Introduction

Chagas disease, caused by Trypanosoma cruzi (T. cruzi) protozoan infection, affects more than 6 million people worldwide, with more than 70 million people at risk of infection [1,2,3]. In the absence of curative treatment, approximately 30% of those chronically infected will present Chagas cardiomyopathy (CCM) as a phenotypic outcome. CCM is a life-threatening condition characterized by myocardial dilatation, conduction abnormalities, and heart failure [4]. The molecular pathophysiology and indicators of CCM progression have been least described among other common forms of dilated cardiomyopathy, such as those related to ischemic/hypoxic stress (mitochondrial dysfunction and impaired calcium sensitivity of cardiomyocytes) or viral infections (viral RNA persistence and subsequent apoptotic loss of cardiomyocyte) [5,6,7,8,9,10,11,12,13,14,15].

The diverse genetic make-up of parasite virulence factors shall evoke unique host responses of variable magnitude [16,17,18,19,20,21], collectively affecting the phenotypic expression of CCM. Thus, the host response to specific T. cruzi genotypes may provide clues to the molecular pathophysiology of CCM progression [22,23,24]. The objective of the current study was to elucidate the transcriptomic signatures of the cardiomyocyte response to infection with genetically discrete T. cruzi strains namely, Colombian, Y, and Tulahuen to explore differential host responses and their relationships with CCM pathogenesis and progression.

Materials and methods

Cell lines and culture conditions

Mouse CH3/10T1/2, clone 8 murine fibroblasts (RRID: CVCL_0190, JCRB0003), obtained from the Japanese Collection of Research Bioresources Cell Bank (JCRB), were cultured in high-glucose Dulbecco’s modified DMEM medium (Fujifilm, Japan) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin (Fujifilm, Osaka, Japan). The cells were seeded in T-25 vent-cap flasks and incubated at 37 °C, under 5% CO2.

Mouse HL-1 cardiomyocytes (RRID: CVCL_0303) were cultured in Claycomb medium (Merck, Darmstadt, Germany) supplemented with 10% fetal bovine serum (Gibco, Maryland, USA), 0.1 mM norepinephrine (Sigma), 0.2 mM L-glutamine (Fujifilm, Osaka, Japan), and 1% penicillin/streptomycin (Fujifilm, Osaka, Japan). Cells were seeded on T-25 vent-cap flasks pre-coated with fibronectin (Sigma, St. Louis, MO, USA) and incubated at 37 °C, under 5% CO2.

T. cruzi epimastigotes of the Colombian, Y, and Tulahuen strains were obtained from NEKKEN Bio-Resource Center (NBRC) and cultured in LIT medium at 28 °C, under 5% CO2. Per strain, 1 × 108 epimastigotes harvested in the exponential growth phase were subjected to metacyclogenesis, following a pre-established protocol [25]. Debris sedimentation was performed by centrifugation at 700 ×g for 5 min followed by trypomastigote collection of the supernatant at 1700 ×g for 10 min. Trypomastigotes were used to infect CH3/10T1/2 cells at a multiplicity of infection (MOI) 1:10. After 5–7 days, successful infection was confirmed by the presence of intracellular amastigotes on light microscopy. Cultured cell-derived trypomastigotes (CCdT), obtained from the supernatant of these stable cultures under the aforementioned centrifugation condition, were used for cardiomyocyte infection experiments.

In vitro cardiomyocyte infection

The 6-well plates were seeded with HL-1 cardiomyocytes as described previously and incubated for 24 h to ensure full attachment and confluency. CCdT of each strain was collected to infect HL-1 cardiomyocytes at an MOI of 10:1. Mouse TNF-α (100 ng/mL; Sigma-Aldrich H8916, Missouri, USA) was added to the culture medium of the positive control wells [19, 21, 26,27,28,29,30,31]. CCdT were allowed to interact with the HL-1 cells for 24 or 48 h depending on the intended condition and were not washed during this interaction period. Non-infected conditions for both 24 and 48 h time points were prepared separately to control for unexpected changes during time course. Each condition was prepared in duplicates.

