Molecular characterisation and potential pathogenicity analysis of Acanthamoeba isolated from recreational lakes in Peninsular Malaysia

The opportunistic parasite Acanthamoeba is classified as a cosmopolitan free-living amoeba (FLA) and inhabits various environments, such as freshwater accumulations, damp and wet soil, sewage, drinking water, cooling towers and contact lens storage containers (Fanselow et al. 2021). Acanthamoeba species are distinguished at the genus level based on their varying trophozoite and cyst characteristics, particularly their double-walled cysts morphology (Sente et al. 2016). Initially, the morphology of Acanthamoeba species comprises three distinct classes (I, II and III) with over 25 nominal species (Pussard & Pons 1977). Nevertheless, the proposed morphological-based classification is imprecise in distinguishing Acanthamoeba species (Alves et al. 2000). Moreover, it is challenging to classify a cyst based on its morphological characteristics since the varying morphologies depend on the medium employed (Balczun & Scheid 2017). With advanced research, the species and genotype of Acanthamoeba have been widely identified using molecular approaches, which offer the most precise Acanthamoeba classification (Azizan & Yusof 2021).

The advanced classification of Acanthamoeba using the 18S ribosomal RNA gene (18S rRNA) was pioneered by Stothard et al. (1998). Acanthamoeba strains that showed differences in the 18S rRNA gene region were less than 5% and classified as a single genotype (Siddiqui & Khan 2012). This genotyping technology facilitates the documentation of strain occurrence in the environment or clinical samples and for pathogenicity purposes (Maciver et al. 2013). To date, the Acanthamoeba genus has been classified into 23 genotypes (T1–T23) (Rayes-Battle et al. 2022), with only a few of these genotypes pathogenic (Nagyova et al. 2010). The T4 isolate is the predominant genotype in environmental samples, followed by T1, T2, T3, T5, T6, T10, T12, T15 and T18 (Basher et al. 2018; Kalra et al. 2020).

Acanthamoeba cyst environmental reservoirs can rapidly spread the parasite to humans and other mammals (Lass et al. 2014) since they are found in seawater, thermal waters, surface waters, dam lakes and chlorinated swimming pools. The majority of genotypes known to date have been reported to infect at least one human (Azizan & Yusof 2021). Specifically, the T4, T5, T6, T11 and T15 have been identified as AK-causing genotypes that may cause health risks to humans if exposed to environmental waters contaminated with Acanthamoeba cysts (Lorenzo-Morales et al. 2015). They were also found to cause Acanthamoeba keratitis (AK) in non-contact lens individuals (Juárez et al. 2018). Therefore, a fast and effective method to identify this amoeba from environmental sources is crucial for diagnosing and treating AK.

Meanwhile, numerous scientific approaches and tools have been developed to evaluate water pollutants (Dissmeyer 2000), which determine various parameters, including pH level, turbidity, conductivity and heavy metals. According to Onichandran et al. (2013), these parameters are associated with the prevalence of waterborne parasites and influence their proliferation. In situ measurements of physical parameters, such as total dissolved solids (TDS), dissolved oxygen (DO), temperature (°C), turbidity, salinity (Sal), conductivity and pH level, are essential for indicating nutrient availability and describe the water quality relative to Malaysia's Interim National Water Quality Standard (INWQS) (DOE 2010). In addition, chemical and microbiological parameters, such as total coliform, chemical oxygen demand (COD), sulphate nitrate levels and the presence of Escherichia coli, provide vital information on water contamination, which may influence the occurrence of Acanthamoeba growth. Although Acanthamoeba may not directly utilise the nutrients in the water, other bacteria absorb them as their food source and lead to other health risks (Azlan et al. 2016).

Malaysia has diverse water sources, most of which are exploited for human activities. For instance, the majority of the population utilises recreational water resources for leisurely pursuits. As a result, they are highly exposed to the unintended splashing of Acanthamoeba-contaminated water in the face or bruises, allowing for rapid transmission and potentially contracting Acanthamoeba-causing diseases (Bunsuwansakul et al. 2019). In 2010, Ithoi et al. isolated Acanthamoeba from various swimming pools in Kuala Lumpur and Petaling Jaya, Selangor, Malaysia. Similarly, Onichandran et al. (2013) reported 100% isolation of Acanthamoeba from two artificial recreational lakes near Petaling Jaya. Moreover, the molecular characterisation of Acanthamoeba genotypes has identified for the first time the presence of pathogenic (T3, T4 and T15) and non-pathogenic (T5, T11, T17, T18 and T20) genotypes from recreational hot springs and marine waters in Peninsular Malaysia (Mohd Hussain et al. 2019, 2022).

