Entamoeba histolytica clones presenting transient virulent phenotypes: subsequent fate of forcibly induced virulent clones by hamster’s liver abscess model Authors: Yumiko Saito-Nakano, Shinji Izumiyama, Yasuaki Yanagawa, Seiki Kobayashi, Yuko Umeki, Takashi Makiuchi, Kumiko Nakada-Tsukui, Koji Watanabe Published: 2026 ABSTRACT AbstractBackgroundAmebiasis, which is caused by Entamoeba histolytica, is prevalent worldwide. Clinical presentation varies from asymptomatic infection to life-threatening extraintestinal disease, and the clinical form can change within one infectious episode. It is also known that periodical animal passages are required for maintaining virulence of E. histolytica in vivo model, such as a hamster liver abscess or murine colitis model. However, it remains uncertain whether or for how long virulence induced by environmental stimuli persists during in vitro passage at the clonal level.ResultsWe generated clones from a hamster liver-passaged E. histolytica strain and periodically checked their liver abscess-forming capability during subsequent in vitro passages. We obtained three clones (ALA-1, ALA-5, and ALA-6) that were highly virulent (abscess weight accounting for > 30% of total liver weight in a hamster model). Interestingly, all clones showed reduced virulence, which disappeared completely after 18 months of in vitro passage. Furthermore, the time to loss of virulence varied among the three clones (18, 12, and 6 months for ALA-1, ALA-5, and ALA-6, respectively). Whole-genome sequencing revealed a small number of single nucleotide polymorphisms at different stages in culture, none of which were shared among clones. Transcriptome analysis comparing gene expression between the highly virulent and avirulent states revealed several differentially expressed genes in each clone. However, there was minimal overlap in differentially expressed genes among the three clones, emphasizing the high complexity of virulence expression in E. histolytica.ConclusionsE. histolytica can transiently adapt their phenotypes according to culture conditions at the clone level. Further studies are needed to clarify the underlying mechanisms of this “protean” infectious disease.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13099-026-00850-4.Keywords: Entamoeba histolytica, Virulence, Transcriptome, Whole genome sequencing, Animal model Background Entamoeba histolytica is the causative agent of amebiasis, the second leading cause of parasite-related deaths worldwide [1]. Infection occurs via the fecal–oral route and is mainly found in developing countries, where poor sanitation leads to frequent contamination of food and water with feces. However, recent epidemiological studies have identified an increase in cases of amebiasis as a sexually transmitted infection, even in developed countries [2–4]. Only 10% of amebiasis cases involve symptomatic “invasive disease,” with most infections presenting as asymptomatic “intestinal colonization” [5]. In rare instances, amebiasis leads to serious extraintestinal conditions, such as liver abscesses and fulminant disease [6–8]. Moreover, clinical forms can change within one infectious episode. It is reported that 10%–20% of asymptomatic patients infected with E. histolytica will develop invasive amebiasis within 1 year [6–9]. Conversely, even after the successful treatment of amebic liver abscess with metronidazole, more than half of patients remain asymptomatic cyst carriers [10]. E. histolytica can adapt to various gut environments in humans, thereby creating elusive clinical forms of this infectious disease. The severity of amebiasis is affected by three main factors: host factors (the immune response), pathogen factors (the virulence of E. histolytica), and factors related to the intestinal environmental (the intestinal microbiome and metabolome). Moreover, these factors influence each other [11]. In previous studies, we investigated pathogen virulence using a laboratory strain (HM1:IMSS), as well as clinical strains derived from patients who developed amebiasis [12–15]. In one study, a clinical strain isolated directly from a liver abscess showed high virulence, causing enormous liver abscesses in a hamster model. By contrast, strain isolated from asymptomatically infected human patient did not cause liver abscess, even with high-dose inoculation [13]. Notably, the transcriptome profile of the virulent clinical strain was completely different from those of avirulent or laboratory strain, even under the same culture conditions. These results indicate that innate genetic factors in a pathogen are critically important for its virulence. However, it is also known that virulence in E. histolytica can be induced by environmental stimuli, such as through passage in a hamster liver abscess model or murine colitis model [16–19]. Although liver abscess formation can be reinforced by the intermittent liver animal passages, this capability will be diminished unless the stimulus is repeated [13]. To sustain virulence, animal passage every 2 to 3 months is necessary. Our past study revealed that the “virulent” HM1:IMSS strain, which was preserved by regularly passage through the liver, had transcriptomic profiles that differed completely from the avirulent HM1:IMSS strain, which was maintained in vitro without any stimulation. In another study using a gerbil liver abscess model, the virulent strain HM1:IMSS was found to comprise a mix of virulent and avirulent clones [20]. It remains unclear whether high diversity among E. histolytica phenotypes is exclusively determined by genetic (or genomic) heterogeneity. In the present study, we created virulent E. histolytica clones via passage through hamster liver, and we examined whether this forcibly acquired virulence could be maintained with in vitro passage, to investigate the elusive clinical manifestations of amebiasis. Methods E. histolytica strain E. histolytica strain HM-1:IMSS clone 6 was cultured axenically [21, 22] and passaged through hamster liver every 3 months to maintain its virulence (virulent HM1:IMSS, v-HM1) [23]. The liver-passed v-HM1 strain was cultured in BI-YIMDHA medium, a modified version of YIMDHA-S medium in which yeast extract was replaced with BBL Biosate Peptone (BD Biosciences, CA, USA), supplemented with live Crithidia