Occurrence and molecular characterisation of Acanthamoeba isolated from recreational hot springs in Malaysia: evidence of pathogenic potential

The genus Acanthamoeba is free-living amoebae, which is ubiquitous in natural and human-made aquatic environments, including hot spring water, tap water, lakes, ponds, freshwater sources and swimming pools (Huang & Hsu 2010). Founded on their cyst characteristics, Acanthamoeba can be classified into three markedly different groups (I, II and III) (Pussard & Pons 1977) and encompassing more than 25 nominal species (Booton et al. 2005). Furthermore, consecutive research investigating molecular and biochemical techniques have highlighted numerous examples of incongruity concerning the genotype or Acanthamoeba grouping (Corsaro & Venditti 2010). As a result, the genus classification and taxonomy of Acanthamoeba have presently been categorised into 21 diverse genotypes that are arranged between T1 and T21 by employing molecular methods based on 18S rRNA nucleotide sequencing (Corsaro et al. 2017). Potentially pathogenic T4 is the most predominant genotype isolated from corneal, skin, nasal, river, hot springs, swimming pools and beaches (Booton et al. 2005; Niyyati et al. 2016; Basher et al. 2018).

There are some diseases like pneumonitis, sinusitis, granulomatous amoebic encephalitis (GAE) and Acanthamoeba keratitis (AK) including disseminated tissue diseases which are caused by Acanthamoeba (Visvesvara & Schuster 2008). Consequently, there has been increasing interest by researchers in studying and identifying harmful and infectious diseases in different species of Acanthamoeba. In 1973, the first reported case of human infection was linked to bathing in non-sanitised water caused by AK (Illingworth & Cook 1998). Likewise, several cases have been linked to using contact lenses while swimming (Lorenzo-Morales et al. 2005). On the other hand, lower respiratory tract infections caused by GAE can also reach the central nervous system transmitted via the blood causing severe infection (Gianinazzi et al. 2009). Notably, the genus of Acanthamoeba can result in a consequential hazard to the well-being of humans, given it can multiply and disseminate as a carrier of infectious diseases such as Legionella, Mycobacterium and Pseudomonas which can withstand or tolerate extreme pH, osmolarity and temperature (Iovieno et al. 2010). In hotter seasons, such as in summer months, the thermotolerance of Acanthamoeba in water is vastly multiplied, which also includes groundwater that is naturally heated resulting in the spread of disease (Khan et al. 2002). Therefore, given the extent and resilience of Acanthamoeba, additional research and tests should be a primary concern and focus for researchers regarding its osmotolerance capability or aptitude.

Although, sadly, given the devastation caused by this disease, it has seldom been documented in Malaysia. The first AK case reported in Malaysia was in 1995 which involved a patient with a long history of wearing contact lenses (Kamel & Norazah 1995). Since the diagnosis of this first case, many AK cases have since been diagnosed indicating a dramatic increase in the aggregate of incidences over the last two decades from using contact lenses (Ghani et al. 2013). Nevertheless, based on the number of reported incidences of the disease caused by Acanthamoeba, it has also been discovered distinctly isolated from other ecological sources such as domestic tap water, seawater, swimming pools, dust from air conditioning systems and drinking water treatment plants (Ithoi et al. 2010; Chan et al. 2011; Richard et al. 2016). For instance, in ecological samples, T4 was frequently isolated in dry soil, dust, wet debris, wet soil and water followed by T5, T15 and T3. In the case of trauma-related AK, a particular risk factor has also been identified from wearing contact lenses where in this case, the patient had washed their eyes using water from an open tank located at a construction site (Mohamed Kamel et al. 2005). Indeed, there are additional factors which can pose significant risks to humans such as sand, soil, stone, etc., which can be lodged in the eyes creating frustration and causing a person to continually rub their eyes possibly incurring some damage which can invariably result in AK if left untreated (Basher et al. 2018).