Mitigating potential infections of Acanthamoeba is crucial, given their ability to thrive in diverse environments and the risk of spreading from anthropogenic activities. However, there is limited information regarding the occurrence of Acanthamoeba in recreational lakes and the existence of potentially pathogenic Acanthamoeba in this country. Therefore, the present study aimed to perform molecular characterisation of Acanthamoeba and evaluate their potential pathogenicity in three recreational lakes in Peninsular Malaysia. Briefly, physiological and microbiological water quality parameters were assessed to aid in interpreting the distribution of Acanthamoeba in the lake water. Positive samples were then grown and underwent advanced molecular analysis for species identification. Physiological tolerance assays were also utilised to determine the pathogenic potential of the isolates. Based on Sustainable Development Goal 6: Clean Water and Sanitation, the collected data could supplement the current baseline information on recreational lakes and implement practical policies for protecting and restoring water-related ecosystems.

A portable multi-parameter (Hanna HI9828, USA) was employed to measure the physical characteristics of the water samples in situ, including TDS (mg/L), water temperature (°C), pH level, DO (mg/L) and oxygen reduction potential (ORP) (mV). An ITS-manufactured portable DM-TU Digimed Turbidity Meter was also applied to measure the turbidity. Meanwhile, additional water samples were collected in 500 mL sterile borosilicate Schott bottles from each sampling location and delivered to the laboratory in chilled containers within 24 h for chemical and microbiological analysis. A Hach spectrophotometer (HACH DR 2800™, USA) was then employed to determine the chemical parameters, including COD and sulphate level, based on the Hach Method (Protocol 480, 385N and 680). Furthermore, Colilert^® and Colilert Quanti-Tray/2000^® (IDEXX, USA) were used to determine the total coliform and E. coli based on the standard most probable number (MPN) technique (Painter et al. 2013), where the water sample was added to liquid broth media in tenfold dilutions and the concentration of viable microorganisms was estimated. The large and small positive wells in the Colilert Quanti-Tray/2000^® were counted and the total number of positive wells was compared to the MPN table.

A weak manifold vacuum system equipped with a sterile bottle-top filter system and fixed with a 0.45 μm pore-size nitrate filter membrane (Gottingen, Germany) was set up to filter each 1 L of water sample at a 1.3 mL/min flow rate. After filtering, the nitrate filter membrane was flipped, cut into four equal parts and laid onto the surface of 1.5% non-nutritive agar (NNA) plates (Sigma-Aldrich A7002, USA), which is composed of Page's Amoeba Saline (PAS) solution lawned (final pH level was adjusted to 6.9) with ultraviolet (UV)-inactivated E. coli. The NNA plates were tightly sealed with Parafilm^® before incubating for 14 days at 30 °C under 85% relative humidity (Ithoi et al. 2010).

The culture plates were observed daily throughout the incubation period using a bright-field microscope to detect visible morphological structures of Acanthamoeba cysts or trophozoites based on taxonomic criteria (Visvesvara & Schuster 2008). Subsequently, these cultures, labelled Acanthamoeba-positive samples, were cloned using a migration technique (Gianinazzi et al. 2009) and cultured at 30 °C for 3–4 days. However, cultures that did not develop morphological amoeba characteristics within 3 weeks of incubation were labelled negative. In addition, a small portion of the agar media (1 cm²) with a minimum number of amoebae was transferred at least one to three times to fresh NNA plates lawned with UV-inactivated E. coli to minimise fungal growth. The transfer process was carried out based on the fungal growth rate. Note that no antibiotics were employed during the isolation process or subsequent cultivation.

Prior to the DNA extraction, the grown Acanthamoeba cells from the cultured clones were harvested by transferring 1 mL of the PAS solution onto the surface of the agar plates. Then, a sterile L-shaped rod was used to scrape off the amoeba from each plate carefully. The amoeba-containing liquid suspension was collected in an Eppendorf tube and centrifuged in a centrifuge at 3,500 rpm for 10 min. The supernatant was collected, while the pellet was used for the DNA extraction using a QIAamp^® DNA mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Once the yield and purity of the DNA extract were measured using a NanoDrop 2000c spectrophotometer (Thermo Scientific, USA), the sample was stored at −20 °C until further analysis.

