Identification and Characterization of microRNAs during Retinoic Acid-Induced Regeneration of a Molluscan Central Nervous System

Retinoic acid (RA) is the biologically active metabolite of vitamin A and has become a well-established factor that induces neurite outgrowth and regeneration in both vertebrates and invertebrates. However, the underlying regulatory mechanisms that may mediate RA-induced neurite sprouting remain unclear. In the past decade, microRNAs have emerged as important regulators of nervous system development and regeneration, and have been shown to contribute to processes such as neurite sprouting. However, few studies have demonstrated the role of miRNAs in RA-induced neurite sprouting. By miRNA sequencing analysis, we identify 482 miRNAs in the regenerating central nervous system (CNS) of the mollusc Lymnaea stagnalis, 219 of which represent potentially novel miRNAs. Of the remaining conserved miRNAs, 38 show a statistically significant up- or downregulation in regenerating CNS as a result of RA treatment. We further characterized the expression of one neuronally-enriched miRNA upregulated by RA, miR-124. We demonstrate, for the first time, that miR-124 is expressed within the cell bodies and neurites of regenerating motorneurons. Moreover, we identify miR-124 expression within the growth cones of cultured ciliary motorneurons (pedal A), whereas expression in the growth cones of another class of respiratory motorneurons (right parietal A) was absent in vitro. These findings support our hypothesis that miRNAs are important regulators of retinoic acid-induced neuronal outgrowth and regeneration in regeneration-competent species.


Introduction
Following injury to the central nervous system (CNS), the ability of damaged neurons to repair and regenerate functional connections is limited in most species. Only a few vertebrates are able to regenerate lost or damaged CNS tissues and cell types, though numerous invertebrates possess this intrinsic ability of self-repair [1,2]. Interestingly, many trophic and chemotropic factors that mediate neuronal outgrowth and connectivity in regeneration-competent species are highly conserved and functional in many vertebrates and invertebrates [3][4][5]. However, a thorough understanding of the molecular and biochemical mechanisms underlying the production and utilization of these factors that results in functional CNS regeneration is lacking.
One critical factor that has been suggested to play a role in CNS regeneration in many species is the vitamin A metabolite, all-trans retinoic acid (RA). RA signaling mediates neuronal outgrowth [6] and differentiation [7] during both CNS development and regeneration [8][9][10]. To exert such effects, RA binds to two classes of nuclear receptors, the retinoic acid receptors (RAR) and the retinoid X receptors (RXR). Following ligand binding, specific subtypes of these receptors typically heterodimerize The Phred scores identified the base call accuracy, and determined the probability of an incorrect base reading. A perfect score of 40 corresponds to 99.99% accuracy of the reads, while a score greater than 30 corresponds to 99.9% accuracy. Our data displayed typical Phred scores of 38 and 39 ( Figure 1D), suggesting a high accuracy (>99.9%) and providing confidence in our miRNA sequencing data analysis.  Figure 1D), suggesting a high accuracy (>99.9%) and providing confidence in our miRNA sequencing data analysis.  Mature miRNA sequence information in molluscs is currently very limited. At present, miRBase release 22 (March 2018; available online: http://www.mirbase.org/) reports the sequences of only three molluscan species, Lottia gigantea, Melibe leonina, and Haliotis rufescens. To identify miRNAs in Lymnaea, the sequences generated from our miRNA sequencing analysis were matched to known miRNAs in the mollusc, Lottia gigantea. The miRNA sequencing analysis identified 482 miRNA sequences in the Lymnaea CNS, including 97 precursor miRNAs, 166 mature miRNAs and 219 novel miRNAs ( Figure 1E; See Supplementary Material for list of all identified miRNA sequences). The novel miRNAs represent a group of mature miRNAs that were not matched to any other known species in miRBase. As most miRNAs are highly conserved, even between distantly related species such as invertebrates and vertebrates [22], these novel miRNA sequences may be Lymnaea-specific miRNAs, or possibly, miRNAs that have not yet been identified or sequenced in other species. It is also possible that some of these novel miRNAs represent RNA fragments.

