In this study we show for the first time that male mice with CVB3 myocarditis have reduced mitochondrial transcription compared to females using an autoimmune model of CVB3 myocarditis that is highly translational to human disease [36]. We show that females with myocarditis have higher expression of several master regulators of mitochondrial homeostasis including PGC1a, NRF1 and ERRa compared to males. Females with CVB3 myocarditis had transcriptional evidence of better mitochondrial function and significantly less myocardial inflammation than males. A sex-specific effect of ERRa on inflammation and cardiac function suggests a potential regulatory mechanism for our observed sex differences in mitochondrial gene transcription.
PGC1a was originally identified as a regulator of mitochondrial function in brown adipose tissue, but was later also found to be expressed at high levels in cardiac tissue where it influences cardiovascular health and disease [25, 37]. PGC1a globally regulates mitochondrial pathways in response to stresses such as cold, fasting and infection [35, 38–41]. Thus, the metabolic stress of CVB3 infection is a likely explanation for the elevated levels of PGC1a that we observed in males and females with myocarditis compared to controls (Fig. 10a,b). Additionally, a study by Dufour et al. using ERRα deficient mice, found that when ERRa was low in the heart PGC1a was elevated as a compensation mechanism [21].
ERRa has been found to be critical in regulating mitochondrial homeostasis in the heart demonstrated by Dufour et al. using male ERRa deficient (KO) C57BL/6 mice [21]. They found that ERRa targeted mitochondrial NRF1, cyclic AMP-response element binding protein (CREB), and STAT3 [42]. Surprisingly, we observed an inverse relationship between mRNA and protein levels of ERRα in the heart with increased levels by ELISA and IHC in males and females with myocarditis. Regardless of mRNA levels, ERRα protein levels can be highly regulated by post-translational modifications and metabolic stress [43]. Previously, it was notably shown that under specific cellular metabolic stress conditions, such as reactive oxygen species (ROS) exposure, ERRα protein levels can be dramatically altered in a proteasome-dependent manner [43]. Similarly, insulin or glucose stimulation increased ERRα protein levels without altering mRNA expression in hepatocytes [44], further strengthening the hypothesis that ERRα protein levels can be altered under metabolic pressure independently of gene expression. Indeed, transcription factor expression often does not necessarily provide detailed information as to the direct actions of that transcription factor; and in this case, how ERRα activity may differ by sex during myocarditis. Future studies utilizing methods to characterize the genomic interactions of ERRα would be useful in elucidating sex-specific transcriptional activity.
Interleukin (IL)-1a, IL-1b, and tumor necrosis factor (TNF)a are known to activate the transcriptional activity of PGC1a through direct phosphorylation of p38 mitogen-activated protein (MAP) kinase [24, 45]. We found previously that IL-1b levels are increased in the heart of males with myocarditis, while cardiac levels of TNFa are increased in females in our CVB3 model of myocarditis [10, 46]. Most cardiac inflammatory cells during acute myocarditis at day 10 pi are CD11b+ (macrophages and mast cells) that express TLR4 and release IL-1b [16]. We showed previously that elevated IL-1b levels in the heart directly correlate to elevated cardiac inflammation in males with myocarditis and poor cardiac function by echocardiography [10]. Importantly, Remels et al. showed that elevated TNFa levels in cardiomyocytes in culture following CVB3 infection were directly associated with decreased PGC1a mRNA levels [47].
Additional evidence of the negative effect that IL-1b can have on mitochondrial gene expression in males was found in studies by Ge et al. [19, 48, 49]. Calpain is a calcium-dependent protease that facilitates apoptotic signaling and localizes to the mitochondria during CVB3 infection to proteolyze mitochondrial substrates, leading to increased mitochondrial fission (mitochondrial fragmentation due to pathological or physiological stress). Inhibition of calpain reduced mitochondrial fission and cardiomyocyte apoptosis during myocarditis [48]. Liu et al. showed that mitochondrial calpain-1 induces mitochondrial dysfunction and ROS production which activated the NLRP3 inflammasome, which leads to IL-1b production [49]. Macrophages, which are the predominant infiltrating immune cells during myocarditis, were found to respond to CVB3 infection by upregulation of calpain-4; RNA sequencing of CVB3 infected macrophages in vitro revealed predominant enrichment for pathways related to macrophage maturation and interleukin signaling, and loss of calpain-4 reduced IL-1b expression [19]. Although we did not specifically examine IL-1b in this study, our previous findings may be relevant to the current results that suggest that elevated inflammatory cells and cytokines, especially IL-1b, in the heart of males during acute CVB3 myocarditis [10] may directly contribute to lower PGC1a levels in males than females leading to decreased mitochondrial gene expression in the heart at that timepoint.
In general, sex differences are known to exist in mitochondrial bioenergetics [50, 51], but we provide sex-specific information in the context of viral myocarditis. Similar to previous studies that examined gene changes in the heart during CVB3 myocarditis in male mice [10, 47, 52], we found that the predominant gene expression changes aside from immune pathways were mitochondrial genes. Previously, Remels et al. reported that PGC1a mRNA and NRF1 protein levels were significantly decreased in the heart of male mice with CVB3 myocarditis compared to controls from day 4 to 7 pi [47]. They also found decreased gene expression profiles for ETC genes during myocarditis in males [47], similar to our results, but they did not examine females with myocarditis.
