C57BL6/J mice (Jackson Laboratory, Bar Harbor, ME) were maintained under 12h:12h light:dark cycles at standard temperature and humidity with food and water provided ad libitum at the Bartlett Animal Facility (Rutgers-New Brunswick). Mice were monitored daily, and cages were routinely changed. Mice were bred for litters to use in generating cultures. All animal experiments were approved by the Rutgers Institutional Animal Care and Use Committee (IACUC Protocol # 10-024).

Cultures of Medium Spiny Neurons (MSN) at ∼95% purity [27, 28], were prepared following well-established protocols [29–31]. At day 5 after birth (P5), pups were decapitated, and brains were immediately removed and placed into a 60 mm plate containing ice-cold CMF-HBSS (100 mL of final solution made with 10 mL 10x HBSS (Life Technologies), 0.7 mL 5% NaHCO₃ (Sigma), final pH 7.1, brought to final volume with ultrapure water, then filter sterilized and stored at room temperature). The Nucleus Accumbens (NAc) was removed using a mouse brain atlas for visual reference [32]. Equal numbers of male and female pups were used for each preparation to limit the bias of using a single gender. Striatal tissue was diced into smaller fragments, ∼1 mm in diameter in 3.15 mL cold CMF-HBSS, and trypsinized with the addition of 0.35 mL of 2.5% trypsin at 37°C. After 10 min, 8 mL DMEM-FBS medium (178 mL DMEM (high glucose, no sodium pyruvate, no glutamine (Irvine Scientist), with 20 mL FBS, 0.5 mL 10,000 u Pen/Strep and 2 mL 200 mM glutamine (Life Technologies), stored at 4°C in the dark) was added, and the solution was centrifuged at 300 x g for 5 min to pellet the cells. The medium was aspirated from the tube, and 5 mL of Growth Medium (100 mL DMEM/F12 (with Glutamax; Life Technologies), with 2 mL FBS (1.9% v/v), 2 mL B-27 (1.9% v/v; Life Technologies), and 1 mL penicillin/streptomycin (0.95% v/v), stored at 4°C in the dark) was added. Trituration to further break apart aggregates was conducted using a fire-polished Pasteur pipet, and the tube was spun again as above. Excess media was removed, and the cells were resuspended in 10 mL of Growth Medium. Preparations were conducted under a sterile hood to help maintain sterility, except for centrifugation. The concentration of live cells/mL was estimated by hemocytometer counts of live:dead cells using trypan blue, and the cell stock solution was diluted to a final concentration of 0.5 × 10⁶ cells/mL. Plates of striatal cells were prepared by seeding 2 mL of the cell stock solution onto 35 mm cell culture plates that had been coated with ornithine (Sigma-Aldrich, St. Louis, MO) and laminin (Life Technologies, Carlsbad, CA) for cell adherence and enrichment for neurons. Preparations were rapidly conducted as speed was critical for cell viability. Twenty-four hours post-seeding, after allowing cells to adhere, the media was replaced with a 2 mL Neurobasal Medium (NB)/plate (100 mL Neurobasal A Medium supplemented to 2.0 mM glutamine final concentration; Life Technologies), with 2 mL FBS (1.9% v/v), 2 mL B-27 (1.9% v/v; Life Technologies), and 1 mL penicillin/streptomycin (0.95% v/v) stored at 4°C in the dark). Cultures were maintained at 37°C/5% CO2 in a cell culture incubator with saturated humidity for another week before starting exposures, and as a result, the neurons were nearly 2 weeks old since birth (5 days in vivo + 8 days in vitro = 13 days total) at the start of the experiments. Sometimes in microRNA studies alpha-amanitin is added to cultured cells to inhibit RNA polymerases II and III, which process microRNA. Since alpha-amanitin can also cause widespread transcriptional stress and apoptosis [33, 34] we did not add it to our cultures.