fasciculata at 35.5 °C [24]. HM1:IMSS was originally isolated from a patient with severe amebic colitis in Mexico City in 1967 [25]. One cloned line, HM1:IMSS clone 6 [26], was obtained from Dr. L. Diamond in 1991 and has been maintained through in vitro passage and in vivo challenge in our laboratory [23] via Keio University [27]. Experimental animals Female Syrian hamsters (Mesocricetus auratus) weighing approximately 70 g and aged 5 weeks were purchased from Japan SLC Inc. (Shizuoka, Japan). Hamsters were anesthetized with isoflurane, the peritoneal cavity was opened via surgical laparotomy, and 1 × 106 trophozoites were injected into the left lobe of the liver. Hamsters were killed 6 days post-infection, and both the whole liver and liver abscesses were weighed [13, 23]. Cloning of v-HM1 The recovery of v-HM1 from hamster liver abscesses was performed as previously [28]. The monoxenic v-HM1 strain, recovered from a liver abscess, was cloned by limiting dilution within 2 weeks of recovery, as previously described [29]. Amoeba dilutions at a concentration of 2.5 trophozoites/mL were prepared in BI-YIMDHA medium supplemented with C. fasciculata, and 280 µL aliquots were dispensed into each well of a 96-well plate. After a 2-hour incubation in an anaerobic chamber using Anaerocult (Merck, Germany), a single trophozoite per well was observed microscopically. One trophozoite was then transferred to a 13 × 100 mm Pyrex screw-cap culture tube (Corning, NY, USA). Three weeks after cloning, five clones were subjected to an in vivo challenge to assess virulence. After selecting three virulent clones, we defined the start of in vitro culture as month 0, corresponding to the date on which liver abscess formation was confirmed. Whole-genome sequencing Genomic DNA was extracted from cultures of the three virulent clones described above at various time points between 0 and 24 months after these were established in vitro. DNA was isolated from 1 × 107 trophozoites using a Blood & Cell Culture DNA Maxi kit (Qiagen, Hilden, Germany), following a previously described procedure [14]. A DNA-seq library was constructed using the NEBNext Ultra DNA Library Prep Kit for Illumina. This was then sequenced using the Illumina NovaSeq 6000 platform (Rhelixa Co., Japan). Read mapping and single nucleotide polymorphism (SNP) analysis SNP calling was performed as previously reported [29], with some modifications using E. histolytica HM-1:IMSS as the reference genome (https://amoebadb.org/amoeba/) and a gene model provided by Dr. Hon [30]. CLC Genomics Workbench software (Qiagen, Hilden, Germany) was used for quality assessment, trimming, subsampling, mapping, and SNP analysis of each sample. DNA read mapping was performed using a slightly different setting: length fraction 0.8. SNP detection was performed using the Basic Variant Detection tool with the following modified program settings: ploidy = 4, ignore broken pairs = no, minimum coverage = 50, minimum count = 10, minimum frequency = 20%, and relative read direction filter = no. Based on this setting, more than 5000 SNPs, including pseudo-positive and detection errors, were detected after filtering (reference allele = no). Most pseudo-positives at month 0 were excluded using the Filter Against Control Reads program with other mappings at month 0. The pseudo-positives at months 12 and 24 were excluded using the program with the mappings of the same clone at month 0. Finally, SNP candidates remaining after filtration were manually inspected. RNA sequencing (RNA-seq) Total RNA was extracted from cultures of the three virulent clones described previously, at various time points between 0 and 24 months after they were established in vitro. In total, 2 × 106 trophozoites were pelleted by centrifugation, lysed, and total RNA was purified using TRIZOL reagent (Thermo Fisher Scientific, MA, USA), according to the manufacturer’s instructions. Total RNA was treated with DNase I (Invitrogen, MA, USA) to remove contaminating genomic DNA. An RNA-seq library was constructed using the NEBNext Poly(A) mRNA Magnetic Isolation Module and NEBNext Ultra II Directional RNA Library Prep Kit. This was then sequenced using the Illumina NovaSeq 6000, and 150 bp paired-end reads were analyzed by Rhelixa Co. (Japan). Bioinformatic analysis of RNA-seq data Bioinformatic analysis was conducted following the established workflow for clinical E. histolytica strains, with specific modifications outlined below. Raw RNA-seq reads were processed using the CLC Genomics Workbench for quality trimming and mapping. Reads were mapped against the E. histolytica HM-1:IMSS reference genome (AmoebaDB v.1.7) using the gene model provided by Hon et al. [30]. To normalize the data and calculate fold changes between time points, the Trimmed Mean of M-values (TMM) method was applied. We performed principal component analysis (PCA) to visualize global gene expression variability among the clones (ALA-1, ALA-5, and ALA-6) and across serial in vitro passages. Hierarchical clustering based on Spearman’s rank correlation was used to generate heatmaps comparing the transcriptomic profiles of the virulent early-phase and avirulent late-phase states. Differentially expressed genes (DEGs) were identified for each clone using volcano plots, defined using the threshold of an adjusted p-value < 0.05 and log 2-fold change. For this analysis, time points were categorized into virulent and avirulent states based on the specific attenuation speed of each clone. ALA-1: 0 and 3 months versus 18 and 24 months. ALA-1: 0 and 3 months versus 18 and 24 months. ALA-5: 0, 3, and 6 months versus 12, 18, and 24 months. ALA-5: 0, 3, and 6 months versus 12, 18, and 24 months. ALA-6: 0 months versus 6, 12, 18, and 24 months post-treatment. ALA-6: 0 months versus 6, 12, 18, and 24 months post-treatment. The commonality and uniqueness of DEGs across the three independent clones were further analyzed and visualized using a Venn diagram to identify the shared molecular signatures of attenuation. Statistical analysis The pathogenicity of clones was assessed according to the liver abscess size, classified into three categories, as follows: highly virulent (abscess size > 30% of total liver weight), intermediately virulent (10%–30%), and avirulent (no abscess or < 10%) [13, 23]. Differences