Assessment of trypanosome infectivity by light microscopy

At 24- and 48-h post-infection (hpi), cells were fixed and stained with Diff-Quick solution (Sysmex, Kobe, Japan). For each cardiomyocyte infection condition, 'percentage of infected cells', as well as the 'intracellular amastigote count per infected cell', were determined by light microscopy to serve as surrogates of the strains infectivity rate. Eight microscopic fields were counted after staining. The results are representative of two separately performed experiments. Pairwise comparisons were performed with the Mann–Whitney test.

Microarray analysis of differentially expressed genes (DEGs)

At 24 and 48 hpi, the culture medium was removed, and cells were gently washed with sterile PBS. Total RNA was isolated using the NucleoSpin RNA kit (Macherey–Nagel, Düren, Germany), according to the manufacturer’s instructions. RNA concentration and purity were assessed using Nanodrop 2000 (ThermoFisher, Wilmington, USA). RNA samples were analyzed using the Clariom S Assay for Mouse (Applied BioSystems, CA, USA) microarray platform to identify DEGs.

Data analysis was done on the Microarray Data Analysis Tool software version 3.2.0.0 (Filgen, Nagoya, Japan). Background signal was defined as any signal intensity values below the median of the signal intensity of a set of negative control probes on the microarray. The threshold for background level was determined as 43 for the infected conditions and 48 for the negative control at 24 hpi; 52 for the infected conditions and 48 for the negative control at 48 hpi. Valid gene probes were the set of probes which surpassed these thresholds of intensity in every infection condition at each time point. Probes not fulfilling the criteria were considered invalid and excluded. Therefore, 11,041 and 11,000 valid gene probes for each respective time point were included for further analyses. Fold change (FC) was defined as the ratio between gene expression level of the infected conditions (or the positive control) and the negative control. P-value (P) was calculated by the ‘Student’s t-test’ to determine the significance of FC of a gene per comparison. Multiple comparison-adjusted P-values were estimated by the Benjamini–Hochberg method using GraphPad Prism version 9.5.1 for Windows (GraphPad Software, Massachusetts, USA). FC ≥ 2 or ≤ 0.5, and P < 0.05 were used as the threshold to define DEGs.

FC of the valid genes shared across all cardiomyocyte infection conditions (total 10,784 genes) were used as input for the principal component analysis (PCA). Analysis and plotting were performed with R 4.1.0 software, and the ‘FactoMineR’ and ‘Factoextra’ R packages [32,33,34]. The principal components 1 and 2, which together described 58.9% of the total variance in the dataset, were selected to plot the PCA results.

Pathway and gene ontology enrichment analysis

Pathway and gene ontology (GO) category analysis was performed using the Microarray Data Analysis Tool software version 3.2.0.0 (Filgen, Nagoya, Japan). Pathways were enriched using the WikiPathways system [35].

Enrichment of both pathways and GO categories was performed using the gene sets (common to all three strains, or selectively shared between Colombian and Y strains) of upregulated and downregulated DEGs as input. The P-value for each pathway and GO category was calculated using the two-tailed Fisher’s exact test. Significantly (Z-score (Z) > 0, P < 0.01) enriched pathways and the top GO terms (up to ten, in the order of lowest P) were selected for discussion. Bar plots depicting GO and pathway enrichment were generated using the ‘GOplot’ R package [36].

Assessment of reactive oxygen species (ROS) intracellular content

The 96-well plates were seeded to a 50% confluency with HL-1 cardiomyocytes as described previously and incubated for 24 h to ensure full attachment. CCdT of T. cruzi Y strain were used to infect HL-1 cardiomyocytes at an MOI of 10:1. Each condition was performed in triplicates. After 24 h, sterile PBS was used to wash non-internalized parasites. Cells were incubated with CellRox green (ThermoFisher, Wilmington, USA) at 10 µM for 30 min, then, gently washed and counterstained with Hoechst dye solution (Abcam, Cambridge, UK) at 4 µM for 15 min. Image acquisition was performed with a BZ-X700 series Keyence microscope (Keyence, Osaka, Japan). Image analyses were performed with the ImageJ software (National Institute of Health, USA) [37]. For each biological replicate, four cell images were randomly chosen to measure their fluorescence intensity for the CellRox green probe, and three random background measurements were also performed. Background fluorescence was subtracted from the cell measurements to calculate the cell fluorescence. Mann–Whitney test was used for group comparisons.

