Abstract
6-methyladenosine (m6A) RNA methylation, regulated by writer, eraser, and reader proteins, modulates mRNA stability, splicing, and translation, thereby influencing key cellular processes. Environmental stressors, such as alcohol, may disrupt this epitranscriptomic machinery and contribute to disease vulnerability. In this study, we investigated how chronic exposure to ethanol, its toxic metabolite acetaldehyde, and subsequent withdrawal affect the expression of m6A regulatory genes. Neuron-like (SH-SY5Y) and non-neuronal (SW620) cells were exposed for 3 weeks to ethanol (40 mM) or acetaldehyde (30 μM) (concentrations comparable to blood levels after heavy drinking), followed by a 24-h withdrawal period. Gene expression of seven writers (KIAA1429, METTL3, METTL4, METTL14, RBM15, RBM15B, and WTAP), two erasers (ALKBH5, FTO), and nine readers (YTHDF1/2/3, YTHDC1/2, IGF2BP1/2/3, and HNRNPA2B1) was quantified by RT-qPCR. Concurrently, RNA-seq data from eight reward-related brain regions of 24 individuals of European ancestry (12 with alcohol use disorder [AUD] and 12 controls) were analyzed for AUD-associated expression changes in m6A regulatory genes. In cell models, ethanol broadly suppressed the expression of most m6A regulatory genes, whereas withdrawal largely restored their levels. Acetaldehyde induced subtler gene expression changes, likely reflecting its lower exposure concentration and rapid metabolism. Postmortem brain analysis revealed trends toward altered expression of m6A regulatory genes across multiple brain regions in individuals with AUD. Collectively, these findings suggest that chronic alcohol exposure dysregulates m6A regulatory gene expression and may impact downstream RNA regulatory pathways involved in AUD pathophysiology. Further studies are warranted to elucidate the mechanisms by which alcohol-induced dysregulation of m6A regulators influences AUD risk.
Keywords: m6A writers, m6A erasers, m6A readers, ethanol/acetaldehyde exposure/withdrawal, RT-qPCR, cellular model, human postmortem brains, RNA-seq
Introduction
6-methyladenosine (m6A) is the most prevalent internal RNA modification in eukaryotes, regulating gene expression by influencing RNA splicing, translation, and stability [1]. m6A RNA methylation is essential for the development and function of multiple biological systems, including the hematopoietic, reproductive, and central nervous systems. This modification occurs across diverse RNA species, including messenger RNA (mRNA), transfer RNA (tRNA), ribosomal RNA (rRNA), microRNA (miRNA), long noncoding RNA (lncRNA), and circular RNA (circRNA), and is now recognized as a key post-transcriptional regulatory mark [2]. Next-generation sequencing (NGS) studies have revealed that about one-third of mammalian mRNAs harbor m6A sites, with an average of three to five modifications per transcript [3]. These marks are enriched within the conserved DRA*CH consensus motif (where A* denotes the methylatable adenosine; D represents A, G, or U; R represents A or G; and H represents A, C, or U) and are frequently located near stop codons, as shown by transcriptome-wide m6A mapping [4, 5].
The discovery in 2011 that m6A is reversible highlighted its dynamic nature. Methyl groups can be added to or removed from the N6 position of adenosine, allowing RNA regulation to adapt to cellular and environmental cues [1]. This process is mediated by three classes of proteins: writers, erasers, and readers. Writers (e.g. METTL3, METTL14, WTAP, KIAA1429, METTL16, RBM15, and ZC3H13) form methyltransferase complexes that install m6A marks. Erasers (FTO and ALKBH5) remove these modifications [3]. Readers, including YTH domain-containing proteins, heterogeneous nuclear ribonucleoproteins (HNRNPs), IGF2 mRNA-binding proteins (IGF2BP1/2/3), and eukaryotic initiation factors (eIFs), recognize m6A-modified RNAs and mediate downstream effects [2, 6].
m6A methylation regulates transcripts involved in neurobiological processes critical for mental health. Dysregulated m6A patterns and altered expression of regulatory proteins have been implicated in psychiatric disorders, including alcohol and drug addiction. m6A regulators influence learning, memory, stress response, and synaptic plasticity, i.e. the key processes underlying addiction-related behaviors [7, 8]. Evidence from animal and cellular studies shows that addictive substances alter m6A regulatory gene expression. For example, morphine downregulated the demethylases FTO and ALKBH5 in rat cortical cultures [9], and the reader YTHDF1 was upregulated in the ventrolateral periaqueductal gray of mice undergoing morphine withdrawal [10]. Cocaine exposure reduced hippocampal FTO expression and increased global m6A levels in a conditioned place preference (CPP) model [11]. METTL14 and YTHDF1 are also critical for cocaine-related behavioral plasticity: conditional deletion of them in D1/D2 receptor-expressing neurons impaired learning and memory that are associated with dopaminergic signaling [12]. Likewise, FTO deficiency disrupts D2/D3 receptor signaling and blunts dopamine release after cocaine administration [13]. Alcohol exposure also alters m6A methylation in the brain and peripheral tissues. In the nucleus accumbens (NAc) of individuals with alcohol use disorder (AUD), 26 mRNAs were hypermethylated and three hypomethylated, many mapping to immune-related pathways such as IL-17 signaling; several lncRNAs and miRNAs also showed differential m6A methylation [14]. Beyond the brain, m6A regulators contribute to alcohol-related liver pathology. For example, METTL3 promotes alcohol-induced liver inflammation via m6A methylation of pri-miR-34a, which suppresses the anti-inflammatory factor SIRT1 [15], while YTHDF1 stabilizes collagen transcripts, promoting hepatic stellate cell activation and fibrosis [16]. Despite this growing evidence, little is known about the direct impact of ethanol and its toxic metabolite, acetaldehyde, on the expression of m6A regulatory genes. Acetaldehyde is rapidly metabolized to acetate but is highly reactive and toxic, and its impact may differ substantially from ethanol.
In this study, we examined the effects of chronic intermittent exposure to ethanol and acetaldehyde, at physiologically relevant concentrations comparable to blood levels after heavy drinking, on the expression of core m6A writer, eraser, and reader genes in neuron-like (SH-SY5Y) and non-neuronal (SW620) cells. Ethanol broadly suppressed the expression of these genes, with withdrawal largely restoring their levels, whereas acetaldehyde exerted more modest effects. To extend these findings, we analyzed AUD-associated m6A writer, eraser, and reader gene expression changes across eight reward-related or alcohol-responsive brain regions using RNA-seq data from our prior study [17]. Notably, postmortem brain analysis revealed trends toward altered expression of m6A regulatory genes across multiple brain regions in individuals with AUD
Results
m6A writer gene expression changes induced by ethanol exposure and withdrawal
One-way analysis of variance (ANOVA) and Tukey’s HSD post hoc test results for m6A writer genes following ethanol exposure and withdrawal are summarized in Table 1 (SH-SY5Y) and Table 2 (SW620). In SH-SY5Y cells, ethanol exposure significantly affected all seven writer genes (P < 0.00001–0.0006). Chronic intermittent ethanol (CIE) exposure markedly downregulated these genes, whereas a 24-h withdrawal partially restored their expression (Fig. 1a; Table 1). In SW620 cells, six writer genes (KIAA1429, METTL3, METTL4, METTL14, RBM15, and WTAP) exhibited significant ethanol effects (P = 0.00001–0.002). Pairwise comparisons revealed significant downregulation of four genes (KIAA1429, METTL3, METTL4, and RBM15) after CIE, with a trend toward recovery following withdrawal (Fig. 1b; Table 2).