We chose 20 mM ethanol for alcohol exposures as it represents a physiologically relevant dose of alcohol while maintaining cell viability. 20 mM ethanol corresponds to a 0.092% Blood Alcohol Content (BAC), which can be achieved in humans by a quick (30–60 min) consumption of 3–4 standard drinks of alcohol by a 150-pound individual [35] causing disinhibition, impaired thinking, and potential DWI/DUI in the US [35]. Previously we have shown that 20 mM ethanol can upregulate the expression of miR-9-5p within 15 min after exposure of the rat brain organotypic cultures containing supraoptic nucleus (SON) neurons leading to alcohol tolerance [17]. Importantly, 20 mM causes minimal neuronal cell death in culture as shown by us [12] and others [36].

Seven days after seeding neurons the cells were subjected to the alcohol exposure and withdrawal with the following collection time points: Control = 0 min exposure +0 h withdrawal, 15 min 20 mM ethanol exposure +0 h withdrawal, 15 min 20 mM ethanol exposure +1 h withdrawal, 15 min 20 mM ethanol exposure +6 h withdrawal, 15 min 20 mM ethanol exposure +12 h withdrawal, 15 min 20 mM ethanol exposure +24 h withdrawal, 6 h 20 mM ethanol exposure +0 h withdrawal, 6 h 20 mM ethanol exposure +6 h withdrawal, 6 h 20 mM ethanol exposure +24 h withdrawal. Collection at each time point was conducted in triplicate. For each control, 5–7 plates were prepared. Cells were treated by aspirating off media and replacing with either a neurobasal medium (“media only” control) or a neurobasal medium with alcohol (NBE with 20 mM final ethanol concentration). Ethanol evaporation was minimized by maintaining NBE plates in a semi-sealed container in the incubator with saturated humidity and additional plates of medium containing the same concentration of ethanol, based upon the methods of Pietrzykowski [12, 17].

After a defined length of exposure, NB or NBE media were removed. For plates without a withdrawal period, cells were collected immediately. For cells with a withdrawal period, the NB medium replaced the NBE medium for a defined length of time after which cells were collected.

Cell collection was carried out by quickly rinsing plates with 2 mL ice-cold PBS followed by scraping cells from the plate with a cell scraper in 200 uL PBS. Cells were immediately flash-frozen in liquid nitrogen and stored at −80°C until processed for total RNA isolation as described previously [16].

Media samples were gathered at each collection point in the experimental process (before and after ethanol addition, during exposure and withdrawal) to verify alcohol concentration. Alcohol measurements were conducted using an AMI Analyzer according to the manufacturer’s instructions (Analox Instruments Ltd., Lunenburg, MA). 10, 20, and 50 mM ethanol standards in media were used to calibrate the instrument prior to reads to ensure accuracy. Alcohol measurements confirmed that ethanol loss was minimized using our methodology as previously described in more detail [12, 17].

Total RNA or Small RNA (for precursor assays) was isolated by miRVana kit according to the manufacturer’s instructions (Life Technologies). Concentration and purity were analyzed by a Nanodrop 1000 Spectrophotometer (Thermo Fisher Scientific Inc., Wilmington, DE), and aliquots of each sample were used to prepare 10ng/ul dilutions in nuclease-free water for RT-qPCR. All samples were kept at −80°C.

Expression data for statistical analysis were obtained using oligos in a standard curve method for mature miR-9-5p and miR-9-3p, 2^−ΔΔCT method for pre-precursors, and Poisson distribution for pri-precursors. The data were analyzed using unpaired, two-tailed t-tests. Data were expressed as fold-change to visualize the relationship between exposure condition and molecule expression. p-value below 0.05 (p < 0.05) was set as statistically significant.

miR-9-5p is a prominent brain microRNA regulated by alcohol. Some reports describe the stimulatory effect of alcohol on miR-9-5p expression [17, 40], while others report the opposite effects [41]. To better understand the intricacies of miR-9-5p regulation by alcohol, we first exposed murine primary neuronal cultures to physiologically relevant 20 mM ethanol for 15 min (the short exposure) and measured its expression at various times after alcohol withdrawal up to 24 h post-exposure (Figure 2A).