Results

Assessment of infectivity by strain

At 24 hpi, Colombian, Y and Tulahuen infections showed 37.2%, 45.1%, and 17.2% infected cardiomyocytes, respectively (Fig. 1, left). A similar trend persisted at 48 hpi when the infected cardiomyocytes counted 70.5%, 85.2%, and 26.0% for Colombian, Y and Tulahuen infections, respectively (Fig. 1, right). At 24 hpi, each infected cell contained a mean (standard deviation): 1.6 (0.8), 1.6 (0.9), and 1.2 (0.4) intracellular amastigotes for Colombian-, Y-, and Tulahuen-infected conditions, respectively. A similar trend in differential infectivity persisted at 48 hpi, showing intracellular amastigote counts of 2.8 (1.9), 3.3 (1.8), and 2.1 (1.6) per Colombian, Y, and Tulahuen strain-infected cells, respectively. Pairwise comparisons found significant differences between Colombian and Y against Tulahuen strain (P < 0.01), in both 24 and 48 hpi. Distribution and exact P values are depicted in Fig. 2. A representative image from each infection condition is depicted in Additional file 2: Fig. S1.

Fig. 1
figure 1

Difference in cardiomyocyte infectivity among Trypanosoma cruzi infection conditions. Percentage of T. cruzi-infected cells (containing intracellular amastigotes) at 24- and 48-h post-infection. hpi hours post-infection

Full size image
Fig. 2
figure 2

Cardiomyocyte infectivity among different Trypanosoma cruzi infection conditions. Intracellular amastigote counts per infected cell are plotted at A 24- and B 48-h post-infection. Pairwise comparisons were performed with Mann–Whitney test. Black horizontal bars indicate the mean and SD

Full size image

Host transcriptomic remodeling per infecting T. cruzi strain

In the Colombian-, Y-, and Tulahuen 24-h-infected conditions, 1097 (4.9%), 1122 (5.1%), and 806 (3.6%) DEGs were identified. 32 DEGs were identified in the TNF-α-treated control condition (Additional file 2: Fig. S2). The proportions of up- and down-regulated DEGs per cardiomyocyte infection condition are depicted in Fig. 3. Colombian and Y infections led to increased numbers of upregulated DEGs at 24 hpi. A similar trend persisted at 48 hpi but to a lesser extent. At 24 hpi, a significant overlap of upregulated DEGs (438 of 666 (66%) in Colombian-, or of 538 (81%) in Y-infection) was observed between Colombian and Y infection, of which 107 (24%) were also shared with Tulahuen-infection (Fig. 4A). Uniquely in Tulahuen-infection, downregulated genes over-represented the total DEGs at 24 hpi. A large proportion of downregulated DEGs (315 of 431 (73%) in Colombian-, 584 (54%) in Y-, and 684 (46%) in Tulahuen-infection) were shared among all infection conditions, suggesting a common pathway of cardiomyocyte functional deterioration (Fig. 4B). As apparent in the PCA plot, the Colombian- and Y-infection conditions were more closely related in terms of the gene expression pattern (Fig. 5), potentially reflecting the significant overlap of upregulated DEGs between the strains and the distinctive over-representation of downregulated DEG in Tulahuen-infection. At 24 hpi, the three infection conditions showed more discrete host cardiomyocyte gene expression patterns compared at 48 hpi. This indicated that 24 hpi would serve as a more favorable time frame to investigate the uniqueness of the host transcriptional remodeling triggered by each infecting T. cruzi strain. Not only were we more interested in the earlier determinants of CCM progression, but the PCA findings also supported our selection of the 24-hpi stage of the in vitro cardiomyocyte infection model as the main focus of the following analyses.

Fig. 3
figure 3

Remodeling of gene expression profiles in cardiomyocytes infected with three different Trypanosoma cruzi strains at two time points post-infection. Y axes indicate the gene expression level of the non-infected control cardiomyocyte, while X-axes indicate gene expression levels of Trypanosoma cruzi-infected cardiomyocytes. The black dashed line, and upper and lower red dashed lines show the line of identity, and the 0.5- and twofold change thresholds, respectively. Each plot represents a differentially expressed gene (DEG). Blue and pink shaded areas represent down- and up-regulated DEGs, respectively. hpi hours post-infection

