human patients. Such disease typically arises secondary to traumatic inoculation of fungal elements into the subcutis. The infection may spread to contiguous subcutaneous tissues or via the lymphatics in a sporotrichoid manner. The first reports of this organism causing disease occurred just over 10 years ago, and since then an increasing number of cases have been encountered, but so far only in cats and people. In cats, lesions are most consistently encountered on their distal extremities, viz. on or near the toes.

In 2004, Kluger et al. reported the first Microsphaeropsis arundinis infection in a mammalian host¹. The patient was a seven-year-old cat living in suburban Sydney. It had a granulomatous lesion within the deep tissues of the distal forelimb. The cat had a concurrent Fusarium chlamydosporum infection affecting another limb. A few months later, Pendle et al. from Royal North Shore Hospital reported the same organism as a cause of disease in two immunocompromised human patients, with limited archival information on a third case, a patient with acute myeloid leukaemia seen 23 years earlier². Is it a coincidence that the first reports of a new mammalian fungal pathogen occurred at virtually the same time, and in the same city, in both human and veterinary (feline) patients? It may be, but it would neglectful not to look further for factors that may explain why humans and cats were becoming infected by this hitherto non-pathogenic fungus. These events also emphasise the ‘One Medicine – One Health’ approach to infectious disease investigation, with animals representing sentinels for the occurrence of human disease. This is particularly the case for fungal diseases acquired from the environment.

In 2009, the first M. arundinis infection in the USA was reported in a human patient receiving immunosuppressive therapy for a renal transplant³. The man was domiciled in Florida, an area with a subtropical environment likely favourable to this organism. Sydney, while potentially temperate in climate by latitude, is classified as subtropical in rainfall distribution and temperature, with summer distribution of rainfall and mild winter temperatures.

In 2010, our group again reported disease caused by M. arundinis infection affecting the distal extremity of a cat (Figure 1), although this animal also had a lesion on its face⁴. Halliday molecularly characterised the internal transcribed spacer (ITS1), 5.8S and ITS2 regions and the D1/D2 region of the 28S rDNA gene cluster of the available human (four) and feline (two) isolates^4,5. These included a case (contributed by Tom Gottlieb) of refractory dermal plaques in a renal transplant recipient (Figure 2; Table 1). The isolates were deposited in various culture collections and the merged ITS and LSU sequences of the six isolates were deposited in the GenBank database (www.ncbi.nlm.nih.gov/genbank/).

Figure 1. Appearance of a localised Microsphaeropsis arundinis infection affecting the interdigital web of a cat. The lesion is highlighted by a red arrow.

Figure 2. Localised Microsphaeropsis arundinis infection on the right elbow of a renal transplant recipient. The infection developed after the patient fell on concrete. The initial lesion is shown in A, while B, C and D show progressive improvement during itraconazole therapy. Note the smaller satellite lesions in the vicinity of the large primary lesion, suggesting sporotrichoid spread via the lymphatics.

Table 1. Summary of all human Microsphaeropsis arundinis infections for which we have information (1981–2014).

Since the first human report by Pendle et al.², there have been at least six additional M. arundinis infections reported in human patients, individual cases being seen at St George Hospital, Wollongong Hospital, Concord Hospital (Figure 2), Westmead Hospital and Prince of Wales Hospital, and a further case from Florida in the USA (Table 1). The additional five Australian cases have been collated and submitted for peer review, including the case in Table 1 and Figure 2.

In the veterinary arena, we continue to see M. arundinis infections in cats along the East coast of Australia. It is now probably the most common cause of feline subcutaneous phaeohyphomycosis in this region, with five additional cases between 2009 and 2012 (Table 2). There does not appear to be any age predisposition in cats (range 5–20 years) and no gender preponderance. Geographically, two cats were domiciled in Sydney, two cats resided in the Central Coast of NSW, one was from Newcastle, another from Wollongong, while the last cat was from Brisbane. While there is the potential of geographical bias due to the catchment area of our pathology services, all these cats lived in coastal environments, which are becoming increasingly warmer and more humid. Lesions were invariably present on distal extremities, with either forelimbs or hind limbs being affected. Microscopy of needle aspirates or crush preparations from lesions were suggestive of phaeohyphomycosis, with pigmented bulbous septate hyphae or pseudo-hyphae of variable length, and occasional yeast-like forms evident in the tissues (Figure 3B–D). Histopathological specimens showed pyogranulomatous inflammation (Figure 3A), with occasional multi-nucleate giant cell formation. The organism grows well on routine fungal media such as Sabouraud dextrose agar (containing antibiotics) and microscopically shows irregularly-shaped pigmented septate hyphae, but no conidia. Since species identification is made difficult by the inability to induce sporulation, PCR amplification and sequence analysis was typically used to establish a specific diagnosis (as outlined above). This could be done using not only colonial material or fresh biopsy specimens from representative lesions but also paraffin-embedded formalin-fixed tissues⁵. Unlike the situation in people where most patients appear to be immunosuppressed by comorbid disease (renal failure, diabetes), corticosteroids or immunomodulatory drugs, most cats appear immune-competent. Cats with co-infections with other fungi (e.g. Fusarium spp.) are postulated to have had penetrating injuries contaminated by multiple fungi normally residing in soil.