Table 1.
Statistical analysis (one-way ANOVA and Tukey’s test) of ethanol exposure- and withdrawal-induced alterations in m6A regulatory gene mRNA expression in neuron-like SH-SY5Y cells.
| m6A Regulatory | ∆Ct CTL | ∆Ct CIE | ∆Ct CIE + WD | CTL vs. CIE | CTL vs. CIE + WD | CIE vs. CIE + WD | |
|---|---|---|---|---|---|---|---|
| Genes | (n = 6) | (n = 6) | (n = 6) | One-way ANOVA | Tukey’s test | Tukey’s test | Tukey’s test |
| Writers | |||||||
| KIAA1429 | 6.1718 ± 0.1082 | 7.2488 ± 0.0586 | 6.8648 ± 0.1532 | F(2,14) = 147.10, P < 0.00 001 | Q = 23.21, P < 0.00 001 | Q = 14.93, P < 0.00 001 | Q = 8.27, P = 0.0001 |
| METTL3 | 6.0163 ± 0.1003 | 7.2462 ± 0.1337 | 6.7880 ± 0.2376 | F(2,15) = 82.40, P < 0.00 001 | Q = 17.96, P < 0.00 001 | Q = 11.27, P < 0.00 001 | Q = 6.69, P = 0.0007 |
| METTL4 | 8.8268 ± 0.0899 | 9.6820 ± 0.1228 | 9.4805 ± 0.2450 | F(2,15) = 43.25, P < 0.00 001 | Q = 12.58, P < 0.00 001 | Q = 9.62, P = 0.00 002 | Q = 2.96, P = 0.124 |
| METTL14 | 6.9060 ± 0.1559 | 8.2393 ± 0.0575 | 7.8832 ± 0.2507 | F(2,15) = 94.84, P < 0.00 001 | Q = 18.81, P < 0.00 001 | Q = 13.79, P < 0.00 001 | Q = 5.02, P = 0.008 |
| RBM15 | 4.9725 ± 0.1325 | 5.6122 ± 0.1065 | 5.1247 ± 0.1043 | F(2,15) = 50.54, P < 0.00 001 | Q = 13.61, P < 0.00 001 | Q = 3.24, P = 0.088 | Q = 10.37, P = 0.00 001 |
| RBM15B | 5.0718 ± 0.0892 | 5.4618 ± 0.1165 | 5.2633 ± 0.1827 | F(2,15) = 12.47, P = 0.0006 | Q = 7.06, P = 0.0004 | Q = 3.47, P = 0.066 | Q = 3.59, P = 0.056 |
| WTAP | 5.2257 ± 0.1220 | 6.0540 ± 0.1217 | 5.5053 ± 0.2582 | F(2,15) = 33.17, P < 0.00 001 | Q = 11.32, P < 0.00 001 | Q = 3.82, P = 0.041 | Q = 7.50, P = 0.0002 |
| Erasers | |||||||
| FTO | 6.2952 ± 0.2176 | 7.2152 ± 0.2505 | 6.6326 ± 0.1930 | F(2,14) = 25.95, P = 0.00 002 | Q = 9.76, P = 0.00 002 | Q = 3.58, P = 0.058 | Q = 6.18, P = 0.002 |
| ALKBH5 | 8.1490 ± 0.1403 | 9.1393 ± 0.1062 | 8.7767 ± 0.1682 | F(2,15) = 76.26, P < 0.00 001 | Q = 17.26, P < 0.00 001 | Q = 10.94, P < 0.00 001 | Q = 6.32, P = 0.001 |
| Readers | |||||||
| YTHDF1 | 6.0045 ± 0.1487 | 6.6542 ± 0.2133 | 6.4797 ± 0.0952 | F(2,15) = 26.54,P =0.00 001 | Q = 9.95, P = 0.00 001 | Q = 7.28, P = 0.0003 | Q = 2.67, P = 0.176 |
| YTHDF2 | 4.5807 ± 0.0959 | 5.2345 ± 0.1754 | 5.0488 ± 0.2700 | F(2,15) = 18.10,P =0.0001 | Q = 8.26, P = 0.00 009 | Q = 5.91, P = 0.002 | Q = 2.35, P = 0.253 |
| YTHDF3 | 6.5292 ± 0.0864 | 7.2565 ± 0.1926 | 6.6122 ± 0.1004 | F(2,15) = 52.21, P < 0.0001 | Q = 13.20, P < 0.00 001 | Q = 1.51, P = 0.549 | Q = 11.69, P < 0.00 001 |
| YTHDC1 | 6.1602 ± 0.2309 | 6.8390 ± 0.1155 | 6.7240 ± 0.3178 | F(2,15) = 14.17,P =0.0004 | Q = 7.03, P = 0.0005 | Q = 5.84, P = 0.002 | Q = 1.19, P = 0.683 |
| YTHDC2 | 8.1320 ± 0.2434 | 8.8312 ± 0.1288 | 8.8665 ± 0.1709 | F(2,15) = 29.41, P < 0.00 001 | Q = 9.15, P = 0.00 003 | Q = 9.62, P = 0.00 002 | Q = 0.46, P = 0.943 |
| IGF2BP1 | 13.4212 ± 0.6871 | 12.9423 ± 0.6602 | 14.1708 ± 0.9798 | F(2,15) = 3.69, P = 0.050 | Q = 1.49, P = 0.557 | Q = 2.33, P = 0.258 | Q = 3.81, P = 0.041 |
| IGF2BP2 | 14.2035 ± 1.5433 | 14.9008 ± 2.4208 | 13.0512 ± 2.6271 | F(2,9) = 0.691, P = 0.526 | Q = 0.62, P = 0.900 | Q = 1.02, P = 0.755 | Q = 1.65, P = 0.502 |
| IGF2BP3 | 8.4410 ± 0.3830 | 9.1197 ± 0.1968 | 8.9370 ± 0.3355 | F(2,15) = 7.45, P = 0.006 | Q = 5.27, P = 0.005 | Q = 3.85, P = 0.039 | Q = 1.42, P = 0.586 |
| HNRNPA2B1 | 1.0763 ± 0.2237 | 1.4167 ± 0.1838 | 1.3780 ± 0.0867 | F(2,15) = 6.84, P = 0.008 | Q = 4.78, P = 0.011 | Q = 4.24, P = 0.023 | Q = 0.54, P = 0.922 |
CTL, control (no ethanol treatment); CIE, chronic intermittent exposure to ethanol; and CIE + WD, CIE followed by 24-h withdrawal.
∆Ct (delta Ct) represents the difference in cycle threshold (Ct) values between an m6A regulatory gene and a reference (housekeeping) gene [i.e. the beta-actin gene (ACTB)] within the same sample.
Table 2.
