Twenty-four patients were enrolled in this study after prior approval of the Institutional Ethics Committee (Hospital das Clínicas de Ribeirão Preto—USP, São Paulo, Protocol number 2285/2007; Brazil). Signed informed consent was obtained from all patients who were positive for at least two serology tests for Chagas disease, as determined by hemagglutination, immunofluorescence or ELISA techniques. Another cohort of 395 patients was used to study the importance of Ebi3 polymorphisms to the outcome of Chagas disease (Under the approval of the Institutional Ethics Committee—CAPpesq–HC-FMUSP—Protocol Number 0265/10, Brazil).

All the experiments using mice model were conducted with the approval of the Institutional Animal Care and Use Committee (IACUC)/Ethics Committee for Animal Care and Research (CETEA-FMRP/USP) under protocol 192/2011. The protocol was approved by the National Committee for Control of Animal Research (CONCEA).

Based on a detailed clinical evaluation and the results of 12-lead rest electrocardiogram (EKG), chest X-ray, and a 2D-echocardiogram, such patients with Chagas disease were stratified into five clinical groups: asymptomatic, mild heart dysfunction, moderate heart dysfunction, and severe heart dysfunction. Blood samples were collected from the eligible patients and from seven healthy subjects to be used as controls. IL-27 production was evaluated in the plasma of patients using ELISA kit for the subunit IL-27p28. All the technical procedures were performed according to the manufacturers’ instructions. Results were expressed as nanograms per milliliter.

Another cohort of 395 patients was used to study the importance of Ebi3 polymorphisms to the outcome of Chagas disease. In this cohort, patients with Chagas disease were stratified into two groups based on the symptoms of the disease: patients with no symptoms were classified as indeterminate (asymptomatic) patients, whereas those with cardiac symptoms and complications were classified as CCC patients. CCC patients were further stratified based on their ejection fraction (EF): those with EF >0.4 were classified as patients having a moderate cardiopathy and those with EF ≤0.4 were classified as patients having severe cardiopathy. DNA of patients with Chagas disease was extracted, and probes specific for the SNPs rs4740 (Assay ID: C_2591150_20) and rs4905 (Assay ID: C_2591151_10) (Life Technologies) of Ebi3 were used.

Mice aged 6–8 weeks were maintained in animal facilities at the Ribeirão Preto Medical School (FMRP-USP), housed with a 12 h light-darkness cycle and used to perform all experiments described in this study. We used female mice for all experiments comparing C57BL/6 (WT) mice with TLR2, TLR3, TLR4, TLR9, TRIF, MyD88, NRLP3, IL-17R, IL-23, IL-6, and Ebi3-deficient mice. All the experiments were conducted by using sex- and aged-matched mice. After approval by the Ethics Committee for Animal Care and Research of the FMRP-USP (CETEA-FMRP/USP, animal protocol 192/2011), animals were intraperitoneally infected with 1,000 trypomastigotes forms of T. cruzi (Y strain). To perform in vitro experiments, trypomastigotes forms were maintained in LLC-MK2 line (Rhesus Monkey Kidney Epithelial Cells).

After infecting mice with the parasites, the survival rate was evaluated for 40 days post infection (d.p.i.). To quantify the bloodstream parasitemia, 5 µL of peripheral blood from infected animals were smeared on a 22 × 22 mm microscope slide, and the number of parasites determined. Parasitemia was analyzed from 5 to 15 d.p.i.

Fragments with approximately 10 mg of heart tissues from WT and Ebi3-deficient mice were obtained from both non-infected and infected animals, 7 d.p.i. After collecting heart tissue of five animals per group, the samples were extracted, quantified, and examined by qPCR assay according to RT²ProfilerPCRArray™ kit of Mouse Innate Immune Responses and Adaptive (PAMM 052Z—Qiagen) on the ABI StepOnePlus™ equipment. RNA integrity was assessed by Agilent Bioanalyzer™ equipment following the manufacturer’s instructions. The non-normalized and normalized data were deposited in the Gene Expression Omnibus (GEO) database under the accession number GSE103330 (Link to the website over the accession number: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE103330).

