Identification of tRNA-Derived Small RNAs and their Potential Roles in the Hippocampus of Nicotine Exposure Rats


 Nicotine use is highly prevalent and brings a huge burden on individuals, society, health-care systems and economic development. The existing mechanisms underlying nicotine’ actions can’t illuminate all basic and clinical problems thoroughly. Transfer RNA-derived small RNAs (tsRNAs) is a novel class of small non-coding RNAs (sncRNAs), possessing potential regulatory functions in various diseases. However, the roles of tsRNAs in nicotine exposure have not been determined yet. In this study, firstly we established nicotine exposure model by subcutaneously injecting (sc.) with 0.5mg/kg of nicotine twice daily for 14 consecutive days, and conducted some behavioral observations (the pain threshold and body weight gains). Secondly, we identified the differentially expressed profiles of tsRNAs in rat hippocampus on saline or nicotine delivery conditions by using ncRNA-Seq, and then predicted the promising functions of the putative genes of the tsRNAs by bioinformatic method. The results shown that there were 26 differentially expressed tsRNAs (7 up-regulated and 19 down-regulated tsRNAs) (Fold change > 1.5; P < 0.05), of which the tRF-5 was the most common type. Eight tsRNAs were selected to validate the sequencing result by RT-qPCR. Then, based on the sequencing and RT-qPCR data, five candidate tsRNAs (tRF-1-T28-His-GTG-1, tRF-5c-Glu-CTC-1, tRF-5c-Glu-CTC-3, tRF-5c-Gly-GCC-2-M2, tRF-5c-Glu-TTC-4) were finally selected for further bioinformatic analysis. The GO and KEGG pathway enrichment analysis suggested that the five candidate tsRNAs might play regulatory roles through the cholinergic synapse pathways, dopaminergic synapse pathways, etc. In conclusion, our results indicated that tsRNAs were dysregulated in the rat hippocampus after nicotine exposure, and among them, tRF-5c-Glu-CTC-1 was the most promising one, which might lay a novel foundation for further research into nicotine’ actions.