Statistical analysis (one-way ANOVA and Tukey’s test) of ethanol exposure- and withdrawal-induced alterations in m6A regulatory gene mRNA expression in non-neuronal SW620 cells.
| m6A Regulatory | ∆Ct CTL | ∆Ct CIE | ∆Ct CIE + WD | CTL vs. CIE | CTL vs. CIE + WD | CIE vs. CIE + WD | |
|---|---|---|---|---|---|---|---|
| Genes | (n = 6) | (n = 6) | (n = 6) | One-way ANOVA | Tukey’s test | Tukey’s test | Tukey’s test |
| Writers | |||||||
| KIAA1429 | 7.0162 ± 0.1830 | 7.7583 ± 0.2030 | 7.3645 ± 0.1452 | F(2,15) = 25.90, P = 0.00 001 | Q = 10.17, P = 0.00 001 | Q = 4.77, P = 0.011 | Q = 5.40, P = 0.004 |
| METTL3 | 7.0162 ± 0.1830 | 7.7583 ± 0.2030 | 7.3645 ± 0.1452 | F(2,14) = 20.75, P = 0.00 006 | Q = 8.76, P = 0.00 007 | Q = 1.48, P = 0.561 | Q = 7.28, P = 0.0004 |
| METTL4 | 9.5317 ± 0.7082 | 10.7256 ± 0.3734 | 9.6717 ± 0.0431 | F(2,14) = 10.28, P = 0.002 | Q = 6.04, P = 0.002 | Q = 0.71, P = 0.872 | Q = 5.33, P = 0.005 |
| METTL14 | 9.4182 ± 0.1828 | 9.5382 ± 0.1728 | 9.3645 ± 0.1149 | F(2,14) = 20.75, P = 0.00 006 | Q = 1.84, P = 0.415 | Q = 0.82, P = 0.832 | Q = 2.67, P = 0.177 |
| RBM15 | 4.8718 ± 0.0579 | 5.2652 ± 0.0781 | 4.7852 ± 0.2019 | F(2,14) = 21.88, P = 0.00 005 | Q = 7.02, P = 0.0006 | Q = 1.54, P = 0.534 | Q = 8.56, P = 0.00 008 |
| RBM15B | 6.2866 ± 0.1946 | 6.3152 ± 0.3433 | 6.1845 ± 0.0952 | F(2,14) = 0.500, P = 0.618 | Q = 0.29, P = 0.978 | Q = 1.02, P = 0.784 | Q = 1.31, P = 0.634 |
| WTAP | 5.0722 ± 0.1133 | 5.2187 ± 0.1092 | 5.3907 ± 0.1050 | F(2,15) = 12.78, P = 0.0006 | Q = 3.29, P = 0.083 | Q = 7.14, P = 0.0004 | Q = 3.86, P = 0.039 |
| Erasers | |||||||
| FTO | 7.6920 ± 0.2275 | 8.3075 ± 0.3841 | 7.7128 ± 0.1812 | F(2,15) = 9.47,P =0.002 | Q = 5.42, P = 0.004 | Q = 0.18, P = 0.991 | Q = 5.24, P = 0.006 |
| ALKBH5 | 8.1293 ± 0.1464 | 8.8203 ± 0.2503 | 8.5042 ± 0.1141 | F(2,15) = 22.18,P =0.00 003 | Q = 9.41, P = 0.00 002 | Q = 5.10, P = 0.007 | Q = 4.30, P = 0.021 |
| Readers | |||||||
| YTHDF1 | 6.0595 ± 0.0760 | 6.2835 ± 0.2742 | 6.3335 ± 0.0906 | F(2,15) = 4.30,P =0.033 | Q = 3.18, P = 0.095 | Q = 3.89, P = 0.037 | Q = 0.71, P = 0.871 |
| YTHDF2 | 5.2960 ± 0.4331 | 5.5818 ± 0.1150 | 5.4647 ± 0.0986 | F(2,15) = 1.76,P =0.205 | Q = 2.64, P = 0.182 | Q = 1.56, P = 0.527 | Q = 1.08, P = 0.729 |
| YTHDF3 | 7.1375 ± 0.8724 | 7.3835 ± 0.1057 | 7.2478 ± 0.1055 | F(2,15) = 0.35,P =0.711 | Q = 1.18, P = 0.688 | Q = 0.53, P = 0.926 | Q = 0.65, P = 0.891 |
| YTHDC1 | 7.0828 ± 0.1777 | 6.7902 ± 0.2961 | 5.9977 ± 0.1645 | F(2,15) = 38.78, P < 0.00 001 | Q = 3.25, P = 0.087 | Q = 12.04, P < 0.00 001 | Q = 8.79, P = 0.00 005 |
| YTHDC2 | 8.8367 ± 0.1913 | 8.8147 ± 0.2620 | 8.1938 ± 0.2603 | F(2,15) = 13.86,P =0.0004 | Q = 0.22, P = 0.986 | Q = 6.56, P = 0.0009 | Q = 6.33, P = 0.001 |
| IGF2BP1 | 6.4103 ± 0.3757 | 6.2887 ± 0.1534 | 5.992 ± 0.2288 | F(2,15) = 3.84, P =0.045 | Q = 1.11, P = 0.719 | Q = 3.81, P = 0.042 | Q = 2.70, P = 0.170 |
| IGF2BP2 | 7.5858 ± 0.3373 | 8.3943 ± 0.3430 | 7.8927 ± 0.3747 | F(2,14) = 7.43,P =0.006 | Q = 5.43, P = 0.005 | Q = 2.06, P = 0.340 | Q = 3.37, P = 0.077 |
| IGF2BP3 | 8.1084 ± 0.9564 | 9.3192 ± 0.8556 | 8.9885 ± 0.3658 | F(2,14) = 3.67,P =0.052 | Q = 3.80, P = 0.044 | Q = 2.76, P = 0.160 | Q = 1.04, P = 0.748 |
| HNRNPA2B1 | 2.4145 ± 0.4103 | 1.4468 ± 0.2989 | 1.5625 ± 0.1538 | F(2,14) = 17.08,P =0.0002 | Q = 7.48, P = 0.0003 | Q = 6.59, P = 0.001 | Q = 0.89, P = 0.805 |
CTL, control (no ethanol treatment); CIE, chronic intermittent exposure to ethanol; and CIE + WD, CIE followed by 24-h withdrawal.
∆Ct (delta Ct) represents the difference in Ct values between an m6A regulatory gene and a reference (housekeeping) gene [i.e. the beta-actin gene (ACTB)] within the same sample.
Figure 1.
One-way ANOVA and Tukey’s test results showing m6A writer gene mRNA expression differences among three ethanol treatment groups in SH-SY5Y (a) and SW620 (b) cells. CTL, control (no ethanol treatment); CIE, chronic intermittent ethanol exposure; and CIE + WD, CIE exposure followed by a 24-h withdrawal.
m6A writer gene expression changes induced by acetaldehyde exposure and withdrawal
One-way ANOVA and Tukey’s HSD post hoc test results for m6A writer genes following acetaldehyde exposure and withdrawal are summarized in Table 3 (SH-SY5Y) and Table 4 (SW620). Acetaldehyde exposure (30 μM) had a more modest effect than ethanol (40 mM). In SH-SY5Y cells, acetaldehyde significantly affected three genes (METTL3, RBM15B, and WTAP; P = 0.016–0.040). In SW620 cells, significant effects were observed for four genes (KIAA1429, METTL4, METTL14, and WTAP; P values ranging from <0.00001 to 0.0009). Pairwise comparisons are shown in Fig. 2. In SH-SY5Y cells, acetaldehyde exposure or withdrawal significantly reduced the expression of METTL3, RBM15B, and WTAP (Fig. 2a and Table 3). Similarly, in SW620 cells, acetaldehyde exposure or withdrawal significantly downregulated KIAA1429, METTL4, and WTAP (Fig. 2b and Table 4).
Table 3.
