Pathogenic Spectrum of Fungal Keratitis and Specific Identification of Fusarium solani

Abstract

Purpose.: To investigate the predominant causative pathogens and epidemiologic features of fungal keratitis and establish a rapid, specific molecular method to detect fungal keratitis caused by Fusarium solani.

Methods.: A total of 174 patients with presumed fungal keratitis and 174 affected eyes were examined. Isolates from corneal specimens were identified according to morphologic and physiological characteristics. The primers that were designed for F. solani were tested to confirm whether they had species specificity. Multiplex PCR with universal fungal and F. solani–specific primers was performed with fungal and bacterial strains and was used to detect microorganisms in the clinical specimens.

Results.: A total of 160 patients (92.0%) were diagnosed with fungal infection by either potassium hydroxide wet-mount or microbiologic culture. Fungal cultures were positive in 128 patients (73.6%) with 139 fungal isolates. Fusarium (48.2%) was the most frequently isolated genus, in which F. solani (35.2%) was the most common species, followed by the Aspergillus (18.7%) and Candida (16.6%) genera. The PCR results showed that the designed primers were species specific and suitable for specific identification of F. solani. The multiplex PCR of 3-day broth cultures could identify and distinguish F. solani from other pathogens rapidly and specifically from clinical specimens.

Conclusions.: Fusarium species, especially F. solani, were found to be the predominant cause of fungal keratitis in northeast China. The established multiplex PCR method could have potential advantages for rapid detection of F. solani. These findings might have significance for early diagnosis and treatment of fungal keratitis.

Fungal keratitis is a refractory and potentially blinding fungal infection, with corneal ulceration and suppurative infection. It occurs in association with host hypoimmunity, ¹ inadequate use of topical antibiotics and steroids, ^2,3 previous corneal injury, ^2,4,5 wearing contact lenses, ^{3,6 –8} and ocular surgery or transplantation. ^8,9 Fungal keratitis is an insidious, rapidly progressive disease that is difficult to diagnose, and it can be resistant to treatment. ¹⁰ Sometimes it can lead to blindness or loss of the affected eye. ² To institute appropriate measures for prevention, accurate diagnosis, and effective treatment of fungal keratitis, the etiologic and epidemiologic characteristics should be determined. 

Fusarium species are opportunistic pathogens that are found widely as soil saprophytes and plant pathogens and commonly cause fungal keratitis. ^{11 –13} Fusarium species, especially Fusarium solani, can also cause other superficial emerging infections such as onychomycosis ^14,15 and sometimes serious, invasive, and even life-threatening systemic infection of immunocompromised patients. ^{16 –18}  

Traditional identification and classification methods based primarily on morphologic characteristics require much time to obtain a final result, and sometimes it is difficult to identify fungi because of atypical cultural findings and a lack of sporulation. Therefore, further research needs to focus on the development of rapid, reliable, and specific identification and diagnostic methods. As previously reported, ^19,20 the mitochondrial cytochrome b (mt cyt b) gene is useful for identification, classification, and phylogenetic analysis of fungi. Therefore, we attempted to establish a rapid, specific PCR method to identify the common pathogen of fungal keratitis, with specific primers designed according to the alignment of the mt cyt b gene. 

The present retrospective study of 174 patients with suspected fungal keratitis was conducted over a 5-year period from October 2004 to September 2009 in Jilin Province, northeast China. We documented the predominant pathogenic fungi that were responsible for infection, and we established a rapid detection method for specific identification of F. solani from clinical specimens. 

Methods

Patient Information

All patients who presented with clinically presumed fungal keratitis with corneal ulceration were included at the China–Japan Union Hospital of Jilin University during the 5 years from October 2004 to September 2009. Each patient was examined with a slit lamp, and sex, age, origin, onset of symptoms and clinical signs, medical history, ocular conditions, and associated predisposing factors were recorded. All subjects were treated in accordance with the Declaration of Helsinki. 

Fungal Isolation and Culture

Corneal specimens were obtained by an ophthalmologist under aseptic conditions using standard clinical techniques. First, the specimens were smeared on slides and examined by potassium hydroxide (KOH) wet-mount and light microscopy. Then they were directly inoculated in potato dextrose broth (PDB; Becton Dickinson, Sparks, MD), incubated aerobically at 25°C and 110 rpm, examined daily, and discarded at 2 weeks if there was no growth present. Strains with yeast colonies were inoculated onto Sabouraud dextrose agar (SDA; Becton Dickinson) and filamentous colonies onto potato dextrose agar (PDA; Becton Dickinson) in C-shaped streaks, and cultured at 37°C and 25°C for 3–5 days, respectively. At the same time, we recorded the morphologic characteristics of the colonies in individual fungal cultures, such as shape, color, growth, and exudation. 

