INTRODUCTION
The fungal order
Mucorales, belonging to a section of lower fungi that until recently was referred to as zygomycetes, constitutes a phylogenetically ancient group of organisms. In the fungal tree of life the group encompasses a number of widely spaced, ancestral lineages. Over time, mutations are hypothesized to have accumulated, which is reflected, e.g., in an immense degree of sequence diversity of evolutionary markers such as the ribosomal operon. By assessment of identical genes, mucoralean species are separated from each other at branches much longer than those of species of more recent fungi, such as
Aspergillus or the dermatophytes. As a result, ancestry is difficult to reconstruct, leading to phylogenetic trees with poorly resolved backbones. For similar reasons, the phylum
Zygomycota has been abandoned: phylogenetic distances are so large that no taxonomic hierarchy can be constructed and no umbrella group defined that would unite all fungi attributed to the
Zygomycota in the classical sense (
23).
The opportunistic members of the
Mucorales are classified in the families
Cunninghamellaceae,
Lichtheimiaceae,
Mucoraceae,
Saksenaceae, and
Syncephalastraceae, with the great majority of human infections being caused by members of
Mucoraceae and
Lichtheimiaceae. In molecular phylogenetic analyses (
31,
43), the genus
Rhizomucor was positioned outside the
Mucoraceae, and in the present article the Index Fungorum (
http://www.indexfungorum.org) is followed, classifying the genus in the
Lichtheimiaceae.
Infections generally occur in severely debilitated patients and are acute, destructive, and with a rapid course and fatal outcome (
14,
37). In general, different types of underlying conditions predispose for different types of infection. Major skin abrasion and burn wounds may lead to erosive subcutaneous infection. Rhinocerebral and pulmonary infection are linked to ketoacidotic diabetes and severe neutropenia, respectively, while immunosuppression and prolonged deferoxamine therapy predispose for disseminated infection (
11). Chronic disorders observed in individuals without severe immune or metabolic dysfunction are exceptional cases (
25). Also, renal mucormycosis tends to occur in immunocompetent individuals (
21,
27).
Given the enormous phylogenetic diversity of the
Mucorales (
31,
43), it is remarkable that frequent case reports appear referring to the etiologic agent without proper species identification (e.g., references
19 and
26). Practical reasons for this are the difficulties of cultivating these fungi from biopsy samples (
22,
34) and of differentiating zygomycete species by classical mycological techniques in clinical microbiology laboratories. Previous comprehensive studies on susceptibility profiles against antifungal drugs in the
Mucorales revealed considerable variation among and within genera and species, as defined either by applying classical parameters (
5,
16,
35,
39,
40) or more recently by using molecular taxonomic methods (
1,
12). A recent review of
in vitro activity of antifungals against zygomycetes was provided by Alastruey-Izquierdo et al. (
2).
Application of molecular methods for species identification in
Mucorales frequently leads to the unexpected detection of novel sibling species (
3,
6,
9). For this reason, and because of the difficulty of morphological species identification, studies on the taxon specificity of susceptibility profiles in this group of fungi need to be preceded by molecular identification of the strains under study (
1). Otherwise, in clinical practice, when the only intention is to determine the most appropriate antifungal as quickly as possible, susceptibility tests could also be performed directly. We chose the internal transcribed spacer (ITS) region for taxonomy at the species level because it earlier has been shown to be the species marker of choice in
Mucorales (
36).
Classification of
Mucorales above the species level is in a state of flux, since molecular phylogenetic analyses found polyphyly of the majority of morphology-based families and genera (
31,
41). The D1/D2 region of the nuclear ribosomal large subunit (LSU) was chosen to reconstruct phylogeny because it could be sequenced directly, while all protein-encoding genes tested revealed paralogs in numerous species (
3). Furthermore, the LSU is alignable over the entire order. A robust molecular phylogenetic hypothesis is necessary to address the main question of this study: do phylogenetic taxa (species, genera, and families) of the
Mucorales possess more or less characteristic susceptibility profiles?
RESULTS
The LSU tree (
Fig. 1) comprises members of the clinically significant genera
Apophysomyces,
Cokeromyces,
Cunninghamella,
Lichtheimia,
Mucor,
Rhizomucor,
Rhizopus,
Saksenaea, and
Syncephalastrum, all belonging to the order
Mucorales. Sequences could be aligned with relative confidence but showed considerable diversity. In contrast, morphological identification of species appeared to be nonpredictive, with extreme differences occurring between neighboring molecular taxa. Ribosomal DNA (rDNA) distances between species belonging to a single genus were considerable. ITS diversity was up to 20% in
Lichtheimia and 35% in
Mucor, while
Rhizopus arrhizus and
Rhizopus microsporus deviated 29.8% (data not shown). ITS sequences showed very limited similarity over the entire data set and could not be aligned. Separate trees were therefore constructed for molecular identification (
Fig. 2).
