The World Health Organization has reported about 214 million new cases of malaria with about 438,000 deaths worldwide in 2015 (WHO, 2015). A vaccine for an effective and reliable anti-malaria prophylaxis is not yet available, despite enormous efforts during the last 35 years (Halbroth and Draper, 2015; Hoffman et al., 2015; Miura, 2016). Until recently, the RTS,S/AS01 has been regarded as the most advanced vaccine candidate. Indeed, the European Medicines Agency has approved RTS,S/AS01 for immunizations of very young children, however the WHO has not recommended its inclusion in the Expanded Programme of Immunizations (Birkitt, 2016; Gosling and von Seidlein, 2016). This points to the urgency of more basic research to uncover those mechanisms of the host defense, which a vaccine has to activate to provide protection against malaria.

The experimental malaria P. chabaudi in mice shares several characteristics with P. falciparum, the most dangerous malaria species for humans (Longley et al., 2011; Stephens et al., 2012). It is a convenient model to study blood-stage malaria. The outcome of P. chabaudi blood-stage malaria is controlled by genes of the mouse-MHC, the H-2 complex, genes of the non-H-2 background, and environmental factors such as testosterone (Wunderlich et al., 1988d). Moreover, there is a vaccination technique available by which survival of mice, being naturally susceptible to malaria, can be raised from 0% to over 80% (Wunderlich et al., 1988a; Krücken et al., 2009). This technique, which uses a non-infectious vaccine consisting of surface membranes isolated from P. chabaudi-infected erythrocytes containing parasite-synthesized proteins (Wunderlich et al., 1988b,c; Fontaine et al., 2012), converts lethal blood-stage infections to take a self-healing course (Krücken et al., 2009). Upon challenging with 10⁶ P. chabaudi-parasitized erythrocytes, all non-vaccinated mice succumb to malaria during the crisis phase of infection, whereas the majority of vaccinated mice survive the infections and generate long-lasting immune mechanisms against homologous re-challenge (Krücken et al., 2009).

The spleen is currently envisioned as the key effector organ against blood-stage malaria (Engwerda et al., 2005; Del Portillo et al., 2012). Indeed, the spleen is equipped with a unique system to eliminate senescent and aberrant erythrocytes including P.-infected erythrocytes. However, there is accumulating evidence that, also the liver, in particular the erythrophagocytotic Kupffer cells are able to remove senescent and other aberrant erythrocytes including P.-infected erythrocytes (Terpstra and van Berkel, 2000; Otogata et al., 2007; Delic et al., 2010; Lee et al., 2011; Wunderlich et al., 2014). In self-healing infections of P. chabaudi, it has been described that the liver dramatically increases its uptake of injected particles at peak parasitemia and during subsequent crisis, whereas, concomitantly, the spleen dramatically declines its uptake of injected particulate material including fluorescently labeled P. chabaudi-infected erythrocytes (Krücken et al., 2005, 2009). It is therefore rather likely that the liver, rather than the spleen, massively removes P. chabaudi-infected erythrocytes from circulation during crisis in vaccination-induced self-healing infections of P. chabaudi (Wunderlich et al., 2014). Very recently, the liver has been even described as the primary organ for a unique mechanism of rapid on-demand erythrocyte disposal (Theurl et al., 2016).

Current evidence demonstrates that miRNAs are critical for maintaining phenotype and functionality of the liver (Chen and Verfaillie, 2014), and that diverse diseases of the liver are associated with dysregulations in the miRNA signature (Szabo and Bala, 2013; Bandiera et al., 2015). MicroRNAs are non-coding single-stranded RNAs of 22 nts in length, which posttranscriptionally regulate gene expression by affecting both stability and translation of mRNAs. At least one third of human genes are estimated to be targeted by miRNAs (He and Hannon, 2004). MicroRNAs are involved in regulating epigenetic networks within cells and via exosomes between cells (Vyas and Dhawan, 2016). They are increasingly recognized as to be critical for the outcome of diverse infectious diseases (Bettencourt et al., 2016; Das et al., 2016; Verma et al., 2016). Also, miRNAs appear to be critical for efficacy of vaccination (Corral-Fernández et al., 2016; Shim et al., 2016; Wang et al., 2016). However, there is still little information available about host miRNAs in malaria, especially with respect to course and outcome of malaria infections and vaccination efficacy (Cohen et al., 2015; Rubio et al., 2016). At least, it is known that miRNAs of the liver are responsive to blood-stage infections of P. chabaudi malaria (Delic et al., 2011). The aim of this paper is therefore to analyze as to whether protective vaccination against P. chabaudi malaria leads to changes in the overall expression of miRNAs in the liver during self-healing infections of P. chabaudi in comparison to lethal infections in non-vaccinated female Balb/c mice.