Introduction
According to the 2019 National Health Interview Survey (NHIS), an estimated 50.6 million U.S. adults aged ≥18 years (20.8%) reported currently using any tobacco product. And Chinese CDC reported that in 2018, approximately 26.6% of people aged ≥15 years were current smokers in China [1,2]. Unfortunately, cigarette smoking remains the leading cause of preventable disease and death in the United States [3].
From 1964-2012, an estimated 17.6 million deaths were related to smoking, more than 480,000 deaths yearly [4]. Cigarette smoking is linked with diseases of nearly all organs of the body, such as various cancers, respiratory diseases, cardiovascular diseases, neurological disorders, etc. In addition to public health, smoking imposed a heavy burden on social economic. Experts estimated that the annual societal costs attributable to smoking in the United States were approximately $300 billion, including $168 billion for healthcare [5,6].
Nicotine, the primary component of tobacco, exerts multiple biological functions through which smoking increases risk for diseases. Nicotine exposure brings a series of changes in body, including stimulating the adrenal glands, activating the reward circuits, modulating neurotransmitters and receptors, and so on.
The most well-known mechanisms of nicotine' actions are the role of nicotinic cholinergic and dopaminergic system. In brief, nicotine can bind and activate nicotinic acetylcholine receptors (nAChRs), which are widely distributed throughout the nervous system and body, and thus result in the release of a variety of neurotransmitters (dopamine, gamma-aminobutyric acid (GABA), serotonin, glutamate etc) in the brain, particularly dopamine [7,8]. Besides, nicotine exposure can also produce additional neuroadaptations, including neuronal homeostatic mechanisms, neuronal scaffolding proteins expression and epigenetic regulations [9][10][11]. However, there is still much to be learned and done in order to fully understand how nicotine functions.
As the non-coding RNAs (ncRNAs) have been the hot topic in many areas in recent years, many researchers pay their attention to the possibility of ncRNAs in nicotine use. Chen et al reported that ncRNAs (miRNAs, lncRNAs and circRNAs) appear to form complex interactions and impact the nicotine biosynthesis in tobacco [12], and Aliso suggested that a pattern of miRNA/mRNA regulation occurs in the habenulo-interpeduncular circuit during nicotine withdrawal [13]. As the major source of small noncoding RNAs (sncRNAs), tRNA-derived small RNAs (tsRNAs) produced by cleavage at different sites from mature or pre-tRNAs are gradually becoming the emerging research direction [14,15]. tsRNAs were rst discovered in cancer patients until 1970s and at that time, they were considered to be merely degradation products of tRNAs and have no biological functions [16]. With the rapid development of high-throughput sequencing technology, tsRNAs were revealed to function under various diseases, such as cancer, neurodegenerative diseases, metabolic disorders, and others [15,17,18]. As to the speci c biological roles of tsRNAs, lots of researches have been done. For instance, Ivanov et al suggested that tsRNAs could inhibit translation by displacing eIF4G/A from mRNA [19]; Kim and his colleagues reported that LeuCAG3′tsRNA could modulate apoptosis by binding at least two ribosomal protein mRNAs (RPS28 and RPS15) [20]. In general, several regulatory mechanisms have been reported to date: 1) act like miRNAs to regulate gene expression, 2) regulate translation process, 3) function as epigenetic factors, 4) modulate apoptosis and immune response, etc [14,15,21]. Although tsRNAs have been the hot area in recent years, the majority of studies mainly focused on the roles of tsRNAs in cancer and immune [17,22,23], and there have no relevant researches about the relationship between tsRNAs and nicotine use or other addictive drugs thus far. Given its various roles under physiopathological conditions, it's no doubt that exploring tsRNAs will provide a bran-new insight into nicotine exposure research.
Therefore, the aim of our study is to identify the expression pro le and potential roles of tsRNAs in hippocampus after nicotine exposure. Using RNA-sequencing technologies, we identi ed the expression pro les of tsRNAs, and then predicted the potential roles of the candidate tsRNAs by bioinformatic methods. Our ndings might reveal the potential mechanisms and novel molecular targets underlying nicotine exposure, and provide a new direction for future research in nicotine use.

Materials And Methods
Animals and experiment groups 24 Adult male Sprague-Dawley rats (200±10g) were obtained from the Hunan SJA Laboratory Animal Company (Hunan, China). And then the rats were housed in groups and maintained for at least 5 days before the study in a temperature-(22±2℃) and humidity controlled (50% relative humidity) environment with 12-h light/dark cycles. The rats had food and water freely available. The experimental procedures were approved by the Animal Care and Use Committee of Hunan Cancer Hospital and conducted in strict accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
To establish the nicotine exposure model, the rats in nicotine group (NI, n = 12) were subcutaneously injected (sc.) with 0.5mg/kg (0.5 mg/ml) of nicotine twice daily at 08: 00-09:00 am and 16:00-17:00 pm for 14 consecutive days. The rats in normal saline group (NS, n = 12) were injected with the same volume of 0.9% saline at the same time points.
To analyze the effects of nicotine on sensation, hind paw mechanical withdrawal thresholds were measured every other day using von Frey laments test (Italy, UGO Basile) and the test was always performed between 13:00-16:00 pm [24]. The results were converted to the 50% of paw withdrawal threshold (50%PWT). The body weight was tested every day before the morning injection.

Tissue Collection and RNA isolation
On day 14, 1h after the morning injection, the rats in both groups were decapitated under deep anesthesia by 4% iso urane inhalation for 3 minutes. The hippocampus (n=6, each group) were quickly dissected and collected on ice and fresh frozen in liquid nitrogen. Total RNA was extracted from frozen tissues using Trizol (Invitrogen, CA, USA) according to the manufacturer's instructions. The quantity and concentration of each RNA sample were assessed using a NanoDrop ND-1000 Spectrophotometer (Thermo, USA) and the integrity was checked by agarose gel electrophoresis.

