Circulating miRNome of Trachemysscripta after elective gonadectomy under general anesthesia

Post-surgical management is an important issue in veterinary medicine, requiring biomarkers with high sensitivity and specificity for timely and effective treatment. Emerging evidence suggests that miRNAs are promising stress- and pain-related markers. The aims were to profile the circulating miRNA signature in plasma of turtles (Trachemys scripta) and point out potential candidate biomarkers to assess the status of the animal. The plasma of female turtles underwent surgical gonadectomy were collected 24 h pre-surgery, and 2.5 h and 36 h post-surgery. The expression of miRNAs was profiled by Next Generation Sequencing and the dysregulated miRNAs were validated using RT-qPCR. The diagnostic value of miRNAs was calculated by ROC curves. The results showed that 14 miRNAs were differentially expressed over time. RT-qPCR validation highlighted that 2-miR-499-3p and miR-203-5p-out of 8 miRNAs tested were effectively modulated. The Area Under the Curve (AUC) of miR-203-5p was fair (AUC 0.7934) in discriminating pre- and 36 h post-surgery samples and poor for other time points; the AUC of miR-499-3p was excellent (AUC 0.944) in discriminating pre-surgery and 2.5 h post-surgery samples, and fair in discriminating pre-surgery and 36 h post-surgery (AUC 0.7292) and 2.5 h and 36 h post-surgery (AUC 0.7569) samples. In conclusion, we demonstrated for the first time that miRNAs profile changes in plasma of turtles underwent surgical oophorectomy and identified miR-203-5p and miR-499-3p as potential candidate biomarkers to assess animals' status. Further studies are necessary to confirm their diagnostic value and to investigate functional and mechanistic networks to improve our understanding of the biological processes.

MiRNAs are short, non-coding RNA, widely conserved across species, exerting often similar functions in the different animal lineages 1,2 . MiRNAs are epigenetic modulators able to interfere with gene expression by binding with full or partial complementarity to targeted mRNA, leading to either its inhibition of translation or degradation 1 , without inducing any change to the genome. Painful events promote several molecular swings, including the modulation of miRNA expression, which in turn can silence the expression of pro-or anti-nociceptive genes; the ability of miRNAs to affect neuronal plasticity and thus to play a pivotal role in pain is reported in several animal models of inflammatory and neuropathic pains and is associated with hyperalgesia and allodynia 3,4 . Pain management is a highly important issue in veterinary medicine, not only for ethical reasons related to animal welfare but also to ensure proper clinical procedures. Nevertheless, animals exhibit complex behavioral and physiological responses to pain, which are ameliorated by analgesics 5 . Pain assessment is challenging in animals and, although several guides based on physical and biological alterations in dogs and cats 6 , horses 7 , and pig 8 have been developed, the ability of animals to experience pain remains controversial 9 . Pain recognition and treatment in chelonians brings additional challenges: although reptiles have been proved to possess all anatomical and physiological structures necessary to perceive and elaborate nociceptive pain, and to display painful behaviours [10][11][12][13] , pain diagnosis and assessment of the efficacy of most analgesic drugs is still rudimentary 10,[13][14][15] . Behavioral changes include reduced mobility and depression, reduced or increased aggression, reduced food intake, increased respiratory rate, unwillingness to swim in aquatic animals. Such changes are not specifically pain-related and can be altered by a wide range of conditions depending on the species, environmental conditions, concurrent diseases, pathophysiological status, and administered therapies 10,12,13 . For these reasons, and unlike in mammalian practice, behavioral ethograms are not completely reliable and are seldom used to assess the status of reptiles in clinical practice, so that clinicians most often provide analgesia miRNA extraction using two commercial kits. To identify the best kit to isolate microRNAs from turtle plasma, a comparison of two commercial kits for microRNA extraction was performed. Small RNAs were extracted using miRNeasy Serum/Plasma kit (Qiagen, Cat. No 217184) and microRNA Concentrator kit (A&A Biotechnology, Cat. No 035-25). Only plasma collected at 2.5 h post-surgery has been included in this analysis. Plasma was thawed on ice and centrifuges at 3000 × g for 5 min at 4 °C. An aliquot of 100 µL per sample was transferred in a new tube and smallRNA was extracted using the two kits, in accordance with the manufacturer's instruction. The Caenorhabditis elegans miRNA cel-miR-39 (25 fmol final concentration) (Qiagen, Cat. No 219610) was used as synthetic spike-in control. To compare the performance of the kits, the spike-in cel-miR-39 was quantified by RT-qPCR as explained below, and the Cq values were compared. After NGS, the analysis was carried out on the same samples targeting DE-miRNAs, to confirm the results. Computational analysis. The output of NextSeq500 Illumina sequencer was demultiplexed using bcl2fastq