Full size image
Fig. 4
figure 4

Identification of differentially expressed genes (DEGs) of the host cardiomyocyte upon Trypanosoma cruzi infection. Numbers of up- (left column) and down-regulated (right column) DEGs at 24- (upper row) and 48-h post-infection (lower row) are shown in the respective Venn diagrams. hpi hours post-infection

Full size image
Fig. 5
figure 5

Principal component analysis plot of the transcriptional profile of the in vitro cardiomyocyte infection models. Principal component analysis plot, based on the fold-change (vs control) values of the 10,784 valid genes. Each dot represents a biological condition at 24- or 48-h post-infection. The principal components 1 and 2 together describe 58.9% of the total variance in the dataset. Dim dimension

Full size image

The common cardiomyocyte response to T. cruzi infection

Of the 107 commonly upregulated genes, eight had FC ≥ 5. The top upregulated genes were Arresting Domain Containing 4 (Arrdc4), thioredoxin interacting protein (Txnip), and growth differentiation factor 15 (Gdf15). (Fig. 6 A).

Fig. 6
figure 6

Top differentially expressed genes of the cardiomyocyte upon Trypanosoma cruzi infection. Extensively A upregulated (fold-change > 5) and B downregulated (fold-change < 0.2) genes of the cardiomyocyte at 24-h post-infection are listed. Δ Gene expression, the fold-change or ratio between gene expression level of the infected cardiomyocytes and the non-infected negative control. Arrdc4 arrestin domain containing 4, Txnip thioredoxin interacting protein, Gdf15 growth differentiation factor 15, Atf3 activating transcription factor 3, Soat2 sterol o-acyltransferase 2, Dusp8 dual specificity phosphatase 8, Chac1 cation transport regulator 1, Gtpbp2 GTP binding protein 2, Capn5 Calpain 5, Dhcr24 24-dehydrocholesterol reductase, Lgals9 lectin galactose binding soluble 9, Acat2 acetyl-coenzyme A acetyltransferase 2, Hmgcs1 3-hydroxy-3-methylglutaryl-Coenzyme A synthase 1, Mvd mevalonate (diphospho) decarboxylase, Tpcn1 two pore channel 1, Medag mesenteric estrogen dependent adipogenesis, Sulf2 sulfatase 2, Fam64a family with sequence similarity 64 member A, Midn midnolin, Ldlr low density lipoprotein receptor, Acat3 acetyl-coenzyme A acetyltransferase 3, Triobp TRIO and F-actin binding protein

Full size image

The 107 upregulated DEGs shared among all three infection conditions enriched pathways related to ‘Hypertrophy model’ (Z = 7.90, P < 0.001), ‘Apoptosis’ (Z = 4.8 P < 0.01), ‘Aminoacid metabolism’ (Z = 4.3, P < 0.01), ‘MAPK signaling pathway’ (Z = 3.9, P < 0.01), ‘ErbB signaling pathway’(Z = 4.9, P < 0.01), ‘Spinal cord injury’ (Z = 4.2, P < 0.01), ‘Selenium metabolism-selenoproteins’ (Z = 4.8, P < 0.01), and ‘Homologous recombination’ (Z = 6.5, P < 0.01) (Fig. 7A). The same gene list was enriched in genes associated with the Biological Process GO terms of ‘Response to reactive oxygen species’ (Z = 8.1, P < 0.001), ‘Response to ionizing radiation’ (Z = 8.2, P < 0.001), ‘Response to oxidative stress’ (Z = 6.7, P < 0.001), ‘Response to hydrogen peroxide’ (Z = 8.1, P < 0.001), ‘Cellular response to stress’ (Z = 5.3, P < 0.001), ‘Primary metabolic process’ (Z = 3.3, P < 0.001), ‘Regulation of metabolic process’ (Z = 3.9, P < 0.001), and ‘Regulation of intracellular protein kinase cascade’ (Z = 5.4, P < 0.001) (Fig. 7C).