DNA from the heart, liver, spleen, and skeletal muscle of WT and Ebi3-deficient mice was purified using the SV Total DNA Isolation System kit (Promega, Madison, WI, USA) according to the manufacturer’s instructions. Real-time PCR was performed using the SYBR Green reagents (Invitrogen, Carlsbad, CA, USA) with 100 ng of total DNA. The sequences of primers used were TCZ-F 5′-GCTCTTGCCCACAMGGGTGC-3’ and TCZ-R 5′-CCAAGCAGCGGATAGTTCAGG-3′. The samples were amplified in a StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) with the following PCR conditions: first step (2 min at 50°C), second step (10 min at 95°C), and 40 cycles (30 s at 95°C, 15 s at 60°C, and 30 s at 72°C), followed by a dissociation stage. The results were based on a standard curve constructed with DNA from culture samples of T. cruzi trypomastigotes (n = 3). To determine heart and liver damage, the release of CK-MB (Creatine Kinase-MB), AST (Aspartate Aminotransferase), and ALT (Alanine Aminotransferase) into the bloodstream of normal and infected animals was measured using specific kit (Labtest^®) following the manufacturer’s recommendations.

To evaluate whether T. cruzi parasites were able to induce IL-27, 2 × 10⁶ splenocytes from normal WT mice were stimulated with 6 × 10⁶ trypomastigotes in 48-well plates. To identify cellular sources of IL-27, bone marrow-derived macrophages (BMDM) and bone marrow-derived dendritic cells (BMDC) were differentiated and cultured with T. cruzi. The BMDMs were obtained as described previously (26). Briefly, the total bone marrow cells were cultured in RPMI 1640 medium (Sigma-Aldrich), supplemented with 10% FBS (Invitrogen) and 30% of L-929 cell conditioned media or with 20% GM-CSF (PeproTech), and maintained at 37°C and 5% CO₂ for 7 days. BMDMs and BMDCs were cultured with trypomastigotes of T. cruzi (MOI 3:1). After 7 h, splenocytes and T. cruzi-primed cells were collected for IL-27 mRNA quantification by real-time PCR and after 48 h the supernatants were collected for measuring IL-27. In order to evaluate which receptors and adapter molecules were involved in the production of IL-27, macrophages of WT and deficient mice of the molecules: NLPR3, TLR9, TLR3, TLR2, TLR4, TRIF, and MyD88 macrophages were differentiated according to the protocol above described. The ability of invasion of T. cruzi inside WT, TLR2^−/− and TLR4^−/− macrophages was evaluated after stimulation of cells with 10 ng/mL of IFN-γ followed by infection with trypomastigote forms of the parasite for 4 h. Then, macrophages were fixed and colored using panótico rapido^® according to manufacturer’s recommendations.

To assess the production of cytokines in the heart tissue of infected animals, myocardial fragments were collected into vials with a protease inhibitor cocktail (Complete, Roche), macerated, centrifuged, and the supernatant collected for cytokine quantification. The ELISA sets were IL-27p28, IL-10, TNF-α, IFN-γ, IL-4, and IL-13 (R&D, Minneapolis, MN, USA). All the technical procedures were performed according to the manufacturers’ instructions. The limits of sensitivity for the different assays were as follows: 15.6 pg/mL for IL-10 and IL-27p28; 31.2 pg/mL for TNF-α, IL-4 and IFN-γ; and 62.4 pg/mL for IL-13.

Heart samples were fixed in 10% buffered formalin solution and, after 72 h, dehydrated in crescent concentration of ethanol solutions and embedded in paraffin. Blocks were sectioned at 5 thick and stained with hematoxylin-eosin (H&E). Using a light microscopy, the number of amastigote nests was counted in 25 images of each animal of the group, the inflammation was estimated with ImageTool 2.0 software (University of Texas Health Science Center, TX, USA).