To explore possible changes of hepatic miRNome by protective vaccination and/or subsequent challenge infections with P. chabaudi blood-stage malaria, microarrays were used to analyze miRNA expression in livers from vaccination-protected (V) and non-protected non-vaccinated mice (N) before infection on day 0 p.i. (Vd0, Nd0), and during different phases of the infections, i.e., at the beginning of infection on day 1 p.i. (Vd1, Nd1), during the mid-precrisis phase on day 4 p.i. (Vd4, Nd4), at peak parasitemia on day 8 p.i. (Vd8, Nd8), and toward the end of the crisis phase on day 11 p.i. (Vd11, Nd11). The parasitemia is about the same in self-healing infections of vaccination-protected mice and in lethal infections of non-protected mice (Figure 1). Only at peak parasitemia, there is a significant lower parasitemia in self-healing infections (Figure 1). Figure 2 shows the heat map of miRNA expression profiles during self-healing and lethal infections in vaccination-protected and non-protected mice, respectively. The profiles on day 0, 1, and 4 p.i. cluster separately from those on day 8 and 11 p.i., respectively, which are closer together.

Among the overall detectable 1247 miRNA species on a given array, 71 miRNA species are found to be robustly expressed (>100 light units) in the liver of both vaccinated and non-vaccinated mice. The most abundant expressions show the miRNAs miR-122-5p, miR-6366, miR-3963 and miR-5100, while relatively low expression is observed for the miRNAs miR-1196-5p, miR-468-3p and miR-669n. None of these miRNA species is affected by protective vaccination per se, evidenced as the expressions of the 71 miRNAs in vaccinated non-infected mice in comparison with the corresponding constitutive expressions in non-vaccinated mice. Upon infection with P. chabaudi, however, 65 miRNA species change their expression pattern, and 41 miRNA species out of the 65 malaria-responsive miRNAs significantly (p < 0.05) differ with respect to their expression levels in the liver between vaccinated and non-vaccinated mice. However, this difference is only evidenced toward the end of the crisis phase on day 11 p.i..

The lethal outcome of P. chabaudi malaria is associated with either increasing or decreasing expressions of miRNAs in the liver during the infections, reaching their minimal or maximal expression levels toward the end of the crisis phase on day 11 p.i. (cf. also Figures 3, 4). Decreasing expressions reveal the 18 miRNA species miR-122-5p, miR-142-3p, miR-148-3p, miR-26a-5p, miR-27a-5p, miR-29b-3p, miR-2861, miR30a/c-5p, miR-3968, miR-5097, and seven members of the let miRNA family (Figure 3), whereas increasing expressions occur for the 14 miRNAs miR-188-5p, miR1187, miR-1196-5p, miR-211-3p, miR-32-3p, miR-3082-5p, miR-3960, miR-466i-5p, miR-468-3p, miR-574-5p, miR-669n, miR-709, mir-5126, and miR-6538 (Figure 4). In vaccination-induced self-healing infections, however, the decreasing expression of the 18 miRNAs is stopped or even slightly turned to be up-regulated only by the end of crisis (Figure 3), whereas the increasing expression of the 14 miRNA species is impaired or even turned down (Figure 4). The up- or down-regulated expressions amount to more than 50% at the end of the crisis phase between vaccination-induced self-healing infections and lethal infections in non-vaccinated mice.

Tables 1, 2 summarize the annotated functions of the 18 and 14 miRNA species, whose expression is either up- or down-regulated in the liver of vaccination-protected mice toward the end of the crisis phase of P. chabaudi infections in relation to non-vaccinated mice, respectively. Remarkably, all the seven up-regulated members of the let-7 miRNA family are known to be involved in liver regeneration.