Next generation sequencing (NGS
Illumina software embedded in docker4seq package 22 (https:// github. com/ kendo maniac/ docke r4seq). miRNA expression quantification was performed using the workflow previously described 21 (https:// github. com/ kendo maniac/ docke r4seq). In brief, fastq files were quality checked (QC) using FastQC software (http:// www. bioin forma tics. babra ham. ac. uk/ proje cts/ fastqc/). Reads shorter than 14 nt were discarded. The QC-passed reads were clipped from the adapter sequences using Cutadapt 23 , by imposing a maximum error rate in terms of mismatches, insertions, and deletions equal to 0.15. Since Trachemys scripta miRNA sequences are not available in a public repository, sequences were mapped to all saurians (aca, cpi, ami, oha) precursors miRNAs available in miRBase 22.0 (http:// www. mirba se. org/) based on phylogeny 24 . The alignment was performed using BWA 25 algorithm v. 0.7.12 with the default settings. Annotation and quantification of miRNAs were done as described 26 . The detected counts were organized in a table including all analysed samples. For visualization purposes, only CPM (counts per million reads) were used. Differential expression analysis was evaluated using DESeq2 Bioconductor package 27 implemented in docker4seq package (https:// github. com/ kendo maniac/ docke r4seq). Differential expression analysis was done using the above-mentioned counts' www.nature.com/scientificreports/ using as threshold an adjusted P-value ≤ 0.1 and an absolute log2 Fold Change (log2FC) ≥ 1. In case a miRNA was detected in Chrysemys picta (cpi) and in other saurian only the Chrysemys picta was reported.
miRNA quantification by qPCR. Reverse transcription was performed using TaqMan Advanced miRNA cDNA Synthesis Kit (Cat. No A28007, Applied Biosystems), following the manufacturer's instruction.
The quantitative reaction was performed in a scale down-reaction volume (15 µl Data were tested for normality using the Shapiro-Wilk test; as the data were not normally distributed, the nonparametric Wilcoxon signed-rank test for paired samples was applied. Statistical significance was accepted at P-value ≤ 0.05. Receiver Operating Characteristic (ROC) analysis was performed to determine the diagnostic accuracy of target that statistically differed between groups. The diagnostic value was calculated for that miRNA that showed a significant differential expression in the blood of turtles, namely miR-499-3p and miR-203-5p. The ROC analysis was carried out by plotting the true positive (sensitivity) versus the false positive (1-specificity).

Results
Physiological and clinical parameters remained stable throughout the procedure for all patients. All the animals returned to normal levels of activity and spontaneous feeding within 12 h after surgery, and no post-surgical complication occurred during the 1-week follow-up period.
A&A Biotechnology kit was selected to extract miRNAs from turtle plasma. To assess the success of miRNA extraction using the two kits, cel-miR-39 was quantified by RT-qPCR on samples collected at 2.5 h post-surgery, as a preliminary step before NGS. At the end of the study, the good performance of the selected kit was confirmed comparing the Cq values of DE-miRNAs at 2.5 h post-surgery starting from the same extracts. Overall, the kit by A&A Biotechnology was the best with the lower Cq values across the 5 miRNAs (P < 0.031) ( Supplementary Fig. 1). In detail, the differences of Cq values between Qiagen and A&A Biotechnology kits were 7.33, 6.72, 4.36, 8.66 and 2.84 for cel-miR-39, miR-499-3p, miR-203-5p, miR-92b-5p and miR-138-1-3p, respectively. Based on the results on cel-miR-39 the kit by A&A Biotechnology was selected.

Validation of DE-miRNAs
Diagnostic performance of circulating miR-203-5p and miR-499-3p. Since circulating miR-203-5p and miR-499-3p were differentially expressed over time, further analysis was performed on these miR-NAs. To evaluate the diagnostic value, ROC curve analysis was performed and the associated area under the curve (AUC), where an area of 1 represents a perfect test and an area of 0.5 a worthless test, was used to confirm   (Fig. 3B).