Total RNA from cardiac tissue was isolated using TRIZOL (Invitrogen) and SV Total RNA Isolation System (Promega, Madison, WI, USA) according to the manufacturer’s instructions. cDNA was synthesized using 500 ng of RNA through a reverse transcription reaction (ImProm-IITM Reverse Transcriptase, Promega). Real-time PCR quantitative mRNA analyses were performed in a StepOnePlus Real-Time PCR System (Applied Biosystems) using SYBR Green reagents (Invitrogen) for quantification of the amplicons. The standard PCR conditions were as follows: 50°C (2 min), 95°C (10 min); 40 cycles of 94°C (30 s), 58°C (30 s), and 72°C (1 min); followed by a standard denaturation curve. Primers were designed by using the Primer Express software package v2.0 (Applied Biosystems), based on the nucleotide reference sequences available at GenBank database. Platinum SYBR Green qPCR SuperMix UDG with ROX reagent (Invitrogen), 1 mg/mL of each specific primer and a 1:20 dilution of cDNA were used in each reaction. The mean Ct values from duplicate measurements were used to calculate the expression of the target gene, with normalization to internal controls (GAPDH, β-actin, and HPRT).

Spleen and heart fragments were collected to evaluate IFN-γ-producing CD4⁺ T cells, IL-10-producing Tregs, and IL-10-producing Tr1 cells. To isolate mononuclear cells from cardiac tissues, the hearts were removed at 15 d.p.i. washed (to remove blood clots), minced with scissors into small fragments, extensively washed, and subjected to enzymatic digestion with 500 mg/mL of Liberase solution (Roche Applied Science, Indianapolis, IN, USA) for 1 h at 37°C. The tissues were washed with RPMI 10% FCS and total cells that passed through the cell strainer were subsequently counted. Two million leukocytes from heart tissue or from spleen were stimulated in vitro with PMA/Ionomycin/Golgi Stop for 6 h. The leukocytes were stained with specific antibodies for Cd11b (clone MI/70 BD), I-A/E (clone 269 BD), CD3 (clone 145-2C11), CD4 (clone RM4-5 BD), Foxp3 (clone MF 23 BD), IFN-γ (clone XMG 1.2 BD), IL-10 (clone JESS-16E3 BD), and IL-27p28 (clone MM27-7b1 Biolegend), and evaluated by flow cytometry.

IFN-γ was neutralized in vivo by intraperitoneal injection of rat anti-mouse IFN-γ mAb (clone R4-6A2). The animals were treated with 2,000μg kg⁻¹ of neutralizing IFN-γ antibodies at day 9, followed by 1,000μg kg⁻¹ at days 12 and 15 d.p.i. To block NOS, animals were treated with 1,000 μg kg⁻¹ of L-NMMA at 9, 12, 15, and 18 days d.p.i. Control mice were injected with equal quantities of vehicle. This delayed kinetics of blockade was chosen to allow an initial IFN-γ or NO response required to avoid uncontrolled parasite spreading and early mortality. After treatment, the mice were used for survival monitoring.

Data were expressed as means ± SEM. The Kaplan–Meier method was used to compare survival curves among groups. Student’s t-test was used to analyze the statistical significance of the observed differences in infected vs control assays, followed by Mann–Whitney post-test. In time course studies, one-way ANOVA was used, followed by Tukey–Kramer or Bonferroni post hoc analysis. All analyses were performed using PRISM 5.0 software.

Depending on the milieu, Th17 lymphocytes can be reprogrammed to Th1 lymphocytes (27). Upon T. cruzi infection, the IL-23–IL-17A axis was shown to be crucial for the maintenance of an optimal Th1 immune response (28). We have also previously reported that Th17 lymphocytes were protective cells during T. cruzi infection. Such cells antagonized the development of Th1 lymphocytes, thus protecting the hosts from Th1 cell-mediated myocarditis (29). Supported by these pieces of evidence, we used genetically deficient mice for Th17-related molecules to create an appropriate microenvironment that facilitate the polarization to Th1 lymphocytes after T. cruzi infection. Taking advantage of deficient mice allowed us to promote intense myocarditis characterized by a robust recruitment of Th1 lymphocytes in the myocardium of T. cruzi-infected mice. IL-17R-, IL-23-, and IL-6-deficient mice rapidly succumbed to the infection at 22–25 d.p.i. compared to mice of the control group (Figure 1A). Thus, although there is cardiac damage in the absence of Th17-related molecules, these animals could efficiently control the cardiac parasitism and minimize the consequent inflammatory changes as compared to the control group (Figure 1B). The strong inflammatory response in the absence of Th17-related molecules in the genetically deficient mice led to intense cardiac damage, as evaluated by increased serum CK-MB levels (Figure 1C). In addition, we found a robust production of TNF-α (Figure 1D), an intense recruitment of Th1 lymphocytes in the cardiac tissue in the absence of Th17-related molecules compared to WT counterparts at 17 d.p.i., as well as an increase in the IFN-γ production (Figures 1E,F). Also, the absence of Th17-related molecules led to a significant increase of IL-27p28 production (Figure 1G). These murine models prompted us to evaluate a possible regulatory function of IL-27 in this inflammatory condition. By contrast, the increase in IL-17 production in the absence of IL-27 signaling as expected was not observed (data not shown).