Identification of miRNAs That Were Differentially Regulated during Retinoic Acid (RA)-Induced Regeneration
From our miRNA sequencing data, 219 putative novel sequences represented a large proportion of total sequenced reads (~45%). Of these novel sequences, 38 exhibited at least a 2-fold increase in RA-treated samples, while 48 displayed at least a 2-fold reduction ( Figure 2A). However, all of the differentially regulated novel sequences were of extremely low abundance, typically with less than 10 sequence reads in the entire CNS. Due to their extremely low level of expression, we did not investigate any novel sequences further.
Additional candidate miRNAs identified by our miRNA sequencing included miR-124 and miR-9, both of which are enriched in the nervous systems of many species and contribute to neuronal differentiation [45][46][47][48][49] and axonal guidance [51,52]. Importantly, both of these miRNAs are upregulated by RA in human and mouse cell lines [47,58,59]. However, we found that miR-9 was downregulated in RA-treated CNS ( Figure 2C), indicating it may have distinct roles and responses to RA in the snail, compared to vertebrates. As described in vertebrates, however, miR-124 was upregulated in RA-treated Lymnaea CNS ( Figure 2C). As miR-124 has been associated with neurite outgrowth, not only in mammals [60], but also in amphibians [52], we focused on examining a potential role for miR-124 in regenerating neurons of the invertebrate, Lymnaea. Mature miRNAs that were differentially expressed between regenerating CNS incubated in RA, or non-regenerating CNS incubated in EtOH; (A) Pie chart depicts the total number of novel sequences identified by miRNA sequencing. Of these sequences, 38 exhibited at least a 2-fold increase in CNS incubated in RA (red), while 48 displayed at least a 2-fold reduction in RA-treated samples (white); (B) Pie chart depicts the total number of mature miRNA sequences identified by miRNA sequencing analysis. Of the identified sequences, a large proportion did not exhibit differential expression between CNS incubated in RA and EtOH (black). However, a small subset exhibited at least a 2-fold change between treatment groups, and were either upregulated (red) or downregulated (white) in RA-treated CNS; (C) A complete list of differentially expressed mature miRNAs following RA treatment. Table indicates the mature miRNA name and its corresponding expression pattern in regenerating RA-treated CNS. Mature miRNAs that were differentially expressed between regenerating CNS incubated in RA, or non-regenerating CNS incubated in EtOH; (A) Pie chart depicts the total number of novel sequences identified by miRNA sequencing. Of these sequences, 38 exhibited at least a 2-fold increase in CNS incubated in RA (red), while 48 displayed at least a 2-fold reduction in RA-treated samples (white); (B) Pie chart depicts the total number of mature miRNA sequences identified by miRNA sequencing analysis. Of the identified sequences, a large proportion did not exhibit differential expression between CNS incubated in RA and EtOH (black). However, a small subset exhibited at least a 2-fold change between treatment groups, and were either upregulated (red) or downregulated (white) in RA-treated CNS; (C) A complete list of differentially expressed mature miRNAs following RA treatment. Table indicates the mature miRNA name and its corresponding expression pattern in regenerating RA-treated CNS.