Discussion
Novel biomarkers with high sensitivity and specificity are required for assessing the status of post-surgical animals. Emerging evidence suggests that miRNAs are promising stress-and pain-related markers with the advantage of high sensitivity and specificity in different animal species 28,29 . However, no investigations have been conducted on circulating miRNAs associated with post-surgical chelonians so far.
In the present study, prefemoral gonadectomy of female pond sliders (Trachemys scripta) has been selected as a surgical model, based on the hypothesis that it may affect plasma molecular markers. A multistep approach, using Next Generation Sequencing as a first step to identifying differentially expressed miRNAs in plasma at three-time points, then validating them by RT-qPCR on the sequenced samples and a separate independent set of samples, was adopted. Our results provided for the first time the evidence that (I) the expression level of circulating miRNAs changes in the plasma of pond sliders underwent surgical castration; (II) the expression level of miR-203-5p decreased 4.3-folds at 36 h post-surgery compared with pre-surgery level, and that of miR-499-3p increased 21.9-folds at 2.5 h post-surgery compared with pre-surgery and decreased to a level similar to that of pre-surgery at 36 h post-surgery; (III) diagnostic accuracy of miR-203-5p was fair in discriminating pre-and 36 h post-surgery animals and that of miR-499-3p was excellent in discriminating animals over time.
The successful purification of miRNAs from small volumes of biofluids, including plasma, is challenging, especially in turtles, for which, to the best of our knowledge, no investigations on circulating miRNAs have been previously performed. As previously reported 30,31 , miRNA recovery can be influenced by the isolation method and a careful selection of the RNA isolation method should be explored to identify the best approach www.nature.com/scientificreports/ for own samples. We compared two column-based commercially available kits developed for miRNAs extraction from biofluids and then we evaluated the qPCR performance using Cq values. The results showed that the A&A Biotechnology kit allowed us to obtain lower Cq values than the Qiagen kit; thus, we identified A&A Biotechnology kit to isolate miRNAs from the plasma of turtles. Sequencing data supplied a list of differentially abundant miRNAs but the validation step using RT-qPCR showed that the plasma miRNAs rate at three-time points was not statistically significant for 6 out of 8 miRNAs tested. The technical bias inherent in sequencing technologies may provide misshape results. A substantial distortion between miRNA levels in NGS data and true miRNA abundancy may occur using Illumina sequencing technology 32 , thus the validation step by qPCR is always advisable. Another issue concerns the analysis of miRNAs in plasma, which is more challenging than conventional intracellular miRNAs. The miRNAs library preparation from plasma suffers from two main critical issues: (i) lower number of miRNAs compared to the intracellular ones, and (ii) higher background due to the presence of degraded RNA fragments in the blood. Especially the contamination of fragmented unspecific RNA massively reduces the number of reads mappable on miRBase precursors. The protocols for preparing miRNAs' libraries need high-quality total RNA to obtain high recovery of miRNA reads from intracellular total RNA. Unfortunately, when working with circulating miRNA it becomes impossible to avoid the contamination of fragmented unspecific RNA. To limit the impact of this issue on the number of identified miRNAs, our sequences were mapped to all saurian precursors miRNAs available in miRBase 22.0; nevertheless, the fraction of reads mapping represented only a very small part of the total sequenced reads. Since few specific diagnostic tests are available and few drugs have been investigated in reptiles, dosages are most of the time empirically extrapolated without proper notions on their effectiveness, which is a particular concern for painkilling drugs, especially as it is harder to understand the welfare status of reptiles 33 . It has been recognized that reptiles are sentient, being able to feel emotion, pleasure, fear, stress, grief, and pain, as recently well summarized by Lambert and colleagues 34 . Trachemys scripta has been introduced in Europe from North America since the mid-'80s. It has competitive advantages over native species such as the European pond turtle (Emys orbicularis), including lower age at maturity, higher fecundity, and larger adult body size 35 . Turtles may compete for food, egg-laying sites, or basking places, the best of which are occupied by T. scripta 35 , thus this species is included in the list of invasive alien species by the EU regulation 2016/1141. Official eradication programs have not yet been established. Surgical gonadectomy has been proposed as