IL-27 seems to be a key anti-inflammatory mediator in different inflammatory settings, including its relatively known role on the control of the liver immunopathology after T. congolense (23) or T. cruzi (24) infections and its suppressive effects on CD4 T cell subpopulations (25). Since CCC is induced and maintained by inflammatory CD4 T cell subpopulations, mainly Th1 cells, we evaluated whether IL-27 also mediates suppressive effects on the inflammation of the myocardium after T. cruzi infection. In vivo findings revealed that mice deficient for Ebi3, a subunit of IL-27, were extremely susceptible to infection with 10³ trypomastigote forms of T. cruzi, a quantity that was not lethal to WT mice (Figure 2A). These Ebi3-deficient mice died from T. cruzi infection within 15–16 d.p.i., whereas WT mice survived until 40 d.p.i. In addition, the bloodstream parasitemia levels in Ebi3-deficient mice were lower compared to WT mice, notably at the peak of the infection (9 d.p.i.) (Figure 2B). Since IL-35 is composed of the subunits Ebi3 and p35, we confirmed that IL-27 participates of the immune response during T. cruzi infection by assessing the expression of Ebi3, p28, and p35 in our model. In agreement with the importance of IL-27 to the T. cruzi infection, we observed that the gene expression of the subunits p28 and Ebi3 was higher than p35 at the peak of the parasitemia (Figure 2C). This finding does not rule out a potential participation of IL-35 upon T. cruzi infection.

IL-27 is thought to protect against tissue damage and regulate inflammatory conditions driven by effector T cells (23, 24). Based on the potential regulatory role of IL-27 upon inflammation, we screened the expression of inflammation-related genes in the heart tissue of WT and Ebi3-deficient mice after T. cruzi infection by performing a PCR array (Figures 2D,E). As expected, we found genes associated with markers of acute inflammatory response, such as serum amyloid P component (Apcs), C-reactive protein (Crp), and TNF-α (Tnf), that were highly expressed in the heart tissue of Ebi3-deficient mice compared to WT mice 7 d.p.i. Casp1, Il6, and Nlrp3 genes were also upregulated in the cardiac tissue of Ebi3-deficient mice compared to WT mice. We also observed that Ebi3 was able to control the expression of chemokine-related genes along with the expression of the Cd40lg gene. The expression of Ccl5, Ccr8, and Cxcl10 was increased in the heart of Ebi3-deficient mice compared to WT mice. We further explored whether Ebi3 downregulated effector T cell pathways. Although Ebi3 was not able to regulate Th2 lymphocytes-related genes, such as Gata3, Stat6, Il4, and Il5, we found that Th1 lymphocytes-related genes, such as Il2, Stat1, and Tbx21, were upregulated in the heart of Ebi3-deficient mice compared to WT mice. Of note, Ifng was one of the most strikingly expressed genes in Ebi3-deficient mice. The expression of Foxp3 and IL17a was not changed 7 days after infection in both groups.