Ebermann et al. also examined gene expression in males with CVB3 myocarditis comparing C57BL/6 (B6) to A.SW/SnJ mice [53]. They used a tissue culture CVB3-induced model that produces similar inflammation in these two strains of mice but different cytokine profiles [53]. This tissue-culture CVB3 model produces a completely different myocardial immune profile than our model of autoimmune CVB3-myocarditis comparing BALB/c to B6 mice [9, 54, 55]. However, Ebermann et al. found that A.Sw/SnJ male mice with myocarditis have significantly lower ETC gene expression compared to controls that was directly related to the level of viral replication in the heart [53]. In our model of CVB3 myocarditis there are no sex differences in VP1 RNA levels or viral replication based on plaque assay during acute myocarditis [16]. The findings of Ebermann et al. may reflect, however, the finding of Sin et al. who showed in cultured cardiomyocytes that CVB3 localizes to mitochondria, induces mitophagy, and disseminates from the cell in an extracellular autophagosome-bound virus-laden mitochondrial complex [56]. Sin et al. showed that upstream suppression of the mitophagy pathway in HL-1 cardiomyocytes using small interfering RNA (siRNA) targeted to dynamin-related protien-1 (DRP1) or mitochondrial division inhibitor (Mdivi-1) significantly reduced virus production from cardiomyocytes [56] (as mitochondrial fission is an early stage of mitophagy). Other viruses that cause myocarditis such as human immunodeficiency virus (HIV), hepatitis B and C, influenza, Epstein-Barr virus and SARS-CoV-2 have been found to localize to mitochondria and hijack aspects of the mitochondrial machinery for replicationn [5, 57–59]. This might explain why so many diverse viruses without specific tropism for cardiac tissue (i.e., murine cytomegalovirus/ MCMV, SARS-CoV-2, CVB3) are able to cause myocarditis, since they can target a mitochondria-rich environment for replicatory advantage.
TRANSFAC analysis identified IRFs and ERRs as key transcription factors that could mediate sex differences in gene expression in our model. CVB3 infection strongly activates type I interferons (IFNas and IFNb) and type II (IFNg) IFN production during myocarditis to reduce viral replication via Toll-like receptor (TLR) activation including TLR3, TLR4, TLR7 and TLR9 and the transcription factor TIR domain-containing adaptor inducing interferon-β (TRIF) which is downstream of TLR3 and TLR4 [10, 46, 60–62]. Although IFNg is increased in our model of CVB3 myocarditis in males [16, 18], we showed that elevated IFN levels in male BALB/c mice with myocarditis are mediated by IL-18, which is downstream from TLR4, rather than traditional STAT4/IL-12 transcriptional activity [18, 63]. We do not observe a sex difference in viral levels in the heart during myocarditis in our CVB3 mouse model and sex differences in IFNg are not mediated by classic IFN signaling. Therefore, is not surprising that we did not observe a significant difference by sex of the nine IFN transcription factors (Fig. 5b).
Sex hormones are known to strongly drive the innate and adaptive immune response to infections in general and during myocarditis [8, 64, 65], and to confer sex differences in mitochondrial morphology and function via estrogen receptor (ER) nuclear and mitochondrial transcription factor activity [39, 66, 67]. The heart of females is known to have greater mitochondrial efficiency, fatty acid utilization during exercise, and calcium retention whereas males have more mitochondrial content, reactive oxygen species production, and higher calcium uptake rate for example [39, 66]. A summary of these known sex differences in mitochondrial-related genes and pathways can be found in Additional File 1: Table S2. ERRα was originally named based on its sequence homology to ERα [68]. Although sex differences in some mitochondrial gene expression pathways during CVB3 myocarditis may be explained by sex hormones, specifically estrogen via ERs, 17b-estradiol and other natural estrogens are not endogenous ligands for ERRα [68, 69]. Two groups have provided evidence that support the hypothesis that cholesterol is the endogenous ligand for ERRα with in vitro and in vivo data [28–30]. During nominal cellular states and unbound by its ligand, ERRα displays some transcriptional activity [25, 32, 70]. Based on structural homology, ERRs are speculated to share target genes, coregulatory proteins, and sites of action with ERs and therefore actively influence the estrogenic response [71]. The genotype-tissue expression (GTEx) project identified ERRα as a “sex-biased” transcriptional regulator in humans [72]. Overall, this could explain sex differences in ERRa expression.
Lee et al. found sex differences in ERRa levels in the brains of 4-week-old immature female but not male mice that had been treated with a chemical known to reduce mitochondrial function [73], suggesting sex differences in ERRa function prior to the production of circulating hormone production. De Jesus-Cortez et al. found that ERRa deficient adult female mice had defects in neural function in a mouse model of eating disorders, which mainly affect women, which was not observed in ERRa deficient male mice, and they concluded that ERRa was required for optimal mitochondrial function in females [74]. Watson et al. found sex-specific effects of ERRa expression in the hearts of female but not male mice in a model of heart failure [75]. We are the first to report sex differences in ERRα expression in the hearts of healthy mice and mice with viral myocarditis. Subsequent studies are needed to further characterize sex-specific effects of ERRα on mitochondrial function during CVB3 myocarditis. However, to fully characterize the sex differences in ERRa effects on gene regulation in healthy and mice with myocarditis, an analysis of gene-specific transcription factor (TF)-DNA interaction of ERRa is needed using chromatin immunoprecipitation (ChIP) or similar methods.