Library preparation and sequencing
The following treatments were performed before library preparation for total RNA samples to remove RNA modi cations that interfered with small RNA-seq library construction: 3'-aminoacyl (charged) deacylation to 3'-OH for 3adaptor ligation, 3'-cP (2,3 -cyclic phosphate) removal to 3'-OH for 3' adaptor ligation, 5'-OH (hydroxyl group) phosphorylation to 5'-P for 5'-adaptor ligation, m1A and m3C demethylation for e cient reverse transcription. Sequencing libraries were size-selected for the RNA biotypes to be sequenced using an automated gel cutter, which were quali ed and absolutely quantified using an Agilent Bio-Analyzer 2100 (Invitrogen, CA, USA). To generate the mature tRNA libraries, we removed the predicted intronic sequences (if present) and added an additional 3-terminal "CCA" to each tRNA. To generate the precursor tRNA libraries, we included 40 nucleotides of anking genomic sequence on either side of the original tRNA sequence.

Sequencing data analysis
Image analysis and base calling were performed using Solexa pipeline v1.8 software (Off-Line Base Caller software, v1.8). Sequencing quality was examined by FastQC software and trimmed reads (pass Illumina quality lter, trimmed 3-adaptor bases by cutadapt) were aligned to mature-tRNA and pre-tRNA sequence getting from the Genomic tRNA Database using Novo Align software (v2.07.11). The remaining reads were aligned to transcriptome sequences (mRNA/rRNA/snRNA/snoRNA/piRNA/miRNA). The abundance of tsRNAs were evaluated using their sequencing counts and normalized as counts per million of total aligned reads (CPM). The tsRNAs differentially expressed are screened based on the count value with R package edgeR. Pie plots, Venn plots, Hierarchical clustering, Scatter plots and Volcano plots are performed in R or perl environment for statistical computing and graphics of the expressed tsRNAs. The ncRNA-seq was performed by Kangcheng Bio-tech (Shanghai, China).

RT-qPCR validation
RT-qPCR was conducted to con rm the sequencing data. Eight differentially expressed tsRNAs were chosen for RT-qPCR. U6 small nuclear RNA (snRNA) was used as a reference. The RNA was reversely transcribed to cDNA using an rtStar™ tRF&tiRNA Pretreatment Kit (Arraystar, MD, USA) and an rtStar First-Strand cDNA Synthesis Kit (3' and 5' adaptor) (Arraystar, MD, USA) according to the manufacturer's instructions. Then, RT-qPCR was performed using 2XPCR master mix (Arraystar, MD, USA) with a ViiA7 Real-time PCR System (Applied Biosystems, CA, USA). The parameter settings were as follows: 95℃ denaturation (10 min), 40 ampli cation cycles at 95℃ (10 s), and 60℃ (60 s). After the ampli cation reaction was nished, the procedure was performed as follows: 95℃ (10 s), 60℃ (60 s), and 95℃ (15 s). All reactions were performed in triplicate. And the sequences of all primers are presented in supplementary table S1.

Bioinformatic Prediction
The ve signi cantly differentially expressed tsRNAs (Fold change > 1.5 and P-value < 0.05) selected according to the results of the sequencing and RT-qPCR were analyzed using bioinformatic methods.
According to a previous study [25], two common algorithms were used to predict the tsRNA targets, including TargetScan and miRanda. A graph of the tsRNA/mRNA network was derived using the Cytoscape software (version 3.5.1, the Cytoscape Consortium, San Diego, CA, USA) to visualize these relationships. Then, to forecast the biological annotation and pathway of the putative targets, we used Gene Ontology (GO) to reveal the biological process, cellular component, and molecular function of the target mRNAs. Signi cant pathways were identi ed using the pathways in the Kyoto Encyclopedia of Genes and Genomes database (KEGG).

Statistical analysis
Statistical analysis was performed using Graphpad prism software (version 3.0, Chicago, IL, USA). The behaviors and PCR results are shown as the mean ± standard error of the mean (SEM). Two-tailed Student's t-test was used to compare the signi cant differences between the two groups. The level of signi cance was set at P < 0.05.

Database and Accession Numbers
The raw data of the tsRNA-Seq in our study were deposited at the NCBI Gene Expression Omnibus (GEO) under the accession number GSE162402.