In Malaysia, hot spring facilities, such as swimming pools, have quickly become a popular form of recreational past-time, given their ability to promote health by improving blood circulation, accelerating the healing process, soothing tired and aching muscles along with special dermatological effects (Yazdi et al. 2015). However, as mentioned earlier, the resilience of Acanthamoeba in tolerating and surviving in warm tepid temperatures in which people bathe is a significant issue and must be detected early to safeguard human health and well-being (Behets et al. 2007). It is important to have the capacity to rapidly detect Acanthamoeba in hot spring facilities where people swim and bathe, as it has become a popular pastime among people in Malaysia. To date, little serious consideration has been given to the potential for large Acanthamoeba infection cases in Malaysia. As a result, there is no molecular epidemiological data on the natural occurrence of Acanthamoeba in hot springs in this country. Hence, this study aims to provide insights into the distribution and genotypes of the potentiality pathogenic species of Acanthamoeba present in five hot springs in Malaysia using culture enrichment, polymerase chain reaction (PCR) amplification and DNA sequencing methods to enable taxonomic identification of the Acanthamoeba species. The pathogenic potential of the isolates will also be assessed employing osmotolerance and thermotolerance tests. To assist with the interpretation regarding the existence of Acanthamoeba, the parameters of physicochemical and microbiological water quality are measured.

Fifty water samples, with temperatures ranging between 32 and 51 °C were initially collected from five hot springs in Malaysia: (i) Sungai Klah, Perak (3°59′ N latitude, 101°23′ E longitude), (ii) Hulu Tamu, Selangor (3°27′ N latitude, 101°41′ E longitude), (iii) Selayang, Selangor (3°15′ N latitude, 101°38′ E longitude), (iv) Bentong, Pahang (3°24′ N latitude, 101°53′ E longitude) and (v) Gadek, Malacca (2°24′ N latitude, 102°14′ E longitude) (Figure 1). The samples were collected between April and September 2018, in four regional areas of Peninsular Malaysia (Northern, East Coast, Central and Southern). The samples were taken from surface water not exceeding a depth of 10 cm. For each respective water sampling location, 1 L of water was poured into a sterile borosilicate Schott bottle, stored at 4 °C before being transported within 24 h to the Integrative Pharmacogenomics Institute, Universiti Teknologi MARA, Malaysia, for subsequent analysis. In total, 10 water samples (10 L) were collected at various sampling points from each of the hot springs.

Next, the physical parameters of the water quality were measured at each sampling location at the time of collection. The water temperature (°C), pH level, total dissolved solids (TDS) (mg/L), dissolved oxygen (DO) (mg/L) and electrical conductivity (μS/m) were measured in situ using a portable multiparameter (Hanna HI9828, USA). The turbidity was measured using a portable DM-TU Digimed Turbidity Meter manufactured by ITS. Additional water samples to examine chemical and microbiological indicators were also collected from each sampling location in 500 mL sterile borosilicate Schott bottles before being transported within 24 h to the laboratory in cooled containers. Testing of the chemical parameters, including the chemical oxygen demand (COD) and sulphate, was carried out in the laboratory employing the Hach method (Protocol 480, 385N and 680) and measured using a Hach spectrophotometer (HACH DR 2800™, USA). For the detection of Escherichia coli (E. coli), Colilert^® and Colilert Quanti-Tray/2000^® (IDEXX, USA) were utilised based on the conventional most probable number (MPN) method (Painter et al. 2013) by estimating the concentration of viable microorganisms in a sample using replicate liquid broth growth in 10-fold dilutions. The principles were applied in Colilert Quanti-Tray/2000^® by counting the large and small positive wells. The total positive wells were then referred to the MPN table to obtain the results.

The membrane filtration method was applied for each 1 L water sample collected utilising a sterile bottle top filter system through a cellulose 0.45 μm pore diameter nitrate filter (Gottingen, Germany), with a weak manifold vacuum system (flow rate of 1.3 mL/min). Following filtration, the membrane was then inverted and sliced into four sections and placed onto 1.5% non-nutritive agar (NNA) plates (Sigma Aldrich A7002, USA) containing Page's amoeba saline (PAS) solution lawn with UV-inactivated E. coli. The final pH of the PAS solution was then adjusted to 6.9. The plates were accordingly sealed with Parafilm^® and incubated upside down at 30 °C, with relative humidity 85% up to 14 days (Ithoi et al. 2010).