Trypanosoma cruzi parasites have preferential tropism for the skeletal muscle tissue. Locally, an enhanced recruitment of inflammatory cells is required to control and eliminate parasitemia and tissue parasitism, but it eventually culminates in nearly fatal T. cruzi-induced myocarditis (1). It was observed that Ebi3-deficient mice exhibited an increased heart parasitism (Figure 3A) with concomitant alongside increased blood levels of CK-MB at day 15 of infection, indicating intense myocardial damage (Figure 3B). We also observed a prominent cardiac inflammatory infiltrate in the myocardium of 15 day-infected Ebi3-deficient mice compared to their WT counterparts, although, as expected, the infiltrate is also higher in the 15-day-infected WT mice when compared to non-infected mice (Figures 3C,D). In addition, we found no myocardial damage (Figure 3B) or immune cell infiltration (Figure 3C) into the heart at day 10 post-infection, suggesting that the cardiac inflammation arises at the late time points and is highly destructive. The levels of AST and ALT were also assessed as indicators of liver damage (Figures 3E,F). The histology showed a lower inflammatory infiltrate in the liver of WT mice in comparison to Ebi3-deficient mice (Figures 3G,H), as previously shown elsewhere (23–25), suggesting that Ebi3 is also required to prevent liver immunopathology in T. cruzi infection. These results suggest that Ebi3 restricts the recruitment of inflammatory cells to the target tissues, besides controlling the T. cruzi-induced myocarditis. These data reveal that Ebi3 plays a crucial role for host survival and parasite control during the experimental T. cruzi infection.

IL-27 is thought to be preferentially produced by macrophages (14) and dendritic cells (15) to control inflammation. To identify the cell populations involved in IL-27p28 production, we analyzed their production after in vitro infection with T. cruzi parasites. We found that wild type (WT) BMDC infected with T. cruzi produced increased levels of IL-27p28 compared to the non-infected WT BMDC (Figure 4A). Macrophages are another essential cell source of IL-27. Similarly, T. cruzi-infected WT macrophages produced significantly higher levels of IL-27p28 than their non-infected WT counterparts (Figure 4B). We also found that IL-27p28 was strongly induced at both mRNA and protein levels by in vitro T. cruzi-infected spleen cells when compared to non-infected spleen cells (Figures 4C,D). Both macrophages and dendritic cells appear to be sources of IL-27p28 during the experimental T. cruzi infection. We also assessed IL-27p28 production in the myocardium of WT mice. We found a higher percentage and number of IL-27-producing MHII⁺CD11b⁺ myeloid cells in the myocardium of infected compared to non-infected mice (Figures 4E,F). These data suggest that myeloid cells are important sources of IL-27p28 under in vivo and in vitro stimulation with T. cruzi.

Some innate immune receptors are known to recognize T. cruzi-expressed molecules or T. cruzi-derived danger signals, including TLRs 2 (30), 4 (31), and 9 (32), as well as NOD1 (33) and NLRP3 (34). It has earlier been reported that IL-27 production is also dependent on MyD88 and TRIF (35). Here, we assessed whether T. cruzi recognition by innate sensors could trigger IL-27p28 expression and production in macrophages. We found that T. cruzi-infected BMDM deficient for TLR2, TLR3, and TLR4 receptors, and MyD88 and TRIF adaptors had a lower expression of the gene IL27p28 compared to BMDM of WT mice (Figure 5A). We also found that T. cruzi-infected BMDM deficient for TLR2, TLR4, MyD88, or TRIF had significant impairment of IL-27p28 production (Figure 5B). Since signaling through TLR2 and TLR4 had a strong impact on IL-27p28 production, we analyzed whether the lack of both receptors would abolish IL-27p28 production by macrophages. Indeed, the IL-27p28 levels produced by TLR2⁻TLR4⁻ (DKO) BMDM were much lower than by their WT counterparts and are equivalent to the basal levels produced by non-stimulated macrophages (Figure 5C). To prove that the IL-27p28 production was a direct effect of TLR2 and TLR4 signaling rather than a higher susceptibility to the T. cruzi infection caused by the absence of these receptors, we evaluated whether the macrophages were equally infected by the parasite regardless the expression of TLR2 or TLR4. As expected, the absence of TLR2 or TLR4 did not alter the susceptibility of the macrophage to the infection (Figure 5D), suggesting that the cellular signaling through these innate immune receptors, mainly TLR2 and TLR4, is required for IL-27p28 production during T. cruzi infection.