Fungal Identification

Identification of fungal pathogens was performed based on colony morphology, microscopic characteristics, and physiological and biochemical properties, using standard mycological techniques. Yeast isolates were identified by direct microscopy with Gram's stain, germ-tube formation in serum, and cultural characteristics (CHROMagar Candida; CHROMagar, Paris, France). Identification of filamentous fungi was based on microscopic appearance on slide cultures stained with lactophenol cotton blue and included septate and branching hyphae, color, size, shape, texture, and formation of conidia, as well as physiological characteristics of colony morphology. 

Extraction of Fungal Mycelial Genomic DNA

Thirty reference strains were used to establish the rapid PCR method (Table 1). Besides two strains of F. solani, another 25 fungal strains and three representative bacterial strains were used as controls. One milliliter of culture broth with fungal mycelia was used for extraction of total cellular genomic DNA (Solutions I, II, and III of the GenTLE for yeast kit; TaKaRa Shuzo Co., Otsu, Shiga, Japan) as described previously. ²¹  

Table 1.

 

Fungal and Bacterial Strains Used to Establish Multiplex PCR

Species	Strain No.	Source
Fusarium solani	IFM 51104	Man
Fusarium solani	IFM 48449	Man
Fusarium verticillioides (Fusarium moniliforme)	IFM 49276	Blood
Fusarium oxysporum f. sp. crotalariae	IFM 48450	Unknown
Fusarium culmorum	IFM 51062	Populus nigra
Fusarium subglutlnans	IFM 51089	Corn
Fusarium proliferatum var. proliferatum	IFM 51092	Unknown
Fusarium inflexum	IFM 49467	Vicia faba
Fusarium larvarum	IFM 49468	Prunus domestica
Fusarium lumulosporum	IFM 49469	Citrus paradisi
Fusarium anthorphilum	IFM 51073	Leaf of Hippeastrum sp.
Fusarium dimerum	IFM 51070	Blood
Fusarium torulosum	IFM 50433	Soil
Fusarium tricinctum	IFM 51297	Wheat
Candida albicans	JLCC 30364	Blood
Candida glabrata	JLCC 30367	Blood
Cryptococcus neoformans	JLCC 30210	Cerebrospinal fluid
Aspergillus fumigatus	IFM 40808	Lung
Aspergillus flavus	IFM 55648	Soil
Aspergillus niger	JLCC 40445	Soil
Penicillium chrysogenum	JLCC 30780	Air
Paecilomyces lilacinus	JLCC 31764	Soil
Cladosporium carrionii	JLCC 31423	Clinical isolate
Cylindrocarpon sp.	JLCC 31299	Soil
Acremonium sp.	JLCC 30819	Soil
Absidia corymbifera	JLCC 30807	Air
Mucor circinelloides	JLCC 30827	Soil
Staphylococcus aureus	ATCC 25923	Clinical isolate
Escherichia coli	ATCC 25922	Clinical isolate
Pseudomonas aeruginosa	ATCC 27853	Unknown

 ATCC, American Type Culture Collection, USA; IFM, Institute for Food Microbiology (at present the Medical Mycology Research Center, Chiba University), Japan; JLCC: Culture Collection of Jilin University Mycology Research Center, China.

DNA Extraction

Corneal specimens were cultured in 2 mL PDB at 25°C for 72 hours. One milliliter was used to extract genomic DNA, and another 1 mL was cultured on PDA to confirm the PCR result. The reaction time between sample and Solution I was extended to 15 minutes, and after addition of Solution II, the mixture was heated at 95°C for 20 minutes. Other steps were the same as those for extraction of fungal mycelial DNA. 

PCR Primer Design

Universal fungal primers that targeted the mt cyt b gene sequence, E1M4 (5′-TGR GGW GCW ACW GTT ATT ACT A-3′) and rE2M4 (5′-GGW ATA GMW SKT AAW AYA GCA TA-3′), were designed as described previously. ^19,22 The species-specific primers of F. solani were chosen from divergent regions by alignment with the relevant fungal species inside the universal fungal amplicon using genetic information processing software (GENETYX-MAC; Genetyx Software Development Co., Tokyo, Japan): fFuso1 (5′-CTC TGT TAA TAA TGC AAC TC-3′) as forward primer, and rFuso2 (5′-TGG TAC TAT AGC TGG AGG A-3′) as reverse primer, with a PCR fragment of ∼330 bp. 