We observed that after the short exposure expression levels of miR-9-5p increased almost two-fold (Figure 2A, left bars), in accordance with previously published findings [17]. Alcohol withdrawal caused a fast decrease of the elevated levels of miR-9-5p even below the pre-exposure, normal levels within 1 h after the start of the exposure (Figure 2A, left bars). In the alcohol-free environment, miR-9-5p levels decreased even further with time, reaching the lowest levels of around 40% of the pre-exposure levels at the 6 h post-exposure mark. Somewhere between 6 h and 12 h of the withdrawal miR-9-5p levels started to rebound from their nadir point and went back to the pre-exposure levels (Figure 2A, left bars). They reached the pre-exposure levels 12 h after the exposure and maintained normal levels up to 24 h after the exposure (Figure 2A, left bars).

Although miR-9-5p is the most recognized final product of miR-9 biogenesis, miR-9-3p also plays an important role in neural development [42] and neuronal differentiation [43] with more predicted targets then miR-9-5p (Supplementary Table S1, miR-9-5p: 1242 targets; Supplementary Table S2, miR-9-3p: 4334 targets). Interestingly, there is a quite large overlap of targets between these two microRNAs: over 34% of miR-9-5p targets are also targeted by miR-9-3p (425 targets, Supplementary Table S3).

We observed that short alcohol exposure also regulates the expression of miR-9-3p. The short exposure increased expression of miR-9-3p (Figure 2A, right bars) similar to its effect on the miR-9-5p expression. In contrast to miR-9-5p however, after the removal of alcohol, miR-9-3p levels continue to rise, reaching significantly higher levels 6 h post-exposure (Figure 2A, right bars).

After reaching the peak of expression, miR-9-3p levels return down to pre-exposure levels at the 12 h post-exposure timepoint and maintain that normal level up to the 24 h post-exposure, mimicking temporal dynamics of miR-9-5p expression changes within 12–24 h post-exposure time interval (Figure 2A, right bars).

It seems that, based on changes in the expression of both microRNAs, two withdrawal periods triggered by short alcohol exposure could be distinguished: the early period starting immediately after alcohol withdrawal and lasting around 6 h, and the late period following the early one and lasting up to the 24-hour post-exposure timepoint (Figure 2A).

During both time periods, the expression of miR-9-5p and miR-9-3p seems to be tightly associated with each other as determined by correlation analysis. During the early withdrawal period changes in miR-9-3p and miR-9-5p expression are strongly and negatively correlated (Figure 2B; Table 1, correlation coefficient r = −0.775). During the late withdrawal period, changes in the expression of miR-9-3p and miR-9-5p are moderately and positively correlated (Figure 2C; Table 2, correlation coefficient r = 0.55).

We compared the short exposure results with the expression of miR-9-5p under the continuous presence of the drug for up to 24 h (continuous exposure).

We assumed that the continuous presence of the drug would maintain the elevated plateau of miR-9-5p since exposure to alcohol increased miR-9-5p levels in the first place. However, it was not the case. We observed that despite alcohol presence, after the initial increase, miR-9-5p levels dropped within 6 h post-exposure (Figure 2B, left bars) and then increased (Figure 2B, left bars) with a similar temporal dynamic seen with the short exposure. Interestingly, in the continuing presence of alcohol beyond 6 h the miR-9-5p expression pattern shifted upwards above the pre-exposure levels presumably trying to set a new, higher equilibrium (Figure 2B, left bars).

