Genetic Diversity among Clinical Isolates of Acremonium strictum Determined during an Investigation of a Fatal Mycosis

Abstract

Primarily saprophytic in nature, fungi of the genus Acremonium are a well-documented cause of mycetoma and other focal diseases. More recently, a number of Acremonium spp. have been implicated in invasive infections in the setting of severe immunosuppression. During the course of routine microbiological studies involving a case of fatal mycosis in a nonmyeloablative hematopoietic stem cell transplant patient, we identified a greater-than-expected variation among strains previously identified as Acremonium strictum by clinical microbiologists. Using DNA sequence analysis of the ribosomal DNA intergenic transcribed spacer (ITS) regions and the D1-D2 variable domain of the 28S ribosomal DNA gene (28S), the case isolate and four other clinical isolates phenotypically identified as A. strictum were found to have <99% homology to the A. strictum type strain, CBS 346.70, at the ITS and 28S loci, while a sixth isolate phenotypically identified only as Acremonium sp. had >99% homology to the type strain at both loci. These results suggest that five out of the six clinical isolates belong to species other than A. strictum or that the A. strictum taxon is genetically diverse. Based upon these sequence data, the clinical isolates were placed into three genogroups.

Serious infections in severely immunocompromised patients due to filamentous fungi belonging to genera other than Aspergillus have become increasingly common (17). The anamorphic genus Acremonium is a case in point. Members of this genus are hyaline, septate, filamentous fungi that reproduce by phialidic conidiation. While Acremonium spp. can be readily isolated from various environmental sources and are a known cause of eumycotic mycetoma and other focal infections in otherwise healthy individuals, they have in the past been generally considered to be minimally invasive human pathogens (6). However, as treatment modalities for malignancy and other diseases have led to increased levels of immunosuppression, so too have Acremonium spp. been increasingly implicated in invasive systemic mycotic disease (6, 15, 27, 30).

The genus Acremonium is known to be a polyphyletic grouping of genetically distantly related fungi (8). As a result of our investigation into a fatal disseminated mycosis in a hematopoietic stem cell transplant (HSCT) patient, we demonstrate that mould isolates phenotypically identified as Acremonium strictum by established clinical mycology laboratories exhibit wide genetic diversity.

CASE REPORT

The patient was a 59-year-old male who received an HSCT from a human leukocyte antigen-matched sibling following nonmyeloablative conditioning therapy 5 months after an initial diagnosis of acute myelogenous leukemia. His course was uncomplicated until day 92 posttransplant, when he developed gastrointestinal graft-versus-host disease (GVHD) manifested by severe gastrointestinal bleeding. At that time, he received therapy with steroids and anti-thymocyte globulin for GVHD and itraconazole for antifungal prophylaxis. Beginning on day 120, the patient experienced several episodes of altered mental status associated with hepatic transaminitis, attributed to GVHD and/or itraconazole. His steroid dose was increased and itraconazole was discontinued, and the patient's mental status markedly improved to the point where he was able to begin physical therapy on day 138. On day 148, skin lesions were first noted on his left thigh, which then rapidly progressed over his body. The lesions were initially maculopapular with necrotic centers, some of which subsequently developed into bullous lesions (Fig. 1A). At that time, the patient's medications included anti-thymocyte globulin, prednisone, foscarnet, levofloxacin, trimethoprim-sulfamethoxazole, vancomycin, and fluconazole. Notable laboratory results included an absolute neutrophil count of 6.8 × 10²/μl (normal, 1.8 × 10³ to 7.0 × 10³/μl). Six blood cultures were collected between days 142 and 152, all of which yielded a fungus. Despite therapy with Ambisome and then an investigational triazole, the patient died on day 155 posttransplant. A fungus resembling the blood isolates was identified in multiple organs by histopathology (Fig. 1B) and culture at autopsy.

FIG. 1.

(A) Cutaneous lesions of lower extremities. (B) Fungal hyphae, myocardium (modified Gomori methenamine silver stain; 40× objective). (C) Fungal form resembling yeast with hyphal element; blood culture (Gram stain; 100× objective). (D) Hyphal mass with terminal conidia; blood culture (Gram stain; 100× objective). (E) A. strictum case isolate; slide culture displaying cylindrical conidia, subtly thickened phialide walls, and basal septum (100× objective; Nomarski differential interference).

MATERIALS AND METHODS

Fungal strains.