Statistical analysis (one-way ANOVA and Tukey’s test) of acetaldehyde exposure- and withdrawal-induced alterations in m6A regulatory gene mRNA expression in neuron-like SH-SY5Y cells.
| m6A Regulatory | ∆Ct CTL | ∆Ct CIA | ∆Ct CIA + WD | CTL vs. CIA | CTL vs. CIA + WD | CIA vs. CIA + WD | |
|---|---|---|---|---|---|---|---|
| Genes | (n = 6) | (n = 6) | (n = 6) | One-way ANOVA | Tukey’s test | Tukey’s test | Tukey’s test |
| Writers | |||||||
| KIAA1429 | 7.2730 ± 0.3094 | 6.9580 ± 0.2396 | 6.8793 ± 0.2632 | F(2,15) = 3.51, P = 0.056 | Q = 2.83, P = 0.146 | Q = 3.54, P = 0.060 | Q = 0.71, P = 0.872 |
| METTL3 | 6.6795 ± 0.2450 | 6.7242 ± 0.3233 | 7.0535 ± 0.1257 | F(2,15) = 4.16, P = 0.036 | Q = 0.45, P = 0.947 | Q = 3.74, P = 0.046 | Q = 3.29, P = 0.083 |
| METTL4 | 9.2905 ± 0.3530 | 9.4270 ± 0.2046 | 9.4145 ± 0.0944 | F(2,15) = 0.58, P = 0.570 | Q = 1.38, P = 0.601 | Q = 1.26, P = 0.656 | Q = 0.13, P = 0.996 |
| METTL14 | 8.4147 ± 0.2195 | 8.1952 ± 0.2384 | 8.2633 ± 0.1337 | F(2,15) = 1.85, P = 0.192 | Q = 2.66, P = 0.179 | Q = 1.86, P = 0.419 | Q = 0.82, P = 0.831 |
| RBM15 | 7.1202 ± 0.1413 | 7.0817 ± 0.4045 | 7.3788 ± 0.2644 | F(2,15) = 1.85, P = 0.191 | Q = 0.32, P = 0.971 | Q = 2.18, P = 0.301 | Q = 2.52, P = 0.213 |
| RBM15B | 5.6475 ± 0.0514 | 5.8515 ± 0.1551 | 5.7988 ± 0.1008 | F(2,15) = 5.48, P = 0.016 | Q = 4.51, P = 0.016 | Q = 3.34, P = 0.077 | Q = 1.16, P = 0.694 |
| WTAP | 5.5830 ± 0.3803 | 5.6922 ± 0.3014 | 6.0332 ± 0.1096 | F(2,15) = 4.01, P = 0.040 | Q = 0.93, P = 0.790 | Q = 3.84, P = 0.040 | Q = 2.91, P = 0.133 |
| Erasers | |||||||
| FTO | 7.3890 ± 0.0671 | 7.3803 ± 0.1227 | 7.5382 ± 0.2351 | F(2,15) = 1.89,P =0.185 | Q = 0.13, P = 0.995 | Q = 2.31, P = 0.262 | Q = 2.45, P = 0.226 |
| ALKBH5 | 8.6330 ± 0.1305 | 8.5508 ± 0.1451 | 8.2802 ± 0.1128 | F(2,15) = 12.08,P =0.0008 | Q = 1.55, P = 0.532 | Q = 6.64, P = 0.0008 | Q = 5.09, P = 0.007 |
| Readers | |||||||
| YTHDF1 | 6.8037 ± 0.0892 | 6.7132 ± 0.1948 | 6.7035 ± 0.0987 | F(2,15) = 0.99,P =0.395 | Q = 1.63, P = 0.499 | Q = 1.80, P = 0.430 | Q = 0.17, P = 0.992 |
| YTHDF2 | 5.6087 ± 0.2304 | 5.5733 ± 0.1230 | 5.4723 ± 0.1269 | F(2,15) = 1.07, P =0.368 | Q = 0.52, P = 0.929 | Q = 1.99, P = 0.361 | Q = 1.48, P = 0.562 |
| YTHDF3 | 6.8230 ± 0.0841 | 6.9203 ± 0.0903 | 6.8237 ± 0.1004 | F(2,15) = 1.07,P =0.368 | Q = 2.59, P = 0.192 | Q = 0.02, P = 1.000 | Q = 2.58, P = 0.196 |
| YTHDC1 | 6.8152 ± 0.2986 | 6.9752 ± 0.1480 | 6.9270 ± 0.1718 | F(2,14) = 0.82,P =0.459 | Q = 1.81, P = 0.428 | Q = 1.27, P = 0.652 | Q = 0.55, P = 0.922 |
| YTHDC2 | 9.3177 ± 0.1841 | 9.1853 ± 0.2685 | 9.2775 ± 0.1052 | F(2,15) = 0.71,P =0.508 | Q = 1.64, P = 0.493 | Q = 0.50, P = 0.934 | Q = 1.14, P = 0.704 |
| IGF2BP1 | 14.866 ± 0.6844 | 15.018 ± 0.6414 | 14.9138 ± 0.8251 | F(2,14) = 0.07,P =0.932 | Q = 0.51, P = 0.932 | Q = 0.16, P = 0.993 | Q = 0.35, P = 0.968 |
| IGF2BP2 | 14.8247 ± 0.7311 | 14.6120 ± 1.1847 | 14.4485 ± 0.5746 | F(2,15) = 0.28,P =0.758 | Q = 0.60, P = 0.906 | Q = 1.06, P = 0.739 | Q = 0.46, P = 0.943 |
| IGF2BP3 | 6.9710 ± 0.1500 | 7.0018 ± 0.3347 | 7.0345 ± 0.0518 | F(2,15) = 0.13,P =0.877 | Q = 0.35, P = 0.966 | Q = 0.73, P = 0.866 | Q = 0.37, P = 0.962 |
| HNRNPA2B1 | 2.2104 ± 0.1505 | 2.5163 ± 0.1997 | 2.3328 ± 0.0838 | F(2,13) = 5.35,P =0.020 | Q = 4.50, P = 0.018 | Q = 1.80, P = 0.434 | Q = 2.70, P = 0.176 |
CTL, control (no acetaldehyde treatment); CIA, chronic intermittent exposure to acetaldehyde; CIA + WD, CIA followed by 24-h withdrawal.
∆Ct (delta Ct) represents the difference in cycle threshold (Ct) values between an m6A regulatory gene and a reference (housekeeping) gene [i.e. the beta-actin gene (ACTB)] within the same sample.
Table 4.
Statistical analysis (one-way ANOVA and Tukey’s test) of acetaldehyde exposure- and withdrawal-induced alterations in m6A regulatory gene mRNA expression in non-neuronal SW620 cells.