miR-124 Is Highly Enriched in the Adult Lymnaea CNS
In vertebrates, miR-124 has been characterized as a neuronally enriched miRNA [61] that contains multiple variants, including miR-124a, miR-124b, and miR-124c [62]. We obtained sequences for two conserved subtypes of miR-124 in the Lymnaea CNS, miR-124a and miR-124c. However, the number of reads for each variant was notably low ( Figure 3A). Rather than focus on each individual subtype, we instead characterized the entire miR-124 family of microRNAs (henceforth referred to as miR-124).
PCR was used to determine whether miR-124 shared a similar neuronally enriched expression in Lymnaea as it does in vertebrates. Indeed, we found miR-124 was enriched in the CNS, but largely absent or undetectable from other adult tissues, including the heart, albumen, prostate, and buccal mass ( Figure 3B). We next determined whether this miRNA was developmentally regulated, as it is in vertebrates. In vertebrates, miR-124 is shown to temporally increase in abundance throughout development, until reaching maximal expression in the adult CNS [61,63,64]. To determine whether the same is true in Lymnaea, we performed qPCR on developing Lymnaea embryos and compared miR-124 expression levels to that of the adult CNS. In developing Lymnaea embryos, the first neurons of the CNS are born approximately 4 days following egg laying [65]. In the days following, ganglia begin to form and continue to increase in size until~10 days, when the embryos hatch [66]. As such, we examined miR-124 expression at 6, 8, and 10 days after egg laying, at stages when the CNS has begun to develop [65,66]. miR-124 was expressed at low levels over the course of Lymnaea development, and did not significantly increase in expression across these developmental time points (F (3,8) = 17.63; p = 0.0007)( Figure 3C). However, miR-124 is more abundant in the adult CNS, as shown in vertebrates [61,63,64] (Figure 3C), and was significantly higher than at 6 (p = 0.0013), 8 (p = 0.0014), and 10 days (p = 0.0022) of development.
After confirming miR-124 was indeed a neuronally enriched miRNA in adult Lymnaea, we next used qPCR to confirm its differential expression between regenerating (RA-treated) and non-regenerating (EtOH-treated) CNS. Although the mean relative level of expression of miR-124 in RA-treated brains was~25% greater than in those treated with EtOH, this difference approached, but did not quite reach, statistical significance (p = 0.07; Figure 3D). in vertebrates. In vertebrates, miR-124 is shown to temporally increase in abundance throughout development, until reaching maximal expression in the adult CNS [61,63,64]. To determine whether the same is true in Lymnaea, we performed qPCR on developing Lymnaea embryos and compared miR-124 expression levels to that of the adult CNS. In developing Lymnaea embryos, the first neurons of the CNS are born approximately 4 days following egg laying [65]. In the days following, ganglia begin to form and continue to increase in size until ~10 days, when the embryos hatch [66]. As such, we examined miR-124 expression at 6, 8, and 10 days after egg laying, at stages when the CNS has begun to develop [65,66]. miR-124 was expressed at low levels over the course of Lymnaea development, and did not significantly increase in expression across these developmental time points ( After confirming miR-124 was indeed a neuronally enriched miRNA in adult Lymnaea, we next used qPCR to confirm its differential expression between regenerating (RA-treated) and non-regenerating (EtOH-treated) CNS. Although the mean relative level of expression of miR-124 in RA-treated brains was ~25% greater than in those treated with EtOH, this difference approached, but did not quite reach, statistical significance (p = 0.07; Figure 3D). PCR demonstrates that miR-124 is enriched within the adult CNS, but appears diminished or completely undetectable in other tissues. eIF4α was used as the positive loading control; (C) miR-124 expression during Lymnaea development. At 6, 8, and 10 days post-egg laying, miR-124 expression remained relatively uniform, and was not statistically significant across developmental days. However, miR-124 exhibited a significant increase in the adult CNS in comparison to earlier developmental days (** = p < 0.01); (D) miR-124 expression in the regenerating Lymnaea CNS. The mean relative normalized expression is 26% greater in RA-treated CNS, in comparison to EtOH controls. However, this increase did not reach statistical significance (p = 0.07). For RT-qPCR, data was made relative to the expression of the acutely isolated CNS (control) and normalized to β-tubulin, actin, and eIF4α. PCR demonstrates that miR-124 is enriched within the adult CNS, but appears diminished or completely undetectable in other tissues. eIF4α was used as the positive loading control; (C) miR-124 expression during Lymnaea development. At 6, 8, and 10 days post-egg laying, miR-124 expression remained relatively uniform, and was not statistically significant across developmental days. However, miR-124 exhibited a significant increase in the adult CNS in comparison to earlier developmental days (** = p < 0.01); (D) miR-124 expression in the regenerating Lymnaea CNS. The mean relative normalized expression is 26% greater in RA-treated CNS, in comparison to EtOH controls. However, this increase did not reach statistical significance (p = 0.07). For RT-qPCR, data was made relative to the expression of the acutely isolated CNS (control) and normalized to β-tubulin, actin, and eIF4α.

miR-124 Is Expressed in Both the Pedal and Right Parietal Ganglia
miRNAs are generally cell type-specific and, importantly, have exhibited differential expression patterns in specific regions within the CNS of both mice [67] and zebrafish [68]. Our next aim was to determine whether specific patterns of expression of miR-124 exist within different ganglia of the molluscan CNS. To this end, we utilized the pedal ganglia and right parietal ganglion ( Figure 4A), which are known to contain different classes of motorneurons [69,70]. In situ hybridization indicated a perinuclear distribution of miR-124 in cells of both the pedal ganglia ( Figure 4(Bi)) and the right parietal ganglion (Figure 4(Bii)). Importantly, RT-qPCR analysis indicated that the overall expression levels of miR-124 did not differ between these ganglia (p = 0.9535; Figure 4(Biii)).

miR-124 Is Expressed in Both the Pedal and Right Parietal Ganglia
miRNAs are generally cell type-specific and, importantly, have exhibited differential expression patterns in specific regions within the CNS of both mice [67] and zebrafish [68]. Our next aim was to determine whether specific patterns of expression of miR-124 exist within different ganglia of the molluscan CNS. To this end, we utilized the pedal ganglia and right parietal ganglion ( Figure 4A), which are known to contain different classes of motorneurons [69,70]. In situ hybridization indicated a perinuclear distribution of miR-124 in cells of both the pedal ganglia (Figure 4(Bi)) and the right parietal ganglion (Figure 4(Bii)). Importantly, RT-qPCR analysis indicated that the overall expression levels of miR-124 did not differ between these ganglia (p = 0.9535; Figure 4(Biii)).  A different miRNA, miR-133, has previously been shown to regulate RA-induced regeneration of the newt spinal cord [14]. We thus also conducted in situ hybridization to examine its presence or absence in these same ganglia. No detectable signal was obtained for miR-133 in either ganglion (Figure 4(Ci,Cii)). miR-133 was, however, detectable by RT-qPCR (Figure 4(Ciii)), which showed no difference in expression across ganglia (p = 0.1862). These data also indicated very low expression levels of miR-133, confirmed also by our miRNA sequencing results. Hence, no further analysis was conducted for miR-133.