PCR Amplification

A Taq PCR amplification kit (TaKaRa Dalian Co., Dalian, China) was used for amplification. Each PCR was done in a 50-μL reaction volume that contained 5 μL 10× PCR buffer, 4 μL dNTP mixture (2. 5 mM each of dATP, dCTP, dGTP, and dTTP), 1.5 U Taq polymerase, 10 pmol of each primer, and 20 ng extracted template DNA. The multiplex PCR was done with the universal fungal and F. solani–specific primers. The amplification protocol consisted of an initial denaturation step at 95°C for 2 minutes, followed by 30 cycles of denaturation at 95°C for 1 minute, annealing at 50°C for 1 minute, and extension at 72°C for 2 minutes, with a final extension at 72°C for 10 minutes. 

Electrophoresis and Imaging

Amplified fragments were electrophoresed on 1.2% (w/v) Tris/borate/EDTA agarose gel at 100 V for 25 minutes, visualized (GelRed Nucleic Acid Gel Stain; Biotium Inc., Hayward, CA), and UV-illuminated using a gel image analysis system (GIS-2008; Tanon Technology, Shanghai, China). A 100-bp DNA ladder marker (TaKaRa Dalian Co.) was used as a molecular weight marker. 

Statistical Analysis

Statistical analysis was performed with commercial software (SPPS version 13.0; SPSS Inc., Chicago, IL). The McNemar test was used to compare the results of KOH wet-mount and microbiologic culture of corneal specimens. P < 0.05 was considered be significant. 

Results

Clinical Data Analysis

From October 2004 to September 2009, 174 patients with a clinical diagnosis of suspected fungal keratitis were examined. Most patients had typical clinical symptoms, including pain, photophobia, corneal epithelial defect, suppurative inflammation, conjunctival hyperemia, visual impairment, and ineffective antibiotic therapy. 

A total of 138 corneal specimens (79.3%) that were detected by KOH wet-mount were positive for septate and branching hyphae. Microbiologic culture showed that 128 specimens (73.6%) were positive for morphologic and physiological characteristics of Candida, Aspergillus, Fusarium, and other fungi. Overall, 160 of 174 patients (92.0%) were diagnosed with fungal keratitis by microbiologic culture (128 cases) or KOH wet-mount (32 cases) (Table 2). Comparison between these two techniques showed that there was no significant difference (P = 0.220; McNemar test). 

Table 2.

 

Detection Results of KOH Wet-Mount and Microbiologic Culture

Detection Method	Number of Cases
KOH (+), culture (+)	106
KOH (+), culture (−)	32
KOH (−), culture (+)	22
KOH (−), culture (−)	14

 KOH, potassium hydroxide wet-mount; culture, microbiologic culture; (+), positive; (−), negative.

The 160 patients with confirmed fungal keratitis comprised 92 men and 68 women, with a mean age of 48.7 ± 11.8 years (range, 41–60 years). There were 137 farmers from rural areas of Jilin Province. Infection of 83 patients was associated with corneal injury that was caused by hay, corn stalks, straw, tree branch, dust, soil, stones, tinsel, glass, needlepoints, or fingernails. Eighty-three patients were previously treated topically with antibiotics and two with steroids. These results were similar to previously published epidemiologic studies. ^{23 –25}  

Microbiologic Diagnosis

In 174 eyes from the same number of patients, 128 (73.6%) were culture positive. Among these, 107 eyes were infected with fungi alone, 10 with fungi and bacteria, and 11 with two species of fungi. The other 46 (26.4%) eyes were culture negative. 

In total, 139 stains of pathogenic fungi were isolated. Thirteen genera, 26 species, and 136 isolates were identified according to their morphologic and physiological characteristics, whereas three other filamentous fungi could not be identified because of a shortage of information provided by slide culture. Strains of the genus Fusarium were isolated most frequently (48.2%), followed by Aspergillus (18.7%), Candida (16.6%), and other genera (16.5%), such as Acremonium, Phialophora, and Absidia (Table 3). F. solani (35.2%) was the predominant causative pathogen in damaged cornea and caused fungal keratitis. 

Table 3.