During the continuous exposure to alcohol, the miR-9-3p expression did not change sufficiently to achieve standard statistical significance (p < 0.05) except for the last timepoint (Figure 3A, 24 h exposure). However, the changes of the miR-9-3p expression tightly followed the changes of the miR-9-5p expression, showing a strong and positive correlation at each timepoint studied (Figure 3A). We think that two periods with similar time frames can be distinguished here as well based on changes in the expression pattern: the early exposure period starting soon after alcohol addition and lasting about 6 h (Figure 3A) with a correlation coefficient r = 0.720 (Figure 3B; Table 3), and the late exposure period following the first one up to the 24-hour of alcohol exposure (Figure 3C; Table 4) with the correlation coefficient r = 0.853.

Both miR-9-5p and miR-9-3p are final products of miR-9 biogenesis (Figure 1, and ref 19). Three separate biogenesis pathways of the miR-9-5p/miR-9-3p pair start with each miR-9 gene generating its own pri-mir-9 precursor, and subsequently pre-mir-9 precursor, which ultimately contributes to the mature miR-9-5p and the mature miR-9-3p pools (Figure 1). We decided to determine the effects of both, the short and the continuous alcohol exposure, on the expression of all of these precursors.

We observed that the short alcohol exposure (15 min) had no effect on the expression of all three pre-mir-9 precursors (Figure 4A). Expression levels of none of the precursors changed immediately after the alcohol exposure. Since they remained consistently at the same, unchanged level for 6 h following the alcohol withdrawal (Figure 4A) we did not explore further time points.

In contrast, within the same timeframe of the early period of withdrawal, we observed a robust, over 2-fold upregulation of pri-mir-9-2 precursor expression by short alcohol exposure immediately following the exposure (Figure 4B). The pri-mir-9-2 precursor expression upregulation was sustained for at least 6 h after the alcohol withdrawal (Figure 4B). This effect was not observable for the other two pri-precursors: pri-mir-9-1 and the pri-mir-9-3 (Figure 4B).

Overall, it seems that a short, 15 min alcohol exposure elicited changes in the expression of pri-miR-9-2 precursor only and that these changes were quick, robust, and unceasing in alcohol absence.

The long (6 h) alcohol exposure affected the expression of both, pre- and pri-mir-9 precursors.

Both, pre-mir-9-1, and pre-mir-9-2 were significantly downregulated after 6 h of alcohol exposure, with pre-mir-9-3 following this trend but not reaching a statistical significance at p < 0.05 yet (Figure 5A). Withdrawal of alcohol for 6 h after the 6 hr-long exposure to the drug did not restore expression levels of any of the pre-mir-9 precursors with all of them being decreased. The decreased expressions of all three pre-miR-9 precursors continued in the absence of alcohol for up to 24 h after alcohol withdrawal (Figure 5A).

The effects of the long (6 h) alcohol exposure on the expression levels of pri-miR-9 precursors also affected all of these precursors but each in a different way (Figure 5B). The expression of the pri-mir-9-1 precursor was consistently downregulated to about 50% of its pre-exposure levels, and this downregulation persisted in the absence of alcohol for up to 24 h after alcohol withdrawal (Figure 5B). In contrast, the expression of the remaining two pri-precursors (pri-mir-9-2, pri-mir-9-3) was significantly upregulated by the long (6 h) alcohol exposure to about 1.5-fold above their pre-exposure levels. After alcohol withdrawal, the upregulated levels of both pri-precursors were sustained (Figure 5B). The pri-miR-9-3 precursor maintained its 1.5-fold upregulation at both, 6 h and 24 h after alcohol withdrawal (Figure 5B), while the pri-miR-9-2 precursor expression levels 6 h after alcohol withdrawal went even further up, reaching above 2-fold upregulation, and maintaining their higher expression levels 24 h post-exposure (Figure 5B).

Overall, it seems that longer alcohol exposure elicited wider changes in the expression of miR-9 precursors, affecting the expression of all precursors. Nevertheless, it seems that the miR-9-2 biogenesis pathway responded in the most striking way.