During incubation, daily inspection was performed using a bright field microscope until detecting the morphological features of the Acanthamoeba trophozoites or cysts according to taxonomic criteria (Visvesvara & Schuster 2008). Cultures were considered as negative if they were observed to be deficient in morphological amoeba features within 3 weeks. If these amoeba features were detected, clones were produced using the migration method (Gianinazzi et al. 2009). In addition, to prevent fungal contamination, a block of agar (size 1 cm²) consisting of a minimal number of amoeba was moved to the new NNA plates comprising a lawn of UV-inactivated E. coli at least one to three times. On this occasion, the transfer times were determined by considering the fungal growth situation. The clones were then incubated at a temperature of 30 °C for between 3 and 4 days. Additionally, no antibiotics were used during isolation nor during further culturing of the isolates.

The agar plates were covered with 1 mL of PAS solution and the amoeba was carefully scraped from each plate using an L-shaped rod. The liquid containing the amoeba was then collected and transferred to an Eppendorf tube, centrifuged at 2500 rpm for 10 min. The supernatant was then removed, and the pellet was utilised for DNA extraction using a QIAamp^® DNA Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's instruction manual. The DNA yield and purity were then determined using the NanoDrop 2000c spectrophotometer (Thermo Scientific, USA) and stored at −20 °C until further use.

The PCR assay genus-specific primers set JDP1 and JDP2 used in the present study were designed for Acanthamoeba genotyping as previously described (Schroeder et al. 2001). They were established as the following sequences, 5′-GGCCCAGATCGTTTACCGTGAA-3′ and 5′-TCTCAC-AAGCTGCTAGGGGAGTCA-3′, to amplify a 450 bp fragment of the 18S rRNA stretch Acanthamoeba-specific amplimer ASA.S1 of Acanthamoeba genotypes. Next, the PCR solution was prepared with 1 μL of the DNA template (50 ng/μL) along with the PCR mixture in order to establish a total volume of 50 μL. The PCR mixture comprised 25 μL of TopTaq Master Mix (2×) (Qiagen, USA), 2 μL each of the oligonucleotide primers along with 20 μL of DNase-free deionised water. The cycling conditions used a temperature of 94 °C for 3 min for the initial denaturation step, followed by 35 cycles of 30 s at 94 °C for denaturation, 1 min at 57 °C for annealing, 1 min at 72 °C for extension and a final extension at 72 °C for 10 min. Positive control (Acanthamoeba castellanii ATCC 50492), negative DNA control (template DNA replaced with distilled water) and sample DNA were analysed in triplicates during each PCR run. The PCR products of Acanthamoeba were detected using gel electrophoresis on 1.5% agarose gel (Vivantis) which was carried out with 10 μL of the reaction solution. The DNA fragments were then confirmed using ethidium bromide staining (0.5 μg/mL, 10 min). A 100 bp DNA ladder (Biolabs, USA) was used as a DNA size marker.

The sequence analysis was performed using a BigDye^® Terminator v.3.1 Cycle Sequencing Kit (Thermo Fisher Scientific, USA). Gene trees were phylogenetically constructed employing neighbor-joining distance trees that produced 1,000 bootstrapped duplicates. The GenBank database was used to assign the 18S rRNA gene sequences in order for Blast searching and orientation utilising the MEGA software application, v.6.0.6 (Mega Software, Tempe, AZ, USA) (Tamura et al. 2013). The highest similarity percentage was then recorded to identify the species.

In order to assay the pathogenicity potential of the amoebae, two tests were performed. In conducting the first test (thermotolerance), two sets of culture plates were made ready with each plate centrally positioned with a small block of NNA, previously soaked with cysts of Acanthamoeba isolates. Here, the first collection of plates were incubated at a temperature of 37 °C and the second collection of culture plates were incubated for 7 days at a temperature of 42 °C following cultivation. During testing, all plates were examined daily using a bright field microscope (400× magnification). The experiments were repeated in triplicate.