The polymerase chain reaction (PCR) assay was conducted to amplify the 450 bp of the 18S rRNA fragment Acanthameoba-specific amplimer ASA.S1 of the Acanthamoeba genotype. The genus-specific primers set JDP1 (5′-GGCCCAGATCGTTTACCGTGAA-3′) and JDP2 (5′-TCTCACAAGCTGCTAGGGGAGTCA-3′) were explicitly designed for Acanthamoeba genotyping, as previously described (Schroeder et al. 2001). A 50 μL PCR reaction mixture was prepared by mixing 1 μL of the DNA template (50 ng/μL) with 25 μL of TopTaq Master Mix (2X) (Qiagen, USA), 2 μL of forward and reverse genus-specific primers (each) and 20 μL of DNase-free deionised water.

The PCR protocol was set up as follows: Initial denaturation at 94 °C for 3 min; 35 cycles of denaturation at 94 °C for 30 s, annealing at 57 °C for 1 min and extension at 72 °C for 1 min; and final extension at 72 °C for 10 min. The DNA extract sample was compared to the positive control (Acanthamoeba castellanii ATCC 50492) and negative control (template DNA replaced with distilled water) during each PCR run in triplicate. Subsequently, 10 μL of the PCR product was transferred in a 1.5% agarose gel (Vivantis), followed by a 100 bp DNA ladder (Biolabs, USA) as the DNA marker. Next, the gel electrophoresis was run. Once completed, the DNA fragments were identified by staining the gel with ethidium bromide (EtBr) (0.5 μg/mL) for 10 min.

The sequence analysis was conducted using a BigDye^® Terminator v.3.1 Cycle Sequencing Kit (Thermo Fisher Scientific, USA). Phylogenetic trees were created based on the neighbour-joining distance tree method, which generated 1,000 bootstrapped duplicates. The GenBank database was utilised to allocate the 18S rRNA gene sequences from the Blast searching and orientation using the MEGA software tool, v.11 (Mega Software, Tempe, Arizona, USA) (Tamura et al. 2013). Finally, the largest similarity percentage was analysed to determine the Acanthamoeba species.

The temperature-tolerance and osmo-tolerance tests were performed to evaluate the pathogenic potential of the Acanthamoeba isolates. For the temperature-tolerance test, two sets of culture plates were prepared by soaking a small piece of NNA block containing Acanthamoeba cysts in the centre of each plate. Following cultivation, the initial collection of plates was incubated at 37 °C for 7 days, whereas the second collection of plates was incubated at 42 °C for 7 days. Daily inspections were performed throughout the incubation period for each plate using a bright-field microscope (400× magnification). The experiments were performed in triplicate.

For the osmo-tolerance test, small blocks of NNA containing Acanthamoeba cysts were cut and placed in the centre of fresh 1.5% NNA supplemented with 0.5 M or 1 M mannitol and lawned with E. coli. The plates were then incubated for 7 days at 30 °C to evaluate the growth. The mannitol-free NNA culture plates served as a negative control for comparison purposes. The growth performance at this stage was determined by counting the number of Acanthamoeba cysts or trophozoites grown approximately 20 mm from the centre of each plate and given the following score: 0 (₋), 1–15 (₊), 16–30 (₊₊) and > 30 (₊₊₊) (Landell et al. 2013). The growth after the incubation period was evaluated similarly to the temperature-tolerance test, and the experiment was repeated three times. As a reference, Acanthamoeba castellanii (ATCC 50492) was grown to represent a potentially pathogenic isolate.

Statistical Package for Social Sciences (SPSS) software for Windows, version 28 (SPSS, Chicago, IL, USA) was employed to analyse the collected data. All water samples subjected to clone culture and PCR assay were evaluated descriptively to determine the prevalence rate and distribution of Acanthamoeba genotypes. In addition, Fisher's exact test was utilised to compare the occurrence of Acanthamoeba between the sampling locations. Spearman's rho correlation coefficient (r) was also applied to evaluate the correlation between the physicochemical parameters and the presence of Acanthamoeba. A probability (P) value of less than 0.05 was considered statistically significant.