Fig. 7
figure 7

Signatures of the common cardiomyocyte response to Trypanosoma cruzi infection. A, B ‘WikiPathways’ and C, D ‘Biological Process gene ontology Terms’ enriched by commonly A, C upregulated and B, D downregulated genes of the host cardiomyocyte at 24-h post-infection are sorted based on P-value. Bar lengths represent enrichment Z-scores

Full size image

Of the 315 commonly downregulated genes, 14 had an FC ≤ 0.2 (Fig. 6B). The 315 downregulated DEGs shared among all three infection conditions enriched pathways related to ‘Cholesterol metabolism’ (Z = 16.8, P < 0.001), ‘Cholesterol biosynthesis’ (Z = 16.1, P < 0.001), ‘Integrin-mediated cell adhesion’ (Z = 4.2, P < 0.01), ‘Apoptosis’ (Z = 4.3, P < 0.01), ‘p53 signaling’ (Z = 4.2, P < 0.01), ‘SREBF and miR33 in cholesterol and lipid homeostasis’ (Z = 5.3, P < 0.01) and ‘Omega-9 FA synthesis’ (Z = 4.8, P < 0.01) (Fig. 7B). The same gene list was enriched in genes associated with the Biological Process GO terms of ‘Sterol biosynthetic process’ (Z = 17.9, P < 0.001), ‘Cholesterol biosynthetic process’ (Z = 16.5, P < 0.001), ‘Sterol metabolic process’ (Z = 12.9, P < 0.001), ‘Cholesterol metabolic process’ (Z = 11.8, P < 0.001), ‘Cytoskeleton organization’ (Z = 7.9, P < 0.001), ‘Cellular component organization’ (Z = 6.0, P < 0.001), ‘Steroid biosynthetic process’ (Z = 10.5, P < 0.001), ‘Cellular component organization or biogenesis’ (Z = 5.9, P < 0.001), ‘Actin cytoskeleton organization’ (Z = 8.1, P < 0.001), and ‘Alcohol biosynthetic process’ (Z = 10.0, P < 0.001) (Fig. 7D).

A biased search of inflammation-related genes returned five cytokines and chemokines that were differentially expressed among all three cardiomyocyte infection conditions, namely Gdf-15, Cxcl10, Cxcl2, Il-6 and Tnf. The expression levels of these inflammatory cytokines were significantly higher in the three cardiomyocyte infection conditions compared to the non-infected control: Colombian strain P = 0.0024, Y strain P = 0.001 and Tulahuen strain P = 0.0014 (Fig. 8).

Fig. 8
figure 8

Heatmap of expression levels of inflammation-related genes in HL-1 cells infected with T. cruzi at 24-h post-infection. Genes are represented in rows and infection conditions are represented in columns. Color legend represents the logarithm 2 of gene expression level. P-values were calculated by paired T-test. Gdf-15 growth differentiation factor 15, Cxcl10 C-X-C motif chemokine ligand 10, Cxcl2 C-X-C motif chemokine ligand 2, Il-6 interleukin-6, Tnf tumor necrosis factor-alpha

Full size image

Transcriptomic remodeling unique to Colombian- and Y-infections

Upregulated genes over-represented the DEGs shared among the Colombian- and Y-infected conditions and thus formed the focus of our interest. Such 331 genes enriched pathways related to ‘DNA replication (Z = 6.8, P < 0.001)’, ‘One Carbon metabolism (Z = 6.8, P < 0.001)’, ‘Nucleotide metabolism (Z = 6.8, P < 0.001)’, ‘Glutathione and one-carbon metabolism (Z = 5.1, P < 0.01)’, and ‘Mitochondrial gene expression (Z = 4.9, P < 0.01)’ (Fig. 9A). The same gene list was enriched in genes associated with the Biological Process GO terms of ‘Nucleic acid metabolic process’ (Z = 8.3, P < 0.001), ‘Nucleobase-containing compound metabolic process (Z = 7.2, P < 0.001), ‘Heterocycle metabolic process’ (Z = 7.1, P < 0.001), ‘Cellular nitrogen compound metabolic process’ (Z = 6.9, P < 0.001), ‘Cellular aromatic compound metabolic process’ (Z = 6.9, P < 0.001), ‘Organic cyclic compound metabolic process’ (Z = 6.8, P < 0.001), ‘Nitrogen compound metabolic process’ (Z = 6.7, P < 0.001), ‘RNA metabolic process’ (Z = 6.9, P < 0.001), and ‘Gene expression’ (Z = 6.1, P < 0.001) (Fig. 9B).