in virulence between E. histolytica at different time points of in vitro passage were determined using uncorrected Fisher’s least significant difference. Statistical analyses were conducted using GraphPad Prism (GraphPad Software, La Jolla, CA, USA). Results Establishment of virulent clones from liver-passed strain (v-HM1) Genomic mutations can accumulate with extensive in vitro cultivation [12, 29]. The pathogenic strain (v-HM1) has been continuously cultured and has undergone periodic hamster liver passage since the 1990 s [23, 27]. Thus, v-HM1 may contain heterogenetic clones owing to accumulated mutations. For the present study, E. histolytica clones were newly isolated from the v-HM1 strain using limiting dilution to avoid heterogeneity. Six clones (clones 1 to 6) were obtained initially from the v-HM1 strain (Fig. 1). One of these six clones (clone 4) was excluded from further analysis owing to its slow growth in vitro. The remaining five clones were inoculated into hamster livers to assess their virulence. On day 6, hamsters inoculated with clones 1, 2, 5, and 6 developed large abscesses accounting for more than 30% of the total liver weight at the time the animal was killed. Hamsters inoculated with clone 3 developed only small liver abscesses (Fig. 2, Supplemental Fig. 1); thus, clone 3 was excluded owing to loss of virulence. Clone 2 failed to recover from frozen stock. We therefore selected the remaining three virulent v-HM1 clones (ALA-1, ALA-5, and ALA-6) for further analysis. Schematic representation of the isolation of virulent clones. Trophozoites were recovered from a hamster liver abscess, and an axenic culture was established within 2 weeks of recovery. Cloning was immediately performed using the limiting dilution method. Of the six clones obtained, three clones were excluded from further analysis for reasons shown in the gray boxes. These three virulent clones were designated ALA-1, ALA-5, and ALA-6 and were used in this study Attenuation of the three virulent clones at different stages during in vitro culture. Virulent clones designated ALA-1 (A), ALA-5 (B), and ALA-6 (C) were surgically injected into hamster livers, and the entire liver was recovered on day 6. Left panels show representative images of liver abscesses (white dotted circles) at 0, 6, 12, and 24 months of cultivation. Right graphs show the percentage weight of liver abscesses versus total weight of the liver. Virulent and avirulent states were defined as liver abscesses accounting for > 30% and < 10% of the total liver weight, respectively, shown as pink and blue areas (n = 3 hamsters per group). The X-axis shows the month of in vitro culture. Statistical significance relative to the baseline month 0 was determined using uncorrected Fisher’s least significant difference. * p < 0.05 Pathogenicity of ALA clones during in vitro passage The virulence of three virulent clones (ALA-1, ALA-5, and ALA-6) was assessed at regular intervals over several months to determine the degree of attenuation occurring during repeated in vitro passages without routine liver inoculation (Fig. 2). Day 0 of the cultivation period was set as the day when clone pathogenicity was confirmed by the first liver inoculation after obtaining clone. The pathogenicity of clones was classified into three categories based on liver abscess size, as follows: highly virulent (abscess size > 30% of total liver weight), intermediately virulent (10%–30%), and avirulent (no abscess or < 10%) [13, 23]. At month 0, all three clones were highly virulent. After 3 months of in vitro passage, ALA-6 was intermediately virulent, with a significant reduction in abscess size (p < 0.05, ALA-6–0 m vs. ALA-6–3 m). By 6 months, ALA-6 had become avirulent. ALA-5 became avirulent at 12 months (p < 0.05, ALA-5–0 m vs. ALA-5–12 m). ALA-1 showed avirulence at 18 months (p < 0.05, ALA-1–0 m vs. ALA-1–18 m). According to these results, all three clones had become avirulent by 18 months of in vitro culture, although the time of attenuation varied among the three clones. Changes in the genomic sequence of ALA clones Given their distinct periods of attenuation, it was presumed that each of the three ALA clones had its own unique set of SNPs. Accordingly, we investigated the differences in genomic sequence among the three clones. We used whole-genome sequencing to determine how many mutations had occurred in each of the three clones and to compare differences in their mutation profiles. Compared with many other eukaryotes, the genome of E. histolytica is aneuploid, typically tetraploid [12, 31]. Consequently, SNPs were defined as base substitutions among those clones with an allele frequency greater than 20% in the read-mapping data [29] (see details in the Methods). At month 0, when compared with the reference sequence, a single synonymous SNP was detected in the coding region of ALA-1 (EHI_130930T300A), and another SNP was identified in the non-coding region of ALA-6 (contig DS571556: region 7358) (Table 1). Furthermore, no SNPs were found in ALA-5 at 0 months. These findings indicated that ALA-1, ALA-5, and ALA-6 are distinct clones originating from the same v-HM1 strain. Notably, these SNPs were no longer detected after 24 months of in vitro passage. Instead, three distinct SNPs each emerged in ALA-1 (contig DS571436: regions 16032, 16034, and 16037), ALA-5 (EHI_174090G953T, contig DS571815: regions 2320 and 2326), and ALA-6 (EHI_164190A863G, contig DS571307: region 27349, contig DS571931: region 2001) at month 12 (total of nine SNPs). By month 24, two SNPs in ALA-1, four in ALA-5, and eight in ALA-6 were detected. We identified two SNPs in common between months 12 and 24: EHI_174090G953T in ALA-5 and EHI_164190A863G in ALA-6. Throughout the entire process, SNPs were predominantly found in the non-coding regions across all three clones (Supplemental Fig. 2 A). The attenuation process of liver abscess formation was not linked to any particular SNPs, as there were no overlapping SNPs among the three clones. In other words, those SNPs were unstable and prone to loss during long-term in vitro maintenance. The role of three independent SNPs with higher allele frequencies (more than 40%, presumed to include two of four alleles) at month 24 was unknown (Table 1, indicated as ##). Table 1List of SNPs