| m6A Regulatory | ∆Ct CTL | ∆Ct CIA | ∆Ct CIA + WD | CTL vs. CIA | CTL vs. CIA + WD | CIA vs. CIA + WD | |
|---|---|---|---|---|---|---|---|
| Genes | (n = 6) | (n = 6) | (n = 6) | One-way ANOVA | Tukey’s test | Tukey’s test | Tukey’s test |
| Writers | |||||||
| KIAA1429 | 7.4467 ± 0.0829 | 7.5977 ± 0.1294 | 7.7567 ± 0.1167 | F(2,15) = 11.61, P = 0.0009 | Q = 3.32, P = 0.080 | Q = 6.81, P = 0.0006 | Q = 3.49, P = 0.063 |
| METTL3 | 7.3000 ± 0.1840 | 7.1445 ± 0.3034 | 7.4030 ± 0.2562 | F(2,15) = 1.59, P = 0.236 | Q = 1.51, P = 0.549 | Q = 1.00, P = 0.764 | Q = 2.51, P = 0.212 |
| METTL4 | 9.6820 ± 0.0846 | 10.4662 ± 0.1946 | 10.5567 ± 0.2562 | F(2,15) = 37.62, P < 0.00 001 | Q = 10.00, P = 0.00 001 | Q = 11.15, P < 0.00 001 | Q = 1.15, P = 0.699 |
| METTL14 | 9.7127 ± 0.0871 | 9.8120 ± 0.1778 | 9.8060 ± 0.1670 | F(2,15) = 37.62, P < 0.00 001 | Q = 1.63, P = 0.500 | Q = 1.53, P = 0.540 | Q = 0.10, P = 0.997 |
| RBM15 | 6.9462 ± 0.2777 | 7.1043 ± 0.2024 | 7.0437 ± 0.2047 | F(2,15) = 0.72, P = 0.504 | Q = 1.68, P = 0.479 | Q = 1.03, P = 0.749 | Q = 0.64, P = 0.893 |
| RBM15B | 6.1988 ± 0.1602 | 6.3222 ± 0.0865 | 6.3417 ± 0.1179 | F(2,15) = 2.29, P = 0.135 | Q = 2.41, P = 0.235 | Q = 2.80 P = 0.152 | Q = 0.38, P = 0.961 |
| WTAP | 5.2470 ± 0.1858 | 5.8562 ± 0.3285 | 5.8570 ± 0.1700 | F(2,15) = 13.01, P = 0.0005 | Q = 6.24, P = 0.001 | Q = 6.25, P = 0.001 | Q = 0.01, P = 1.000 |
| Erasers | |||||||
| FTO | 8.0262 ± 0.1398 | 8.1002 ± 0.4308 | 7.9725 ± 0.1718 | F(2,15) = 0.32,P =0.734 | Q = 0.65, P = 0.891 | Q = 0.47, P = 0.941 | Q = 1.12, P = 0.714 |
| ALKBH5 | 8.5565 ± 0.2333 | 8.6577 ± 0.1616 | 8.9223 ± 0.1900 | F(2,15) = 5.51,P =0.016 | Q = 1.26, P = 0.655 | Q = 4.54, P = 0.015 | Q = 3.29, P = 0.083 |
| Readers | |||||||
| YTHDF1 | 6.5383 ± 0.0964 | 6.7523 ± 0.2597 | 6.5697 ± 0.2962 | F(2,15) = 1.46,P =0.263 | Q = 2.24, P = 0.283 | Q = 0.33, P = 0.971 | Q = 1.91, P = 0.390 |
| YTHDF2 | 6.0047 ± 0.1550 | 5.9268 ± 0.2811 | 6.0034 ± 0.1991 | F(2,14) = 0.24,P =0.789 | Q = 0.84, P = 0.825 | Q = 0.01, P = 1.000 | Q = 0.83, P = 0.830 |
| YTHDF3 | 7.5147 ± 0.1768 | 7.3410 ± 0.1092 | 7.4022 ± 0.3182 | F(2,15) = 0.97,P =0.403 | Q = 1.94, P = 0.380 | Q = 1.26, P = 0.656 | Q = 0.68, P = 0.880 |
| YTHDC1 | 7.5487 ± 0.2701 | 7.3647 ± 0.1160 | 7.4967 ± 0.2963 | F(2,15) = 0.93,P =0.416 | Q = 1.87, P = 0.405 | Q = 0.53, P = 0.926 | Q = 1.34, P = 0.619 |
| YTHDC2 | 8.7823 ± 0.0899 | 8.7528 ± 0.1142 | 8.7338 ± 0.2160 | F(2,15) = 0.16,P =0.855 | Q = 0.48, P = 0.938 | Q = 0.79, P = 0.843 | Q = 0.31, P = 0.974 |
| IGF2BP1 | 6.0382 ± 0.1659 | 6.1898 ± 0.2362 | 6.0405 ± 0.2756 | F(2,15) = 0.85,P =0.446 | Q = 1.61, P = 0.506 | Q = 0.02, P = 1.000 | Q = 1.59, P = 0.516 |
| IGF2BP2 | 7.4968 ± 0.0749 | 7.8263 ± 0.1131 | 7.7828 ± 0.1749 | F(2,15) = 11.77,P =0.0008 | Q = 6.32, P = 0.001 | Q = 5.48, P = 0.004 | Q = 0.83, P = 0.828 |
| IGF2BP3 | 7.5403 ± 0.0615 | 7.5680 ± 0.0677 | 7.4410 ± 0.0791 | F(2,15) = 5.49,P =0.016 | Q = 0.97, P = 0.774 | Q = 3.48, P = 0.064 | Q = 4.46, P = 0.017 |
| HNRNPA2B1 | 1.8342 ± 0.0940 | 1.6993 ± 0.0823 | 1.8332 ± 0.1992 | F(2,15) = 1.96,P =0.176 | Q = 2.43, P = 0.230 | Q = 0.02, P = 1.000 | Q = 2.42, P = 0.234 |
CTL, control (no acetaldehyde treatment); CIA, chronic intermittent exposure to acetaldehyde; CIA + WD, CIA followed by 24-h withdrawal.
∆Ct (delta Ct) represents the difference in Ct values between an m6A regulatory gene and a reference (housekeeping) gene [i.e. the beta-actin gene (ACTB)] within the same sample.
Figure 2.
One-way ANOVA and Tukey’s test results showing m6A writer gene mRNA expression differences among three acetaldehyde treatment groups in SH-SY5Y (a) and SW620 (b) cells. CTL, control (no acetaldehyde treatment); CIA, chronic intermittent acetaldehyde exposure; and CIA + WD, CIA exposure followed by a 24-h withdrawal.
m6A eraser gene expression changes induced by ethanol exposure and withdrawal
One-way ANOVA and Tukey’s HSD post hoc test results for m6A eraser genes following ethanol exposure and withdrawal are summarized in Table 1 (SH-SY5Y) and Table 2 (SW620). Significant effects of ethanol exposure and withdrawal were observed for both eraser genes in both cell lines [FTO: P = 0.00002 (SH-SY5Y); ALKBH5: P < 0.00001 (SH-SY5Y); FTO: P = 0.002 (SW620); ALKBH5: P = 0.00003 (SW620)]. Pairwise comparisons are presented in Fig. 3. In SH-SY5Y cells, CIE exposure significantly downregulated the expression of both eraser genes (FTO: Q = 9.76, P = 0.00002; ALKBH5: Q = 17.26, P < 0.00001), However, ethanol withdrawal partially restored their expression levels (Fig. 3a; Table 1). Similarly, in SW620 cells, CIE exposure suppressed both FTO (Q = 5.42, P = 0.004) and ALKBH5 (Q = 9.41, P = 0.00002) expression, while ethanol withdrawal increased their expression compared to CIE-exposed cells (Fig. 3b; Table 2).
Figure 3.
One-way ANOVA and Tukey’s test results showing m6A eraser gene mRNA expression differences among three ethanol treatment groups in SH-SY5Y (a) and SW620 (b) cells. CTL, control (no ethanol treatment); CIE, chronic intermittent ethanol exposure; and CIE + WD, CIE exposure followed by a 24-h withdrawal.
m6A eraser gene expression changes induced by acetaldehyde exposure and withdrawal
One-way ANOVA and Tukey’s HSD post hoc test results for m6A eraser genes following acetaldehyde exposure and withdrawal are summarized in Table 3 (SH-SY5Y) and Table 4 (SW620). Significant acetaldehyde effects were observed only for ALKBH5 in both cell lines [ALKBH5: P = 0.0008 (SH-SY5Y); ALKBH5: P = 0.016 (SW620)], whereas FTO expression was not significantly affected. Pairwise comparisons are presented in Fig. 4. Neither chronic intermittent acetaldehyde (CIA) exposure nor CIA followed by withdrawal (WD) significantly altered FTO expression in either cell line. CIA alone also did not significantly affect ALKBH5 expression. However, acetaldehyde withdrawal (WD) differentially regulated ALKBH5 expression: it was upregulated in SH-SY5Y cells (Q = 6.64, P = 0.0008; Fig. 4a and Table 3) but downregulated in SW620 cells (Q = 4.54, P = 0.015; Fig. 4b and Table 4).