means for population control, but several obstacles limit successful discomfort recognition and treatment, such as subjectivity in pain assessment, and limited knowledge of analgesic efficacy 10,15,36 . Usually, discomfort in non-verbal patients is assessed by semi-objective scales, using both behavioral and physiological parameters believed to be correlated with a stressful or painful condition 10,12 ; unfortunately, in reptiles, such parameters aren't exclusive nor strongly correlated to discomfort 12 , requiring careful and often time-consuming evaluation. Thus, the identification of proper biomarkers to evaluate the chelonians' status may support the clinical decision-making process. The here selected anesthetic protocol consisted of a combination of a benzodiazepine (midazolam), an α 2 -agonist www.nature.com/scientificreports/ (dexmedetomidine), a neuroactive steroid (alfaxalone), and a dissociative agent derived from phencyclidine (ketamine). Benzodiazepines and alfaxalone both act on inhibitory neurotransmitter gamma aminobutyric acid complex (GABA) receptors 37,38 : midazolam enhancing central and peripheral GABA A receptors 39 , and alfaxalone acting on central GABA receptors 37 . Alpha-2 agonists are used in combination with other molecules for sedation and myorelaxation 40,41 , stimulating central and peripheral α 2 -adrenergic receptors decreasing excitability, and provide analgesia 42 ; their effects can be antagonized with specific antagonists such as atipamezole 40 . Ketamine, acting as N-methyl-d-aspartate (NMDA) glutamate central and peripheral receptor antagonist, is a dissociative anesthetic drug frequently used in reptilian practice in combination with other drugs (multimodal anesthesia regimen) to reduce side effects and recovery time 13,40,[43][44][45] . Ovariectomy in animals is a relatively standardized source of moderate soft tissue pain, making it suitable for clinical studies 46 . In the present study, for ethical reasons, it was decided to provide analgesic drugs, and since GABA-agonists do not display effective analgesic property, selected molecules for this purpose was dexmedetomidine and ketamine. Analgesic efficacy of ketamine and α 2 -agonist has been well demonstrated in mammalian practice 47 ; in reptiles, such properties are still controversial, but indications are pointing to the potential use of α 2 -agonist as analgesic supplement [11][12][13]42,[48][49][50] . The increase in miR-499-3p expression at T1 can be thus explained by the depletion of the activity of dexmedetomidine after reversal with atipamezole, while the decrease 36 h after surgery is likely correlated to the normal decrease of acute pain caused by the procedure. The absence/low level of discomfort at this time point could be confirmed by the fact that, even if the behavioral ethogram proposed by Kinney and co-workers 10 was not applied, all the animals showed normal behavior and feeding 24 h after the surgery. The miR-203-5p can be considered as a "negative biomarker". Indeed, it is found in a considerable amount at pre-surgery, lowers at 2.5 h post-surgery, and becomes significantly lower (P = 0.021) at 36 h post-surgery compared to the pre-surgery level. Considering that no NSAIDs were administered in the post-operative period in this study, this negative biomarker could be influenced by surgery, inflammation, drugs, pain and, more in general, by stress and discomfort. Focusing on surgical inflammation, it usually refers to the initial physiological response to tissue damage. Acute inflammation begins within seconds to minutes following an injury to tissues and increases in the following hours 51 . The miR-203-5p production, high in normal conditions, may be suppressed by several factors, including the presence of pro-inflammatory molecules in post-surgical course. Further studies are necessary to explore and confirm this hypothesis, for example by the administration of NSAIDs to turtles in the post-operative period.

Conclusions
To support the clinicians in the management of post-surgical chelonians, we explored the potential use of miRNAs as markers to assess and monitor the post-surgical status of turtles (T. scripta), demonstrating that miR-499-3p and miR-203-5p might be further explored as promising candidate biomarkers. MiR-499-3p and miR-203-5p are modulated after surgery, representing potential candidate biomarkers of chelonians' discomfort. Nevertheless, the pathophysiological role played by miR-499-3p and miR-203-5p must be yet fully elucidated.
Although this study has provided new and important insights, further experiments are required to identify and validate the specific elements these miRNAs are useful for measuring. Since the nucleotide sequence of miR-499-3p and miR-203-5p are not conserved between the turtle and human and other species for which potential target genes and functional information are available on public databases, Gene Ontology could not be carried out.