miR-124 Is Expressed within the Cell Bodies and Neurites of Two Populations of Regenerating Motorneurons
We next examined miR-124 expression patterns within cultured, regenerating motorneurons from both pedal and right parietal ganglia. This examination of cultured cells allowed a better resolution of cellular compartmentalization of the miRNA within regenerating neurons, and provided more detailed information on cell type-specific expression patterns. Importantly, different functional classes of motorneurons were used. Pedal A (PeA) motorneurons innervate the cilia of the foot musculature and are involved in locomotion [69]. These PeA ciliary motorneurons have previously been shown to exhibit robust outgrowth and chemotropic responses upon application of RA [10,18]. Another class of motorneuron, with as yet unknown responses to RA, were also included; these right parietal A (RPA) motorneurons control movement of the pneumostome and are, thus, required for aerial respiration [30].
In the cultured regenerating motorneurons ( Figure 5A), miR-124 was consistently expressed within the cell bodies of both PeA (n = 46, Figure 5(Bi)) and RPA (n = 41, Figure 5(Bii)) cell types. Once again, miR-124 clearly exhibited a perinuclear distribution within the soma, similar to the expression pattern we found in the ganglia. miR-124 was also detected within regenerating neurites of both PeA (Figure 5(Ci)) and RPA motorneurons ( Figure 5(Cii)). Interestingly, miR-124 was expressed as individual punctae along the length of PeA and RPA neurites ( Figure 5C). However, miR-124 was not consistently expressed in all neurites of either cell type, with its expression in RPA neurites being less frequent in comparison to PeA neurites ( Figure 5C). Interestingly, miR-124 expression was often abundant in branch points of both PeA ( Figure 5(Di)) and RPA ( Figure 5(Dii)) neurites.
RT-qPCR, data was made relative to the expression in the entire Lymnaea CNS, and normalized to β-tubulin, actin, and eIF4α; (D) Representative image of ganglia sections incubated with a scrambled probe (negative controls). No signal was detected in either pedal (Di) or right parietal (Dii) ganglia. Scale bars (B-D) = 50 μm.
A different miRNA, miR-133, has previously been shown to regulate RA-induced regeneration of the newt spinal cord [14]. We thus also conducted in situ hybridization to examine its presence or absence in these same ganglia. No detectable signal was obtained for miR-133 in either ganglion (Figure 4(Ci,Cii)). miR-133 was, however, detectable by RT-qPCR (Figure 4(Ciii)), which showed no difference in expression across ganglia (p = 0.1862). These data also indicated very low expression levels of miR-133, confirmed also by our miRNA sequencing results. Hence, no further analysis was conducted for miR-133.

miR-124 Is Expressed within the Cell Bodies and Neurites of Two Populations of Regenerating Motorneurons
We next examined miR-124 expression patterns within cultured, regenerating motorneurons from both pedal and right parietal ganglia. This examination of cultured cells allowed a better resolution of cellular compartmentalization of the miRNA within regenerating neurons, and provided more detailed information on cell type-specific expression patterns. Importantly, different functional classes of motorneurons were used. Pedal A (PeA) motorneurons innervate the cilia of the foot musculature and are involved in locomotion [69]. These PeA ciliary motorneurons have previously been shown to exhibit robust outgrowth and chemotropic responses upon application of RA [10,18]. Another class of motorneuron, with as yet unknown responses to RA, were also included; these right parietal A (RPA) motorneurons control movement of the pneumostome and are, thus, required for aerial respiration [30].
In the cultured regenerating motorneurons ( Figure 5A), miR-124 was consistently expressed within the cell bodies of both PeA (n = 46, Figure 5(Bi)) and RPA (n = 41, Figure 5(Bii)) cell types. Once again, miR-124 clearly exhibited a perinuclear distribution within the soma, similar to the expression pattern we found in the ganglia. miR-124 was also detected within regenerating neurites of both PeA ( Figure 5(Ci)) and RPA motorneurons ( Figure 5(Cii)). Interestingly, miR-124 was expressed as individual punctae along the length of PeA and RPA neurites ( Figure 5C). However, miR-124 was not consistently expressed in all neurites of either cell type, with its expression in RPA neurites being less frequent in comparison to PeA neurites ( Figure 5C). Interestingly, miR-124 expression was often abundant in branch points of both PeA ( Figure 5(Di)) and RPA ( Figure 5(Dii)) neurites.