After observing previously the biological effects of a short exposure to 20 mM alcohol [17], we questioned whether the upregulated miR-9-5p levels persist after alcohol withdrawal and for how long. We determined here that the short alcohol exposure triggered changes in miR-9-5p expression observable during withdrawal. These changes could be divided into two, subsequent phases: 1/downregulation below the pre-exposure level, and 2/upregulation to the pre-exposure level. These phases seem to follow a pattern of homeostatic regulation, during which miR-9-5p levels thrown off of the steady state equilibrium by alcohol exposure would undergo changes after alcohol withdrawal to return eventually to the pre-exposure status quo. Based on our collection time points we observed that MSN neurons need roughly around 6–12 h of the drug withdrawal to return miR-9-5p levels to normal (i.e., the pre-exposure steady-state equilibrium). It would be of interest, in the next studies, to further narrow down the time window of this homeostatic adaptation.

Recently, miR-9-3p, the passenger strand derived from the same duplex as miR-9-5p, gained recognition as biologically active on its own [43, 55]. Martinez et al. [56] showed that chronic ethanol exposure over the course of 55 days elevated miR-9-3p in the serum of rats. Balaraman proposed that the ratio between these two mature microRNAs is important in the regulation of neuronal differentiation and in the development of cancer [57]. Both microRNAs impact the differentiation of neural stem cells through the co-regulation of a transcription factor, REST (RE1 silencing transcription factor/neuron-restrictive silencer factor). miR-9-5p targets REST directly, while miR-9-3p regulates the expression of coREST, a cofactor of REST [42, 57]. Therefore, miR-9-5p and miR-9-3p working in tandem can create various combinations of REST:coREST, thus influencing neuronal differentiation [58, 59].

Our results postulate an even tighter, joint effect of miR-9-5p and miR-9-3p on gene expression. Simultaneous downregulation of miR-9-5p and upregulation of miR-9-3p observed at some points, and their convergence on a large number of targets (425 transcripts, over 34% of miR-9-5p targets) could be a neuronal attempt to attenuate, at least some acute alcohol effects on miR-9-5p targets, consistent with a homeostatic response and preservation of pre-exposure equilibrium. We also would like to propose that any future studies focused on the regulation of miR-9-5p and its targets by alcohol or other factors should include miR-9-3p and its targets as well.

We expected that in the continuous presence of alcohol, upregulated levels of miR-9-5p and miR-9-3p would be maintained. However, that was not the case. We observed that despite the continuous presence of alcohol, both miR-9-5p and miR-9-3p levels followed a response pattern, similar to one observed in a short exposure/withdrawal experiment, which could be also divided into two phases: 1/initial downregulation, 2/subsequent upregulation, with a demarcation line between these two phases happening after 6–12 h of alcohol exposure. Interestingly, the final outcome after 24 h of alcohol exposure was a significant upregulation of both miR-9-5p and miR-9-3p. One could interpret these results as a neuronal adaptation at the molecular level to the continuous presence of alcohol by attempting to set up a new, overcorrected set-point of miR-9-5p and miR-9-3p expression despite the continued presence of the drug. This is consistent with achieving new stability through change—a tenet of allostasis [60, 61] and the allostatic model of addiction [62].