Table 2. Summary of all feline Microsphaeropsis arundinis infections for which we have information (2001 to 2014).

Figure 3. Representative microscopic appearance of Microsphaeropsis arundinis in periodic acid Schiff (PAS)-stained histological sections (A) and Diff-Quik^®-stained smears from aspirates (B–D) from lesions on the distal extremities of cats. The hyphae have obvious septation, are of variable length, with occasional globose dilatations and yeast-like forms visible in different parts of the smear.

In veterinary practice, frustratingly, repeat samples for culture and antifungal susceptibility are often difficult to obtain. This is usually because serial specimen collection generally requires sedation or anaesthesia with concomitant cost and morbidity. To overcome this potential limitation, we have found that there is sufficient fungal nucleic acid preserved in methanol-fixed, DiffQuik^®-stained smears to permit successful DNA purification, panfungal PCR and sequence analysis using material scraped from the cytological specimens (Table 2). The method⁶, adapted from a similar technique used for diagnosis of veterinary mycobacteria specimens⁷, was successful in 4/4 feline M. arundinis cytology slides in which it was attempted (Table 2).

Antifungal minimum inhibitory concentration breakpoints have not been determined for this organism, although broth microdilution assays suggest most isolates are susceptible to broad spectrum azoles (itraconazole, voriconazole and posaconazole), amphotericin B and terbinafine, with variable susceptibility to fluconazole and resistance to echinocandins and flucytosine. Clinically, itraconazole, posaconazole, ketoconazole and terbinafine all appear to have good in vivo activity.

As voriconazole causes neurotoxicity in many feline patients⁸, we currently recommend posaconazole to treat M. arundinis infections in cats. Although the drug is expensive, it is palatable, available in a liquid form, has minimal hepatotoxicity (unlike itraconazole) and possesses favourable and reliable pharmacokinetics in this species, where once daily administration with meals is convenient for owners⁴. Terbinafine is probably also a good option for cats, with established pharmacokinetics and inexpensive generic formulations available, but we have no experience to date. Despite long courses of therapy, and in some cases cytoreductive surgery (typically toe amputation), there is a tendency for infections in cats to recur months or years after apparently successful therapy. This does not appear to be the case in human patients. Despite this propensity to recur, the organism seems inherently of low virulence, producing indolent infections without dissemination. This is possibly due to the organism favouring lower temperatures for growth, like many fungal saprophytes. The few well documented human cases reported to date have been cured with monotherapy using a triazole or terbinafine, sometimes combined with debridement or amputation of infected tissues^2,3.

We speculate that this infection might be becoming more common due to expansion in the range of ‘elephant grass’ or expansion of other plant hosts capable of supporting its environmental growth. The location of lesions on the toes of cats suggests penetrating injury of distal extremities during digging may play a role in disease pathogenesis. For the same reason, it might be expected that cat scratch injuries (typically to the face) might also be contaminated by this organism^9,10, as in the case reported by Krockenberger et al.⁴. Similar predisposing factors may be operating in people who garden without protective measures (gloves and long sleeved apparel) and therefore may be at risk for development of M. arundinis infections through contamination of skin wounds, especially if they have diabetes or are receiving immunosuppressive therapy. It is interesting that cats have an apparent predisposition, as to date there have been no reported infections in non-feline veterinary patients, i.e. dogs, horses, cows, etc. The reason for this species predilection requires investigation. Time may reveal the answers to these questions as we actively search for more cases!