Assessment of genotypes, endosymbionts and clinical characteristics of Acanthamoeba recovered from ocular infection

During the study period, as per institutional protocol, all patients presenting with clinical features of microbial keratitis underwent a comprehensive slit-lamp examination for investigation of the corneal epithelial defect, size of infiltrate (if present), size of hypopyon, involvement of sclera, absence/presence of foreign particles and documentation of the ocular adnexal status [27]. The microbiological investigation of the corneal scrapings included smear preparation for microscopy using Gram stain and potassium hydroxide with calcofluor white (KOH + CFW) mount, and inoculation into appropriate media [5% sheep blood agar (BA), chocolate agar (CA), Sabouraud dextrose agar (SDA), potato dextrose agar (PDA), non-nutrient agar with Escherichia coli (NNA), thioglycolate broth, and brain heart infusion broth (BHI)]. Smear and inoculation of all media were done (as is standard) by the attending clinician at the slit lamp. Dehydrated culture media were supplied by HiMedia, Mumbai, India and freshly prepared in-house for use. Bandage contact lenses (2/13) were aseptically cut into small pieces and inoculated into BA, CA, SDA, NNA and BHI in the microbiology laboratory. All media were incubated aerobically at 37 °C except for SDA and PDA at 27 °C and chocolate agar which was incubated in 5% CO2 at 37 °C. The media were observed for 14 days for any growth. Acanthamoeba isolates from corneal scraping of 13 patients were confirmed by observation of active trophozoites and double-walled polygonal cysts; confirmed isolates were preserved on NNA and transported to School of Optometry and Vision Science, Faculty of Medicine and Health, UNSW, Sydney, Australia. The thermotolerance of isolates was evaluated by growing isolates at 28 °C, 37 °C, 40 °C and 42 °C. Following initial isolation, the Acanthamoeba strains were grown axenically in peptone yeast glucose (PYG) medium [28]. Each culture plate was monitored to ensure no extracellular microbes or contamination was present. To prevent possible contamination the culture medium was replaced with freshly prepared PYG every two days till trophozoites were harvested. In addition, a separate sterile incubator set at 32 °C was used for this study.

Acanthamoeba strains were further identified by PCR assay. Amoebal cells grown on NNA were collected into 500µL of 1X PBS (1.4 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄ and 1.8 mM KH₂PO₄, pH 6.9) using a sterile scraper (Sigma-Aldrich, St. Louis, Missouri, USA). Genomic DNA (gDNA) was extracted using 10% v/v Chelex lysis solution (Bio-Rad, CA, USA,) in 0.1% v/v Triton X-100 (Sigma-Aldrich) and Tris buffer, pH 8.0 (Thermo Fisher Scientific, Bedford, USA), as previously explained [29]. The quantity of extracted dsDNA was measured using Nano Drop UV–Vis spectrophotometer (Thermo Fisher Scientific) and gDNA vials were stored at -20 °C until further use.

The rns gene region of 18S rRNA was amplified in a PCR reaction using Acanthamoeba genus specific primers, JDPFw (5’-GGCCCAGATCGTTTACCGTGAA-3’) and JDPRv (5’-TCTCACAAGCTGCTAGGGGAGTCA-3’), which encode the highly variable DF3 region and give amplicons of ~ 450 bp [30]. Each PCR assay was performed in 12.5µL of DreamTaq Master Mix (DNA Polymerase, 2X DreamTaq buffer, dATP, dCTP, dGTP and dTTP: 0.4 mM each, and 4 mM MgCl₂; Thermo Fisher Scientific), 1µL of each primer (10 µM), 6.5µL of molecular grade water and 4µL of DNA template. PCR amplification was carried out in a 96-well T100 thermal cycler (Bio-Rad, California, USA) using the following thermal cycling conditions: 95 °C for 5 min for initial denaturation, followed by 35 cycles of amplification (94 °C for 30 s, 56 °C for 30 s and 72 °C for 45 s) and a final extension at 72 °C for 10 min [31]. PCR amplifications were observed by electrophoresis of 4μL PCR product aliquots in 1% agarose gel and amplicon bands were examined using Gel Doc XR + with image lab software (Bio-Rad). PCR positive amplicons were sent to the Ramaciotti Centre for Genomics (UNSW, Sydney, Australia) for Sanger sequencing. PCR products were purified using an EXOSAP-IT kit (Thermo Fisher Scientific) and sequencing was performed with forward primer JDPFw (5’-GGCCCAGATCGTTTACCGTGAA-3’) using BigDye Terminator (V3.1) reaction mix in 3730 DNA analyser (Applied Biosystems, Massachusetts, USA).