In order to accomplish the osmotolerance test, small agar blocks containing Acanthamoeba cysts were sliced and centrally positioned on a fresh 1.5% NNA containing 0.5 or 1 M of mannitol and overlaid with E. coli. The NNA culture plates without mannitol were used as a negative control. Then, to assess the growth, the plates were incubated for 7 days at a temperature of 30 °C. To determine any evidence of growth or nil growth during this stage, the number of cysts or trophozoites that were detected at around 20 mm from the middle of each plate were calculated and scored as 0 (−), 1–15 (+), 16–30 (++ ) and >30 (+++) (Landell et al. 2013). The experiments were executed three times and the growth after the incubation period was then assessed similarly as for the thermotolerance assay. Acanthamoeba castellanii (ATCC 50492) was used as a reference strain for a potentially pathogenic isolate.

To analyse the data, SPSS software (Statistical Package for Social Sciences) for Windows version 24 (SPSS, Chicago, IL, USA) was used. All water samples subjected to both cultivation and PCR were analysed descriptively in determining the prevalence rates and the distribution of Acanthamoeba genotypes. In comparing the occurrence of Acanthamoeba across the sampling sites, Fisher's exact test was employed. In addition, correlation analysis between the presence of Acanthamoeba and the physicochemical parameters was determined using Spearman's rho correlation coefficient (r). A probability (P) value <0.05 was considered as evidence of statistical significance.

The locations of the five hot spring recreation areas (latitude and longitude) as well as the results of Acanthamoeba detection are presented in Table 1. Among the 50 hot spring water samples, 38 (76%) samples were identified as positive using culture along with direct microscopy. The percentages of samples containing Acanthamoeba was highest in Hulu Tamu (100%), followed by Gadek (90%), Sungai Klah (70%), Bentong (70%) and Selayang (50%). The Acanthamoeba obtained by the culture-confirmed method are able to survive, reproduce and were usually visible within 1 week. Acanthamoeba revealed the presence of double-walled cysts measuring 20–25 μm (Figure 2(a)–2(d)) and trophozoites having flat shape and spine-like structures (Figure 2(e)–2(f)). It is also worth mentioning that according to the Fisher's exact test, there was no significant association between sampling sites and the occurrence of Acanthamoeba (P= 0.074).

Correlation between the Acanthamoeba-positive samples based on the culture-confirmed method and physicochemical parameters (water temperature, pH value, turbidity, DO, TDS, COD and sulphate), as well as microbiological parameter (E. coli) are shown in Table 2. Significant positive correlation was observed between the presence of Acanthamoeba with COD (r = 0.314; P = 0.026), sulphate (r = 0.431; P = 0.002) and E. coli (r = 0.585; P < 0.001). Nevertheless, a significant negative correlation was observed between the presence of Acanthamoeba and water temperature (r = −0.299; P = 0.035). No significant correlation was observed between the Acanthamoeba-positive samples with parameters such as DO, pH value, turbidity and TDS.

Acanthamoeba-positive samples detected by culture were confirmed by PCR-based detection and then sequenced to identify the species. The Acanthamoeba reference strains from NCBI showed high similarity percentage (98–100%) with all 38 of the PCR products. Neighbor-joining analysis inferred the relationships between the 38 PCR products and reference strains from the NCBI GenBank, shown in Figure 3. The most frequently identified Acanthamoeba genotype was T4 (n = 24), followed by T15 (n = 9) and then T3 (n = 2). Genotypes T5, T11 and T17 were detected once, respectively (Table 3). Two Acanthamoeba isolated from Sungai Klah detected in this study were clustered into genotype T3, the same genotype as A. griffini (KF010846). Twenty-four Acanthamoeba isolated from Sungai Klah, Selayang, Hulu Tamu, Bentong and Gadek were listed in genotype T4, similar to the species A. castellanii (KC164230) and A. polyphaga (GU596994). The other two Acanthamoeba that were isolated from Bentong and one from Sungai Klah were identified as genotype T11, T17 and T5, respectively. These isolates showed similar genotypes as A. hatchetti (KC164235), Acanthamoeba sp. (GU808277) and A. lenticulata (KT897270). Finally, nine isolates formed a cluster into genotype T15, similar to the genotype A. jacobsi (GU573858 and KJ446981).

All of these isolates could be a possible cause of GAE and AK. Only one isolate from Bentong was identified as Acanthamoeba sp. since the specific species could not be determined (Table 4). The identified genotype of Acanthamoeba-positive samples in the same sampling sites analysed by culture and PCR-based methods demonstrates that the Acanthamoeba-positive samples may include more than one Acanthamoeba species and genotypes. The present study possibly obtained various identified Acanthamoeba species and genotypes through various analytical methods. The genome sequences of the isolates were submitted to GenBank under the accession numbers MH790980–MH791016.