Figure 4.
One-way ANOVA and Tukey’s test results showing m6A eraser gene mRNA expression differences among three acetaldehyde treatment groups in SH-SY5Y (a) and SW620 (b) cells. CTL, control (no acetaldehyde treatment); CIA, chronic intermittent acetaldehyde exposure; and CIA + WD: CIA exposure followed by a 24-h withdrawal.
m6A reader gene expression changes induced by ethanol exposure and withdrawal
One-way ANOVA and Tukey’s HSD post hoc test results for m6A reader genes following ethanol exposure and withdrawal are summarized in Table 1 (SH-SY5Y) and Table 2 (SW620). Significant ethanol effects were observed for seven reader genes in SH-SY5Y cells (YTHDF1: P = 0.00001; YTHDF2: P = 0.0001; YTHDF3: P < 0.0001; YTHDC1: P = 0.0004; YTHDC2: P < 0.00001; IGF2BP3: P = 0.006; HNRNPA2B1: P = 0.008) and six reader genes in SW620 cells (YTHDF1: P = 0.033; YTHDC1: P < 0.00001; YTHDC2: P = 0.0004; IGF2BP1: P = 0.045; IGF2BP2: P = 0.006; HNRNPA2B1: P = 0.0002). Pairwise comparisons are shown in Fig. 5. In SH-SY5Y cells, all seven affected genes (YTHDF1, YTHDF2, YTHDF3, YTHDC1, YTHDC2, IGF2BP3, and HNRNPA2B1) exhibited significantly reduced expression levels after chronic intermittent ethanol (CIE) exposure (0.00001 < P ≤ 0.011). Their expression partially recovered after ethanol withdrawal, except for YTHDC2, which remained downregulated (Fig. 5a and Table 1). In SW620 cells, CIE exposure significantly downregulated IGF2BP2 (Q = 5.43, P = 0.005) and IGF2BP3 (Q = 3.80, P = 0.044), while significantly upregulating HNRNPA2B1 (Q = 7.48, P = 0.0003). Following ethanol withdrawal, YTHDF1 expression was significantly reduced (Q = 3.89, P = 0.037), whereas YTHDC1 (Q = 12.04, P < 0.00001), YTHDC2 (Q = 6.56, P = 0.0009), IGF2BP1 (Q = 3.81, P = 0.042), and HNRNPA2B1 (Q = 6.59, P = 0.001) were significantly upregulated (Fig. 5b and Table 2).
Figure 5.
One-way ANOVA and Tukey’s test results showing m6A reader gene mRNA expression differences among three ethanol treatment groups in SH-SY5Y (a) and SW620 (b) cells. CTL, Control (no ethanol treatment); CIE, chronic intermittent ethanol exposure; and CIE + WD, CIE exposure followed by a 24-h withdrawal.
m6A reader gene expression changes induced by acetaldehyde exposure and withdrawal
One-way ANOVA and Tukey’s HSD post hoc test results for m6A reader genes following acetaldehyde exposure and withdrawal are summarized in Table 3 (SH-SY5Y) and Table 4 (SW620). Compared with ethanol at 40 mM, acetaldehyde at 30 μM had a much smaller impact on m6A reader gene expression. Significant acetaldehyde effects were observed for only one reader gene in SH-SY5Y cells [HNRNPA2B1: F(2,13) = 5.35, P = 0.020] and for two reader genes in SW620 cells [IGF2BP2: F(2,15) = 11.77, P = 0.0008; IGF2BP3: F(2,15) = 5.49, P = 0.016]. Pairwise comparisons are shown in Fig. 6. In SH-SY5Y cells, HNRNPA2B1 expression was significantly downregulated by CIA exposure (Q = 4.50, P = 0.018) (Fig. 6a and Table 3). In SW620 cells, both CIA exposure and withdrawal significantly downregulated IGF2BP2 expression (CIA: Q = 6.32, P = 0.001; CIA + WD: Q = 5.48, P = 0.004) (Fig. 6b and Table 4).
Figure 6.
One-way ANOVA and Tukey’s test results showing m6A reader gene mRNA expression differences among three acetaldehyde treatment groups in SH-SY5Y (a) and SW620 (b) cells. CTL, control (no acetaldehyde treatment); CIA, chronic intermittent acetaldehyde exposure; and CIA + WD, CIA exposure followed by a 24-h withdrawal.
m6A regulatory gene expression changes in eight brain regions of individuals with AUD
Differential expression analyses of seven m6A writer genes (KIAA1429, METTL3, METTL4, METTL14, METTL16, RBM15, and WTAP), two eraser genes (FTO, ALKBH5), and seven reader genes (YTHDF1, YTHDF2, YTHDF3, YTHDC1, YTHDC2, IGF2BP2, and HNRNPA2B1) across eight human brain regions (amygdala, caudate nucleus, cerebellum, hippocampus, nucleus accumbens, prefrontal cortex, putamen, and ventral tegmental area) are summarized in Supplementary Table 3a–h. Genes with unadjusted P-values <0.100 in any brain region are listed in Table 5. In the amygdala, AUD subjects exhibited upregulation of the reader gene YTHDF1 (t = 2.134, Punadjusted= 0.034) and downregulation of the reader gene HNRNPA2B1 (t = –1.669, Punadjusted = 0.097). In the hippocampus, HNRNPA2B1 expression was also reduced (t = –2.533, Punadjusted = 0.012). In the nucleus accumbens, the eraser gene FTO and the reader gene IGF2BP2 showed increased expression (FTO: t = 1.705, Punadjusted = 0.090; IGF2BP2: t = 1.697, Punadjusted = 0.091). In the prefrontal cortex, the writer gene METTL16 was downregulated (t = –1.940, Punadjusted = 0.054). In the putamen, the writer gene KIAA1429 was downregulated (t = –2.076, Punadjusted = 0.039), while the reader gene IGF2BP2 was upregulated (t = 1.835, Punadjusted = 0.068). Finally, in the ventral tegmental area, both the writer gene WTAP (t = 2.863, Punadjusted = 0.005) and the reader gene YTHDF1 (t = 2.700, Punadjusted = 0.008) were upregulated. While several m6A regulatory genes showed nominal differences between cases and controls, these did not survive false discovery rate (FDR) correction and should therefore be interpreted with caution.
Table 5.
Differential mRNA expression of m6A regulatory genes across eight brain regions in individuals with AUD.
| Brain Regions | m6A Regulatory Genes | Log2FC | AveExpr | t | P unadjusted | P adjusted | B |
|---|---|---|---|---|---|---|---|
| Amygdala | YTHDF1 | 0.217 | 5.458 | 2.134 | 0.034 | NS | −3.816 |
| HNRNPA2B1 | −0.140 | 8.313 | −1.669 | 0.097 | NS | −4.262 | |
| Caudate Nucleus | |||||||
| Cerebellum | |||||||
| Hippocampus | HNRNPA2B1 | −0.213 | 8.313 | −2.533 | 0.012 | NS | −2.897 |
| Nucleus Accumbens | FTO | 0.156 | 7.119 | 1.705 | 0.090 | NS | −4.282 |
| IGF2BP2 | 0.531 | 1.496 | 1.697 | 0.091 | NS | −4.437 | |
| Prefrontal Cortex | METTL16 | −0.179 | 5.270 | −1.940 | 0.054 | NS | −4.122 |
| Putamen | KIAA1429 | −0.200 | 5.926 | −2.076 | 0.039 | NS | −3.689 |
| IGF2BP2 | 0.569 | 1.496 | 1.835 | 0.068 | NS | −4.069 | |
| Ventral Tegmental Area | WTAP | 0.247 | 5.319 | 2.863 | 0.005 | NS | −2.675 |
| YTHDF1 | 0.275 | 5.458 | 2.700 | 0.008 | NS | −2.924 |
Log2FC: Log₂ fold-change (FC) between cases and controls.