 

Distribution of 139 Fungal Isolates from 128 Corneal Tissue Specimens

Species	Number of Isolates	Percentage (%)
Fusaium solani	49	35.2
Fusaium moniliforme	6	4.4
Fusaium poae	5	3.6
Fusaium oxysporum	2	1.4
Fusaium avenaceum	1	0.7
Fusarium spp.	4	2.9
Aspergillus fumigatus	17	12.2
Aspergillus flavus	5	3.6
Aspergillus niger	3	2.2
Aspergillus terreus	1	0.7
Candida glabrata	15	10.8
Candida guilliermondii	3	2.2
Candida albicans	1	0.7
Candida krusei	1	0.7
Candida spp.	3	2.2
Acremonium spp.	7	5.0
Alternaria alternata	5	3.6
Scedosporium apiospermum	1	0.7
Arthrinium phaeospermum	1	0.7
Phialophora verrucosa	1	0.7
Absidia corymbifera	1	0.7
Rhizopus stolonifer	1	0.7
Cladosporium carrionii	1	0.7
Bipolaris sp.	1	0.7
Neurospora sp.	1	0.7
Mycelia sterilia	3	2.2
Total	139	100

Specificity of Species-Specific Primers

To confirm that our species-specific primers for F. solani were indeed specific, we amplified genomic DNA from Fusarium species, other relevant fungi, and representative bacteria. Only F. solani was able to yield a clear, stable, and specific PCR fragment of ∼330 bp, whereas other fungi and bacteria could not be amplified (Fig. 1). 

Figure 1.

 

Specificity of the designed primers of F. solani. M: 100 bp marker, Lane 1–2: Fusarium solani, Lane 3: Fusarium verticillioides, Lane 4: Fusarium oxysporum, Lane 5: Fusarium culmorum, Lane 6: Fusarium subglutlnans, Lane 7: Fusarium proliferatum var. proliferatum, Lane 8: Fusarium inflexum, Lane 9: Fusarium larvarum, Lane 10: Fusarium lumulosporum, Lane 11: Fusarium anthorphilum, Lane 12: Fusarium dimerum, Lane 13: Fusarium torulosum, Lane 14: Fusarium tricinctum, Lane 15: Candida albicans, Lane 16: Candida glabrata, Lane 17: Cryptococcus neoformans, Lane 18: Aspergillus fumigatus, Lane 19: Aspergillus flavus, Lane 20: Aspergillus niger, Lane 21: Penicillium chrysogenum, Lane 22: Paecilomyces lilacinus, Lane 23: Cladosporium carrionii, Lane 24: Cylindrocarpon sp., Lane 25: Acremonium sp., Lane 26: Absidia corymbifera, Lane 27: Mucor circinelloides, Lane 28: Staphylococcus aureus, Lane 29: Escherichia coli, Lane 30: Pseudomonas aeruginosa, Lane 31: specimens from normal cornea, Lane 32: negative control. The amplified fragment of F. solani was ∼330 bp. Other strains and specimens from normal cornea could not be amplified.

Multiplex PCR with the universal fungal and F. solani–specific primers was performed. As predicted, multiplex PCR could distinguish F. solani from other species (Fig. 2). All fungi could be amplified with one fragment of ∼430 bp, but F. solani had another fragment of ∼330 bp, whereas bacteria and specimens from normal cornea could not be amplified. The results indicated that the species-specific primers were suitable for specific identification of F. solani. 

Figure 2.

 

Results of multiplex PCR with the universal fungal and F. solani–specific primers. M: 100 bp marker, Lanes 1–2: Fusarium solani, Lane 3: Fusarium verticillioides, Lane 4: Fusarium oxysporum, Lane 5: Fusarium dimerum, Lane 6: Fusarium subglutlnans, Lane 7: Aspergillus fumigatus, Lane 8: Acremonium, Lane 9: Mucor circinelloides, Lane 10: Penicillium chrysogenum, Lane 11: Candida glabrata, Lane 12: Staphylococcus aureus, Lane 13: Escherichia coli, Lane 14: Pseudomonas aeruginosa, Lane 15: normal cornea, Lane 16: negative control. F. solani had two amplified fragments of ca. 330 bp and ca. 430 bp. Other fungi had only one fragment of ca. 430 bp, while bacteria and specimens from normal cornea had no fragments. The bands <100 bp in Lanes 12–16 represented primer dimer.

Detection of Fungi in Corneal Specimens by Multiplex PCR

For rapid detection of clinical corneal specimens, genomic DNA was directly extracted from 3-day broth cultures and amplified with the established multiplex PCR method. This method detected F. solani and fungal infection at the same time (Fig. 3). The confirmatory test with culture of the specimens showed that the PCR results were correct. Therefore, the multiplex PCR with universal fungal and F. solani–specific primers could have potential advantages for rapid detection of clinical specimens. 

Figure 3.

 

Detection of corneal specimens by multiplex PCR. M: 100 bp marker, Lanes 1–14: corneal tissue specimens, Lane 15: normal cornea, Lane 16: negative control. The patients of No.1, 3, 7, 9, 11, and 14 were infected with F. solani, the patients of No. 2, 6, 8, and 12 were infected with fungi other than F. solani, while other patients were not infected with fungi.

