Curcumin improves atrial fibrillation susceptibility by regulating tsRNA expression in aging mouse atrium

Animals

All procedures and protocols in this study have been approved by the animal experiment ethics committee of Guangdong Provincial People’s Hospital, Guangdong Academy of Medical Sciences (No. GDREC2016128A). All animals received care in compliance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication no. 85-23, revised 1996). Male C57BL/6 mice were purchased from Guangzhou University of Traditional Chinese Medicine Laboratory Animal Center under the license number SCXK2018-0034. The mice were housed at 22 ± 2 °C with a humidity of 55 ± 10% and a 12 h light/dark cycle with adibitum access to food and water. They were randomly divided into three groups, namely the 5 month group, the 18 month group, and the 18 month + curcumin group (n = 6). Curcumin (100 mg/kg/d) was administered once daily from the age of 12 to 18 months via food feeding. To reduce the pain during the experiment, the mice were sacrificed by cervical dislocation after anesthesia by intraperitoneal injection of pentobarbital sodium, and the hearts were dissected to collect the atrial appendages, which were immediately placed into a liquid nitrogen tank, and then transferred to a freezer at −80 °C for later inspection.

Rapid transjugular atrial pacing

The mice were weighed, anesthetized by intraperitoneal injection of 1% pentobarbital sodium (60 mg/kg), and fixed on a heating pad (RWD Life Science Company, Shenzhen, China) at a constant temperature (37 ± 1) °C. At least six mice from each group were tested. Surface ECG and intracavitary ECG were recorded using the iWorx system (Dover, NH, USA). A sheath was placed in the right jugular vein through a small incision, and the electrode catheter was inserted along the sheath to the right atrium. After recording the intracavity electrogram, rapid pacing of the endocardium was performed. The pacing threshold was measured and the atrial burst stimulation (short burst electrical stimulation) mode was used to induce AF and record its duration, sinus node recovery time (SNRT), and other indicators. On the ECG, AF is characterized by the disappearance of P waves and the appearance of f waves; the restoration of sinus rhythm is denoted by the recurrence of P waves and the disappearance of f waves. The duration of AF was defined as the time from AF occurrence to AF termination.

After completing the electrophysiological experiment, the mice were sacrificed by cervical dislocation after deep anesthesia. The hearts were dissected, and the atrial appendages were collected for histological examination and sequencing.

PANDORA-seq

Small RNA annotation

Small RNA sequences were annotated using the pipeline SPORTS (a small non-coding RNA annotation pipeline optimized for rRNA and tRNA-derived small RNAs, https://github.com/junchaoshi/sports1.1). Small RNA annotation was performed for annotating miRNAs, tRNAs, rRNAs, and other small non-coding RNAs from TIGR, miRBase, tRFdb, MINTbase, and piRBase. Subsequently, the expression difference of sncRNAs between the 5, 18 and 18 month+curcumin group was analyzed. TsRNA reads were normalized to RPM (reads per million). Differentially expressed tsRNAs were detected by using the R package edgeR. TsRNAs were considered significantly differentially expressed when the P-value was <0.05 and |log₂-fold change| ≥ 1.

Target mRNA prediction and correlation analysis of differential tsRNA

To verify the reliability of the PANDORA-seq, we numbered tsRNAs containing the ends in turn and randomly selected four numbers. Results extracted four tsRNA containing ends as mature-mt_tRNA-Val-TAC_CCA_end, mature-mt_tRNA-Glu-TTC_CCA_end, and mature-tRNA-Asp-GTC_CCA_end. To further reveal the functions of differential tsRNAs, we performed a bioinformatics analysis of these selected tsRNA. miRanda (http://www.microrna.org) and RNAhybrid (https://bibiserv.cebitec.uni-bielefeld.de/) were used to predict differentially expressed target mRNAs (screening criteria: miRanda, score ≥150, energy <−20). The miRanda score is the combined score of tsRNA and its target gene; a higher score indicates more accurate target gene prediction. Based on the regulatory relationship between tsRNA and predicted target genes, the tsRNA and target gene network were constructed using Cytoscape software (version 3.7.2, https://cytoscape.org/). Furthermore, enrichment analysis was performed on the target genes of tsRNAs on metascape (http://metascape.org/). Smaller P values indicated more significant effects, and terms with P < 0.05 were considered statistically significant.

Cell culture and mock transfection

Mouse atrial fibroblasts (MAF) were used in the study. The cells were obtained from the atrial tissues of 12 mice and were isolated by 50 mL 0.25% trypsin-EDTA (GIBCO, Grand Island, NY, USA) digestion. The cells were then cultured at 37 °C in DMEM F12 medium (GIBCO, Grand Island, NY, USA) containing 10% FBS, 100 U/ml penicillin, and 100 U/ml streptomycin. In order to induce senescence, passage 4 (P4) MAF was treated with 100 μM Tert-Butyl hydroperoxide (TBHP) for 2 h, then replaced with normal medium, and cultured for 2 days, while the control group was treated with PBS. To knock down and overexpress mature-mt_tRNA-Val-TAC_CCA_end, cells were seeded into 6-well plates 24 h before transfection with Lipofectamine 3000 (Invitrogen, Waltham, MA, USA), which was performed according to the manufacturer’s instructions. A negative control (GenePharma, Shanghai, China) at a concentration of 1 nM and tsRNA mimics or inhibitors (GenePharma, Shanghai, China) at a concentration of 1 nM were used in transfection. Eight hours after transfection, the cells were used for further experiments.