Since we observed the presence of the initial phase triggered by a short or continuous exposure lasting about 6 h, we decided to determine whether alcohol differently affects upstream elements of the miR-9 biogenesis pathway (precursors) during that window using two scenarios. First, we used this window as a withdrawal window preceded by the short exposure, second, we used it as an exposure window followed by a 24-hour withdrawal period. As we observed no measurable effect on all three miR-9 pre-precursors’ expression levels during the first scenario, we concluded that the likelihood of alcohol affecting the steps of microRNA biogenesis responsible for the production of pre-precursors from pri-precursors is rather low. However, in scenario 2, alcohol downregulated all three miR-9 pre-precursors suggesting a possibility that alcohol could interfere with one or some of the steps producing pre-precursors from pri-precursors. Production of microRNA pre-precursors starts with pri-precursors cleavage by the Microprocessor machinery, followed by export from the nucleus to the cytoplasm by the exportin5 complex, and capture by Dicer for further processing by the RISC complex [63]. There are many proteins involved in microRNA biogenesis as each microRNA processing complex consists of several proteins. The microprocessor contains Drosha, DGCR8, RIIa and RIIb proteins, and Exportin-5, a mediator of nuclear export that needs a cofactor RanGTP protein [63], while Dicer works with auxiliary proteins TRBP and members of the Argonaute protein family (AGO) to form the RISC complex [63]. It is possible that some of the regulation of precursors by alcohol reported here is due to an alcohol effect on some of these proteins. Indeed, Mulligan [64] showed an association between Drosha and Dicer expression and response to alcohol, while Prins [65] determined that, in the rat hippocampus, alcohol alters Drosha and Dicer expression (also see 18). Moreover, Gedik [66] reported a genetic association of DGCR8, AGO1, and AGO2 alleles with alcohol dependence risk. It would be of great interest to gain a full picture of alcohol regulation of activity of the key elements of the microRNA biogenesis complexes which process precursors.

In order to better understand the temporal regulation of miR-9 expression by alcohol we should also focus our future efforts on the initial steps of the biogenesis, namely, the production of the miR-9 pri-precursors from their respective genes. Our results revealed that even the short alcohol exposure triggered upregulation of pri-mir-9-2, while longer exposure affected the expression of all three miR-9 pri-precursors. At this point we cannot rule out any mechanisms regulating miR-9 gene expression; however, we suspect epigenetic control of the mir-9-2 gene expression by alcohol. Pappalardo-Carter et al. [67] showed that alcohol increases CpG dinucleotide methylation of the mir-9-2 gene promoter. We believe that further, comprehensive studies of the epigenetic regulation of all miR-9 genes by alcohol are fully warranted.

Further studies should also shed some light on the differential regulation of miR-9 expression not only by the temporal aspect of alcohol exposure but also by alcohol concentration. Pappalardo-Carter [67] reported that a high alcohol concentration (130 mM) suppressed miR-9 expression, while Tapocik [68] showed that an alcohol concentration of 70 mM inhibited the expression of miR-9, creating a lower steady-state level in alcohol-dependent rats.

AUD is known to have a genetic component [69, 70]. Because human miR-9-1 and miR-9-3 genes are located near or within the AUD susceptibility loci [8], we believe that exploring the differential effects of alcohol on each miR-9 biogenesis pathway is also of great importance in understanding the genetic predisposition to AUD. We would hypothesize that the first response to alcohol exposure is mostly through the mir-9-2 biogenesis pathway. However, with continuous exposure (longer than 6 h) or possibly multiple exposures (mimicking frequent drinking characteristic of the AUD) the mir-9-2 gene may be eventually substantially turned down, with the remaining contribution shifting to mir-9-1 and mir-9-3 genes. As these two genes combined produce less miR-9 than mir-9-2, this hypothesis would explain lower levels of miR-9 observed in chronic alcohol exposure experiments [67, 68]. This hypothesis would also rationalize the presence of miR-9-1 and miR-9-3 genes in the AUD susceptibility loci.

It is worth mentioning that a deeper understanding of alcohol regulation of miR-9 biogenesis would also benefit research focused on cancer and neurodegenerative diseases. Aberrant levels of miR-9 (either miR-9-5p, miR-9-3p, or both) have been reported in many types of cancer [71]: breast cancer [25, 72, 73], Burkitt’s Lymphoma [24], hepatocarcinoma [74], prostate cancer [75], gastric cancer [76], colorectal cancer [77], as well as Alzheimer’s and Huntington’s diseases [78]. Chronic heavy alcohol consumption increases the risk of all of these cancers and neurodegenerative diseases [79–86].