Body temperature-dependent microRNA expression analysis in rats: rno-miR-374-5p regulates apoptosis in skeletal muscle cells via Mex3B under hypothermia

Forensic diagnosis of fatal hypothermia is considered difficult because there are no specific findings. Accordingly, exploration of novel fatal hypothermia-specific findings is important. To elucidate the molecular mechanism of homeostasis in hypothermia and identify novel molecular markers to inform the diagnosis of fatal hypothermia, we focused on microRNA expression in skeletal muscle, which plays a role in cold-induced thermogenesis in mammals. We generated rat models of mild, moderate, and severe hypothermia, and performed body temperature-dependent microRNA expression analysis of the iliopsoas muscle using microarray and quantitative real-time PCR (qRT-PCR). The results show that rno-miR-374-5p expression was significantly induced only by severe hypothermia. Luciferase reporter assay and qRT-PCR results indicated that Mex3B expression was regulated by rno-miR-374-5p and decreased with decreasing body temperature. Gene ontology analysis indicated the involvement of Mex3B in positive regulation of GTPase activity. siRNA analysis showed that Mex3B directly or indirectly regulated Kras expression in vitro, and significantly changed the expression of apoptosis-related genes and proteins. Collectively, these results indicate that rno-miR-374-5p was activated by a decrease in body temperature, whereby it contributed to cell survival by suppressing Mex3B and activating or inactivating Kras. Thus, rno-miR-374-5p is a potential supporting marker for the diagnosis of fatal hypothermia.

Expression of target mRNAs was altered with decreasing body temperature. Abca8a, Ccnl1, Slc25a33, and Zfp423, which belong to "mitochondrial part", "cell differentiation", and "regulation of transcription" GO analysis terms (Table 2), are target gene candidates of rno-miR-374-5p. Expression of these genes was significantly altered with decreasing body temperature ( Fig. 2A-D). In particular, expression of Ccnl1 and Slc25a33 was significantly increased only by severe hypothermia compared with control, mild, and moderate hypothermia animals.
Expression of rno-miR-374-5p was significantly increased only by severe hypothermia, suggesting that these genes might not be controlled by rno-miR-374-5p under hypothermic conditions. In our study, rno-miR-30c-1-3p was downregulated in the severe hypothermia group compared with the control group (Supplementary Table 1). TargetScan predicted Slc25a33 as a target gene of rno-miR-30c-1-3p (https ://www.targe tscan .org/cgibin/targe tscan /vert_72/view_gene.cgi?rs=ENST0 00003 02692 .6&taxid =10116 &showc nc=0&shown c=0&shown c_nc=&shown cf1=&shown cf2=&subse t=1). This result indicates that Slc25a33 might be primarily regulated by rno-miR-30c-1-3p under hypothermic conditions. Accordingly, it is possible that in hypothermic conditions, factors such as other miRNAs promote the expression of these genes, and this mechanism is even more effective during severe hypothermic conditions. This issue will be a focus of future research. expression of Mex3B exhibited an inverse correlation with rno-miR-374-5p expression. Expression of Mex3B was significantly decreased by moderate and severe hypothermia compared with control and mild hypothermia animals (Fig. 3A). In addition, the level of Mex3B protein after moderate and severe hypothermia was significantly decreased compared with the other groups (Fig. 3B). Expression of rno-miR-374-5p was gradually increased with decreasing body temperature, suggesting that this gene might be controlled by rno-miR-374-5p under hypothermic conditions. rno-miR-374-5p directly regulated Mex3B translation in vitro. To verify that the mRNA we identified were bona fide targets of rno-miR-374-5p, we employed a luciferase reporter assay, qRT-PCR, and western blotting. rno-miR-374-5p is predicted to bind with high affinity to Mex3B. In luciferase reporter assays, a decrease in luciferase activity indicates binding of the miRNA mimic to the 3′-UTR of the target sequence. Luciferase reporter assay, qRT-PCR, and western blotting results showed that the rno-miR-374-5p mimic could effectively inhibit Mex3B expression (Fig. 3C,D); thus, we concluded that rno-miR-374-5p directly regulates Mex3B expression in vitro.
rno-miR-374-5p and Mex3B were involved in regulation of GTPase activity. As Mex3B is involved in the induction of apoptosis by cellular stress and belongs to the GO term "positive regulation of GTPase activ-  The statistical significance of differences between means was assessed by one-way ANOVA, followed by Tukey's multiple comparison test. *P < 0.05, **P < 0.01, ***P < 0.001.
Scientific RepoRtS | (2020) 10:15432 | https://doi.org/10.1038/s41598-020-71931-w www.nature.com/scientificreports/ ity" (Table 2), we focused on Ras, a GTPase that functions as a molecular switch for signalling pathways regulating cell survival, growth, and so on. We examined the expression of Kras, Pik3ca, Akt1, Bad, and Bcl2l1, which are involved in apoptosis according to the KEGG pathway. qRT-PCR using the SYBR Green I assay revealed that Kras expression in iliopsoas muscle increased with decreasing body temperature (Fig. 4A). Expression of Pik3ca was slightly increased with decreasing body temperature, but did not significantly change (Fig. 4B). Expression of Akt1 was not significantly changed ( Fig. 4C), but AKT1 expression significantly increased with decreasing body temperature (Fig. 4D). Expression of Bad significantly decreased with decreasing body temperature (Fig. 4E). In contrast, Bcl2l1 expression significantly increased with decreasing body temperature (Fig. 4F).
To examine whether Mex3B regulated Kras expression, we synthesised siRNA for Mex3B and transfected iliopsoas muscle cells. As a result, Mex3B siRNA induced decreased Mex3B expression, followed by inactivation of Kras expression (Fig. 4G). This result contradicted observed Kras expression in the animal model. When we co-transfected iliopsoas muscle cells with the rno-miR-374-5p inhibitor and Mex3B siRNA, or transfected only rno-miR-374-5p inhibitor, Kras expression showed the same expression pattern as Mex3B (Figs. 3D, 4H and Supplementary Fig. 1A). However, when we transfected iliopsoas muscle cells with the rno-miR-374-5p mimic, Mex3B expression decreased and was followed by an increase in Kras expression ( Fig. 3D and Supplementary  Fig. 1B), consistent with the observed expression pattern in the animal model. Accordingly, these results suggest that Kras expression might be regulated not only by Mex3B, but also indirectly by activated rno-miR-374-5p in iliopsoas muscle cells.