To further substantiate these differences in miRNA expression toward the end of the crisis phase, we have also performed quantitative PCR of arbitrarily selected miRNAs. Figure 5 shows that the miRNAs let-7a-5p, let-7c-5p, miR-122-5p, miR-142-3p, miR-29b-3p, and miR-30c-5p are up-regulated and the miRNA species miR-669n and miR-709 are down-regulated in livers of vaccination-induced self-healing infections on day 11 p.i. in comparison to lethal infections in non-vaccinated mice, thus confirming our microarray analyses. By contrast, the expressions of the malaria-responsive miR-92-3p and miR-126-3p are not significantly affected by infections, as it was also found by microarrays.

This is the first study providing evidence that protective vaccination against blood-stage malaria of P. chabaudi leads to changes in the expression of distinct miRNA species in the liver during subsequent homologous challenge infections. These changes, however, are evidenced only toward the end of the crisis phase on day 11 p.i. in relation to lethal infections in non-vaccinated mice. Lethal malaria induces increasing expressions of 18 miRNA species and decreasing expressions of 14 miRNA species during the precrisis phase reaching highest and lowest expression levels toward the end of the crisis phase. In vaccinated mice, however, the course of the decreasing and increasing expressions of the 18 and 14 miRNA species, respectively, is apparently stopped at peak parasitemia and even partly reversed during crisis, resulting in changes by more than 50 % in relation to non-vaccineted mice toward the end of the crisis phase. This suggests that vaccination-induced self-healing infections of P. chabaudi blood-stage malaria are associated with a remodeling of miRNA-controlled processes in the liver of Balb/c mice, which otherwise would lead to a lethal outcome of the infections during crisis.

Such a remodeling is presumably associated with an accelerated liver regeneration. Indeed, seven members of the let-7 miRNA family, found here to be up-regulated in the liver of vaccination-protected mice during crisis, have been all described to be involved in liver regeneration by targeting mRNAs of dicer and trbp (Forman et al., 2008; Shu et al., 2011). Also, the mir-26a up-regulated during crisis is known to be involved in liver regeneration and hepatocyte proliferation (Zhou et al., 2016). Moreover, the down-regulated expression levels of miR-1187 may reflect diminished hepatocyte apoptosis through targeting casp-8 (Yu et al., 2012). Furthermore, the down-regulated levels of miR-669n during crisis may conceivably contribute to accelerated liver regeneration by dampening malaria-induced Kupffer cell activation (Long et al., 2015). All these data support the view that liver regeneration may be accelerated in vaccination-protected mice during the crisis phase of self-healing infections of otherwise lethal P. chabaudi malaria. Accelerated liver regeneration in vaccinated mice is presumably not a direct effect of vaccination, but rather it may be associated with the process of self-healing, i.e., it may be an indirect consequence of the vaccination-induced conversion of lethal to self-healing infections of P. chabaudi.

Accelerated liver regeneration may contribute to overcome dysfunctions of the liver, presumably due to injuries induced by the P. chabaudi blood-stage infections. Indeed, the acute phase of P. chabaudi malaria has been described to be associated with pathological damages and heavy injuries in the liver with distant effects on other organs, as e.g., hepatoencephalopathy (Wunderlich et al., 2005; Delic et al., 2010). Even human patients suffering from malaria with P. falciparum and P. vivax have been shown to be associated with massive dys-functions of the liver (Ananad et al., 1992; Kochar et al., 2003; Nautyal et al., 2005; Rupani and Amarapurkar, 2009). Moreover, it is known that pathogenesis and diseases of the liver are in general associated with dys-regulated expressions of diverse miRNAs (Szabo and Bala, 2013; Chen and Verfaillie, 2014; Bandiera et al., 2015; Murakami and Kawada, 2016). For instance, injuries of the liver in HIV/HCV patients suffering from necroinflammation and portal hypertension have been recently reported to cause elevated levels of the miR-122, the most abundant miRNA species in the liver (Jansen et al., 2015). Our data, however, reveal continuously decreasing levels of miR-122-5p in the liver of both vaccinated and non-vaccinated mice during the precrisis phase of self-healing and lethal P. chabaudi infections, respectively. Only during the crisis phase, vaccination-protected mice exhibit elevated levels of miR-122-5p, whereas the corresponding levels further decrease in non-vaccinated mice with lethal infections. Similar courses of miRNA expression levels are here found for miR-30a known to be involved in manifestation and resolution of liver fibrosis (Roy et al., 2015), for miR-27a described to be involved in fibrosis (Cui et al., 2016), and for mir-29b known to suppress transcription of genes encoding for extracellular matrix proteins (cf. Table 1) (Roderburg et al., 2011; Lambrecht et al., 2015; Kitano and Bloomston, 2016). Thus, the dys-regulations of these different miRNA species may presumably due to the liver injuries induced by P. chabaudi blood-stage malaria occurring during precrisis in both vaccination-protected and non-vaccinated mice. However, the up-regulated levels of mir-122-5p, miR-30a, mir-27a, and miR-29, observed in vaccination-protected mice during crisis, may contribute to the accelerated liver regeneration suggested to occur in these mice.