AveExpr: Average gene expression levels.
t: t-statistics used to assess differential expression.
B: Empirical Bayes log-odds of differential expression.
P unadjusted: Differential expression P values unadjusted for multiple testing.
P Adjusted: Adjusted P values for differential expression (Benjamini–Hochberg method).
N.S.: Not significant.
Discussion
Environmental exposures, such as chronic alcohol consumption, can profoundly reshape brain epigenetic landscapes, including the RNA methylome (or epitranscriptome). In neuronal cells, m6A RNA methylation regulates the expression and function of transcripts critical for synaptic plasticity, neuronal development, and stress responses. Disruptions in RNA methylation can impair these processes, increasing vulnerability to AUD by altering reward, learning, and stress circuits [14]. Because m6A RNA methylation patterns are governed by coordinated activity of writers, erasers, and readers, it is essential to determine whether alcohol exposure and withdrawal alter the expression of these m6A regulatory genes, thereby driving aberrant RNA methylation. Our study addressed this question using complementary cellular models and postmortem human brain tissue.
We found that ethanol broadly suppressed the expression of most m6A regulatory genes, with neuron-like SH-SY5Y cells exhibiting greater sensitivity than non-neuronal SW620 cells. In SH-SY5Y cells, 16 of 18 m6A regulatory genes were significantly altered, whereas only 9 of 18 were affected in SW620 cells (CTL vs. CIE, Tukey’s test, P < 0.050; Tables 1 and 2). This heightened sensitivity of neuronal cells suggests that chronic alcohol exposure may induce more extensive RNA methylation alterations in the brain, potentially leading to long-lasting molecular and functional changes that contribute to AUD susceptibility. The limited sensitivity of SW620 cells may reflect their non-neuronal, colorectal cancer-derived origin rather than brain-resident glial or endothelial cells, which could respond differently under ethanol exposure.
Ethanol withdrawal partially restored the expression of m6A regulatory genes in both cell types. We observed that a 24-h withdrawal after a 3-week chronic intermittent ethanol (CIE) exposure led to increased expression of m6A regulatory genes in both neuron-like SH-SY5Y and non-neuronal SW620 cells. The downregulation of m6A regulatory genes during ethanol exposure suggests that ethanol directly suppresses the machinery responsible for installing, removing, or reading m6A modifications. This could lead to widespread changes in RNA methylation and downstream effects on neuronal gene expression and function. The partial restoration of gene expression after 24 h of ethanol withdrawal indicates that the suppression is at least partially reversible. This suggests that the epitranscriptomic machinery can recover when ethanol is removed, reflecting cellular resilience. Moreover, the partial recovery could involve transcriptional reactivation, relief of ethanol-induced stress pathways, or compensatory feedback mechanisms in the cells that restore m6A regulatory gene expression levels.
Compared with ethanol, acetaldehyde, the toxic metabolite of ethanol, induced subtler changes in the expression of m6A regulators genes. This likely reflects its lower physiological concentration (∼30 μM vs. 40 mM ethanol), rapid metabolism, and high volatility. Alcohol metabolism proceeds through a two-step process: ethanol is first converted to acetaldehyde by alcohol dehydrogenases (ADHs), and acetaldehyde is then metabolized to non-toxic acetate by aldehyde dehydrogenases (ALDHs). Genetic variation in ALDH genes, such as ALDH2, can impair this detoxification step, resulting in the accumulation of acetaldehyde and the alcohol flushing response [18]. Notably, this flushing response is protective, as it discourages excessive alcohol consumption. Acetaldehyde can also upregulate ALDH2 expression to enhance its own clearance, as observed in our cellular model [19]. Furthermore, acetaldehyde is highly volatile, particularly at body temperature, and is therefore rapidly eliminated. Following heavy alcohol consumption, blood acetaldehyde levels reach only ∼30 μM, compared with ∼40 mM ethanol [20]. At such low concentrations, acetaldehyde exposure is unlikely to cause major changes in cellular epigenetic states or functional processes. This likely explains why we observed only modest effects of acetaldehyde (30 μM) on the expression of m6A regulatory genes.
Postmortem brain analysis revealed a trend toward differential expression of m6A regulatory genes across multiple brain regions, although the direction of change did not always align with that observed in ethanol-exposed cells. In both neuron-like SH-SY5Y and non-neuronal SW620 cells, ethanol exposure consistently decreased the expression of all m6A regulatory genes. In contrast, postmortem brains from individuals with AUD showed both up- and downregulated m6A regulatory genes, with some displaying consistent changes across multiple brain regions (Table 2). For example, the reader gene HNRNPA2B1 was downregulated in both the amygdala and hippocampus, the writer gene METTL16 was downregulated in the prefrontal cortex, and the writer gene KIAA1429 was downregulated in the putamen. Conversely, several genes were upregulated, including the reader gene YTHDF1 (amygdala and ventral tegmental area), IGF2BP2 (nucleus accumbens and putamen), the eraser gene FTO (nucleus accumbens), and the writer gene WTAP (ventral tegmental area). These discrepancies between cellular models and postmortem brain findings are likely due to several factors. First, human brain tissue is highly heterogeneous, containing diverse neuronal and glial subtypes with distinct responses to ethanol, whereas cell models are homogeneous. Second, postmortem brains reflect long-term, chronic alcohol exposure and the cumulative effects of withdrawal, relapse, and comorbidities, while in vitro models use defined ethanol doses and timeframes that do not fully capture years of intermittent alcohol intake. Third, cell models lack the systemic influences present in humans, such as liver metabolism, endocrine signaling, and immune responses. Fourth, postmortem studies are subject to confounding variables, including genetic background, sex, age, psychiatric comorbidities, medications, smoking, and nutritional status, that are absent in controlled cell culture systems. Finally, gene expression changes in cell models reflect immediate responses to ethanol exposure or withdrawal, whereas in human brains they likely represent long-term adaptive or compensatory states after years of alcohol use and associated damage. Taken together, these differences help explain why findings from cellular models and human postmortem brains do not always align.
To reconcile discrepancies between cell culture and postmortem brain results, future studies should apply single-cell (or single-nucleus) transcriptomics in combination with iPSC-derived neuronal and glial models. Single-cell profiling of postmortem tissue can identify the specific cell types and circuits in which m6A regulators are dysregulated in AUD, whereas iPSC-derived neurons and glia permit controlled, donor-matched experiments using chronic intermittent ethanol and acetaldehyde exposures. Integrative analysis, mapping in vitro cell populations to cell-type signatures from human tissue and performing targeted m6A profiling on sorted cell types, would clarify whether the discrepancies arise from cell-type composition, chronicity of exposure, metabolic differences, or compensatory adaptations. Such an approach would more directly connect molecular mechanisms observed in defined cellular systems to the cell-type specific pathology in the human brain.
In conclusion, the findings from the present study indicate that chronic intermittent alcohol exposure/withdrawal dynamically disrupts m6A regulatory gene expression and may alter downstream RNA regulatory pathways contributing to AUD pathophysiology. Further studies are warranted to elucidate the mechanisms by which alcohol-induced dysregulation of m6A regulators contributes to AUD risk.