Low quality nucleotide sequences were manually checked and trimmed with Chromas (2.6.6) software (Technelysium Pty Ltd, Brisbane, Australia) [32]. The trimmed sequence reads were blast in the NCBI nucleotide sequences database (BLASTn) to identify the Acanthamoeba genotypes and species. Additionally, isolated Acanthamoeba strains were morphologically characterized to the species based on the classification scheme of Pussard and Pons [33]. All the confirmed and validated sequence reads were submitted to GenBank sequence data repository. Publicly available nucleotide sequences of Acanthamoeba genotypes were retrieved from the NCBI as reference strains for phylogenetic analysis (Additional file 1: Table S1). Sequences were aligned using ClustalW algorithm and a phylogenetic tree was constructed using the maximum likelihood method and Bayesian approach with Kimura-2 parameters by 1,000 bootstraps in MEGA-X [34] and phylogenetic tree was visualised using the interactive tree of life (iTOLv6) [35].

Acanthamoeba were grown axenically and maintained in PYG (peptone yeast glucose, pH 6.5) [36]. All isolates were seeded in separate wells of 24-well culture plate with PYG medium (500 µL/well) and incubated statically at 28 °C until the trophozoites formed 90% confluent layers at the bottom of each well. Aliquots of the PYG were inoculated onto trypticase soy agar (Becton, Dickinson and Company, Sparks, MD, USA) and incubated at 37 °C for 48 h. At the end of the incubation, the agar plates were examined for growth of any bacteria. Before collection of trophozoites, PYG medium was gently replaced by 1 mL of PBS containing gentamicin (100 μg/mL) to kill any extracellular bacteria and the plate was put on ice with gentle agitation to dislodge adhered trophozoites. TRIzol reagent (Invitrogen, Life Technologies, New York, USA) was used for the whole gDNA extraction following the phenol–chloroform separation method as per manufacturer’s protocol. The method was slightly modified to exclude the addition of EDTA which can interfere with PCR assay by sequestering Mg²⁺ ions [37]. To enhance the recovery of gDNA, TRIzol-treated cell suspensions were passed 10 times through 25G syringe needles (BD, New Jersey, USA) to lyse the amoebal cells.

The 16S rRNA gene (V4 region) of intracellular bacteria was amplified using eubacteria 16S rRNA primer; 515Fw/806Rv: Fw (5′- CCTACGGGNGGCWGCAG-3′) and Rv (5’ – GACTACHVGGGTATCTAATCC-3′) [38]. Genomic DNA of E. coli ATCC 10,798 and nuclease free water were included as positive and negative controls, respectively. The presence of intracellular fungi was assessed using fungus-specific forward ITS1Fw (5’- TCCGTAGGTGAACCTGCGG-3’) and reverse ITS4Rv (5’-CCTCCGCTTATTGATATGC-3’) primers targeting the conserved sequences of 18S and 28S rDNAs [39]. A clinical isolate of Candida albicans was included as a control and intracellular fungus was identified by Sanger sequencing of PCR amplicon.