The sources of fungal strains used in this study are listed in Table 1. Strains from outside institutions were graciously provided to us by the following individuals: UWFP940 and -941, Deanna Sutton; UWFP942, James Snyder; and UWFP982, Wiley Schell. The case isolate has been deposited with the University of Alberta Microfungus Collection and Herbarium (Edmonton, Canada) (culture number UAMH 10253).

TABLE 1.

Fungus isolates

Species identity (by phenotype)	Strain no.	Institutiona	Comment
Acremonium sp.	UWFP580	UWMC	Human blood
A. strictum	CBS 346.70	CBS	Type strain; isolate of Triticum aestivum leaf
	UWFP836	UWMC	Case index human isolate
	UWFP940	FTL	Human isolate; originally FTL 98-924
	UWFP941	FTL	Human isolate; originally FTL 00-2492
	UWFP942	ULMC	Human isolate; originally UN 2814
	UWFP982	Duke	Human isolate; originally Duke 136.90
Aspergillus fumigatus	UWFP502	UWMC	Human isolate
N. mauritiicola	CBS 313.72	CBS	Type strain; Linum usitatissimum rhizosphere

^aCBS, Centraalbureau voor Schimmelcultures (Utrecht, The Netherlands); Duke, Duke Medical Center (Durham, N.C.); FTL, Fungus Testing Laboratory, University of Texas Health Science Center (San Antonio, Tex.); ULMC, University of Louisville Medical Center (Louisville, Ky.); UWMC, University of Washington Medical Center.

Culture conditions.

Blood culturing was performed using the BACTEC 9240 automated blood-culturing system (Becton Dickinson Co., Sparks, Md.). Each culture consisted of one each Plus Aerobic/F, Lytic/10 Anaerobic/F, and Myco/F Lytic bottles. Aerobic and anaerobic media were held in the BACTEC cabinet for 5 days; Myco/F bottles were held for 28 days. Aerobic and anaerobic bottles positive for yeast-like fungi were subcultured to chocolate, bromcresol green, and inhibitory mold agar plates and incubated at 35°C supplemented with CO₂ to 5%. Other fungal cultures were performed using Sabouraud dextrose agar (SAB; Emmon's modification); brain heart infusion agar with blood, chloramphenicol, cycloheximide, and gentamicin; and inhibitory mold agar incubated at 30°C. Subcultures for morphological studies were made on potato dextrose agar and incubated at 30°C unless otherwise noted. All plate media were purchased from Remel Inc. (Lenexa, Kans.).

Phenotypic identification.

The identification of Acremonium isolates at the University of Washington was primarily based upon the dichotomous key of Domsch et al. (3). The patient isolate was independently identified by our mycology reference laboratory, the Fungus Testing Laboratory (University of Texas Health Science Center at San Antonio, San Antonio, Tex.). Strains from outside institutions were definitively identified by those institutions; upon receipt by the University of Washington mycology laboratory, these strains were checked for purity and for the expected microscopic and macroscopic morphologies.

Susceptibility testing.

Susceptibility testing was performed by the Fungus Testing Laboratory using the NCCLS broth macrodilution method (24).

Genotypic analysis.

Fungal DNA for sequence analysis was extracted from mature colonies, grown on SAB agar with chloramphenicol and gentamicin (Remel Inc.) at 30°C, using the QIAmp Mini Kit (Qiagen Inc., Valencia, Calif.) following the manufacturer's tissue extraction protocol.

The intergenic transcribed spacer 1 (ITS1) and ITS2 regions of the rRNA operon, flanking the 5.8S rRNA gene, were PCR amplified using the ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) primers (16, 35). The D1-D2 variable domain of the 28S rRNA gene was amplified using the NL-1 (5′-GCATATCAATAAGCGGAGGAAAAG-3′) and NL-4 (5′-GGTCCGTGTTTCAAGACGG-3′) primers (14). The PCR and sequencing protocols were described previously (2). The nucleotide-nucleotide BLAST program (http://www.ncbi.nlm.nih.gov/BLAST/) was used to query the National Center for Biotechnology Information GenBank nucleotide database for homologous sequences. The sequences were aligned and phylogenetic trees were drawn with Clustal X, which uses the neighbor-joining method of Saitou and Nei (28, 34). The aligned sequences were edited with Jalview version 1.3b (M. Clamp, European Bioinformatics Institute [http://circinus.ebi.ac.uk:6543/jalview]). Phylogenetic trees were displayed using Treeview version 1.6.6 (23).