With a single exception, all strains included in susceptibility testing could be identified reliably to the species level by using the ITS region, because of their position in shared clades with ex-type strains (
Fig. 2). The correct identification of
Syncephalastrum racemosum turned out to be problematic because this species has not been typified yet, and strains morphologically identified as
Syncephalastrum racemosum formed two clearly separate clades in the ITS tree (
Fig. 2f): some
S. racemosum strains were part of a well-supported clade with
Syncephalastrum monosporum, while others (including the clinical strain CNRMA03.414), represented by two types of ITS sequences (note different clone numbers of the same strains), formed a second group.
Syncephalastrum racemosum has been described by Schroeter (original reference cited in reference
13) as “inter
Aspergillus oryzae in
Oryza et pane,” i.e., on moldy rice and bread, in Wroclaw, Poland, and no type material is known to be preserved. None of the isolates studied morphologically completely matched the description given by Schroeter, and more-detailed taxonomic studies are needed to select a neotype for this species. Therefore, we designate the putative species in the following as
Syncephalastrum “
racemosum” species I and
S. “
racemosum” species II.
Judging from LSU and ITS phylograms (
Fig. 1,
2)
Rhizomucor regularior and
Rhizomucor variabilis were found to be related to
Mucor but belonged to the species complexes of
M. circinelloides and
M. hiemalis, respectively;
R. variabilis has recently been renamed as
M. irregularis, and
R. regularior has been synonymized with
M. circinelloides (
9). The genus
Rhizomucor is restricted to thermophilic species forming spherical spores, such as
R. pusillus and
R. miehei. Mucor ramosissimus belongs to
M. circinelloides. The varieties
azygosporus,
chinensis,
rhizopodiformis, and
tuberosus of
Rhizopus microsporus possessed identical ITS sequences and could not be discriminated molecularly (
Fig. 2b). Therefore, the morphology-based assignment to the varieties was retained. In
Rhizopus arrhizus (syn.
R. oryzae), small molecular differences between varieties were observed, but because of the limited number of reference strains available representing each of the varieties, identifications were done at the species level only.
A total of eight antifungal compounds were tested on six mucoralean genera:
Mucor and
Rhizopus (
Mucoraceae),
Lichtheimia and
Rhizomucor (
Lichtheimiaceae),
Syncephalastrum (
Syncephalastraceae), and
Cunninghamella (
Cunninghamellaceae). The
Cunninghamellaceae and the
Syncephalastraceae are unigeneric, and conclusions drawn on the genus are also valid for the family. The positions of the genera in
Tables 1 and
2 refer to their positions in the LSU phylogram of
Fig. 1, except for
Rhizomucor. Rhizomucor as a member of the
Lichtheimiaceae is placed with
Lichtheimia because its position closer to
Circinella in the LSU phylogram is not supported. Within the genera, strains are arranged according to their position in the phylogenetic trees of
Fig. 2. Because clinical breakpoints for filamentous fungi have not been assigned to the majority of antifungals, we refer to those given by Almyroudis et al. (
5) and de Hoog et al. (
17),
viz. AMB ≤ 1; CAS ≤ 2; 5-FC ≤ 16; ITC ≤ 0.5; PCZ ≤ 0.5; and VRC ≤ 2.
For 5-FC, MCF, and CAS, little variation in susceptibility was noted: for 5-FC and CAS, all strains showed high MIC/MEC values of 32 to >64 and ≥4 μg/ml, respectively. Except for two strains of Cunninghamella, high MEC values (≥4 μg/ml) were also obtained for MCF. All strains included in this study were resistant or less susceptible to VCZ.