In order to visualise the intracellular bacteria or fungi, FISH in combination with confocal microscopy was performed following a previous protocol [40]. Briefly, 1 ml amoebal cells (> 90% trophozoites) grown axenically were collected in 1.5 ml eppendorf tubes and centrifuged for 5 min at 3,000 g to harvest trophozoites. The cell pellet was washed twice with 1X Page’s saline and 30µL of amoebic suspension was transferred on poly-l-lysine coated slides (Thermo Scientific, Braunschweig, Germany) and left for 30 min at ambient temperature. The adherent cells were fixed by applying 30µL of freshly prepared 4% formaldehyde (buffered, pH 6.9) for 25 min. The attached amoebal cells were washed with 1X PBS, dehydrated in increasing ethanol concentrations (50%, 80%, and 96%, 3 min each) and air-dried. Intracellular bacteria were examined by hybridization using Cy3 labelled bacterial-domain specific probe EUB338 [41], and fungi using the fungus-specific probe PF2 conjugated with Hex and a Cy5 labelled EUK516 probe (Table 1) [42] for eukaryotic 18S rRNA (Biomers, Ulm, Germany). Aliquots (1µL of 50 ng/µL) of each probe were mixed with 9µL of hybridization buffer (20 mM Tris–HCl, pH 7.1, 900 mM NaCl, and 20% v/v formamide, 0.01% SDS) and added to the fixed amoebal cells on slides. Hybridization was carried out for at least 90 min at 46 °C in the dark after which slides were rinsed with 20µL of pre-warmed (48 °C) buffer (180 mM NaCl, 20 mM Tris/HCl, pH 7.2, and 0.01% SDS). The slides were then covered with 200µL buffer and a washing step was performed at 48 °C for 25 min. All slides were quickly immersed in ice-cold MilliQ water, air dried, and were mounted using Prolong Diamond Antifade with DAPI (Thermo Fisher Scientific), then mounted slides were left overnight to cure at room temperature in the dark before imaging. Three independent assays were performed and at least 30 amoebal host cells were visualized under Olympus FV1200 confocal laser scanning microscope in Katharina Gaus Light Microscopy Facility of UNSW and FISH images were analysed in ImageJ [43].

Demographic and clinical features of all 13 AK patients were retrieved in a customised datasheet from hospital records and were reviewed for the following features: initial and final visual acuity (VA); reported symptoms; duration, size and characteristics of infiltrates and epithelial defects (some measurements were transcribed from clinical images); prior treatment (defined as medicines used before the first presentation of patients at LVPEI); clinical treatment regimens and duration under care at LVPEI; need for surgery; treatment outcome; patients’ occupation; and reported risk factors for AK [45].

Identification of the Acanthamoeba strains and their growth at different temperatures are shown in Table 2. Seven isolates (53.8%) grew at 40 °C and 42 °C, demonstrating the thermotolerance of certain strains that cause AK. Figure 2 shows the trophozoites (1A and 1B) and cysts (1C-1E) of an Acanthamoeba isolate (L-2483/20). Figure 3 shows the agarose gel image of the PCR products of the DF3 region of amoebal 18S rDNA gene and the bacterial 16S rRNA. All amoebae were confirmed as being Acanthamoeba, of which eight (61.5%, 8/13) harboured intracellular bacteria and one isolate (L-2429/20) contained fungus. No bacteria grew from the PYG aliquots cultured on trypticase soy agar, suggesting that any bacteria identified in FISH experiments were very likely to be located intracellularly.

The partial nucleotide sequence of amoebal 18S rDNA DF3 region was aligned using ClustalW algorithm and showed the highest inter-strain nucleotide sequence variation (Additional file 1: Fig. S2). The neighbour-joining phylogenetic analysis of 13 isolates formed two major clades, with most isolates (12/13, 92.3%) belonging to genotype T4 (L-552/20, L-1133/20, L-1257/20, L-2482/20, L-2483/20, L-2487/20: A. culbertsoni; L-579/20, L-1326/20: A. polyphaga; L-1137/20, L-2429/20: A. triangularis, and L-604/20, L-1765/20: Acanthamoeba spp. T4) and one isolate (L-2391/20) being genotype T12 (A. healyi; Table 2 and Additional file 1: Table S2). Along with cyst morphology, the sequence producing the highest alignment (% identity) with sequences in NCBI of existing species and genotypes of Acanthamoeba was assigned to each isolate. For two isolates (L-604/20 and L-1765/20), cyst morphology was not obvious to a particular species of Acanthamoeba but the NCBI nucleotide blasting, and phylogenetic analysis confirmed both strains belonged to genotype T4. Genotype T12 was detected from a farmer from Maharashtra with no history of ocular trauma; two other cases from the same state were infected with genotype T4. The remaining ten isolates of T4 variants were recovered from AK patients from Andhra Pradesh, Rajasthan, Telangana, Tripura, and Uttar Pradesh (Additional file 1: Map S1). In the predominant T4 clade, the isolates formed three sub-clusters, T4A (1 isolate), T4B (10 isolates), and T4F (1 isolate). The nucleotide sequence of all 13 isolates have been deposited at GenBank under accession numbers OK042094 to OK042106 (Table 2). Of the genotype T4B, 6 patients were infected with A. culbertsoni, 2 with A. polyphaga, and one each with A. triangularis or Acanthamoeba spp. One patient was infected with T12 A. healyi and another patient (L-1765/20) was with T4A Acanthamoeba spp. (Fig. 4).