Discussion
In forensic practice, various biochemical, pathological, and molecular biological examinations are performed using blood and tissues collected from the corpse to assess the cause of death [8][9][10][11][12][13][14][15][16] . However, because corpses are often found with a postmortem interval, examinations often do not represent the state at the time of death. Accordingly, an index for diagnosis of cause of death that can withstand the postmortem interval is required. We have reported specific molecular markers using mRNA that support the diagnosis of fatal hypothermia, one of the pathologies for which the diagnosis of cause of death is difficult 15,16 . In a cold environment, mRNA is relatively stable, but a molecular marker that can withstand the postmortem interval is required. Recently, microRNA (miRNA) has been proposed as a molecule that can withstand the postmortem interval 18 . The iliopsoas muscle performs shivering-or non-shivering-thermogenesis to maintain body temperature when the body temperature declines extremely 23 , and haemorrhage of the iliopsoas muscle is observed in fatal hypothermia 3 . In our previous study, the iliopsoas muscle might have been activated because haemorrhage was observed in this muscle of rats with extreme hypothermia 16 . Therefore, we focused on miRNA expression in the iliopsoas muscle.
In this study, expression of rno-miR-190a-5p and rno-miR-374-5p in iliopsoas muscle was increased with decreasing body temperature. Previous reports suggest that the expression of both miRNAs fluctuates dynamically in response to hypoxia 24,25 . Hypothermic conditions induce tissue asphyxiation because blood is cooled Table 1. Microarray fold-change analysis. Expression levels of 17 mirnas were more than doubled in severe hypothermia compared with control (Ctrl), mild, and moderate hypothermia.

Symbol
Fold change (ctrl vs severe) Regulation Mirbase accession no. www.nature.com/scientificreports/ by the inhalation of low-temperature air, and oxygen and hemoglobin do not dissociate in a cold environment. Therefore, both miRNAs may be activated by tissue asphyxiation induced by severe hypothermia and, therefore, be potential markers of hypoxia in a cold environment. The expression of target gene candidates (Abca8a, Ccnl1, Slc25a33 and Zfp423) of rno-miR-374-5p increased with decreasing body temperature and showed similar expression to rno-miR-374-5p. These results indicate that these genes may not be regulated by rno-miR-374-5p. Abca8 is a sinusoidal efflux transporter for cholesterol and taurocholate in mouse and human liver 26 , but is function in hypothermia is still unknown. Zfp423 is involved in skeletal muscle regeneration and proliferation after injury 27 . Increased expression of Zfp423 may contribute the promotion of iliopsoas muscle repair. Expression of Ccnl1 and Slc25a33 was only increased by severe hypothermia. Inhibited expression of Ccnl1, a cell cycle regulatory gene, by forced expression of a specific miRNA suppressed cell proliferation and invasion, arrested cell cycle progression, and promoted cell apoptosis 28 . Hence, Ccnl1 may contribute to cell survival during cellular stress such as cold stimulation. Slc25a33 is reportedly involved in mitochondrial oxidative phosphorylation for ATP synthesis 29 . Accordingly, it is suggested that Slc25a33 expression might be induced by the promotion of thermogenesis accompanying a decrease in body temperature. Moreover, Ccnl1 and Slc25a33 represent novel findings that will potentially complement the diagnosis of fatal hypothermia because the expression of both genes was induced only by severe hypothermia.
Our study showed that significantly decreased Mex3B expression in moderate and severe hypothermia was directly regulated by rno-miR-374-5p. Mex3B is reportedly involved in the induction of apoptosis following cell stresses such as heat, cold shock, reactive oxygen species, and radiation 30 . Accordingly, overexpression of miR-374b-5p inhibited apoptosis and contributed to the elevation of cell survival 31 . Our results indicate that Mex3B directly or indirectly regulated Kras expression, which was accompanied by subsequent activation of downstream molecules such as Bcl2l1 in the Ras signalling pathway. Kras is a GTPase that functions as a molecular switch for signalling pathways regulating cell proliferation, survival, growth, and differentiation 32 . Bcl-2 family proteins regulate apoptosis through protein-protein interactions. Previous reports suggest that overexpression of Bcl-2 inhibits apoptosis and plays an important role in the development of inflammation-related disorders 33,34 . Accordingly, these results suggest activation of a novel rno-miR-374-5p/Mex3B/Kras pathway to elevate cell viability in extreme hypothermia.
In conclusion, our results indicate that rno-miR-190a-5p and rno-miR-374-5p may be activated by tissue asphyxiation induced by severe hypothermia and, thus, serve as potential markers of hypoxia in a cold environment. In addition, Ccnl1 and Slc25a33 may be upregulated by cellular stresses such as cold shock and the promotion of thermogenesis accompanying decreases in body temperature. Thus, these miRNAs and mRNAs are potential novel supporting markers for the diagnosis of fatal hypothermia. In addition, a novel rno-miR-374-5p/Mex3B/Kras pathway may be involved in the pathological process of fatal hypothermia. However, further investigation of these genes is necessary before they can be applied to forensic practice.