Fig. 9
figure 9

Signatures of transcriptomic remodeling unique to Colombian- and Y-infections. ‘WikiPathways’ (A) and 'Biological Process gene ontology Terms’ (B), enriched by the genes uniquely upregulated at 24-h post-infection in Colombian- and Y-infected cardiomyocytes are sorted based on P-value. Bar lengths represent enrichment Z-scores

Full size image

Differentially expressed genes corresponding to the ‘Glutathione and one-carbon metabolism’ pathway (Wikipathway ID: WP730) were individually tracked and analyzed, where the genes corresponding to the ‘one-carbon metabolism’ and ‘Glutathione metabolism’ showed an up and downregulation trend, respectively. Shared, as well as Colombian/Y-unique, pathways and GO terms enriched with upregulated DEGs upon infection are summarized in Table 1.

Table 1 Summarization of the pathways enriched by the upregulated genes in all the T. cruzi-infection conditions
Full size table

Transcriptomic remodeling unique to Tulahuen-infections

Tulahuen-infection led to the unique downregulation of 218 DEGs in the cardiomyocyte transcriptome. Significantly enriched pathways were: ‘Proteosome degradation’ (Z = 5.1, P < 0.001), ‘Electron transport chain’ (Z = 4.1, P < 0.01), and ‘Oxidative phosphorylation’ (Z = 3.7, P < 0.01). Top 10 Biological processes GO terms enriched by this gene set are ‘Cellular metabolic process’ (Z = 3.9, P < 0.001), ‘Metabolic process’ (Z = 3.4, P < 0.001), ‘Cellular process’ (Z = 2.3, P < 0.001), ‘Organic substance metabolic process’ (Z = 3.2, P < 0.001), ‘Localization’ (Z = 4.0, P < 0.001), ‘Organic substance transport’ (Z = 4.5, P < 0.001), ‘Macromolecule localization’ (Z = 4.4, P < 0.001), ‘Cellular protein metabolic process’ (Z = 3.9, P < 0.001), and ‘Cellular protein metabolic process’ (Z = 3.9, P < 0.001).

Trypanosoma cruzi infection increases intracellular reactive oxygen species in HL-1 cardiomyocytes

With specific interest in the accentuated ROS mitigation response elucidated from transcriptomic analysis of the cellular models, we next evaluated the intracellular ROS level of infected cardiomyocytes. HL-1 cardiomyocytes showed higher ROS levels upon infection with T. cruzi in comparison to non-infected controls. Signal intensity (mean, interquartile range) of the ROS-specific fluorescent dye, CellRox green, was significantly higher in the T. cruzi-infected condition (16.55, 9.73–19.02; 6.85, 5.70–7.54; P < 0.0001), further validating the relevant role of ROS in CCM pathogenesis (Fig. 10A). Representative images are shown in Fig. 10B.

Fig. 10
figure 10

Elevated intracellular levels of reactive oxygen species in Trypanosoma cruzi-infected cardiomyocytes. A Signal intensity of reactive oxygen species-specific fluorescence was higher in the cardiomyocytes infected with the Y strain at 24-h post-infection, compared with non-infected control cells. Statistical comparison was performed with Mann–Whitney test. Bars represent mean fluorescence intensity and error bars represent standard deviation. B Representative fluorescence microscopy images from reactive oxygen species-specific CellRox green staining. Comparison with the non-infected (upper) control cells shows higher intensity of reactive oxygen species fluorescence signal in the cardiomyocytes 24-h post-infection with Trypanosoma cruzi Y strain (lower). Scale bars represent 100 µm

Full size image

Discussion

In vitro infection using HL-1 cardiomyocytes and T. cruzi from different genetic strains provided mechanistic clues to CCM progression.

Subsequent transcriptomic and functional analyses of the in vitro cardiomyocyte infection models had characterized in-depth the myocardial innate response that is a potential determinant of CCM progression. The cross-strain comparative analysis led to the characterization of the universal hallmarks of the cardiomyocyte response against T. cruzi infection, consisting of DEGs evoked under all three infection conditions. Among the hallmark transcriptomic signatures were pathways related to oxidative stress (the ‘Response to reactive oxygen species’, ‘Response to oxidative stress’, ‘Response to hydrogen peroxide’ GO terms), cell proliferation and cardiac hypertrophy (the ‘ErbB signaling’, ‘Hypertrophy model’ and ‘MAPK signaling’ pathways), and those related to heart energy metabolism (the ‘SREBF and miR33 in cholesterol and lipid homeostasis’, and ‘Omega-9 FA synthesis’ pathways and the ‘Sterol biosynthetic process’, ‘Cholesterol biosynthetic process’, ‘Sterol metabolic process’, ‘Cholesterol metabolic process’ and ‘Steroid biosynthetic process’ GO terms).