Discussion

Fungal keratitis is a disease that occurs worldwide, which is characterized by corneal ulceration and visual loss. In China, it constitutes >60% of cases of severe infective keratitis. ^23,24 It mainly occurs in developing countries in tropical and subtropical regions, such as India, China, Iran, Ghana, and Brazil, ^{1,2,4,5,24,26,27} where the incidence is ∼50% of infective keratitis. ²⁸ In contrast, it is relatively uncommon in developed countries in temperate regions, such as the United Kingdom and United States. ^3,29  

At present, there are more than 56 genera and 105 species of pathogenic fungi that cause fungal keratitis or other ocular infections. ²⁸ The pathogenic spectrum of fungal keratitis differs between countries and regions as a result of different climate, environmental, and other natural conditions. The predominant pathogen of fungal keratitis in India is Aspergillus species, followed by Curvularia species. ³⁰ In Brazil, the most frequent pathogen is Fusarium species, followed by Candida and Aspergillus species. ³¹ In the United Kingdom, it has been shown that over half the cases of fungal keratitis can be attributed to Candida species. ²⁹ In our study, fungal culture and identification showed that the predominant pathogen in Jilin Province, northeast China, was Fusarium species, followed by Aspergillus and Candida species, which was similar to other areas of China. ^24,25  

The results of this study showed that there was no significant difference in direct microscopy with KOH wet-mount (79.3%) versus microbiologic culture (73.6%) for the tentative diagnosis of fungal keratitis. This suggested these two diagnostic tests were equally useful for diagnosis of fungal keratitis. KOH wet-mount was relatively rapid, simple, convenient, and inexpensive relative to the culture method. ^2,30,32 However, the culture method could accurately identify pathogens to the species level, whereas KOH wet-mount could not. Therefore, culture isolation and identification of species in fungal keratitis play an important role in antifungal therapy and epidemiologic studies. 

We found that F. solani was one of the most common isolates in fungal keratitis in Jilin Province, which was difficult to identify and had multiple drug resistance. At present, the identification of this species depends on morphologic characteristics and other methods such as serologic testing, electrophoretic karyotyping analysis, ³³ nested PCR, ¹¹ PCR-Restriction Fragment Length Polymorphism, ²⁷ mass spectrometry, ³⁴ and gene sequencing, ^35,36 but these are seldom used for clinical detection and diagnosis. 

In our present study, one isolate of Fusarium oxysporum JLCC 30687 that was identified by morphology was in fact F. solani, which was identified by mt cyt b sequence homology. Similarly, one isolate of Fusarium poae JLCC 30222 was Fusarium verticillioides according to sequencing. These results indicate that morphologic identification has some limitations. Microbiologic culture-based identification needs more time, sometimes more than 2–3 weeks, which can result in too long a delay for diagnosis and effective treatment. Therefore, it is important to develop a rapid, specific diagnostic strategy for clinical microbiologic detection. 

In this study, we have described a convenient, inexpensive, and reliable method, multiplex PCR of 3-day broth cultures, to identify fungi with a single amplified fragment of ∼430 bp and F. solani with another species-specific amplified fragment of ∼330 bp. Using this method, we reclassified the isolates of Fusarium moniliform JLCC 30864 and F. oxysporum JLCC 30226 and JLCC 30879 to F. solani, which was confirmed by sequence alignment. We tried to use this method to detect fungi in clinical specimens. Because of the size limitation of corneal specimens, we extracted genomic DNA and amplified it by multiplex PCR after it was cultured with PDB for 72 hours. Although this was not a culture-independent method, it only required ∼4 days, which is significantly shorter than conventional culture-based identification technique that needs more than 2–3 weeks. 

Our results indicated that the species-specific primers that we designed were indeed specific for identification of F. solani from other pathogens. Our multiplex PCR method was able to achieve specific identification of F. solani and could have potential advantages for rapid diagnosis of corneal infection, along with conventional culture techniques. These would be useful for accelerating early antifungal therapy, increasing the cure rate, and saving the diseased eye. 

Footnotes

 Supported by grants from the National Natural Science Foundation of China (Projects No. 30571773, 30910103903) and a grant from the Science and Technology Department of Jilin Province, China (Project No. 20080444-2).

Footnotes

 Disclosure: D. He, None; J. Hao, None; B. Zhang, None; Y. Yang, None; W. Song, None; Y. Zhang, None; K. Yokoyama, None; L. Wang, None

The authors thank the participants in this study, the staff of Jilin University Mycology Research Center, and Koji Yokoyama from the Medical Mycology Research Center of Chiba University for his help and guidance.