Results
Behavioral observations: the effects of nicotine on sensation and body weight The mechanical test revealed that on day 5, the 50%PWT of the nicotine group was signi cantly higher than that of the saline group (n = 12 in each group, ***P 0. 001), and this state continued to day 13 (Fig.  1a). The result indicated that nicotine exposure could change the pain sensation and produce an analgesic effect. Weight measure shown that on day 4, the weight increase of the nicotine group was lower, compared with the saline group(n = 12 each group, **P 0. 01), suggesting that nicotine exposure might impact the energy metabolism and appetite. In the days that followed, there was no signi cant difference about weight increase between both groups (Fig. 1b).

Expression pro les of tsRNAs in hippocampus of nicotine and saline rats
The scatter plots were created presenting the tsRNA expression variation between the nicotine group and saline group using the fold change (n=3 in each group, shown in Fig. 2a). According to the criteria of a log2 (fold change) > 1.5 and a P-value< 0.05, a total of 26 differentially expressed tsRNAs were nally identi ed, including 7 up-regulated and 19 down-regulated tsRNAs (Fig. 2b). And the hierarchical clustering heat-map for the 26 differentially expressed tsRNAs was displayed in Figure 2c. The detailed information on the top 5 up-regulated and 10 down-regulated tsRNAs ranked by fold change were presented in Table 1. In Figure 2d, the Venn diagram presented that a total of 354 were commonly expressed tsRNAs in both groups, 73 tsRNAs speci cally expressed in nicotine group, and 24 tsRNAs speci cally expressed in saline group.

Catalog of tRFs/tiRNAs in both groups
The pie chart demonstrates the percentage of each subtype of the tsRNAs expressed uniquely in two groups ( Fig. 3a-b). Overall, except for tRF-2, the expression levels of other tsRNAs (tRF-1, tRF-3, tRF-5, tiRNA-3, tiRNA-5) in Group nicotine were increased, as compared to Group saline. In both groups, tRF-5 were the most common type. The read counts and read length for each unique read in both groups are shown in the bar chart of sequence read length distribution (Fig. 3c-d). We found that the length distribution of tsRNAs in both groups mainly concentrate on 20-24nt. The number of tsRNA subtypes against tRNA iso-decoders and the frequency of subtypes against the length of the tsRNA were shown by the stacked bar charts (Fig. 3e-h). And in these charts, we discovered that Glu-CTC, Glu-TTC, Gly-CCC, Gly-GCC, His-GTG are the most common tRNA iso-decoders of tsRNAs in both groups.
Validation for the differentially expressed tsRNAs (DEtsRNAs) To validate the sequencing data, RT-qPCR was performed. Eight tsRNAs (3 upregulated and 5 downregulated) were chosen according to their raw expression level, the fold change value and the pvalue (shown in Table 2). As a result, tRF-1-T28-His-GTG-1 was signi cantly up-regulated, and tRF-5c-Glu-CTC-1, tRF-5c-Glu-CTC-3, tRF-5c-Gly-GCC-2-M2, tRF-5c-Glu-TTC-4 were down-regulated in Group nicotine, compared with Group saline (n=3 in each group, p < 0.05) (Fig. 4a-b). The comparison of the results of sequencing and PCR were displayed in Table 3, and the overall results were in line with the sequencing data.