Firstly, possible derangements in cardiomyocyte ROS homeostasis, as observed here in our cellular model, has been previously reported to accompany cardiac infection with T. cruzi [27, 29, 31, 38,39,40]. ROS mitigation is not a unique host response but in fact a shared mechanism of the cardiomyocyte against the more common insults leading to dilated cardiomyopathy, namely ischemia/hypoxia and Coxsackie virus infection [41,42,43]. Among the unique remodeling patterns in Colombian and Y infections were the upregulation of glutathione metabolism for ROS mitigation in CCM progression. Previous in vitro [23, 24] and in vivo studies [22] have shown that glutathione and its related metabolites serve as key antioxidants in cardiomyocytes under infectious or hypoxic stress [44,45,46]. Altogether, the concomitant upregulation of the compensatory ‘Glutathione and one-carbon metabolism’ pathways in Colombian- and Y-infected cardiomyocytes further highlights the intriguing role of ROS homeostasis in CCM pathogenesis. Several genes from the ‘Hypoxia signaling’ pathway, notably the hypoxia-inducible factor 1-α (HIF-1α) have been reported to be upregulated in human cardiomyocytes as a response to infection with T. cruzi [18]. HIF-1α has been reported to be positively correlated with myocarditis severity in patients with CCM [47]. Although in our dataset the ‘hypoxia signaling’ pathway was not enriched, a detailed search of HIF-1α-targeted genes revealed the downregulation of Glutathione peroxidase 3 (Gpx3), and the upregulation of Growth Arrest and DNA Damage Inducible Alpha (Gadd45a). An overtime decline of glutathione peroxidase levels has been previously described in a CCM model after an initial burst on its levels triggered by chronic ROS accumulation typical of hypoxic states [48], and might promote progression to heart failure [49].

Secondly, upregulation of cell proliferation pathways, such as ‘PI3K/AKT’, ‘Extracellular signal-regulated kinases 1 and 2 (ERK1/2)’, ‘JAK/STAT signaling’ and ‘The mammalian target of Rifampicin (mTORC)’ pathways, have been shown previously in multiple studies with human and murine cell lines infected with a variety of T. cruzi strains [24, 29, 50,51,52,53,54,55,56,57,58,59]. Cell proliferation seems a shared response mechanism among variable host cells, considering the differential expression of 18 well-known pro-proliferative genes, including Rhob, Atf3, Brca2, Jun, Jund, Myc, as well as the enrichment in the mutually interacting pathways of ErbB and MAPK signaling, observed in the cardiomyocyte response of our data set [56]. A summarization of the DEGs of the ErBb and MAPK signaling pathways is depicted in Additional file 2: Fig. S3. In the context of cardiomyocytic responses, although initially adaptive to stress stimuli, such signaling pathways are the underlying cause of pathological cardiac hypertrophy and mechanical dysfunction when chronically activated [60,61,62]. Thus, these pathways have constituted the ‘double-edged' roles in the development of, and the common potential targets for, ischemic and viral cardiomyopathies [63,64,65].