The response of the Acanthamoeba isolates from recreational hot springs towards the thermo- and osmotolerance assays is shown in Table 5. Through these tolerance tests, it was found that 6 (SD1-SLG9, UTA4-HT14, UTC1-HT20, SKA5-SK35, GB2-GDK44 and GC3-GDK49) out of the 38 isolates (15.79%) were resistant at both 37 and 42 °C temperatures including 0.5 and 1 M of mannitol. The findings also revealed that 100% of the samples tested presented thermotolerance at 37 °C. In fact, 78.94% (30/38) of the isolates managed to overcome stressful environment at 42 °C. For the osmotolerance test, only three isolates (UTA1-HT11, UTB2-HT17 and BTGA3-BTG23) were not resistant towards 0.5 M of mannitol. The reference strain (A. castellanii ATCC 50492) used in this study also survived at 42 °C and 1 M of mannitol but with lower number of cells than it was obtained at 37 °C and 0.5 M of mannitol.

Among the isolated free-living amoebae distributed in the environment, Acanthamoeba is quite common. Similarly, this protozoa is found to be present in high percentages from water environments like hot springs which are popular locations for bathing (Saburi et al. 2017). To the best of our knowledge, this is the first study investigating the occurrence and molecular genotyping of Acanthamoeba in recreational hot springs in Malaysia. Acanthamoeba was detected in 38 of the 50 water samples (76%) that were collected using culture and PCR-based methods. This result was in parallel with a previous study carried out locally by Latiff et al. (2018) also detecting a high prevalence (76%) from selected hot springs in Selangor, Malaysia, using only morphological criteria. Indeed, the higher contamination could be due to the formation of bacterial biofilm on the walls of the hot tub that attract more predatory Acanthamoeba that continually feed on various groups of bacteria (Khan 2006; Bagheri et al. 2010). For instance, several studies have been undertaken globally discovering the range of contamination in geothermally warmed waters ranging between 3.6% and 21.2%, all depending on sample size and site (Amorn et al. 2005; Gianinazzi et al. 2010; Huang & Hsu 2010; Badirzadeh et al. 2011).

The current study has demonstrated a significant correlation of Acanthamoeba with COD (mg/L), E. coli (MPN/100 mL), water temperature (°C) and sulphate (mg/L). High levels of COD in hot spring water will reduce the level of DO and subsequently lead to an anaerobic condition which will, in turn, cause significant harm to higher aquatic life forms, but not E. coli. Moreover, it will contribute to significantly high levels of bacteria in hot spring water as a source of food for Acanthamoeba explaining the reason why the occurrence of Acanthamoeba was associated with E. coli in the current study. Furthermore, even though Acanthamoeba is considered to be thermophilic protozoa, they tend to be more prevalent in hot springs having moderately low-temperature tolerance. This concurs with previous data reported at very low temperature which ultimately leads to high Acanthamoeba trophozoite counts, given that the optimum growth temperature is around 30 °C (Nielsen et al. 2014). Interestingly, Acanthamoeba-positive samples highlighted a significant correlation with sulphate. This study suggests that all five recreational hot springs containing sulphur may present preferable water conditions for this free-living amoeba to flourish and survive. In contrast, limited studies have indicated that the presence or absence of Acanthamoeba did not rely on the physicochemical parameters such as water temperature, pH value and turbidity (Tung et al. 2013; Richard et al. 2016; Ghaderifar et al. 2018).

In the current study, six distinct partial Acanthamoeba sequences belonging to the sequence T-genotype groups T3, T4, T5, T11, T15 and T17 along with six identical (A. griffini, A. polyphaga, A. castellanii, A. hatchetti, A. lenticulata and A. jacobsi) to already know Acanthamoeba sp. sequences from the NCBI comparisons were obtained. Only one Acanthamoeba sp. which was not classified to a particular species level may be unique to Malaysia. The Acanthamoeba strain identified as genotype T4 was reported as the predominant environmental and isolated genotype, identifying it as an important risk to the population in this country. This finding has also been supported in the literature (Niyyati et al. 2009; Solgi et al. 2012; Landell et al. 2013; Shokri et al. 2016). Moreover, Acanthamoeba T4 corresponding to A. castellanii is the primary genotype associated with AK, representing more than 90% of AK cases found in the literature (Kao et al. 2014). Accordingly, this suggests that recreational hot springs in Malaysia may be a significant source of acanthamoebic diseases in humans.