miR-124 Is Differentially Expressed in the Growth Cones of Different Classes of Motorneurons
Neuronal growth cones are structures responsible for initiating and guiding regenerative outgrowth. We, therefore, next examined the expression of miR-124 within growth cones of the regenerating cultured RPA and PeA motorneurons. Interestingly, miR-124 was not detected in the growth cones of any RPA neurons (n = 0 of 43 growth cones; Figure 6A). It was, however, expressed within growth cones of PeA neurons (n = 28 growth cones). miR-124 was expressed along the leading edge of the PeA growth cones, and was most frequently restricted to the lamellipodia (L; Figure 6B). It was, however, noticeably absent from the central domain (CD) of the growth cones ( Figure 6B). miR-124 was also often associated or aligned with the filopodia in a long, fibrillar expression pattern ( Figure 6C).
Interestingly, miR-124 was not expressed in all PeA growth cones. Therefore, we next determined whether exposure of regenerating neurites to retinoic acid might affect the expression pattern and/or number of growth cones expressing miR-124. Cultured PeA motorneurons were incubated in either RA (10 −7 M) or in EtOH (0.001%; vehicle-control) during the first 12-18 h of regenerative outgrowth. However, no significant differences in the proportion of growth cones containing miR-124 were shown between treatment groups (p = 0.4995). miR-124 was found in the growth cones of ~40% of RA-treated neurons and ~25% of EtOH-treated neurons (Figure 6(Di)).
Cultured PeA motorneurons typically contain multiple neurites with numerous growth cones (see Figure 5(Ai) as an example). With this in mind, we re-examined the proportion of growth cones on individual motorneurons expressing miR-124, this time, only analyzing those cells that expressed miR-124 in at least one of its growth cones. Once again, the proportion of growth cones that expressed this miRNA did not differ between cells treated with either RA (~63%) or EtOH (~66%) (p = 0.8747; Figure 6(Dii)).
In summary, these data demonstrate cell type-specific expression of miR-124 in Lymnaea growth cones, and that this pattern of expression is not dependent on prior exposure to RA.

miR-124 Is Differentially Expressed in the Growth Cones of Different Classes of Motorneurons
Neuronal growth cones are structures responsible for initiating and guiding regenerative outgrowth. We, therefore, next examined the expression of miR-124 within growth cones of the regenerating cultured RPA and PeA motorneurons. Interestingly, miR-124 was not detected in the growth cones of any RPA neurons (n = 0 of 43 growth cones; Figure 6A). It was, however, expressed within growth cones of PeA neurons (n = 28 growth cones). miR-124 was expressed along the leading edge of the PeA growth cones, and was most frequently restricted to the lamellipodia (L; Figure 6B). It was, however, noticeably absent from the central domain (CD) of the growth cones ( Figure 6B). miR-124 was also often associated or aligned with the filopodia in a long, fibrillar expression pattern ( Figure 6C).
Interestingly, miR-124 was not expressed in all PeA growth cones. Therefore, we next determined whether exposure of regenerating neurites to retinoic acid might affect the expression pattern and/or number of growth cones expressing miR-124. Cultured PeA motorneurons were incubated in either RA (10 −7 M) or in EtOH (0.001%; vehicle-control) during the first 12-18 h of regenerative outgrowth. However, no significant differences in the proportion of growth cones containing miR-124 were shown between treatment groups (p = 0.4995). miR-124 was found in the growth cones of~40% of RA-treated neurons and~25% of EtOH-treated neurons (Figure 6(Di)).
Cultured PeA motorneurons typically contain multiple neurites with numerous growth cones (see Figure 5(Ai) as an example). With this in mind, we re-examined the proportion of growth cones on individual motorneurons expressing miR-124, this time, only analyzing those cells that expressed miR-124 in at least one of its growth cones. Once again, the proportion of growth cones that expressed this miRNA did not differ between cells treated with either RA (~63%) or EtOH (~66%) (p = 0.8747; Figure 6(Dii)).
In summary, these data demonstrate cell type-specific expression of miR-124 in Lymnaea growth cones, and that this pattern of expression is not dependent on prior exposure to RA.

Discussion
In this study, we performed the first transcriptome analysis of miRNAs expressed during CNS regeneration in the invertebrate, Lymnaea stagnalis. Lymnaea is a useful model organism for the study of adult CNS regeneration due to its extensive regenerative capacity, and ease of isolation of large, identifiable neurons for cell culture. We identified 483 miRNAs in the adult Lymnaea CNS, and discovered a specific subset that may contribute to RA-induced regeneration. In particular, we focused on one neuronally enriched miRNA, miR-124. Using RT-qPCR, we confirmed the upregulation of miR-124 during regeneration, and utilizing in situ hybridization, found that it was

Discussion
In this study, we performed the first transcriptome analysis of miRNAs expressed during CNS regeneration in the invertebrate, Lymnaea stagnalis. Lymnaea is a useful model organism for the study of adult CNS regeneration due to its extensive regenerative capacity, and ease of isolation of large, identifiable neurons for cell culture. We identified 483 miRNAs in the adult Lymnaea CNS, and discovered a specific subset that may contribute to RA-induced regeneration. In particular, we focused on one neuronally enriched miRNA, miR-124. Using RT-qPCR, we confirmed the upregulation of miR-124 during regeneration, and utilizing in situ hybridization, found that it was present within motorneurons. Interestingly, we found miR-124 was enriched in the growth cones of PeA motorneurons, but was restricted to the cell bodies and/or neurites of RPA motorneurons. Together, these data are suggestive of a role for miR-124 in RA-induced CNS regeneration.