Discussion
In this study, we reported the expression pro les of tsRNAs in the rat hippocampus after nicotine exposure for the rst time. The sequencing data revealed that there were 26 differentially expressed tsRNAs, involving 7 up-regulated and 19 down-regulated tsRNAs, in the rat hippocampus after nicotine exposure. Subsequently, 8 DEtsRNAs were chosen to validate by RT-qPCR and among them, 5 candidate DEtsRNAs for further bioinformatic predictions. The GO and KEGG pathway analysis demonstrated that the potential targeted genes and the putative pathways of the 5 DEtsRNAs mainly focused on several critical signaling channels, which played important roles in nicotine' actions. Taken together, our ndings suggest that these dysregulated tsRNAs may play regulatory roles on how nicotine functions.
There is a consensus that nicotine exposure can induce changes in pain perception and body weight in animal models and human studies [26][27][28]. Recent research indicated that nAChRs agonists (such as nicotine, epibatidine, choline) might make up a novel class of analgesics for pain management [29][30][31]. In our previous research, we found that short-term use of nicotine could produce analgesic effects, but longterm use or withdrawal led to hypersensitivity, and further investigation indicated that nicotine could alter pain sensitivity by affecting the expression of the pain related factor [24]. In this study, when the rats were continuously injected with nicotine, the paw withdrawal threshold increased and the body weight gains decreased, which was in accordance with our previous study.
Nicotine exerts its functions largely through the widespread nAChRs. Substantial researches suggested that nicotine exposure could upregulate nAChRs by multiple processes, including changes in receptor assembly, tra cking, and degradation [32][33][34]. In recent years, genetic factors is becoming the research focus about nAChRs, and variants in the CHRNA5-CHRNA3-CHRNB4 gene cluster are the most studied associated with nicotine' actions [35,36]. Besides, Cameli et al pointed out that genetic variations in CHRNA7 and CHRFAM7A were related to nicotine addiction [10].In our study, we found that tRF-5c-Glu-CTC-1, tRF-5c-Glu-CTC-3, tRF-5c-Glu-TTC-4 might be involved in the cholinergic synapse pathway, and in turn in uence the excitability of neurons and glial cells. Further analysis indicated that the putative genes of these above tsRNAs included CHRM1, CHRM3, CHRNA3, CHRNA4 and CHRNA7, which encoded variant subunits of nAChRs. So we presumed that tRF-5c-Glu-CTC-1, tRF-5c-Glu-CTC-3, tRF-5c-Glu-TTC-4 might function in nicotine' actions by modifying the genes expression of nAChRs' subunits, thus in uencing the structure and activity of nAChRs.
In addition to nAChRs, there exited some other possible mechanisms involved in nicotine' actions. On the one hand, nicotine exposure can produce widespread neuroadaptations in nervous system, including dopaminergic (DA) synapses, GABAergic circuitry, glutamatergic synapses, and so on. On the other hand, these neuroadaptations were in turn to participate in the nicotine' actions [37][38][39]. Recent studies shown that genetic variations in DA, GABA could regulate nicotine' actions. For instance, Bühler and his colleagues summarized that apart from nAChRs genes, variations in ANKK1, DRD2 and GABA were likewise associated with nicotine-related phenotypes [35]. Liu et al constructed the networks of candidate genes associated with nicotine addiction, and concluded that cholinergic receptors (CHRNA1, CHRNA4, CHRNA7), dopamine receptors (DRD1, DRD2, DRD3) and GABA receptors (GABRA1, GABRA2, GABRA4) were involved in diverse biological functions [40]. Other researches about genetic modulations in DA (COMT, GCH1, and DRD2) and GABA indicated these changes could affect the modulation of DA in pain pathway [41,42]. In our study, we predicted the target genes of these ve candidate tsRNAs, and found that tRF-5c-Glu-CTC-1 might regulate the expressions of COMT, DRD2, GABRA4, whereas tRF-5c-Glu-CTC-3 might regulate DA related genes COMT and DRD2. The KEGG pathway analysis indicated that tRF-5c-Glu-CTC-1 probably participated in the dopaminergic synapse, GABAergic synapses and glutamatergic synapses pathways, while tRF-5c-Glu-CTC-3 might merely modulate the dopaminergic synapse and glutamatergic synapses pathways, not GABAergic synapses.