identified from ALA clonesClone nameContigRegionReferenceAlleleFrequencyGene IDCoding region changeAmino acid changeGene product0 m12 m24 mALA1DS57117042540AT30.1#--EHI_130930T300A-*Purine nucleoside phosphorylaseDS57120867797CG--43.6##--DS5714168882TC--35.6EHI_055390A618G-*hypothetical proteinDS57143616032CA-22.6---16034AT-21.4---16037AT-22.2---ALA5DS5712144235CT--59.1##EHI_164190G938AArg313GlnDNA polymeraseDS57122015715CA-23.728.3EHI_174090G953TGly318Valhypothetical proteinDS57123838918TA--22.9--DS5718152320CT- 23.5- - - 2326GT-20.0- - - DS572641371AT--25.4--ALA6DS57120363537..63538GCCA--27.8--63540CT--24.3--DS5712144310TC- 30.5#42.2##EHI_164190A863GLys288ArgDNA polymeraseDS57126410865GC--28.8--DS5712829751GT--20.6- - 9764AG--20.6- - 9768TC--21.8- - DS57130727349TA-23.4-EHI_088100T818AMet273LysBspA familyDS5715567358AT28- - - - DS5719312001TC-23.331.4#- -* Synonymous mutationSingle nucleotide polymorphisms (SNPs) with 30%–40% allele frequency are indicated by # and those with > 40% by ## List of SNPs identified from ALA clones * Synonymous mutation Single nucleotide polymorphisms (SNPs) with 30%–40% allele frequency are indicated by # and those with > 40% by ## In terms of mutation type, SNPs detected in ALA-1 and ALA-5 from month 0 to month 24 were predominantly transversions (Supplemental Fig. 2B). Unlike the other clones, ALA-6 exhibited both transition and transversion mutations, suggesting that SNP acquisition varied in frequency and pattern across the three clones. These findings suggested that mutations occurred during in vitro passage and ALA clones lost their virulence; however, there was no overlap among ALA clones. Hence, genomic sequence analysis could not be used to determine the SNPs responsible for pathogenicity. DNA methylation in ALA clones We hypothesized that the methylation status of cytosine underlies the transiently variable pathogenicity observed in ALA clones, as several methylated-cytosine residues have been identified in pathogenic v-HM1 and non-pathogenic laboratory reference strains [32, 33]. Therefore, the DNA methylation status of ALA clones was first evaluated at months 0 and 3 using whole-genome bisulfite sequencing (WGBS). In this method, unmethylated cytosines undergo chemical modification to become uracil [32]. However, manual inspection of the read mapping in CLC Genomics Workbench using the Exhaustive mode (for detection of 5-methylated cytosines independently of their nucleotide context) revealed no evidence of methylated sites. A previous report demonstrated the presence of cytosine methylation in the promoter region of the EhHSP100 gene [34]. Although we carefully analyzed sequencing reads from this region, no methylated cytosines were found (Supplemental Fig. 3). Therefore, we did not perform WGBS for ALA clones at later time points. These results indicated that DNA methylation rarely occurs in E. histolytica and is not the mechanism underlying the gradual loss of virulent phenotypes in these clones. Differential expression analysis of virulent and avirulent states. (A) Two-dimensional plot showing the results of principal component analysis for RNA sequencing (RNA-seq) reads of virulent (red) and avirulent (blue) states. (B) Hierarchical cluster analysis of RNA-seq reads from 0, 3, 6, 12, 18, and 24 months. (C–E) Volcano plot showing differentially expressed genes (DEGs) between virulent and avirulent states. Red and blue dots represent the log 2-fold change value. Dotted lines indicate the cutoff lines. (F–G) Venn diagram of shared DEGs with upregulated (F) and downregulated (G) genes in the avirulent state Comparison of the transcriptome between virulent and avirulent states of ALA clones We used RNA-seq to analyze the gene expression profiles of ALA clones in both virulent and avirulent states. The criteria for virulent and avirulent states were defined as the liver abscess size (Fig. 2, red- and blue-highlighted graph areas). The resulting groupings were used for differential expression analyses: ALA-1 (0 and 3 months vs. 18 and 24 months), ALA-5 (0, 3, and 6 months vs. 12, 18, and 24 months), and ALA-6 (0 months vs. 6, 12, 18, and 24 months). PCA was first applied to all time points to illustrate how gene expression profiles differed among the clones (Fig. 3A). The first and second principal components (PC1 and PC2) explained 18.7% and 14.3% of the variability in the RNA-seq dataset, respectively. There was no obvious separation between virulent and avirulent states (Fig. 3A). This result indicated that global transcriptomic variation did not simply correspond to the experimentally defined virulent and avirulent states. Because the three clones showed different attenuation kinetics, PCA may reflect clone-specific transcriptional backgrounds and culture duration-dependent changes in addition to virulence-associated expression patterns. Thus, PCA was interpreted as an exploratory visualization of the overall RNA-seq dataset rather than as evidence for a distinct transcriptomic boundary between virulent and avirulent states. To visually compare gene expression across clones using a heatmap, we applied a statistical filter and performed hierarchical clustering to compare the different points (Fig. 3B). The resulting dendrogram revealed three distinct clusters: virulent, intermediate, and avirulent clusters. Gene expression patterns at earlier time points resembled those at month 0 whereas patterns at later time points, particularly at month 18, more closely resembled those at month 24. For each clone, DEGs distinguishing the virulent from the avirulent state were detected using a volcano plot displaying log2-fold changes with p-values < 0.05 (Fig. 3C and E). In the avirulent state of ALA-1, nine genes showed increased expression whereas another nine genes exhibited decreased expression (Table 2). ALA-5 exhibited increased expression of 12 genes and decreased expression of 10 genes (Table 3). In ALA-6, 16 genes showed increased expression whereas six genes exhibited decreased expression (Table 4; Fig. 3F and G). Protein-coding sequences were mapped against the SMART research tool [35] to predict functional protein domains. Only two DEGs, EHI_059255 and EHI_077750, were common to all three clones as downregulated and upregulated genes in the avirulent state, respectively (Table S1). The