RESULTS

Microbiology and Antifungal Susceptibility Data.

Moulds with similar morphologies were isolated from 13 cultures: 6 blood cultures, 1 skin biopsy culture, and 6 postmortem cultures (liver, spleen, left and right lungs, kidney, and brain). Gram stains of positive blood culture bottles showed both yeast-like forms with hyphal elements (Fig. 1C) and fully formed hyphal masses suggestive of a sporulating mould (Fig. 1D). Four out of six blood cultures grew the fungus in both the Myco/F and Plus Aerobic bottles, while the other two produced the fungus in Myco/F bottles only. No Lytic Anaerobic bottles signaled positive. The mean time to positive for the Myco/F medium was 4.2 days (range, 3 to 6 days) and 4.3 days (range, 4 to 5 days) for the Plus Aerobic medium. The initial isolate arose from a blood culture collected 142 days posttransplant.

On subculture, all isolates grew within 7 days at 30°C. Young colonies were smooth, moist, and pink, with a colorless reverse on inhibitory mold agar. Mature colonies were raised in the center and slightly velvety but still moist. Lactophenol aniline blue preparations showed conidia and septate hyphae. The conidia were one celled, cylindrical, 3 to 4 by 1 to 1.5 μm, smooth, hyaline to slightly pink, and grouped in slimy heads. The conidiophores were simple, slender, and erect phialides with basal septa arising from the vegetative hyphae, sometimes from fasiculated aerial hyphae (Fig. 1E). While no macroconidia were observed, initial observations nevertheless suggested either an Acremonium or Fusarium species. The initial case isolate grew to 2.2 cm in 7 days on SAB, which is consistent with Acremonium but not Fusarium (31). All case isolates were subsequently identified as A. strictum based upon micro- and macroscopic characteristics. The initial case isolate was referred for identification and antifungal susceptibility testing to the Fungus Testing Laboratory (Table 2), where it was also independently identified as A. strictum.

TABLE 2.

Antifungal susceptibility results for case isolate of A. strictum

Druga	MIC (μg/ml)		MLCb (μg/ml)
	24 h	48 h	24 h	48 h
AMB	2	8	8	>16
5-FC	>64
KETO	4	8
ITRA	>8

^a5-FC, flucytosine; AMB, amphotericin B; ITRA, itraconazole; KETO, ketoconazole.

^bMLC, minimum lethal concentration.

Sequence analysis.

The initial A. strictum case isolate (UWFP836), a number of clinical isolates phenotypically identified as A. strictum, and the A. strictum type strain (CBS 346.70) stratified into three genogroups based upon percent sequence similarities at the ITS and 28S loci (Table 3). UWFP580, phenotypically identified as an Acremonium sp., was the only strain that matched the A. strictum type strain, CBS 346.70, at both the ITS and 28S loci. In contrast, the case isolate had only 78.6 and 91.4% similarities at the ITS and 28S loci, respectively, with the A. strictum type strain (Table 4). The sequence similarities of all genogroups to the A. strictum type strain are given in Table 4 and indicate the diverse genetic nature of moulds phenotypically identified as “A. strictum.”

TABLE 3.

DNA sequence analysis of A. strictum isolates

Geno- group	Genotypic identitya	Phenotypic identity	Strain	ITS lengthb (bp)	% Identity (strain com- pared to)	GenBank accession no.c	28S lengthb (bp)	% Identity (strain com- pared to)	GenBank accession no.c
I	A. strictum	A. strictum	CBS 346.70T	587		AY138845	607		AY138483
		Acremonium sp.	UWFP580	587	99.3 (CBS 346.70)	AY138848	607	99.8 (CBS 346.70)	AY138485
II		A. strictum	UWFP836	574		AY138844	603		AY138482
	A. alternatum		CBS 223.70		99.3 (UWFP836)	U57674		99.8 (UWFP836)	U57349
III		A. strictumd	UWFP940, -941, -942, -982	578		AY138846	607		AY138484
	N. mauritiicola		NRRL 20420			NDf		100 (UWFP940, -941, -942, -982)	U88129
	N. mauritiicolae	N. mauritiicola	CBS 313.72T	573	81.2 (UWFP940, -941, -941, -982)	AY138847	605	91.6 (UWFP940, -941, -942, -982)	AY138481

^aA. alternatum CBS 223.70 (8) and N. mauritiicola NRRL 20420 (20) as given in GenBank (see Discussion). Full annotation of the strains associated with GenBank sequences may be found at http://www.ncbi.nlm.nih.gov/.