Nevertheless, the second genotype commonly found amid the identified Acanthamoeba strains was T15 with nine strains, which was clustered with A. jacobsi. A. jacobsi being sufficiently isolated from other strains of Acanthamoeba and is allocated to a new sequence type (T15) (Hewett et al. 2003). Acknowledged as a non-keratitis-initiating strain, T15 maintains its bacteriological symbionts many times more compared with other strains of Acanthamoeba during the cultivation or incubation process (Flint et al. 2003). Until recently, clinical isolates of genotype T15 were documented in Italy, representing the linkage between human amoebic keratitis and T15 (Cave et al. 2009; Huang & Hsu 2010).

On the other hand, genotype T3 is less prevalent in the environment compared with T4 and T15. A. griffini is allocated to the T3 sequence type, which is harmful to humans and causes keratitis and infrequently GAE (Hewett et al. 2003). This was also evident from the collected sources of hot spring water in this study, in which A. griffini was found, thus suggesting that A. griffini is a native microorganism that thrives in these hot spring waters. Interestingly, only one isolate was detected in this study confirming the pathogenic nature of T5 in producing keratitis belonging to the T5 genotype which was found in the USA and Greece (Spanakos et al. 2006; Ledee et al. 2009). On the contrary, Huang & Hsu (2010) also reported genotype T5 as a common contaminant found in Taiwan spring recreation areas.

It is also interesting to note that six (15.8%) Acanthamoeba isolates were able to grow at high temperatures (42 °C) along with high osmotic stress (1 M). Likewise, 20% of the hot therapeutic springs found in north-western Iran reveal that some Acanthamoeba isolates which were thermotolerant are able to flourish at temperatures exceeding 40 °C (Solgi et al. 2012). This finding is also supported by that of Rohr et al. (1998), revealing that Acanthamoeba isolates are sensitive to temperatures exceeding 40 °C, although their ability to survive at such extreme temperatures is rare. Previous studies have also shown that Acanthamoeba strains that can tolerate elevated temperatures and osmolarity present the significant pathogenic potential for the well-being of both humans and animals (Visvesvara et al. 2007; Wannasan et al. 2009). Nonetheless, the in vitro development and growth of Acanthamoeba isolates at elevated temperatures or under significant osmotic stress is linked to its infectiousness, given this is partially associated with the ability of the isolate to survive and acclimatise in the host tissues of mammals (Khan & Tareen 2003). Moreover, it is important to mention that the genotypes of the mentioned strains belonging to T3, T15 and T4 are predominantly pathogenic Acanthamoeba which may be the result of the infectiousness of the strains undergoing environmental stress and potentially associated with aspects like the discharge of heat shock proteins (HSP70) (Solgi et al. 2012). However, it is proposed that further research is conducted employing in vivo tests which include cell culture tests or animal assays including the examination of the cytopathic consequences which, in terms of pathogenicity, are the preferred tests to be performed (Todd et al. 2015).

In summary, this study highlights the contamination of recreational hot springs in Malaysia by potentially pathogenic Acanthamoeba genotypes, which can result in Acanthamoeba-related disease. Thus, this study supports the requirement to stringently inspect recreational hot springs used for human activities (i.e. swimming) by posting warning signs at these sites to warn people about the risk associated with bathing in these waters. As a result, further study should be undertaken concerning the dissemination of this type of Acanthamoeba genotype in waters associated with human activities.

This study was financially supported by the Integrative Pharmacogenomics Institute (241910/2017/TSA/13) and Geran Inisiatif Penyeliaan (600-IRMI 5/3/GIP [024/2019]), Universiti Teknologi MARA, Malaysia. The authors thank Mr Muhd Hanis Md Idris for his assistance during sample collection.