Correlation between the Acanthamoeba-positive samples based on culture-confirmed method and physicochemical parameters (DO, water temperature, pH value, TDS, ORP, turbidity, COD and sulphate) as well as microbiological parameter (E. coli and total coliform) are shown in Table 2. A significant positive correlation was observed between the presence of Acanthamoeba and oxidation-reduction potential (ORP) (r = 0.638; P < 0.001). Nevertheless, a significant negative correlation was observed between the presence of Acanthamoeba with water temperature (r = −0.754; P < 0.001), pH (r = −0.575; P = 0.002) and DO (r = −0.673; P < 0.001). No significant correlation was observed between the Acanthamoeba-positive with electrical conductivity, TDS, turbidity, COD, sulphate, E. coli and total coliforms.

All of these isolates could be a possible cause of GAE and AK (Table 4). The identified genotype of Acanthamoeba-positive samples in the same sampling sites analysed by culture and PCR-based methods demonstrate that the Acanthamoeba-positive samples may include more than one Acanthamoeba species and genotype. 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 OQ247939–OQ247964.

The response of the Acanthamoeba isolates from recreational lakes towards the temperature-tolerance and osmo-tolerance tests are shown in Table 5. Through these tolerance tests, it was found that 8 (B4, K1, K2, K8, K9, SA1, SA5 and SA10) out of the 26 isolates (30.78%) were resistant at both 37 and 42 °C temperatures including 0.5 and 1 M of mannitol. The findings also revealed that 84.62% (22/26) of the samples tested presented thermo-tolerance at 37 °C. In fact, 30.78% (8/26) of the isolates managed to overcome stressful environment at 42 °C. For the osmo-tolerance test, only one (SA4) and eight (B3, B7, K3, K5, K6, SA3, SA4 and SA6) isolates were not resistant towards 0.5 and 1.0 M of mannitol, respectively. The reference strain (A. castellanii ATCC 50492) used in this study also survived at 42 °C and 1 M of mannitol but with a lower number of cells than was obtained at 37 °C and 0.5 M of mannitol.

Acanthamoeba is one of the vastly distributed FLA in aquatic ecosystems and their potential pathogenicity in humans and animals has attracted the scientific community's interest. In response to the dramatic rise in FLA cases, numerous research has been primarily conducted on environmental samples, including lakes (Ballares et al. 2020; Hagosojos et al. 2020; Aykur & Dagci 2021). To the best of the authors' knowledge, this is the first report on the molecular characterisation and occurrence of Acanthamoeba in three recreational lakes in Peninsular Malaysia, routinely exploited for anthropogenic purposes. Using a combination of culture and PCR-based techniques, this study detected Acanthamoeba in 26 of the 30 water samples (87%). Similar research by Onichandran et al. (2013) and Azlan et al. (2016) used only morphological criteria and discovered a high prevalence of Acanthamoeba (100%) in designated lakes in Selangor, Malaysia. The increased contamination may be attributed to the formation of bacterial biofilm on aquatic vegetation, stones and other sediments, which stimulates a higher proliferation of Acanthamoeba. These protozoa and other bacteria engage in a continuous predator-prey relationship in which Acanthamoeba envelops and consumes innumerable bacterial colonies (Khan 2006; Sente et al. 2016). Conversely, Hagosojos et al. (2020) revealed only a 10% prevalence of Acanthamoeba in Lake Buhi, Philippines, which may be attributed to the use of a distinct technique in the study, where water centrifugation was applied in place of filtration prior to Acanthamoeba cultivation on NNA plates. The result is supported by another previous study by Gabriel et al. (2019), which compared various detection methods of Acanthamoeba in water samples. However, the study revealed that the filtration method was more effective at detecting Acanthamoeba than the centrifugation-based plating assay.