Tissue and cell senescence β-gal staining

The kit (#9860) from Cell Signaling Company was used for SA-β-gal staining according to the manufacturer’s instructions. The activity of β-galactosidase, a marker of senescent cells, was detected in a solution with a pH value of 6. Senescent tissues/cells show a green stain under acidic conditions.

ROS detection

After transfection or treated with TBHP, MAFs were incubated with 10 μM DCFH-DA (Beyotime, Shanghai, China) at 37 °C for 20 min, then washed three times with a serum-free cell culture medium. The intracellular ROS content, the green fluorescence intensity in the cells was detected using a fluorescent confocal microscope under 488 nm excitation wavelength and 525 nm emission wavelength. The stronger the fluorescence, the higher the ROS content.

Validation by qRT-PCR

qRT-PCR was performed on mouse atrial tissue (n = 6) and atrial fibroblasts (n = 3) to validate tsRNA. All experiments were performed in biological replicates, each in technical triplicates for the groups. AG RNAex Pro RNA Reagent (Accurate Ciotechnology, Hunan, China) was used to extract the total RNA of the atrium. Total RNA was reverse transcribed using the miRNA 1st Strand cDNA synthesis kit (Accurate Ciotechnology, Hunan, China). qRT-PCR was performed using 10 μl 2×SYBR Green Pro TaqHS Premix II (Accurate Ciotechnology, Hunan, China), cDNA (40 ng), 0.8 μl miRNA-specific (10 μM), 0.8 μl miRNA qRT-PCR 3′ primer (10 μM) and RNase free water, with a final volume of 20 μl. The reaction mixture of the sample was incubated in the StepOne real-time PCR system (Applied Biosystems, Foster City, CA, USA) at 95 °C for 10 min, 95 °C, 15 s, 60 °C, 1 s, and finally at 72 °C for 30 s, for 35 PCR cycles. Normalization was performed using U6 small nuclear RNA. All analyses were performed with the 2^−ΔΔCq method. The primer sequences used for qRT-PCR are listed in Table 1.

Statistical analysis

Results are expressed as mean ± standard error of the mean (SEM). Statistical analysis was performed by SPSS 23.0 software. Differences between the two groups were determined using Student’s t-test. One-way analysis of variance (ANOVA) was used for comparison among the three groups. P < 0.05 was defined as statistically significant.

Curcumin reduced susceptibility to AF by attenuating atrial aging and inflammation in aged mice

Rapid atrial pacing was performed through the jugular vein to observe the inducible rate of AF (Fig. 1A) and the electrophysiological characteristics of mice (Table 2). Compared with the 5 month mice, a significantly prolonged PR interval was observed in 18 month mice (P < 0.05), and the sinus node recovery time (SNRT) and rate-corrected sinus node recovery time (cSNRT) were also significantly increased (P < 0.05). Curcumin treatment significantly improved the above indicators. Furthermore, the high induction rate of AF in mice in the 18 month group was significantly improved by curcumin (100 mg/kg/d) treatment (Fig. 1B). SA-β-gal staining showed greater green staining in the atrium of 18 month mice than in the 5 month group, suggesting a significantly higher degree of aging in aged mice, which could be improved by curcumin treatment (Fig. 1C). In addition, qRT-PCR showed that curcumin also reduced the inflammatory index IL-6/TNFα in the atrial tissue of 18 month mice (Fig. 1D). Therefore, these results suggest that curcumin might ameliorate the atrial tissue aging through decreasing the inflammatory response, thereby reducing the susceptibility of AF in aged mice.

Exemplary annotation and profiling of PANDORA-seq datasets generated by SPORTS1.1

SPORTS1.1 was used to annotate and analyze the PANDORA-seq data sets, and the sncRNA expression profiles of the atrium tissue in the 5, 18, or 18 month mice treated with 100 mg/kg/d curcumin were shown in Fig. 2. It displayed the length distribution for sncRNAs. A large number of modified sncRNAs were identified, including miRNAs, piRNAs, tsRNAs, rsRNAs, and some unannotated sncRNAs, whose lengths were concentrated in the range of 15–45 nt. However, tsRNA and rsRNA accounted for the largest proportions of these sncRNAs. The role of sncRNA in the aging process was speculated to be mainly mediated by tsRNA and rsRNA, so further experiments were performed on tsRNA.

General characteristics of tsRNAs expression

The tsRNA expression profiles were annotated and analyzed. According to the source of tRNA region, tsRNA can be divided into four groups: 3′ end, 5′ end, CCA end, and other tsRNAs (Figs. 3A–3C). In addition, tsRNAs are mainly derived from tRNAs of glycine (Gly), serine (Ser), valine (Val), and glutamine (Glu) (Figs. 3D–3F). Compared with the 5 month group, a higher total tsRNA expression was observed in the 18 month group, while a reversed trend was found in the 18 month + curcumin group with a different expression profile (Figs. 3A–3C). Moreover, significant differences in tsRNA origin sites were found among the three groups. Compared with the 5 month group, the total tsRNA expression in the 18 month group was up-regulated, while curcumin treatment reversed this trend and changed the expression profile (Figs. 3G–3I).