Among the 13 Acanthamoeba isolates, 9 (69.2%) were positive for intracellular microbes with eight harbouring bacteria (Fig. 3) and one isolate (L-2429/20) harbouring the fungus Malassezia restricta (Additional file 1: Fig. S3). For the eight isolates harbouring bacteria, rod-shaped intracellular bacteria were observed throughout the cytoplasm, and the bacteria were present in all Acanthamoeba cells in the population observed. No extracellular bacteria were observed, as expected due to the protocol. M. restricta was seen as oval-shaped green cells within Acanthamoeba isolate (Fig. 5).

The mean symptom duration of AK patients infected by Acanthamoeba with intracellular bacteria or fungus was 43.0 ± 44.3 days compared to 16.5 ± 5.1 days of patients infected with endosymbiont free Acanthamoeba (Table 3). Among those nine AK cases, six had a history of either ocular trauma (n = 4) or contact lens wear (n = 2) and three were farmers. A severe form of AK with stromal infiltrate (88.9%, 8/9) and hypopyon (55.6%, 5/9) was noted in the presence of amoebal endosymbionts and five cases had both stromal infiltrate and hypopyon. However, these differences in stromal infiltrate (p = 0.2) and hypopyon (p = 1.0) were not significantly more common in AK patients with amoebal endosymbionts compared to those without the endosymbionts. Interestingly, a longer medical treatment duration (median, IQR: 75.0, 43.5–202.5 vs. 30.0, 17.0–91.8 days) was observed in patients without Acanthamoeba endosymbionts (p = 0.2). In the presence of endosymbionts, the proportion of patients with a healing ulcer (66.6%) or an improvement in final VA (44.4%) were lower compared to those without endosymbionts (75% and 50%, respectively) (p > 0.05). Two patients (25%, 2/8) infected by Acanthamoeba with bacterial endosymbionts had received antibiotics along with antiamoebic drugs and the ulcers of both cases were resolved after medication.

Most of the AK patients were farmers (5/13) or students (5/13) and the reported risk factors associated with AK were ocular trauma (4 cases; 30.8%; 95% CI = 9.1—61.4) and contact lens wear (2 cases; 15.4%; 95% CI = 1.9—45.5). Unilateral AK was reported in 12 (92.3%) cases. The majority of AK patients had decreased vision (84.6%) as a major symptom followed by eye pain (69.2%), redness (69.2%), watering (53.8%), and white spot on the cornea (15.4%). A typical AK characteristic ring infiltrate (> 4 mm) was observed in 3 (23.1%) cases. Epithelial defect was noted in 6 cases (46.2%; 6.2 ± 1.7 mm), stromal infiltrates in 10 cases (76.9%; 5.0 ± 2.2 mm), and hypopyon in 6 cases (46.1%; 1.12 ± 0.6 mm). The median duration of symptoms onset was 20 days (IQR = 15—30) and the final visual acuity was not improved by ≥ 2 lines among patients with farming background (5/5, 100%), age > 32 years (4/6, 80%) and patients showing AK symptoms more than 20 days (3/6, 50%) (Additional file 1: Table S3). PHMB and chlorohexidine were the most common treatments in this study (11 cases; 84.6%). Three cases received antibiotics, one case antiviral and one case antifungal as supportive therapy (38.5%); for these there was no improvement (p > 0.05) in BCVA at final presentation compared to cases treated only with PHMB and chlorohexidine. Overall, the median duration of medical treatment was 38 days (IQR = 23—90). Of six cases with surgical treatment, 4 cases had therapeutic penetrating keratoplasty (TPK, n = 4), including one case that had an amniotic membrane transplant (Table 4). Of the remaining two cases, one had photodynamic antimicrobial therapy with rose bengal (RB-PDAT) and one underwent evisceration. Among 6 patients who had to undergo ocular surgery, 66.7% (4/6) were infected by Acanthamoeba strains with intracellular bacteria.