Apart from the above neurotransmitters, previous researches suggested that nicotine shared the similar functional molecules, signal paths and acting sites with some other addictive drug, such as cocaine, morphine [11,43,44]. We found in our study that the candidate tsRNAs could modulate the pathways of cocaine, morphine and amphetamine addiction. To be speci c, tRF-5c-Glu-CTC-1 took part in all the three paths of cocaine, morphine and amphetamine addiction, tRF-1-T28-His-GTG-1 and tRF-5c-Glu-CTC-3 took part in cocaine addiction, and tRF-5c-Gly-GCC-2-M2 in morphine addiction. Hence, we speculated that those candidate tsRNAs might likely contribute to nicotine addiction in a similar way. More researches are needed.
According to the KEGG pathway analysis, we also discovered that except the above pathways, the putative genes of the ve candidate tsRNAs were also enriched in some other critical paths, such as MAPK signaling pathway, mTOR signaling pathway, neurotrophin signaling pathway, which have been suggested to involve nicotine' actions in body [45][46][47]. And further analysis indicated that the majority of the putative pathways of tRF-5c-Glu-CTC-1 were associated with the previously known mechanisms of nicotine' actions, so we regard tRF-5c-Glu-CTC-1 as the most promising candidate for further study in nicotine' actions.
Generally, tsRNAs with lengths of 18-40 nucleotides, include two main types based on the length and cleavage sites on tRNA or pre-tRNA: tRNA-derived fragments (tRFs) and tRNA-derived, stress-induced RNAs (tiRNAs). tiRNAs are usually the products of angiogenin cleavage of mature tRNAs at the anticodon site during stress, which contain two subtypes tiRNA-5 and tiRNA-3. While tRFs derive from cleavage on any sites of mature or pre-tRNAs, which contain four subtypes tRF-5, tRF-3, tRF-1, tRF-2 [14,15,48]. Growing evidence indicated that the functions of tsRNAs depended on their subtype and speci c subcellular localization. Some researchers stated that cytosolic and mitochondrial tiRNAs could repress protein translations and be associated with apoptosis initiation [49,50]. Kumar et al remarked that in HeLa cell line, tRF-5s are mostly nuclear while tRF-3s and -1s are cytoplasmic, and further study in human HEK 293 cells suggested that tRF-5s and tRF-3s are associated with Argonautes 1, 3 and 4, and then target mRNAs in a manner similar to miRNAs [51]. Zhang and his colleagues reported that in monocytes/dendritic Cells, td-piR (Glu) (tRF-5) could interact with PIWI protein and play a role in regulation of chromatin remodeling in somatic cells, just like piRNAs [52]. According to the known literature of tsRNAs as well as our data, the most promising candidate tRF-5c-Glu-CTC-1 belonged to tRF-5, so we postulated that it's likely that tRF-5c-Glu-CTC-1 might function by interacting with Ago proteins or PIWI protein, and have post-transcriptional regulations similarly to miRNAs or epigenetic modulations like piRNAs. In nicotine exposure model, nicotine exposure could upregulate the expression of nAChRs and downregulate tRF-5c-Glu-CTC-1 whose downstream genes CHRM1, CHRM3, CHRNA4 and CHRNA7 could encode the subunits of nAChRs shown in our data. Given the negative regulatory relationship between tRF-5c-Glu-CTC-1 and its putative genes, we conjecture that nicotine exposure may reduce the biosynthesis of tRF-5c-Glu-CTC-1, then alleviate its suppression on the genes expression of nAChRs subunits, and thereby upregulate nAChRs, which lay the foundation for nicotine' actions. Further research is needed in future.
There are some limitations about our study. First, we only detect the pro les of tsRNAs on nicotine exposure condition, not nicotine withdrawal, and lack the dynamic detection of tsRNAs at different time point. Secondly, our research is preliminary and further functional research is needed to identify the speci c locations and functions of the candidate tsRNAs in vitro and in vivo. Finally, detection of tsRNAs in human trial is hopefully to conduct in future.
To sum up, our study provided the differentially expressed pro les of tsRNAs in rat hippocampus after nicotine exposure for the rst time. Further bioinformation analysis revealed that the tsRNAs might be a new class of regulatory molecules in nicotine related researches, which might function by regulating some important signaling pathways, such as the cholinergic synapse, dopaminergic synapse. Among them, tRF-5c-Glu-CTC-1 is the most promising one, and probably act in a manner of miRNAs or piRNAs. In a word, our study paves a new road for research into nicotine' actions, especially in addiction and nicotine-mediated analgesia, and provides novel modulatory targets for this area.