For the remaining compounds, considerable variation was found at the family and generic levels inferred from the LSU phylogram. Overall, AMB
in vitro was the most effective antifungal agent against
Mucorales, but its efficacy proved to be ambiguous in
Rhizopus and
Cunninghamella, where relatively high MICs were obtained. Three out of 19
Rhizopus strains and 5 out of 8
Cunninghamella strains tested had MICs of 2 μg/ml, exceeding the assumed breakpoint for this drug. MICs of these two genera were significantly elevated when compared to the other
Mucorales, for which the amphotericin B MICs ranged between 0.03 and 0.5 μg/ml (Mann-Whitney,
P < 0.0001 for both genera). Together with those for posaconazole, amphotericin B MICs showed the smallest range of variation over the entire set of
Mucorales tested (
Table 2). Among the azoles, PCZ was the most effective antifungal drug, with all strains being inhibited by concentrations of 2 μg/ml. The genera
Mucor (defined in a phylogenetic sense) and
Cunninghamella showed significantly lower degrees of susceptibility to PCZ (Mann-Whitney,
P < 0.0001 for both genera). In
Rhizopus and
Syncephalastrum, resistant strains occurred only sporadically. The
Lichtheimiaceae, represented by
Lichtheimia and
Rhizomucor, completely lacked strains with reduced susceptibility for PCZ.
Itraconazole was significantly less active against the Mucoraceae (defined in a phylogenetic sense) and Cunninghamella, in which especially high MIC values were obtained (Mann-Whitney; P < 0.0001 for the Mucoraceae as well as for Cunninghamella). In contrast, the Lichtheimiaceae and Syncephalastrum (Syncephalastraceae), united in a well-supported clade in the LSU phylogram, did not include strains with reduced susceptibility for ITZ.
For TBF, elevated MIC values were found only in the Mucoraceae, while the remaining families, namely, the Cunninghamellaceae, the Lichtheimiaceae, and the Syncephalastraceae, were fully susceptible. The susceptibility of Cunninghamella to TBF is of special importance because TBF was the only thoroughly active antifungal against this genus in our study. The susceptibility to TBF in the Mucoraceae appeared to be species dependent. Terbinafine MICs were strikingly different between R. microsporus (geometric mean [GM] MIC of 0.12 μg/ml) and R. arrhizus (GM MIC of 21.1 μg/ml) (Mann-Whitney, P = 0.0020) but not between M. circinelloides sensu lato (GM MIC of 5.66 μg/ml) and the remaining Mucor species (MICs of ≤0.5 μg/ml), although a trend was noted (Mann-Whitney, P = 0.0746). Susceptibility values of PCZ and TBF were more heterogeneous in the Mucoraceae than in the remaining families.
Maximum susceptibilities were reached with PCZ (MIC = 0.03 μg/ml) for Rhizopus arrhizus, ITZ (MIC = 0.03 μg/ml) for Lichtheimia ramosa and Syncephalastrum “racemosum” species I, and AMB (MIC = 0.03 μg/ml) for Syncephalastrum “racemosum” species II, while maximum resistance was observed with 5-FC (MIC > 64 μg/ml) for nearly all strains tested and with ITZ (MIC > 64 μg/ml) for 3 Cunninghamella strains.
Intrageneric differences of
in vitro susceptibility in
Mucor and
Rhizopus ranged from 4 to 9 log
2 dilution steps. With respect to PCZ and ITZ, the genus
Mucor seems to contain a higher proportion of strains that are resistant to or show reduced susceptibility than
Rhizopus (Mann-Whitney,
P = 0.0042 and
P = 0.0369, respectively). In contrast MICs for VCZ were lower in
Mucor than in
Rhizopus (Mann-Whitney,
P < 0.0001). Some intraspecific variability of
in vitro susceptibility was found in all genera tested (
Table 1). Strain CNRMA 04.1469 of
R. microsporus deviated from remaining strains of this species in its response to azoles.
DISCUSSION
A robust taxonomy and phylogeny of the Mucorales require species validation by their ex-type materials. These reference materials serve to calibrate additional strains examined, such that a taxonomic scaffold is provided, allowing meaningful comparison of susceptibility data at the levels of species, genera, and families.
The genera
Mucor and
Rhizomucor have been misapplied in some medical publications. Judging from ITS sequence data,
Rhizomucor regularior (as
R. variabilis var.
regularior) was declared to be a synonym of
Mucor circinelloides (
9) as suggested before (
6,
36). The typical variety of
Rhizomucor variabilis var.
variabilis fits elsewhere in the genus
Mucor (
1,
10,
42), and recently Álvarez et al. (
9) proposed a name change to
Mucor irregularis.