We next explored whether Ebi3 could antagonize Th1 cell-induced inflammatory reactions after T. cruzi infection. We found increased number of IFN-γ producing CD4⁺ T cells in the cardiac tissue and spleen of Ebi3-deficient mice compared to WT counterparts 15 days after T. cruzi-infection (Figures 6A,B). In addition, number and frequency of IFN-γ-producing CD8 T cells were also higher in infected Ebi3-deficient mice compared to the control group. Although CD8⁺ T cells are also important to control the heart parasitism, the main source of IFN-γ into the heart of infected mice was CD4 T cells instead of CD8 T cells (4.32 × 10⁴ × 0.62 × 10⁴ in WT and 7 × 10⁴ × 0.88 × 10⁴ in Ebi3^−/− mice) (Figures 6A,C). In accordance with increased number of IFN-γ producing cells, we also found higher levels of IFN-γ in spleen (Figure 6D) and cardiac tissue (Figure 6E) of 15-day-infected Ebi3-deficient animals when compared to WT group. These data suggest that Ebi3 is a key regulator of IFN-γ production during the T. cruzi-induced myocarditis.

We next assessed whether Ebi3 was indirectly suppressing Th1 cells by inducing a regulatory cell population. We firstly assessed whether Ebi3 could be regulating alternatively activated macrophages. Although the infection upregulated the gene expression of markers expressed on alternatively activated macrophages (Arg1 and Chi3l3), Ebi3 did not have any impact on their expression in the heart of 15-day-infected animals (Figures 7A,B). Also, Fizz1 was downregulated in the heart of both groups after the infection with T. cruzi (Figure 7C). Moreover, IL-4 and IL-13, key cytokines associated with M2 macrophages, were not modulated by the infection or the absence of Ebi3 in the hearts of animals (Figures 7D,E). Taken together, these data suggest that Ebi3 cannot regulate key molecules expressed on alternatively activated macrophages in the heart tissue of T. cruzi-infected mice. We also evaluated whether Ebi3 could promote classical Tregs and/or Tr1 cells during the infection. We found that IL-10 production was significantly reduced in the myocardium of Ebi3-deficient mice compared to WT mice at day 15 after infection (Figure 7F). Unexpectedly, the number of classical CD4⁺Foxp3⁺ Tregs producing IL-10 was similar in the cardiac tissue of Ebi3-deficient and WT mice at day 15 after infection (Figures 7G,H), suggesting that Ebi3 was unable to promote the development of classical Tregs. By contrast, the frequency of IL-10-producing CD4⁺Foxp3⁻ T cells was significantly reduced in the myocardium of Ebi3-deficient mice compared to WT mice (Figures 7G,I), suggesting that the main cellular source of IL-10 production induced by IL-27 were CD4⁺Foxp3⁻ T cells (Tr1 cells). These observations show that Ebi3 could mediate Tr1 cell induction in the heart tissue of T. cruzi-infected animals.

Based on our earlier findings that Ebi3 could attenuate inflammatory intensity by inducing Tr1 cells after T. cruzi infection, we evaluated whether the treatment of T. cruzi-infected Ebi3-deficient mice with anti-IFN-γ MAb could control myocarditis and improve host survival. Notably, the blockade of IFN-γ after 9 d.p.i. was essential to improve host survival of Ebi3-deficient mice when compared to untreated Ebi3-deficient mice (Figure 8A). Because IFN-γ is required for parasite control by promoting macrophage release of NO, we assessed whether the blockade of IFN-γ contributed to T. cruzi growth. The treatment with α-IFN-γ MAb did not influence the level of parasitemia (Figure 8B). As IFN-γ also promotes NO production of macrophages, we assessed whether the use of L-NMMA to blockade all isoforms of NOS could similarly prevent myocarditis during the T. cruzi infection. As expected, treatment with L-NMMA was also able to prevent host susceptibility to T. cruzi infection of Ebi3-deficient mice (Figure 8C). Furthermore, treatment with L-NMMA after 9 d.p.i. did not lead to impaired T. cruzi clearance from the bloodstream (Figure 8D).

We next examined the impact of IL-27p28 in human clinical forms of Chagas disease. To perform this, we stratified our Chagas disease sample population into controls and patients with mild, moderate, or severe heart dysfunction caused by T. cruzi infection. As suggested by the results in the experimental model of T. cruzi infection, serum levels of IL-27p28 were higher in patients with indeterminate form or mild heart dysfunction compared to chronic patients with moderate or severe heart dysfunction (Figure 9). These data indicate that IL-27p28 can also be used as a biomarker associated with different clinical forms of CCC.