downregulated gene, EHI_059255, encodes a protein of unknown function consisting of 132 amino acid residues. Although EHI_059255 is not registered in the NCBI database, its decreased expression over time was confirmed using quantitative real-time polymerase chain reaction (PCR; Supplemental Fig. 4). The upregulated gene, EHI_077750, possesses a fungal lipase domain at the C-terminus (Pfam entry 01764, e-value = 1.9e-09) [36]. Because there were three homologous genes (EHI_032470, EHI_170950, and EHI_170940) within the genome that shared 85%–92% sequence identity with EHI_077750, specific oligonucleotides for evaluation using quantitative reverse transcription polymerase chain reaction (qRT-PCR) could not be designed. We observed very little commonality in the biological functions of genes among the three clones, except for those related to lipid metabolism, the androgen-inducible gene (AIG) family, and transcriptional regulation (Tables 2, 3 and 4). First, regarding genes related to lipid metabolism, two fungal lipase-like domain-containing proteins (EHI_077750 and EHI_032470), an acid sphingomyelinase-like protein (EHI_040600), and a geranylgeranyl pyrophosphate synthase (EHI_142720) were upregulated in the avirulent state. EHI_077750 was conserved among the three clones, and EHI_032470 was shared between ALA-1 and ALA-6 (Table S1). Second, six AIG family genes were upregulated in the avirulent states of ALA-5 and ALA-6 (Tables 3 and 4). Among them, EHI_126560 and EHI_126550 were shared by two clones (Table S1). Finally, we identified transcriptional regulatory genes, including deoxynucleotide kinase (EHI_034570), DNA polymerase (EHI_164190), mRNA-binding protein (EHI_083120), and Myb-like DNA-binding protein (EHI_092700). None of these transcriptional regulatory genes were conserved among the three clones. These results suggest that each virulent clone uses a distinct gene repertoire during attenuation. Table 2Genes differentially expressed in ALA-1 between virulent and avirulent stagesgene IDGenBank accession numberAverage expression levels at 0,3 m (TPM)Average expression levels at 18,24 m (TPM)ALA1_0,3 m vs. 18,24 m - Fold changeFDR p-valueproductEHI_169670XP_6503830.54168.45249.950.000N-terminal C2 domain proteinEHI_037710XP_0019143140.168.8348.240.000Coiled-coil domain proteinEHI_193690XP_0019142641.5938.3419.330.010TLDc domain-containing proteinEHI_077750XP_6484390.264.9415.600.000Fungal lipase-like domainEHI_142570XP_0019144892.5844.6714.220.003uncharacterized proteinEHI_034570XP_6479741.0514.6311.490.048Deoxynucleoside kinaseEHI_032470XP_6485594.3528.325.710.000Fungal lipase-like domainEHI_014020XP_6550671.9312.625.680.015uncharacterized proteinEHI_164460XP_00191370119.0172.853.400.048Leucine-rich repeat proteinEHI_006885N/A395.08150.12−3.090.048uncharacterized proteinEHI_074080XP_653548559.90180.61−3.720.037uncharacterized proteinEHI_058330XP_651378317.7980.45−4.460.004Galactose-inhibitable lectin small subunitEHI_181610XP_00191403325.432.21−12.820.000Swi5 super familyEHI_078540XP_00191381211.500.80−16.560.015Coiled-coil domain proteinEHI_059255N/A12.540.64−22.190.006uncharacterized proteinEHI_124620XP_65486418.900.22−91.550.000uncharacterized proteinEHI_089770XP_00191414044.610.00−2075.430.048uncharacterized proteinEHI_042870XP_65263225.540.00−3819.850.009Cell surface protease gp63TPM, transcripts per kilobase million; FDR, false discovery rate Genes differentially expressed in ALA-1 between virulent and avirulent stages TPM, transcripts per kilobase million; FDR, false discovery rate Table 3Genes differentially expressed in ALA-5 between virulent and avirulent stagesgene IDGenBank accession numberAverage expression levels at 0,3,6 m (TPM)Average expression levels at 12,18,24 m (TPM)ALA5_0,3,6 m vs. 12,18,24 m - Fold changeFDR p-valueproductEHI_017550XP_0019137670.096.9563.890.001uncharacterized proteinEHI_025990XP_6481890.032.5250.090.046AIG1 family proteinEHI_027010XP_0019143590.094.7544.460.003uncharacterized proteinEHI_006850XP_6544610.194.7822.300.003uncharacterized proteinEHI_164190XP_0019137000.5410.8719.650.000DNA polymerase type BEHI_078520XP_0019138100.183.5718.200.034uncharacterized proteinEHI_126560XP_0019142242.7926.019.340.000AIG1 family proteinEHI_077750XP_6484391.2010.418.630.000Fungal lipase-like domainEHI_142570XP_0019144899.9059.345.860.014uncharacterized proteinEHI_126550XP_6485556.0435.685.860.000AIG1 family proteinEHI_115160XP_6492062.809.473.300.018AIG1 family proteinEHI_040600XP_6514574.4914.713.230.006Acid sphingomyelinase-like phosphodiesteraseEHI_067210XP_65266125.719.30−2.780.014RNA recognition motif domain containing proteinEHI_030850XP_6485645.681.66−3.360.034TLDc domain-containing proteinEHI_081750XP_0019138919.542.74−3.460.003Leucine-rich repeat proteinEHI_059255N/A14.672.78−5.220.001uncharacterized proteinEHI_095310XP_6492139.851.49−6.330.018uncharacterized proteinEHI_092110XP_65456331.974.60−6.820.000Proteolipid membrane potential modulatorEHI_127220XP_65217113.881.37−9.910.000SynapsinEHI_067200XP_6526626.440.48−13.390.000uncharacterized proteinEHI_151890XP_65739019.400.71−26.140.000EF-hand calcium-binding domain containing proteinEHI_185260XP_6561716.740.13−52.550.000Protein phosphatase 2 A homologuesTPM, transcripts per kilobase million; FDR, false discovery rate Genes differentially expressed in ALA-5 between virulent and avirulent stages TPM, transcripts per kilobase million; FDR, false discovery rate Table 4Genes differentially expressed in ALA-6 between virulent and avirulent stagesgene IDGenBank Accession NumberExpression level at 0 m (TPM)Average expression levels at 6,12,18,24 m (TPM)ALA6_0m vs. 6,12,18,24 m - Fold changeFDR p-valueProductEHI_077750XP_6484390.4625.8255.894.8E-10Fungal lipase-like domainEHI_083120XP_6480840.4622.1347.583.3E-04Nonsense-mediated mRNA decay protein 3EHI_126560XP_0019142240.5421.5336.621.2E-02AIG1 family proteinEHI_176800XP_6498390.166.3235.732.4E-03Zinc finger proteinEHI_072740XP_0019142750.8024.6329.381.3E-07Coiled-coil domain proteinEHI_163240XP_64796960.301307.4621.257.6E-17Dolichol kinaseEHI_144390XP_6488871.0916.8414.949.1E-05AIG1 family proteinEHI_092700XP_6512390.9714.5414.284.0E-03myb-like DNA-binding domain containing proteinEHI_062960XP_0629604.0848.7811.452.4E-03Signal peptide, C-terminal tansmemebrane domainEHI_144490XP_65596011.08102.429.368.3E-05AIG1 