Meanwhile, the present study demonstrated a significant correlation between Acanthamoeba and ORP (mV), water temperature (°C), pH level and DO (mg/L). ORP is a measure of water's purity and ability to decompose contaminants and dead vegetation and animals. As such, a high ORP value implies a large amount of oxygen in the water. Thus, microorganisms that decompose decaying tissue and pollutants can function more efficiently in water with high ORP levels (Horne & Goldman 1994). Consequently, additional bacteria accumulate in the water leading to an increased FLA population, including Acanthamoeba (Sente et al. 2016). Despite the fact that Acanthamoeba is considered a thermophilic amoeba, it tends to be more prevalent in lake waters within a moderately low-temperature tolerance of 29.4–33.7 °C. The occurrence of a high trophozoite count at a low temperature in this study is consistent with previous findings, which reported an optimal growth temperature of approximately 30 °C for Acanthamoeba (Nielsen et al. 2014) as well as in other literature (Sente et al. 2016; Mohd Hussain et al. 2019). On top of that, the pH values measured in this study at 6.9–9.3 is well within the recommended national water quality standard for recreational lakes (Class IIB) (DOE 2019). This finding is similar to the most recent report by Mohd Hussain et al. (2022), where a significant correlation between the presence of Acanthamoeba and the pH level of marine water was observed. The negative correlation in the present study also indicated that lower pH levels of lake water signify a greater likelihood of the presence of Acanthamoeba. Khan (2006) stated that pathogenic Acanthamoeba can thrive in a pH range of 4–12. A change in the DO level would be more complex regarding community interactions. Toxins cause the mortality of other organisms, which would indirectly consume oxygen and affect the survival of oxygen-producing and oxygen-consuming microorganisms. In the current study, the presence of Acanthamoeba increased the DO level (5.2–10.9 mg/L), which exceeds the 5–7 mg/L limit set by the Malaysian national water quality standards for recreational lakes (DOE 2019). Tsai et al. (2020) inferred that such a phenomenon could be related to predation activity (bacteria), indicating poor environmental and recreational lake water quality. In addition, the DO concentration may be altered due to other bacterial activities, such as heterotrophic bacterial respiration, oxygen generation by phototrophic species and other microbial metabolisms (Riedel et al. 2013; Sente et al. 2016).

Based on the molecular approach and NCBI comparison in this study, six of the partially identified Acanthamoeba sequences belonged to the T-genotypes, including T4, T5, T9, T11, T17 and T18, in addition to six identical previously known Acanthamoeba spp. (A. castellanii, A. culbertsoni, A. lenticulata, A. astronyxis, A. hatchetti and A. byersi). Only one Acanthamoeba species was not assigned to a specific species, possibly indicating a unique species to Malaysia. This study also revealed that the reported T4 genotype comprised two species that were 99–100% identical to A. castellanii and A. culbertsoni. This was consistent with previous local studies, which stated that the T4 genotype is frequently detected in various water sources, such as hot springs and marine waters (Mohd Hussain et al. 2019, 2022). The predominance of the T4 Acanthamoeba genotype heightens the risk of infection in humans, which aligns with the frequently reported clinical cases, especially in AK and GAE patients (Booton et al. 2002; Yera et al. 2007). According to Kao et al. (2014) and Ghaderifar et al. (2018), A. castellanii (T4 genotype) is responsible for over 90% of all documented AK cases. Based on this data, it is conceivable that recreational lakes in Malaysia are a significant reservoir of pathogenic Acanthamoeba that can transmit acanthamoebic diseases to humans.

Meanwhile, T17 (six strains) was the second most prevalent genotype among the identified Acanthamoeba strains. However, none of these strains matched any specific species in the NCBI database, suggesting that they may be Malaysia-specific. The presence of the T17 genotype in environmental water samples has also been reported in freshwater sources in Thailand (Nuprasert et al. 2010), hot springs in Malaysia (Mohd Hussain et al. 2019) and swimming pools in Turkey (Değerli et al. 2020). Genotype T17, belonging to the morphological group I, is identified as a non-pathogenic strain but rarely detected in the environment, making it a less research subject (Magliano et al. 2012; Diehl et al. 2021). T18 is another uncommon genotype discovered in Biru Lake and Shah Alam Lake samples at a prevalence rate of 15.4%. Surprisingly, this was the first detection of the T18 genotype in recreational lakes in Malaysia. So far, this genotype has been isolated from sewage and rivers (Possamai et al. 2018), as well as the brain and lungs (Matsui et al. 2018). Moreover, the isolation of the novel T18 genotype from a small number of samples suggests that this genotype may have thrived abundantly in nature. Thus, more novel Acanthamoeba genotypes may exist in the environment and remain to be discovered.