Thirdly, cellular metabolism can drastically influence cellular functions in the failing heart [66, 67]. Previous in vitro studies have replicated downregulation of key genes in lipid metabolism such as Dhcr24 [23, 68]. The uncoupling of heart lipid synthesis and β-oxidation, coined cardiac lipidopathy, has been suggested as a hallmark metabolic feature of CCM [69]. Sterol regulatory element binding factor (Srebf), a transcription factor that regulates cellular lipid homeostasis by controlling the expression of an array of enzymes required for endogenous cholesterol, fatty acid, triacylglycerol, and phospholipid synthesis [70, 71] was found to be one of the commonly downregulated DEGs in our dataset at 24 and 48 hpi (Additional file 1). Srebf has been proposed as a potential target for ameliorating the cardiac lipidopathy in Chagas disease [72, 73]. This imbalance of lipid synthesis and beta-oxidation became further prominent in our in vitro model at 48 hpi, accompanied by the downregulation of ‘Fatty acid oxidation’ and ‘Beta-oxidation’ pathway-related genes: Acss2, Acsl4, and Acsl5 (Additional file 1). The correction of such pathogenic metabolic uncoupling that may potentiate T. cruzi proliferation may exert therapeutic potential against CCM through the restoration of metabolic homeostasis [73,74,75]. While the glycolysis pathway was not significantly enriched in our dataset, a manual search of glucose metabolism-related genes showed the common upregulation of Solute Carrier Family 2 Member 3. The upregulation of 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 and Phosphoglycerate mutase family member 5 in Colombian and Y-infected cells only; and the downregulation of Solute Carrier Family 2 Member 1 and Phosphofructokinase, liver, B-type in Tulahuen infection. Noteworthy, Arrestin Domain Containing 4 (Arrdc4), Thioredoxin interacting protein (Txnip), and Activating transcription factor 3 (Atf3), which were among the most extensively upregulated genes common to all cardiomyocyte infection conditions, may hinder glucose uptake and lead to apoptosis in vitro as shown in murine cardiomyocytes and T. cruzi-infected human fibroblasts [76,77,78,79,80]. Their upregulation in vivo promotes cardiac hypertrophy and contractile dysfunction [81, 82]. Furthermore, Atf3, which is induced by ROS and pathogens including T. cruzi, may promote fibrosis and cardiac hypertrophy [83,84,85,86,87,88,89]. T. cruzi-infected Induced Pluripotent Stem Cardiomyocytes (iPS-CM) display an activation of the glycolysis and oxidative phosphorylation pathway where attenuation of glucose uptake significantly decreased parasites’ invasion and replication within the host [18]. The aforementioned phenomenon has also been observed after restricting the cell from other carbon sources like glutamine [90]. This number of phenotypic effects derived from their dysregulation renders them an interesting group of genes to explore their role in the pathogenesis of CCM. Further functional analysis in vitro on actual glucose utilization using our cellular model may elucidate whether a metabolic compensatory mechanism is operating to mitigate the ongoing infection.

Additionally, among the top upregulated DEGs was the TGF-β superfamily Gdf15, an emerging cardiovascular biomarker. GDF15 protects cardiomyocytes from ROS- and nitric oxide-induced apoptosis in vitro. In addition, the risk stratification potential of GDF15 in patients with heart failure is currently gaining attention [91,92,93]. Given that the upregulation of GDF15 comprises the cardinal feature of cardiomyocyte response in our in vitro model, the behavior of this cytokine in actual CCM patients arises as our next intriguing research question.

Regional factors, including circulating T. cruzi strains, have attracted attention as likely determinants of Chagas disease phenotypic expression and CCM progression [94,95,96,97,98,99]. Experimental data from murine and cellular models of CCM have shown that T. cruzi Colombian and Y strains among others cause severe myocarditis and higher mortality [22, 100, 101]. Epidemiological data from Colombian, Brazilian and Argentinian patients suggest that CCM prevalence and severity is increased in those infected with certain T. cruzi strains [102,103,104,105]. Although not the only causative factor, it is likely that infections caused by parasites of specific genetic traits render the host prone to severe CCM forms.

The limitations of this study include the lack of normalization of the gene expression values by intracellular amastigote burden across the infecting T. cruzi strains. Higher infectivity was observed in the Colombian and Y strains compared with the Tulahuen strain and may have affected the time course of progression of the in vitro pathogenesis. Such a trend towards less infectivity of the Tulahuen strain has been observed in previous studies [106, 107]. Single-cell resolution analyses, to overcome the weaknesses of bulk transcriptomics, may aid in eliminating potential interpretation biases of the observed inter-strain differences from the current approach. It is also noteworthy that since the present study was mainly observational in nature, further functional experiments are to be performed in order to make conclusions about the relevance of the individual pathways that were highlighted here.

Conclusions

In conclusion, the ubiquitously observed hypertrophic and oxidative stress-related transcriptome remodeling are universal hallmarks of the cardiomyocyte response against T. cruzi infection. The Colombian and Y strain elicited an accentuated response in cardiomyocyte ROS mitigation, nitrogen, and glutathione metabolism suggesting a role in CCM progression. The downregulation of the cardiomyocyte’s energy metabolic pathways exclusively observed in the Tulahuen strain infection may enact a protective effect against the infection.