^bas determined in this study.

^cAll sequences beginning with the prefix “AY” were generated in this study.

^dAll four isolates had identical ITS and 28S sequences.

^eN. mauritiicola CBS 313.72 is listed for reference and has not been included in “A. strictum” genogroup III.

^fND, not deposited in GenBank.

TABLE 4.

Percent identity to A. strictum type strain CBS 346.70^a

Group	% Identity
	ITS	28S
I	99.3	99.8
II	78.6	91.4
III	94.4	98.1

^aA. strictum CBS 346.70 was compared to the following strains from each genogroup: UWFP580 (I), UWFP836 (II), and UWFP940 (III), >99% sequence homology at the ITS locus suggests conspecificity (see Discussion for details).

The ITS and 28S sequences of each strain were also compared to those available in GenBank (Table 3). The case isolate displayed 99.3 and 99.8% sequence homologies at the ITS and 28S loci, respectively, with Acremonium alternatum (CBS 223.70). Four other clinical isolates of “A. strictum,” UWFP940, -941, -942, and -982, were identical to one another at the 28S and ITS loci. They were also identical to Nectria mauritiicola (NRRL 20420) at the 28S locus. No ITS data for NRRL 20420 were available in GenBank for comparison. In contrast, UWFP940, -941, -942, and -982 had only 81.2 and 91.6% homologies to the ITS and 28S loci, respectively, of the N. mauritiicola type strain, CBS 313.72.

DISCUSSION

Fusarium spp. are consistently the most common causes of filamentous fungal disease in the HSCT patient after Aspergillus (13, 17-19, 22). This case was instructive because of its similarities to Fusarium-associated mycosis at two levels. The case clinical presentation, particularly the prominent cutaneous involvement, bore a close resemblance to disseminated fusariosis. This is in direct contrast to disseminated aspergillosis, in which cutaneous lesions are less common (10, 36). In fact, a provisional clinical diagnosis of fusariosis had been made before microbiology results became available. The similarities between Fusarium and Acremonium also extend to the microbiology of the two genera. Both are hyaline, septate moulds that usually cannot be distinguished by histopathological examination. Both may produce single-celled conidia of similar shapes on erect phialides, which were a characteristic of the case isolate. While Fusarium also produces sickle-shaped multicellular macroconidia in sporodochia, these are not always observed in the laboratory. Colonies of Fusarium spp. often produce various shades of red, blue, or purple, but these can be absent or subtle; furthermore, Acremonium spp. may produce similar pigments. When this occurs, one must resort to other techniques, including growth rate studies and a detailed analysis of reproductive structures, to accurately distinguish Acremonium from Fusarium. In this case, the growth rate and morphology studies clearly indicated the case isolate to be an Acremonium sp. DNA sequence analysis also clearly placed the isolate in the genus Acremonium.

The initial blood culture was thought to contain a Candida-type yeast forming hyphal elements (Fig. 1C). Upon review of the blood culture Gram stain the next day, forms suggestive of germinating conidia (Fig. 1C) and the so-called “adventitious form” noted by Schell and others (Fig. 1D) were observed (29, 30). Adventitious forms were also found in tissue sections by histopathology (Fig. 1B). Produced by certain members of Fusarium, Acremonium, and several other genera, but not by Aspergillus spp., adventitious forms represent phialidic conidiation in vivo, and in liquid media in vitro, in the absence of atmospheric gases. It has been hypothesized that adventitious conidiation is responsible for the high frequency of isolation in blood culture in cases of disseminated mycoses caused by these fungi. One cannot, therefore, rule out filamentous fungi when “yeasts with hyphal elements” are seen in blood culture or in tissue.