family proteinEHI_144280XP_6488901.019.408.941.8E-02uncharacterized proteinEHI_126550XP_6485553.9533.078.033.1E-03AIG1 family proteinEHI_032470XP_6485596.2045.507.294.0E-03Fungal lipase-like domainEHI_024490XP_65097947.02259.155.392.9E-04Dolichol kinaseEHI_142720XP_00191390029.58130.754.323.2E-02Geranylgeranyl pyrophosphate synthaseEHI_131880XP_65261610.9547.014.173.2E-02Acyl-coA synthetaseEHI_162570XP_65174545.5210.44−4.361.2E-02Ornithine cyclodeaminaseEHI_059255N/A28.242.64−10.784.3E-02uncharacterized proteinEHI_062990XP_00191401514.020.86−16.251.6E-02Deoxynucleoside kinaseEHI_142190XP_00191349088.855.24−16.642.8E-03WD40 repeat protein 2EHI_095840XP_64824713.900.04−240.892.5E-05DNA-damage inducible proteinEHI_159570XP_65024126.170.00−1346.441.6E-03Mob1/phocein family proteinTPM, transcripts per kilobase million; FDR, false discovery rate Genes differentially expressed in ALA-6 between virulent and avirulent stages TPM, transcripts per kilobase million; FDR, false discovery rate Components of virulent E. histolytica strains. Three types of clone comprising genetically determined and undetermined phenotypes (A). E. histolytica strains contain all clone types and may exhibit transient phenotypes in the virulent state (B) Discussion The main objective of this study was to demonstrate transient pathogenicity at the clonal level. For this purpose, virulent clones (ALA clones) were isolated from a polyclonal v-HM1 strain (Fig. 1). We obtained several virulent clones from v-HM1, and another avirulent clone was isolated. This was consistent with a previous study, which stated that virulent HM1 is an aggregate of various phenotypic clones [20]. Whole-genome sequencing analysis confirmed that the sequences of three virulent clones (ALA-1, ALA-5, and ALA-6) were distinct. Thereafter, these clones were separately maintained in vitro without liver passage for up to 24 months. Surprisingly, all clones steadily diminished in virulence with animal passage, and the lost pathogenicity was never recovered during the follow-up period (Fig. 2). Interestingly, ALA-6 rapidly lost its virulence whereas ALA-1 and ALA-5 maintained their virulence for a relatively longer period. Our observation was in contrast to that of a previous study using a single virulent clone that maintained virulence for more than 5 years in vitro [20]. Maintenance of virulence was assessed using “E. histolytica clone” in the previous and present studies; therefore, both results should be valid. This strongly indicates that a single mechanism cannot explain the transient phenotypic changes in E. histolytica at a polyclonal level (Fig. 4A and B). One mechanism proposed in a past study is that E. histolytica culture strains comprise pathogenic and non-pathogenic clonal groups, which induces shifts in the population composition with alternating in vitro/in vivo passages. Virulent clones were selected after liver passage whereas avirulent clones are expanded during in vitro passages. Another possible mechanism according to the present study findings is that E. histolytica clones can transiently adapt their phenotype to culture conditions. To elucidate these mechanisms, approaches such as analyzing genomic data at single-cell level are required. In the present study, the interpretation of transcriptomes for ALA clones was extremely difficult, which might be owing to the high complexity involved with expressing the virulent phenotype of E. histolytica. First, we compared transcriptomic data from time points during the highly virulent state with those during the avirulent state within the clone; we identified 18, 22, and 22 DEGs from ALA-1, ALA-5, and ALA-6, respectively. The rationale for this grouping was based on the central experimental phenotype of this study, namely, liver abscess-forming capacity, rather than on the absolute duration of in vitro culture. Because the timing of attenuation differed among the three clones, direct comparisons of identical culture periods across clones would not necessarily correspond to the same biological state. We therefore considered that phenotype-based comparison between virulent and avirulent states was the most appropriate strategy to identify transcriptomic features associated with the loss of liver abscess-forming capability. We acknowledge that comparisons between time points immediately before and after a marked reduction in virulence, such as 12 versus 18 months for ALA-1 or ALA-6 versus 12 months for ALA-5, may provide information on clone-specific transcriptional changes occurring around the transition phase. However, such single-interval comparisons would be highly dependent on the attenuation kinetics of each clone and could be strongly affected by transient fluctuation, culture-duration effects, and clone-specific transcriptional noise. Because the primary objective of the present study was not to identify the precise temporal trigger of attenuation in each clone but rather to examine whether forcibly induced virulence is maintained or lost at the clonal level, we retained phenotype-based virulent-versus-avirulent comparison as the principal RNA-seq analysis. However, only two DEGs (EHI_059255: upregulated and EHI_077750: downregulated) overlapped across the three clones. Common transcriptional profiles of virulence were not determined. Comprehensive transcriptomic analyses have been widely applied to identify the responsible genes in previous studies. Duplicated (triplicated or more) RNA-seq data are usually obtained from the clone under each condition, which are compared with those under another condition. In such studies, DEGs in common across duplicated data are identified, many of which are defined as virulent genes in subsequent experiments using gene overexpression/silencing of target genes [13, 20, 37, 38]. However, the responsible genes identified in one study are often not found in other studies. This is usually explained by the genomic heterogeneity of E. histolytica owing to prolonged passages under different conditions in different laboratories. In the present study, all virulent clones were simultaneously obtained from the same strain (v-HM1), followed by in vitro passages under the same conditions. Many DEGs in the virulent state were identified in each clone; however, DEGs rarely overlapped