Our study supports AMB as the antifungal of choice for most genera of
Mucorales. High MICs for AMB have been reported for some genera not tested in the present study, such as
Apophysomyces and
Saksenaea (
7,
8,
12,
15,
35,
39). Results for
Rhizopus and
Cunninghamella, however, are ambiguous, with MICs of 2 μg/ml for AMB. Similar results were obtained by other authors (
1,
35,
39,
40). Alastruey-Izquierdo et al. (
1) listed
R. arrhizus and
Cunninghamella bertholletiae with MIC ranges as wide as 0.03 to 32 μg/ml and 2 to 32 μg/ml, respectively. In concordance with previous studies (
1,
5,
16), PCZ was the second most active drug, all strains being inhibited by 2 μg/ml or less (
Table 2). The success rate of therapy was reported to be 79% (
20).
5-FC, MCF, and CAS were ineffective in all or nearly all
Mucorales, showing high MIC/MEC values of ≥32, ≥4 (for 64 out of 66 strains), and ≥4 μg/ml, respectively. This resistance is in accordance with literature data (
1,
5,
16,
40). With MICs consistently ≥1 μg/ml in all
Mucorales analyzed, VCZ was less active than other azoles. The poor activity of this antifungal has been highlighted previously (e.g., references
1,
5, and
16).
Phylogenetic distances within the
Mucorales are reflected in
in vitro susceptibility profiles against antifungal drugs on all levels of family, genus, species, and strain (
Tables 1 and
2). Members of the
Lichtheimiaceae were consistently susceptible to PCZ, ITZ, TBF, and AMB and exhibited smaller MIC ranges than the
Mucoraceae. This matches with MIC values published (
5,
16,
24,
32,
35,
39), although Torres-Narbona et al. (
40) and Alastruey-Izquierdo et al. (
1,
4) reported on individual
Lichtheimia strains that were resistant to PCZ, ITZ, and TBF and to ITZ, respectively.
The consistent
in vitro activity of TBF against
Cunninghamella is of clinical relevance because the proportion of strains that are resistant or have reduced susceptibility to PCZ and AMB is relatively high in this genus, as shown in this study and by other authors (
1,
5,
39). Alastruey-Izquierdo et al. (
1) tested the susceptibility of
Cunninghamella against TBF with a similar result. However, the authors found an individual strain resistant to TBF.
The species-specific differences in the susceptibility to TBF found (
Rhizopus arrhizus [resistant] versus
R. microsporus [susceptible] and
Mucor circinelloides sensu lato [resistant] versus remaining
Mucor species [susceptible]) were also in agreement with earlier studies (
1,
16). Significantly deviating resistance to azoles was found in a single strain of
R. microsporus; Alastruey-Izquierdo et al. (
1) reported that TBF data were variable in that species.
In conclusion, AMB and PCZ were the most effective antifungal agents against
Mucorales. Susceptibility profiles (restricted to the drugs that were at least partly active) differed significantly at the familial, generic, and specific levels and reflected relationships as referred from the phylogram; the
Lichtheimiaceae were fully susceptible to PCZ, ITZ, TBF, and AMB.
Syncephalastrum (
Syncephalastraceae) members, positioned at the shortest distance to the
Lichtheimiaceae in the LSU phylogram, showed similar susceptibility profiles. The only difference was a single strain resistant to PCZ. In contrast, the
Mucoraceae were characterized by a reduced susceptibility to PCZ, ITZ, and AMB and a lack of activity of TBF in some species. While more
Mucor strains were resistant to the azoles, only
Rhizopus strains exhibited a reduced susceptibility to AMB. Susceptibility profiles of
Cunninghamella (
Cunninghamellaceae) resembled those of the
Mucoraceae in terms of the high proportion of strains resistant to PCZ, ITZ, and AMB but differed in terms of the low MIC values for TBF. Judging from MIC values recorded by other authors, the
Sakseneaeceae behave similarly to
Cunninghamella, their closest studied neighbor in the LSU phylogram, in susceptibility tests. Along with
Cunninghamella, they exhibit a lack of or reduced susceptibility against PCZ, ITZ, and AMB (
5,
7,
8,
12,
35) and low MIC values for TBF (
7) (tested only for
Saksenaea).
Obviously, the Mucorales cannot be considered as a single entity from an antifungal perspective, since large differences in susceptibility exist between families, genera, species, and strains. The reduced activity of AMB in Rhizopus and Cunninghamella and the high proportion of Mucor and Cunninghamella strains that were less susceptible to PSZ are of practical importance because they concern deviations for compounds that are otherwise recommended for antifungal therapy of infections due to Mucorales. Molecular identification of the etiologic agent is therefore required unless the susceptibility profile of the strain is known. In genera with high intraspecific variation in their antifungal profiles, such as Mucor and Rhizopus, immediate susceptibility testing is recommended to confirm the most effective and appropriate compound.