To confirm the importance of IL-27 in Chagas disease and understand whether polymorphic sites at Ebi3 gene might contribute to the severity of the disease, we assessed whether the two most prevalent polymorphic regions at the Ebi3 gene were related to the clinical manifestations of Chagas disease. Our cohort consisted of 395 individuals divided into two different groups: 93 indeterminate patients and 302 CCC patients. The indeterminate group included 31 (33.3%) males and 62 (66.7%) females, whereas the CCC group included 150 (49.7%) males and 152 (50.3%) females. Sex ratio was significantly different between the two groups (p = 0.006). The CCC patients were further divided into two subgroups: 174 patients with moderate cardiopathy (EF > 0.4) and 128 with severe cardiopathy (EF ≤ 0.4). The moderate cardiomyopathy subgroup included 67 (38.3%) males and 108 (61.7%) females, whereas the severe cardiomyopathy subgroup included 83 (65.4%) males and 44 (34.6%) females. Sex ratio was also significantly different between the two subgroups (p = 5E−6). Thus, we included sex in the analysis as a covariate.

Regarding the polymorphism at the position +609 of Ebi3 gene (SNP rs4740), we noticed that the genotype AA is less frequent in patients with severe cardiopathy (6.6%) compared to patients with the indeterminate form (19.8%) or those with moderate cardiopathy (20.2%) (Table 1). The following comparisons were done for SNP rs4740: CCC vs indeterminate, severe CCC vs indeterminate, and moderate CCC vs severe CCC. We compared patients carrying AA genotype to those carrying either GG or GA genotypes. By comparing CCC vs indeterminate patients, none of the covariates (sex and polymorphism) remained significantly associated with disease [sex: p = 0.836; OR (95% CI) = 0.933 (0.49–1.80); rs4740: p = 0.501; OR (95% CI) = 0.785 (0.39–1.59)]. By comparing severe CCC vs indeterminate patients, only the rs4740 polymorphism remained significantly associated [sex: p = 0.083; OR (95% CI) = 2.043 (0.91–4.58); rs4740: p = 0.025; OR (95% CI) = 3.157 (1.15–8.64)] after a multivariate binary regression analysis. By comparing moderate CCC vs severe CCC, the two covariates (sex and polymorphism) remained significantly associated with disease [sex: p = 0.004; OR (95% CI) = 2.833 (1.39–5.79); rs4740: p = 0.006; OR (95% CI) = 3.645 (1.45–9.16)]. Thus, this polymorphism distinguishes moderate CCC from severe CCC and severe CCC from asymptomatic controls.

Regarding the polymorphism at the position +680 of Ebi3 gene (SNP rs4905), we found that the genotype GG is also less frequent in patients with severe cardiopathy (8.7%) compared to the other groups (22.6% for indeterminate and 21.1% for moderate cardiopathy) (Table 1). The following comparisons were also done for SNP rs4905: CCC vs indeterminate, severe CCC vs indeterminate, and moderate CCC vs severe CCC. We compared subjects carrying GG genotype to those carrying either AA or AG genotypes. By comparing CCC vs indeterminate patients, only the sex remained significantly associated with the disease [sex: p = 0.008; OR (95% CI) = 0.499 (0.30–0.83)]. By comparing severe CCC vs indeterminate patients, both sex and polymorphism were associated with the disease [sex: p = 1.4E−5; OR (95% CI) = 3.800 (2.08–6.94); rs4905: p = 0.025; OR (95% CI) = 2.699 (1.13–6.43)] after a multivariate binary regression analysis. By comparing moderate CCC vs severe CCC, the two covariates remained significantly associated with disease [sex: p = 3.200E−5; OR (95% CI) = 2.899 (1.76–4.79); rs4905: p = 0.018; OR (95% CI) = 2.536 (1.18–5.47)]. This polymorphism also distinguishes moderate CCC from severe CCC and severe CCC from asymptomatic controls. Our data show that individuals who inherit the genotype AA for the rs4740 polymorphism or the GG for the rs4905 polymorphism have a lower risk to develop severe CCC. In our whole cohort, among the 135 individuals carrying the rs4740GG genotype, 131 of them also carry the rs4905AA genotype, suggesting that these regions are in linkage disequilibrium.