among clones. These results strongly indicated that the virulence of genetically distinct clones diminished in a similar way; however, the mechanism of attenuation was unique to individual clones. In fact, a recent study identified two genes as upregulated genes in the avirulent state (metallopeptidase ehmp8-2 and ehhp127) by comparing gene expression between virulent and avirulent clones. However, the regulation of other genes and related mechanisms is extremely diverse. The authors of that study concluded that the regulation of E. histolytica pathogenicity is achieved via a complex network of molecular mechanisms rather than a single factor [39]. The metallopeptidase ehmp8-2 (EHI_042870) gene was also identified as a downregulated gene in ALA-1 in this study (Table S1) whereas ehmp8-2 was identified as an upregulated gene in a study by Bruchhaus et al.; there are no other reports on EHI_042870 (Table S1). Herein, we also could not identify the molecular mechanism for transient phenotypic modulations in a clone culture. Considered together, other mechanisms, such as phenotypic plasticity, post-transcriptional regulation, and population heterogeneity, could explain the high plasticity of this pathogen. It is important to consider the high complexity of E. histolytica when analyzing the mechanisms underlying its characteristics. The present study has some limitations. First, with short reads, it may be difficult to precisely show genomic sequences and mutations (SNPs) among clones. At the time of our analysis, the structure of the E. histolytica genome and its chromosomes had not yet been fully elucidated. Furthermore, its genome is usually 4 N; however, it frequently changes from 2 N to 8 N [12]. In the present study, SNPs were defined as base substitutions among those clones with an allele frequency greater than 20% because a more sensitive definition for SNPs (e.g., that greater than 10%) could result in a higher frequency of false positives. Owing to the technical difficulty of identifying certain rare SNPs in polyploid genomes, SNPs may be underestimated. Nevertheless, we identified a few SNPs by comparing clones under the assumption of 4 N. The absence of any SNPs that overlapped among the clones suggests that certain genomic mutations are not responsible for changes in the virulence phenotype. Additionally, RNA-seq of clones failed to reveal any common virulence profiles for E. histolytica. No DNA methylation was observed in any clones at any of the tested time points. Thus, genomic analyses, including whole-genome sequencing, WGBS, and RNA-seq, could not identify the underlying mechanisms. Other completely new approaches may be needed to fully understand the flexibility of this pathogen Conclusion E. histolytica may present a transient phenotype in virulence at the clonal level. However, in the present research, we could not identify the underlying mechanisms in genomic analyses, including whole-genome sequencing, WGBS, and RNA-seq. Further studies are needed to elucidate the underlying mechanisms of this “protean” infectious disease. Supporting information Whole-genome bisulfite sequencing (Materials and Methods for Supplemental Fig. 3) The EZ DNA Methylation Gold Kit (Zymo Research, UA) and Accel-NGS Methyl-Seq DNA Library Kit (Swift Biosciences, CA, USA) were used for library construction with purified genomic DNA. The libraries were then sequenced using the Illumina NovaSeq 6000 platform (Rhelixa Co., Japan). Unmethylated lambda phage DNA was included as an internal control during library construction, and the conversion rate was >98%. Supplementary Information Supplementary Material 1: Fig. 1. Loss of virulence in clone 3 Large abscesses were formed with clone 1 (ALA-1) and clone 2, accounting for 50% of the total liver weight. By contrast, clone 3 formed a small abscess, comprising around 5% of total liver weight. Fig. 2. Mutation types detected in three virulent clones. (A) Total number of mutations detected at 0, 12, and 24 months. Mutations were categorized into non-coding and coding regions. Mutations in the coding region were further classified as synonymous substitutions, which do not alter the amino acid sequence, or non-synonymous substitutions, which resulted in amino acid changes. (B) Mutation spectra of the three ALA clones. Complementary mutations, such as A → C and T → G, are combined. SNP, single nucleotide polymorphism. Fig. 3. Loss of DNA methylation sites in whole-genome bisulfite analysis. A representative image of the “Map bisulfite reads to reference” read mapping performed using CLC Genomics Workbench in this whole-genome bisulfite analysis. The EhHsp100 promoter region, which had been reported as a methylation site [34], is shown. The top track, “ALA6_3m_wgbs.,” shows three gray arrows indicating three potential methylation sites detected by the program. As shown in the tip window, each of these three arrows has only one read of a methylated signal (methylated coverage = 1), and these detections appear to be false positives. The second track from the top is the reference sequence for Entamoeba histolytica HM-1:IMSS. The third track shows the read mapping. The color of the mapped reads follows the standard CLC convention: dark or pale blue for paired reads (forward or reverse) and yellow for non-specific matched reads. The highlighted T and A indicate differences from the reference sequence. The highlighted T was converted from a C residue by the bisulfite reaction, and the highlighted A was converted from a G residue on the opposite strand to the C residue. All C residues in the forward read (and all G residues in the reverse read) appeared to have converted to A (and T). According to a report by Agrahari et al. [34], the frequency of methylated C residues within the sequence is 20%–60% in some independent clones. The 11 locations in the genome are identical to the sequence “ATGGA TCCAA ACAAA TGGAC AGATG CAACA GTTCA AATGT TCAAA GAATC ACAAG AAAAA GC” from positions 5801 to 5863 (red frame) in this image. These locations are 100% homologous, and none exhibit methylated C residues. Fig. 4. Verification of downregulated expression of EHI_059255 common to all three clones. Expression of EHI_059255 under axenic cultivation for 0, 6, 12, and 24 months in ALA-1, ALA-5, and ALA-6 clones. Quantitative RT-PCR was performed using mRNA prepared from ALA-1, ALA-5, and ALA-6 with Fast SYBR Green Master Mix (Thermo Fisher Scientific, USA) on a StepOne Plus real-time PCR system (Applied Biosystems, USA). The amount of mRNA for each clone was normalized relative to the amount of RNA polymerase II 15 kDa subunit (EHI_056690). Relative values are shown. Primer sequences were: 5′-ATGAATAACATTCTCGATATTATTAATGAT-3’ and 5’-TTAAATCTCTCCATTTTGCTTGCG-3’(EHI_059255); 5′-gatccaacatatcctaaaacaaca-3′ and 5′-tcaattattttctgacccgtcttc-3′ (EHI_056690, RNA polymerase II). Acknowledgements We thank Eiko Nakasone for her technical assistance. The authors thank the Life Science Support Center, Tokai University, for technical assistance. Author contributions Conceptualization: K. W. Data curation: Y. S.