This case is also notable with respect to the DNA sequence findings, which clearly indicate the genetic diversity of clinical isolates phenotypically identified as A. strictum by clinical microbiologists (Table 4). Two independent laboratories with extensive mycological experience identified the case isolate as A. strictum: however, the ITS and 28S sequences of this strain did not match those of the A. strictum type strain, CBS 346.70, but were 99.3 and 99.8% similar, respectively, to the ITS and 28S loci of a strain designated in GenBank as A. alternatum CBS 223.70 (Table 3, genogroup II). While neither laboratory specifically considered A. alternatum (the dichotomous key of Domsch does not consider this species), several additional lines of evidence did suggest A. strictum. (i) The case isolate, like A. strictum, grew at 35°C, while A. alternatum does not (R. Summerbell, personal communication). (ii) A. alternatum produces conidia predominantly in chains, while both the case isolate and A. strictum do not (7). (iii) A. alternatum produces hyaline conidia, while the case isolate produced pink conidia (7). A number of studies with various yeasts and filamentous fungi have found that, in general, >99% sequence homology at the ITS or 28S loci is indicative of conspecificity and that superspecies differences tend to be much greater (1, 2, 5, 11, 12, 14, 32). (In contrast, O'Donnell found up to a 15% difference at the ITS locus of Fusarium sambucinum [21].) Assuming that the sequence submitted to GenBank correctly represents CBS 223.70, our data suggest by the criterion of equating >99% homology with conspecificity that either the designation of CBS 223.70 as A. alternatum is incorrect or A. strictum genogroup II is composed of phenotypically diverse but genetically closely related fungi (1, 2).

Other discrepancies were noted as well. Genogroup III contains a GenBank entry for N. mauritiicola, U88129, that at the 28S locus is 100% homologous with four clinical isolates phenotypically identified as A. strictum. While members of the anamorphic genus Acremonium are known to have affinities with various teleomorphic Nectria species, N. mauritiicola has been variously associated with Acremonium kashiense and Rhizostilbella hibisci but not A. strictum (20; http://www.cbs.knaw.nl/). The finding that genogroup III is genetically distinguishable from the N. mauritiicola type strain, CBS 313.72, (i) calls into question the identity of N. mauritiicola NRRL 20420, (ii) calls into question the validity of the GenBank sequence entered for NRRL 20420, or (iii) suggests that the taxon may also be polyphyletic or a genetically diverse single species. Only genogroup I, consisting of a clinical isolate identified as Acremonium sp. which was no longer available to us for further evaluation, matched the A. strictum type strain at the ITS and 28S loci (Table 3). Taken together, these results suggest that the A. strictum taxon is polyphyletic, as demonstrated with strong statistical support in Fig. 2. Studies of additional isolates by systematic mycologists will be needed to further clarify the natures of these genogroups.

FIG. 2.

Rooted phylogram of ITS sequence data. A. strictum genogroup I is represented by strain CBS 346.70; genogroup II is represented by UW836; genogroup III is represented by UWFP940. N. mauritiicola CBS 313.72 is included for comparison with genogroup III (see Discussion for details). The numbers at the tree nodes represent bootstrap values (n identical replicates out of 1,000 total replicates). A. fumigatus UWFP502 represents the outgroup. The bar shows relative genetic distance.

In vitro and in vivo susceptibility data for Acremonium spp. and the infections they cause are insufficient to make definitive treatment recommendations. While in vitro data indicate that Acremonium spp. are uniformly resistant to fluconazole and itraconazole but variably sensitive to amphotericin B (ca. 50% of the strains tested are sensitive), failures have occurred with amphotericin B and successes have been reported with itraconazole (6, 9, 33). Reports suggest that the new triazole drugs and caspofungin have in vitro activities against Acremonium spp. as well, but the efficacies of these drugs remain to be determined (4, 26). Antifungal susceptibility testing of the filamentous fungi is still in an early stage of development, which may serve to explain some of these apparent discrepancies. The recent advent of an accepted reference method for susceptibility testing in the clinical laboratory should facilitate the correlation of in vitro susceptibility data with clinical outcomes (25). The in vitro data for our case isolate indicated resistance to amphotericin B and itraconazole and are consistent with reported data. This patient had a rapidly progressive infection despite therapy with liposomal amphotericin B and an investigational triazole. As with many disseminated fungal infections, it is likely that this outcome was influenced by the impaired host defenses of the patient.

In conclusion, we have presented details of the first reported case of a fatal disseminated mycosis in a nonmyeloablative HSCT patient that was caused by a fungus phenotypically identified as A. strictum. However, this identification was not supported by DNA sequence data. We therefore believe that the A. strictum taxon may be polyphyletic or genetically diverse, a question that awaits further studies. Until then, therefore, the identity of this isolate remains A. strictum genogroup II.

While DNA sequence analysis was not definitive in identifying the case isolate, its use was instrumental in distinguishing the isolate from Fusarium. We anticipate that as public and private sequence databases become more robust and the taxonomy of the medically important fungi becomes clearer, the use of molecular methods to identify these fungi will become another accepted technique of the clinical microbiologist.