Lymnaea Stagnalis miRNA Transcriptome
Using miRNA sequencing, we identified a large subset of 483 miRNA sequences in the snail CNS. Of these 483 identified sequences, 264 miRNAs were conserved in other molluscan species, while 219 represented a group of unique miRNAs that may be Lymnaea-specific. Interestingly, the number of novel miRNAs within the Lymnaea CNS was relatively high; representing 45% of all sequenced reads. In comparison, in the regenerating axolotl tail, fewer than 12% of the total sequences were identified as novel miRNAs [71]. Rather than reflecting an overabundance of novel Lymnaea-specific miRNA sequences, this finding may be due to the minimal sequence information available for the molluscan miRNA transcriptome. We expect that these potential novel miRNAs will be identified in other molluscan species as more sequence data becomes available.
Interestingly, many of these novel sequences were differentially regulated during CNS regeneration. Specifically, 86 novel miRNA sequences exhibited at least a 2-fold increase or decrease in RA-treated CNS, corresponding to 39% of all the novel sequences. We found that many of the known or characterized mature miRNAs that were differentially regulated contributed to similar biological processes, including neuronal differentiation, proliferation, neurite guidance, or synaptogenesis. As such, we predict the differentially regulated novel sequences may contribute to similar events during CNS regeneration. However, a majority of the novel miRNAs exhibited an extremely low number of reads, generally much lower than miR-133 (which was undetectable by in situ hybridization). With such low abundance, characterizing the specific functions of these novel sequences will prove to be difficult, as they may be undetectable by most standard molecular techniques.

miR-124 Expression Patterns in Lymnaea CNS
Using both miRNA sequencing and RT-qPCR analyses, we found miR-124 was abundant within the adult Lymnaea CNS, and was upregulated in regenerating CNS, implicating its potential role in molluscan CNS regeneration. miR-124 is a well-characterized miRNA that is predominantly expressed in neuronal cells, and regulates a variety of processes, including neuronal differentiation [47,48], neurite outgrowth [52], neuronal cell fate [47,61], and the transition from neural progenitors to mature neurons [47]. Importantly, miR-124 has also been shown to regulate CNS regeneration in flatworms of the class Turbellaria [72]. When miR-124 was inhibited during planarian brain regeneration, this resulted in a significant reduction of dopaminergic and GABAergic neurons, reducing the overall brain size [72]. Together with our data, these studies support a role for miR-124 in the regeneration of invertebrate nervous systems.
In mice, miR-124 expression is 100 times higher in the CNS than in other tissues [73], and gradually increases in abundance in parallel with neuronal maturation [61]. We found similar trends in Lymnaea, as miR-124 was enriched in the CNS, compared to other tissues and organs. Moreover, we also discovered miR-124 exhibited very low levels of expression during Lymnaea development, but was highly enriched in the mature adult snail CNS. Collectively, these data indicate miR-124 shares similar expression patterns in an invertebrate species as it does in some vertebrates, and is highly enriched in the mature CNS.