In contrast, T9, T5 and T11 are the genotypes with the lowest frequency in this study. Similarly, Ballares et al. (2020) reported the presence of Acanthamoeba T9 in samples collected from Seven Crater Lakes of Laguna, Philippines. The Tasik Titiwangsa-isolated genotype T9 displayed 92% homology with A. astronyxis. Although A. astronyxis was initially identified as a potential cause of a non-fatal case of human meningitis (Callicott et al. 1968), this is no longer believed to be the case. Nonetheless, research on A. astronyxis is not widely available (Magliano et al. 2012). Hajialilo et al. (2016) have linked T9 to AK infection, making it an emerging pathogen from a public health standpoint. Clusters of A. lenticulata (T5) and A. hatchetti (T11) were detected in a sample from Shah Alam Lake, but only a single isolate was identified in the current study. This confirms the presence of T5 and T11 in diverse water reservoirs, as previously reported (Todd et al. 2015; Dendana et al. 2018; Mohd Hussain et al. 2019; Milanez et al. 2020). Genotype T5 could cause keratitis (Ledee et al. 2009) in immunocompromised individuals and disseminated infection, as reported in a case involving a heart transplant patient (Barete et al. 2007). Therefore, the presence of T5 in recreational lakes may be a significant risk factor for Acanthamoeba infection, particularly in immunocompromised individuals. In the meantime, since the prevalence of AK is still uncommon when associated with T11, additional epidemiological research is being conducted to confirm the association between T11 and clinical cases. This genotype has thus far been isolated from clinical specimens in Iran (Niyyati et al. 2009).

It is difficult to determine the correlation between in vitro evaluation and the actual human pathogenic potential of Acanthamoeba, given its diverse genotypes. The profound heterogeneity has been emphasised by the response of the organism to pathogenicity experiments (Tawfeek et al. 2016). Notably, eight isolates (30.8%) of Acanthamoeba comprising the T4, T9 and T11 genotypes were able to thrive at elevated temperatures (42 °C) and intense osmotic stress (1 M), indicating an indirect association between virulence factors with potential pathogenicity (Khan 2001). A recent study in the Philippines also reported that 47% of samples tested from two major water reservoirs (lakes and rivers) were temperature-tolerant and could flourish at 40 °C (Milanez et al. 2020). The ability of Acanthamoeba to survive under high temperatures and osmolarity makes them more potentially virulent than strains that can only exist at 30 °C (Visvesvara et al. 2007; Wannasan et al. 2009; Todd et al. 2015). However, the in vitro growth of Acanthamoeba isolates under elevated temperatures or intense osmotic stress is correlated to their infectiousness, as this is partially associated with their survival and adaptation ability in host tissues of mammals (Khan & Tareen 2003). It is essential to note that the T4, T9 and T11 genotypes are predominately pathogenic, which may be due to the strains being exposed to environmental stress and may be associated with other factors, such as the synthesis of heat shock proteins (HSP70) (Solgi et al. 2012). Contrarily, Kahraman & Polat (2022) isolated T4B and T4E genotypes from keratitis cases and found that three of the four strains exhibited zero growth at 39–41 °C and 1 M mannitol. Thus, Acanthamoeba tolerance to 39 °C and 1 M mannitol does not indicate its pathogenicity. It is therefore necessary to conduct a broad range of pathogenicity analyses on Acanthamoeba species, such as cytopathic effects to confirm the infection impact of each genotype on humans (Mohd Hussain et al. 2022).

In conclusion, this study demonstrated the presence of unique Acanthamoeba genotypes with varying degrees of pathogenicity in recreational lakes across Malaysia. The culture and PCR-based approaches highlighted that the T4 genotype inhabited environmental waters related to human activities and verified their potential threat to human health. Hence, precise identification of risk factors that may cause contamination should be implemented through proactive programmes designed to prevent future infections. Ultimately, substantial engagement should be exercised to raise awareness among the public regarding the hazards of potentially waterborne diseases and sufficient warning should be placed to inform the population of the significant health risk at these locations.

This study was financially supported by the Fundamental Research Grant Scheme (FRGS/1/2022/SKK0/UITM/02/13), Ministry of Higher Education, Malaysia and Geran Inisiatif Penyeliaan (600-RMC/GIP 5/3[088/2022]), Universiti Teknologi MARA, Malaysia.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.