Moreover, the present study shows that the expression level of mir-142-3p is significantly up-regulated during crisis by more than 100% in the liver of vaccinated mice as compared with non-vaccinated mice, i.e., this percental difference during crisis is by far higher than those found for all other malaria-responsive miRNA species in the liver. Recent evidence indicates that the miRNA-142 locus is important for the regulation of macrophage-related processes, as e.g., macrophage and dendritic cell differentiation (Fordham et al., 2015), regulation of cell migration (Kim et al., 2015), control of profibrogenic macrophage program (Su et al., 2015), role in colony-stimulating factor 1-induced monocyte differentiation into macrophages (Lagrange et al., 2013), prevention of macrophage differentiation during cancer-induced myelopoiesis (Sonda et al., 2013). The relevance of these data for the liver of vaccinated mice remains to be shown. Currently, however, the most unequivocal evidence demonstrates that mir-142-3p is critical for megakaryopoiesis (Chapnik et al., 2014). The miRNA-142-3p stringently controls specific cytoskeletal rearrangements required for maturation and function of megakaryocytes, and genetic deletion of miR-142-3p results in impaired megakaryocyte ablation, cytoskeletal dys-integrity, abnormal proplatelet formation and thrombocytopenia (Chapnik et al., 2014). In accordance, thrombocytopenia also occurs in experimental malaria including P. chabaudi malaria (Watier et al., 1992; Piguet et al., 2000, 2002; Gramaglia et al., 2005) and even represents a severe complication in human malaria caused by P. vivax and P. falciparum (Ansari et al., 2009; Lacerda et al., 2011; Gill et al., 2013; Gupta et al., 2013). Thus, our data showing up-regulation of mir-142-3p may suggest that protective vaccination affects hepatic megakaryopoiesis induced by P. chabaudi blood-stage malaria during crisis.

On the other hand, however, there is evidence that miR-142-3p is also involved in the regulation of erythropoiesis (Muhseen and Abbood, 2014). Erythropoiesis in turn is tightly coupled with megakaryopoiesis. Both erythropoiesis and megakaryopoiesis are controlled by the two transcription factors GATA1 and GFI1b (Fatica et al., 2006; Crispino and Weiss, 2014). Moreover, both megakaryocytic and erythroid cells are derived from common myeloid precursors (Fatica et al., 2006), and there exist even bipotent progenitor stem cells generating both erythroid and magakaryoctic cells (Debili et al., 1996; Papayannopoulou et al., 1996; Papayannopoulou and Kaushansky, 2016). However, there is also information available that erythropoiesis is enhanced at the expense of megakaryopoiesis (Bianchi et al., 2015). Under identical experimental conditions as described here, Al-Quraishy et al. (2016) have recently found that protective vaccination augments P. chabaudi-induced extramedullary hepatic erythropoiesis. It appears therefore plausible to assume that the up-regulation of mir-142-3p observed here during crisis is involved in enhanced erythropoiesis at the expense of megakaryopoiesis in the liver of vaccination-induced self-healing infections of P. chabaudi malaria (Al-Quraishy et al., 2016).

Collectively, our data suggest that vaccination-induced self-healing and survival of otherwise lethal blood-stage infections of P. chabaudi malaria activates epigenetic miRNA-regulated mechanisms in the liver, which accelerate liver regeneration and enhance hepatic erythropoiesis, possibly at the expense of megakaryopoiesis, during crisis of P. chabaudi blood-stage malaria.

MD, SA, and FW designed the study; MD, AA, and DD carried out the experiments and analyzed the data. All authors wrote and revised the manuscript.