Materials and methods
All methods were performed according to relevant guidelines and regulation.
The Animal Care Committee of Nagasaki University approved this research protocol (approval number 1606081312-2).

Animals.
Pathogen-free 8-week-old male Wistar rats (300-350 g in body weight) were obtained from Charles River Laboratories (Yokohama, Japan). Prior to experiments, rats were housed for 1 week under a 12/l2-h light/ dark cycle (light on at 07:00 and off at 19:00) at a constant temperature and humidity, and allowed free access to food and water.  www.nature.com/scientificreports/ Experimental groups and thermal treatments. Twenty-three 9-week-old male Wistar rats were divided into four groups of 5-6 rats (control, mild, moderate, and severe hypothermia) and anesthetised by intraperitoneal injection of 0.3 mg/kg medetomidine, 2 mg/kg midazolam, and 2.5 mg/kg butorphanol. After 30 min, rats were exposed to cold (ambient temperature of 4 °C); rectal temperature (Ret) was continuously measured. Control rats were euthanised by cervical dislocation at 30 min after anesthesia. Mild hypothermia rats were euthanised by cervical dislocation when Ret reached 30 °C. Similarly, moderate and severe hypothermia rats were euthanised when Ret reached 22 °C and 12 °C, respectively 16 .

Scientific RepoRtS
Relative quantification of miRNA and mRNA transcripts was performed using the ∆∆Ct method 36 .
Total protein extraction and western immunoblot analysis. Iliopsoas muscles were homogenised using a TissueLyzer II (Qiagen). T-PER Reagent (Thermo Fisher Scientific), consisting of proteinase and dephosphorylation inhibitors, was then added. Debris was removed from the supernatant using an Ultrafree-MC 0.45mm filter (Merck Millipore, Darmstadt, Germany). Filtered protein samples were quantified using a Direct Detect Spectrometer (Merck Millipore), separated on 4-12% NuPAGE Novex Bis-Tris gels (Thermo Fisher Scientific), transferred to polyvinylidene difluoride membranes, and blotted according to standard protocols with rno-miR-374-5p mimic, mutation #2, rno-miR-374-5p inhibitor, or inhibitor control. Graphs show mean ± SD (n = 3). The statistical significance of differences between means was assessed by unpaired t test. *P < 0.05.
In situ hybridisation. In situ hybridisation (ISH) was performed using a microRNA ISH buffer set and miRCURY LNA Detection 5′-and 3′-DIG-labelled probes (Qiagen) in accordance with the manufacturer's instructions. In brief, 4% paraformaldehyde perfusion-fixed tissues were embedded in paraffin. Six-micrometerthick sections were deparaffinised and incubated with Proteinase K solution (DAKO, Glostrup, Denmark) for 10 min at 37 °C. After washing in phosphate-buffered saline, sections were dehydrated. Hybridisation was performed using 40 nM miRNA probe in microRNA ISH buffer (Qiagen) at 50 °C for 3 h. Sections were rinsed in 5 × SSC buffer at 50 °C for 5 min, twice with 1 × SSC buffer at 50 °C for 5 min, twice with 0.2 × SSC buffer at 50 °C for 5 min, and with 0.2 × SSC buffer at room temperature for 5 min. Sections were treated with blocking solution (Nacalai Tesque, Kyoto, Japan) for 15 min at room temperature and then incubated with an anti-DIG antibody (1:800; Roche Diagnostics, Basel, Switzerland) in blocking solution (Nacalai Tesque) overnight at 4 °C. Sections were developed using NTB/BCIP (Roche Diagnostics) at 30 °C. Observations were made using a BZ-9000 (Key- Anti-rabbit IgG HRP-linked whole antibody (GE Healthcare) 1:100,000