Expression of miR-124 in Motorneurons
To further characterize miR-124 expression, we determined the subcellular distribution of miR-124 within individual regenerating motorneurons, and specifically, within the growth cone. The presence of miRNAs within neuronal growth cones has previously been indicative of their role in controlling neurite outgrowth and growth cone guidance [52,[74][75][76]. As such, miRNAs within the growth cone can rapidly downregulate mRNAs that might impede neurite sprouting or impede specific turning responses. In Xenopus laevis, miR-124 is localized to the growth cones of retinal ganglion cells, and regulates neurite outgrowth in response to the guidance cue, Sema3A [52]. Similarly, in our present study, we discovered that miR-124 was expressed in the growth cones of PeA ciliary motorneurons. Interestingly, we determined that the number of PeA growth cones containing miR-124 was not significantly altered when cells were cultured in the presence of RA in comparison to EtOH alone. This may indicate that miR-124 does not play an integral role in RA-induced neurite sprouting. However, miR-124 may instead contribute to fast-acting growth cone turning responses, similar to other vertebrate growth cone-specific miRNAs [52,[74][75][76]. Within vertebrate growth cones, miRNAs have been shown to regulate local protein synthesis to rapidly alter cytoskeleton dynamics in response to specific guidance cues. In Lymnaea, when local protein synthesis is inhibited in PeA growth cones, their attractive chemotropic response to RA is abolished [18]. This may implicate miR-124 as a positive regulator of local mRNA translation during fast-acting growth cone turning responses, as opposed to a substantial role in neurite sprouting.
Interestingly, we discovered that miR-124 did not share a similar expression pattern in the growth cones of RPA motorneurons. miR-124 was expressed in~40% of all PeA growth cones, but completely absent in all RPA growth cones examined. This suggests that miR-124 may not be involved in regulating growth cone guidance in all motorneuron cell types. Indeed, it is likely that different miRNAs may mediate the turning behaviors of different classes of motorneurons, possibly in response to a variety of guidance cues encountered during innervation of different targets during development.
In this study, we detected miR-124 in the cell bodies of both populations of Lymnaea motorneurons studied, though, interestingly, it was not previously detected in the motorneurons of another molluscan species. In Aplysia californica, miR-124 was found to be an essential regulatory molecule at the cultured sensory-motor synapses, where it regulated the transcription factor, CREB [56]. miR-124 was, however, primarily expressed in the sensory neurons and was undetected in the motorneurons [56]. However, the sensory and motorneurons were cultured together, and formed synaptic connections [56], which may have affected expression levels in either cell, compared to neurons cultured in isolation. It is possible that miR-124 may exhibit a higher abundance in motorneurons prior to detecting a synaptic partner. Alternatively, the expression of miR-124 may be species-specific and/or cell type-specific. Due to the limited number of sensory neurons that have been identified in Lymnaea, we did not compare the expression patterns of this miRNA between sensory and motorneurons in this study.

Role of miR-124 during RA-Induced CNS Regeneration
Using miRNA sequencing, we demonstrated that miR-124 was upregulated in RA-treated CNS, a trend that has also been described in vertebrate cell cultures [58,59]. In Lymnaea, the expression of both nuclear receptors that bind retinoic acid, RXR [19] and RAR [20], increase during CNS development. Similarly, a specific RAR subtype, RARβ, has been shown to increase in a stage-specific manner during CNS regeneration in the adult newt [13]. As both RXR and RAR are likely critical during RA-induced regeneration [13,77], and expressed in Lymnaea PeA growth cones [19,20], it is possible that miR-124 could be targeting either mRNA sequence during specific stages of regeneration. Indeed, the 3 UTR of the Lymnaea RAR and RXR mRNAs contain potential binding sites for miR-124, though we have not yet explored whether these mRNAs are found locally in the neurites and/or growth cones. Interestingly, during newt spinal cord regeneration, RARβ protein increases [13], while RXR is instead downregulated [77]. If similar trends are exhibited during invertebrate CNS regeneration, it is feasible that Lymnaea RXR may act as a potential target for miR-124 during Lymnaea CNS regeneration.
Alternatively, miR-124 may target mRNAs responsible for degradation of RA. Specifically, Cytochrome P450 protein 26 (Cyp26) is responsible for degrading all-trans RA, and is expressed in Lymnaea (Genbank Accession No. KF669878). Indeed, the 3 UTR of the Lymnaea Cyp26 mRNA contains multiple binding sites for miR-124, suggesting this mRNA may also act as a potential binding site for miR-124 during RA-induced regeneration. However, its presence within PeA or RPA neurites and growth cones has not yet been determined. Alternatively, miR-124 may target mRNAs that may impede neuronal outgrowth. However, as the Lymnaea genome has not yet been sequenced, it is difficult to obtain an exhaustive list of potential Lymnaea mRNA targets for miR-124 at this time.
In summary, this study provides the first miRNA sequencing analysis of miRNAs in the regenerating CNS of the mollusc, Lymnaea stagnalis and, more importantly, characterizes classes of both novel and conserved miRNAs that are regulated during RA-induced regeneration of the adult CNS in this invertebrate. We also demonstrate that a specific, conserved miRNA, miR-124, is abundant in the adult snail CNS, as it is in vertebrates. Furthermore, we provide new evidence for cell type-specific expression of miR-124 in the growth cones of different classes of motorneurons. In future studies, it will be important to characterize the expression patterns and potential targets of our other miRNAs whose expression is mediated by RA signaling, and to identify mRNA targets and functions of these miRNAs during CNS regeneration.

Isolation of CNS
Lymnaea stagnalis were bred in the laboratory and kept in aerated pond water at room temperature on a 12 h/12 h light-dark cycle. For all experimental procedures, adult snails were anesthetized in 25% Listerine ® (containing menthol; 0.042% w/v, Johnson & Johnson Inc., Markham, ON, Canada) in pond water prior to removal of the central ring ganglia (CNS).