TsRNA expression difference and target gene prediction

Hierarchical clustering heatmaps and volcano plots were used to visualize tsRNA divergence (Figs. 4A–4C). The 5 month group was compared to the 18 month group, yielding 473 significantly altered tsRNA sequences (212 down-regulated, 261 up-regulated), from 127 kinds of tsRNA species (P < 0.05, |log₂FC| ≥ 2), while 947 tsRNA sequences were significantly altered between the 18 month group and the 18 month+curcumin group (647 tsRNAs down-regulated, 300 up-regulated), from 154 kinds of tsRNA species (P < 0.05, |log₂FC| ≥ 2). Overall, 99 common differentially expressed tsRNA sequences (12 kinds of tsRNA species) between the pairwise comparisons in 5, 18 and 18 month + curcumin groups, of which 52 tsRNA sequences from mt-tRNAs and 47 tsRNA sequences from tRNAs (Fig. 4D, Table 3).

To reveal the function of tsRNAs, we predicted potential target mRNAs of tsRNAs (mature-mt_tRNA-Val-TAC_CCA_end, mature-mt_tRNA-Thr-TGT_5_end, mature-mt_tRNA-Glu-TTC_CCA_end, mature-tRNA-Asp-GTC_CCA_end) using miRanda and RNAbrid databases. A total of 102 mRNAs were associated with the 4 tsRNAs. The data was then acquired into Cytoscape to create a brand-new tsRNA-mRNA network to visualize and showed that one tsRNA might be associated with 0 or more mRNAs. Mature-mt_tRNA-Val-TAC_CCA_end was matched with 46 mRNAs, mature-mt_tRNA-Glu-TTC_CCA_end was matched with 17 mRNAs and mature-tRNA-Asp-GTC_CCA_end was matched with 39 mRNAs (Fig. 4E). It was noteworthy to mention that mature-mt_tRNA-Thr-TGT_5_end could not match with target mRNAs (Fig. 4E). Furthermore, enrichment analysis was performed on these tsRNA target genes on metascape (http://metascape.org/), revealing that the target genes of these 4 tsRNAs were mainly involved in DNA damage, protein modification, nutrient metabolism, and other processes, including DNA damage response (U-WP707), R-SUMOylation of transcription cofactors (R-HSA-3899300), and 2-Oxocarboxylic acid metabolism (hsa01210), etc., (Fig. 4F).

qRT-PCR verification of differential tsRNA and target genes

To verify the reliability of the PANDORA-seq, we confirmed the expression of tsRNAs and their target genes in these three groups using qRT-PCR. We found that, compared with the 5 month group, the expression of mature-mt_tRNA-Thr-TGT_5_end was down-regulated, and mature-mt_tRNA-Val-TAC_CCA_end, mature-mt_tRNA-Glu-TTC_CCA_end, and mature-tRNA-Asp-GTC_CCA_end were up-regulated in the 18 month group, which could be improved by the treatment of curcumin (Figs. 5A–5D). Their relative expression levels were the same as those obtained by random sequencing. Compared with the 5 month group, the expression of mature-mt_tRNA-Val-TAC_CCA_end core target genes (Smad3, Mkl1, Ndufs6, Mrpl27) was increased in the 18 month group, which was reversed by curcumin treatment excepting for Mrpl27 (Figs. 5E–5H).

Mature-mt_tRNA-Val-TAC_CCA_end regulated the celluar senecence and the levels of ROS and inflammatory factors

TsRNA mimics and inhibitors were transfected into MAF to verify the target mRNA expression levels of tsRNAs. Subsequently, mature-mt_tRNA-Val-TAC_CCA_end was selected for further verification as it contained the most target genes. We found that mature-mt_tRNA-Val-TAC_CCA_end mimic treated group had more senescent cells, increased ROS levels and IL-6/TNFα expression compared with the control group (Figs. 6A–6C). The findings indicate that overexpression of mature- mt_tRNA-Val-TAC_CCA_end can induce cellular senescence, oxidative stress, and elevated levels of inflammation. Compared with the TBHP group, the inhibitor group showed fewer senescent cells, decreased ROS levels and IL-6/TNFα expression, suggesting that inhibiting mature-mt_tRNA-Val-TAC_CCA_end can reduce the ratio of senescent cells, oxidative stress, and inflammation levels (Figs. 7A–7C). Moreover, increased expression of the core target genes Smad3, Ndufs6, Mkl1, and Mrpl27 was observed in the mimic group, while decreased gene expression in the inhibitor group compared with the control group excepting for Mrpl27. These results suggest that mature-mt_tRNA-Val-TAC_CCA_end may regulate Smad3, Ndufs6, Mkl1, and Mrpl27, which are involved in the process of aging, fibrosis, oxidation and inflammation.