-N, Y. Y, S. I. Formal analysis: Y. S.-N, Y. Y, S. I. Funding acquisition: Y. S.-N, K.W. Investigation: Y. S.-N, S. K, Y. U. Methodology: Y. S.-N, Y. Y, S. I, K. N.-T, T. M, K. W. Project administration: K. W. Software: Y. Y, S. I. Resources: Y. S.-N, S. K, Y. Y. Supervision: K. N.-T, T. M. Visualization: Y. S.-N, Y. Y, S. I. Writing-original draft: Y. S.-N, Y. Y, S. I, K. W. Writing-review & editing: S. K, Y. U, K. N.-T, T. M. Funding This study was conducted using grants for research on emerging and re-emerging infectious diseases from the Japan Agency for Medical Research and Development (AMED, JP23fk0108680h0001 to K.W.; JP24fk0108683 and JP24fk0108680 to Y.S.-N), a grant from JSPS KAKENHI to K. W., a grant from the National Center for Global Health and Medicine (23A2017 to K.W.) and the Shionogi Infectious Disease Research Promotion Foundation (2024S026 to K.W.), and JST, CREST Japan (JPMJCR23B5 to K.N.-T). the Research and Study Project of Tokai University Research Organization (Grant Number 26NP346808 to K.W.). Data availability Datasets supporting our results of animal experiments are available at doi: 10.17632/d7pctcgbk7.1. Also, all raw genome sequencing, RNA-seq, and whole-genome bisulfite sequencing (WGBS) data were deposited in the DNA Data Bank of Japan (DDBJ) under the accession number PRJDB40622. Declarations Ethics approval and consent to participate Animal experiments were approved by the Institutional Animal Care and Use Committee (approval number: 123041-II) and conducted at the AAALAC-accredited National Institute of Infectious Diseases, Japan Institute for Health Security (JIHS), Japan. At the end of the experiments, Syrian hamsters were killed by cervical dislocation under isoflurane anesthesia. This study has adhered to the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Yumiko Saito-Nakano, Shinji Izumiyama and Yasuaki Yanagawa contributed equally to the present study. References Associated Data Supplementary Materials Supplementary Material 1: Fig. 1. Loss of virulence in clone 3 Large abscesses were formed with clone 1 (ALA-1) and clone 2, accounting for 50% of the total liver weight. By contrast, clone 3 formed a small abscess, comprising around 5% of total liver weight. Fig. 2. Mutation types detected in three virulent clones. (A) Total number of mutations detected at 0, 12, and 24 months. Mutations were categorized into non-coding and coding regions. Mutations in the coding region were further classified as synonymous substitutions, which do not alter the amino acid sequence, or non-synonymous substitutions, which resulted in amino acid changes. (B) Mutation spectra of the three ALA clones. Complementary mutations, such as A → C and T → G, are combined. SNP, single nucleotide polymorphism. Fig. 3. Loss of DNA methylation sites in whole-genome bisulfite analysis. A representative image of the “Map bisulfite reads to reference” read mapping performed using CLC Genomics Workbench in this whole-genome bisulfite analysis. The EhHsp100 promoter region, which had been reported as a methylation site [34], is shown. The top track, “ALA6_3m_wgbs.,” shows three gray arrows indicating three potential methylation sites detected by the program. As shown in the tip window, each of these three arrows has only one read of a methylated signal (methylated coverage = 1), and these detections appear to be false positives. The second track from the top is the reference sequence for Entamoeba histolytica HM-1:IMSS. The third track shows the read mapping. The color of the mapped reads follows the standard CLC convention: dark or pale blue for paired reads (forward or reverse) and yellow for non-specific matched reads. The highlighted T and A indicate differences from the reference sequence. The highlighted T was converted from a C residue by the bisulfite reaction, and the highlighted A was converted from a G residue on the opposite strand to the C residue. All C residues in the forward read (and all G residues in the reverse read) appeared to have converted to A (and T). According to a report by Agrahari et al. [34], the frequency of methylated C residues within the sequence is 20%–60% in some independent clones. The 11 locations in the genome are identical to the sequence “ATGGA TCCAA ACAAA TGGAC AGATG CAACA GTTCA AATGT TCAAA GAATC ACAAG AAAAA GC” from positions 5801 to 5863 (red frame) in this image. These locations are 100% homologous, and none exhibit methylated C residues. Fig. 4. Verification of downregulated expression of EHI_059255 common to all three clones. Expression of EHI_059255 under axenic cultivation for 0, 6, 12, and 24 months in ALA-1, ALA-5, and ALA-6 clones. Quantitative RT-PCR was performed using mRNA prepared from ALA-1, ALA-5, and ALA-6 with Fast SYBR Green Master Mix (Thermo Fisher Scientific, USA) on a StepOne Plus real-time PCR system (Applied Biosystems, USA). The amount of mRNA for each clone was normalized relative to the amount of RNA polymerase II 15 kDa subunit (EHI_056690). Relative values are shown. Primer sequences were: 5′-ATGAATAACATTCTCGATATTATTAATGAT-3’ and 5’-TTAAATCTCTCCATTTTGCTTGCG-3’(EHI_059255); 5′-gatccaacatatcctaaaacaaca-3′ and 5′-tcaattattttctgacccgtcttc-3′ (EHI_056690, RNA polymerase II). Data Availability Statement Datasets supporting our results of animal experiments are available at doi: 10.17632/d7pctcgbk7.1. Also, all raw genome sequencing, RNA-seq, and whole-genome bisulfite sequencing (WGBS) data were deposited in the DNA Data Bank of Japan (DDBJ) under the accession number PRJDB40622.