Regenerating CNS Preparation
Isolated Lymnaea CNS were incubated in 3 mL of defined medium (DM; comprised of 50% Leibovitz's L-15 (Gibco, Dublin, Ireland) and additional salts) containing 10 −5 M RA (Sigma, Oakville, ON, Canada) (to induce neural regeneration/neurite sprouting) in a plastic Falcon dish (VWR, Radnor, PA, USA) for 72 h. Different CNS were also incubated in 0.1% EtOH (Greenfield Global, Brampton, ON, Canada) as a vehicle control. Following the 72 h incubation period, any neurite sprouting from individual nerves emanating from the CNS were imaged using Q-Capture imaging software (v2.90.1, Q Imaging, Surrey, BC, Canada).

RNA Sequencing of Lymnaea miRNAs
RA-treated (regenerating) and EtOH-treated (non-regenerating) samples were collected by pooling five-CNS from adult Lymnaea (shell length of 20-25 mm) per sample (

Isolation of Lymnaea Embryos
Egg masses were incubated in pond water at room temperature for 6, 8, and 10 days post-egg laying, corresponding to various stages of Lymnaea development (as described by Nagy and Elekes [66]). At each developmental stage (day 6, 8, and 10), egg capsules were removed from their gelatinous surroundings. Following isolation, all embryos encased in one egg mass were pooled and frozen on liquid nitrogen for molecular analysis.

Cell Culture
Adult snails (16 to 20 mm in length) were used for all cell culture procedures. Following isolation of the Lymnaea CNS, individual ganglia were desheathed to expose cells of interest, including pedal A (PeA) and right parietal A (RPA) motorneurons. Individually identified neurons were removed from the ganglia using suction applied via a fire polished pipette, and then plated on poly-L-lysine (Sigma)-coated Falcon dishes (VWR). Culture dishes contained 3 mL of conditioned medium (CM), which contain unidentified trophic factors that can produce neurite outgrowth [18,78]. In addition, 10 −7 M RA was added to the culture dishes overnight to promote neurite sprouting [10,17]. All culture dishes were maintained at 21 • C overnight. To determine the effects of RA on the proportion of growth cones expressing miR-124, cells were cultured in either 10 −7 M RA or 0.001% EtOH (as the vehicle control).

RNA Isolation and cDNA Synthesis
All Lymnaea organs and tissues were isolated from adult Lymnaea stagnalis with a shell length of 20-25 mm, then immediately flash frozen in liquid nitrogen. For molecular analyses, one pooled sample for each organ contained: 2 CNS, 10 hearts, 10 albumen organs, 4 prostates, or 4 buccal masses. For analysis of individual CNS ganglia, the following were pooled for each sample: 20 pedal ganglia, or 31 right parietal ganglia. For regenerating CNS samples, each biological replicate contained five pooled CNS. For all experiments, 3 biological replicates were utilized.
Total RNA was isolated from these samples using TRI Reagent (Sigma) and Direct-zol RNA MiniPrep kit (Zymo Research). RNA quality was confirmed using spectrophotometry and gel electrophoresis. A total of 750 ng of RNA was utilized from each sample for cDNA synthesis using gene specific stem-loop primers with the SuperScript III Reverse Transcriptase kit (Invitrogen, Burlington, ON, Canada). Stem-loop primers were designed for miR-124 (RT: GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACGGCATT) and miR-133 (RT: GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACAGCTGG) based on sequences provided by miRNA sequencing.

LNA-FISH and Tyramide Signal Amplification
Cultured cells or CNS were fixed in 4% paraformaldehyde (Sigma) for 20 min at room temperature, then stored at 4 • C in phosphate-buffered saline (PBS; Sigma) until required. Prior to staining procedures, both cells and CNS were treated with 3% H 2 O 2 (Sigma) for 1 h at room temperature to remove endogenous peroxidase activity.
CNS were next washed, on rotation, with increasing sucrose concentrations: 10% for 30 min, 20% for 30 min, and 30% overnight at 4 • C. Samples were then embedded in Optimal Cutting Temperature (O.C.T.) compound (Tissue-Tek, Sakura, Osaka, Japan), and 12 µm sections were obtained using a cryostat (Leica Microsystems, Richmond Hill, ON, Canada). Tissue sections were mounted on Superfrost Plus slides (Fisher Scientific, New Hampshire, NH, USA).

Statistics
All data were analyzed utilizing GraphPad Prism, Version 7.0 for Mac OS X (La Jolla, CA, USA), and values were expressed as the mean ± SEM. For statistical analysis investigating miR-124 expression during Lymnaea development, a one-way analysis of variance (ANOVA) was performed, followed by Tukey's post hoc test. For all other statistical analyses, a Students unpaired t-test was performed. For all analyses, a p value less than 0.05 was considered to be statistically significant.