Materials and methods
SH-SY5Y and SW620 cells
The neuron-like SH-SY5Y cell line (neuroblastoma, derived from metastatic bone cancer) and the non-neuronal SW620 cell line (human colorectal adenocarcinoma, derived from a metastatic lymph node) were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). Both cell types were cultured in a 1:1 mixture of EMEM and F-12K medium (ATCC), supplemented with 10% fetal bovine serum (v/v; Corning, Woodland, CA, USA) and 1% penicillin-streptomycin (Corning, Manassas, VA, USA), at 37°C in a humidified incubator with 5% CO₂.
Chronic intermittent ethanol and acetaldehyde exposure
Chronic intermittent ethanol (CIE) and acetaldehyde (CIA) exposures were performed following previously described protocols [19, 21–23] over 3 weeks, followed by a 24-h withdrawal (WD) period (Supplementary Fig. 1). The alternating cycles of exposure and withdrawal in the CIE model parallel the binge/abstinence patterns often seen in individuals with AUD. The 24-h withdrawal following CIE mimics the early abstinence phase in humans, which is associated with increased anxiety-like behavior and drives continued drinking and relapse. Briefly, SH-SY5Y and SW620 cells were seeded in 12 culture dishes at ∼5 × 105 cells per dish, respectively. For CIE, cells in half of the dishes (n = 6) were exposed to 40 mM ethanol, and the medium was replaced with ethanol-free medium after 4 h. This cycle was repeated on Days 2–4, followed by 3 days of ethanol-free medium. The same weekly pattern was applied during Weeks 2 and 3, except for Days 17–18. To simulate withdrawal, on Day 17, treated cells were split: one set continued ethanol exposure, while the other was switched to ethanol-free medium until Day 18. By the end of the 3-week protocol, 18 collections per cell line were obtained: control (n = 6), CIE (n = 6), and CIE + WD (n = 6), and stored at −80°C. CIA exposure followed the same schedule, except that ethanol was replaced with 30 μM acetaldehyde. The chosen concentrations (40 mM ethanol and 30 μM acetaldehyde) approximate blood levels observed in humans after heavy drinking [24–29].
Total RNA extraction
A total of 72 cell collections were prepared [18 collections (6 × controls, 6 × treatment, and 6 × treatment + WD) × 2 treatments (ethanol and acetaldehyde) × 2 cell lines]. Total RNA was extracted using the miRNeasy Mini Kit (QIAGEN, Germantown, MD, USA) with on-column DNase digestion (RNase-Free DNase Set, QIAGEN) to remove genomic DNA. RNA concentration and purity was assessed using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).
m6A regulatory gene expression analysis by RT-qPCR
Primer pairs for 18 m6A regulatory genes, including seven writers (KIAA1429, METTL3, METTL4, METTL14, RBM15, RBM15B, WTAP), two erasers (ALKBH5, FTO), and nine readers (YTHDF1/2/3, YTHDC1/2, IGF2BP1/2/3, HNRNPA2B1), were designed using NCBI Primer-Blast (Supplementary Table 1). A 144-bp amplicon of the β-actin gene (ACTB) was used as the reference, following the recommendation of a published study [30]. PCR products were visualized by agarose gel electrophoresis (Supplementary Fig. 2).
Relative gene expression was quantified by RT-qPCR using SYBR® Green PCR Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) on a QuantStudio™ 12K Flex Real-Time PCR System (Thermo Fisher Scientific). The cycling conditions were: 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min, with a final melting-curve analysis, following the manufacturer’s SYBR® Green PCR Master Mix protocol. For each m6A regulatory gene and its paired house-keeping gene ACTB, a single 96-well PCR plate was used to include all RNA samples (Control: n = 6; CIE or CIA: n = 6; and CIE or CIA + WD: n = 6) across two cell types and both genes (a m6A regulatory gene and ACTB), resulting in a total of 72 PCR reactions per plate. Threshold cycle (Ct) values were analyzed using QuantStudio™ 12K Flex Software. The relative expression levels (∆Ct) of seven reader genes, two eraser genes, and nine reader genes in each sample collected from the three treatment groups (Control, CIE or CIA, CIE or CIA + WD) of two types of cells (SH-SY5Y and SW620) are presented in Supplementary Table 2A-J.
Statistical analyses of CIE- and CIA-induced gene expression
Relative mRNA expression levels (ΔCt) were calculated by normalizing the Ct values of each m6A regulatory gene (writer, reader, or eraser) to that of the reference gene ACTB [ΔCt = Ct(m6A regulatory gene)—Ct(ACTB)]. Differences in gene expression among the three experimental conditions, Control (no ethanol or acetaldehyde treatment), CIE or CIA, and CIE or CIA followed by withdrawal (CIA or CIA + WD), were analyzed using one-way ANOVA. Post hoc pairwise comparisons were performed with Tukey’s test to identify significant group differences. Bar graphs illustrating m6A regulatory gene expression across conditions [Control, CIE (or CIA), and CIE (or CIA) + WD] were generated using GraphPad Prism 10 (GraphPad Software, Boston, MA, USA).
Analysis of AUD-associated m6A regulatory gene expression changes in human postmortem brains
Recently, we profiled the miRNA and mRNA transcriptomes by RNA-seq in 192 postmortem brain tissue samples dissected from eight brain regions (amygdala, caudate nucleus, cerebellum, hippocampus, nucleus accumbens, prefrontal cortex, putamen, and ventral tegmental area) of 12 individuals of European ancestry with AUD and 12 matched controls. Based on these data, we constructed AUD-associated miRNA-mRNA regulatory networks in each brain region [17]. The RNA-seq FASTQ files for these 192 samples are publicly available in the NCBI Gene Expression Omnibus under accession number GSE181982. In the present study, we extracted expression data for 16 m6A regulatory genes and assessed their differential expression across the eight brain regions in individuals with AUD. As described previously, differential expression analysis was conducted using the limma-voom framework, which leverages the strengths of the limma and voom packages in R [31]. A linear regression model implemented in limma [32] was applied to compare m6A regulatory gene expression between AUD cases and matched controls within each brain region.
Supplementary Material
Contributor Information
Ji Sun Koo, Department of Psychiatry, Boston University Chobanian & Avedisian School of Medicine, Boston, MA 02118, United States; The Biomedical Genetics Section, Department of Medicine, Boston University Chobanian & Avedisian School of Medicine, Boston, MA 02118, United States.
Qiansheng Zhan, Department of Psychiatry, Boston University Chobanian & Avedisian School of Medicine, Boston, MA 02118, United States; The Biomedical Genetics Section, Department of Medicine, Boston University Chobanian & Avedisian School of Medicine, Boston, MA 02118, United States.
Huiping Zhang, Department of Psychiatry, Boston University Chobanian & Avedisian School of Medicine, Boston, MA 02118, United States; The Biomedical Genetics Section, Department of Medicine, Boston University Chobanian & Avedisian School of Medicine, Boston, MA 02118, United States.
Author contributions
Ji Sun Koo (Conceptualization [equal], Data curation [equal], Investigation [equal], Methodology [equal], Writing – original draft [lead]), Qiansheng Zhan (Investigation [equal], Methodology [equal], Writing – review & editing [equal]), Huiping Zhang (Conceptualization [equal], Data curation [equal], Formal Analysis [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Project administration [equal], Resources [equal], Supervision [equal], Writing – original draft [equal], Writing – review & editing [equal])
Conflicts of interest
None declared.
Funding
This research was supported by the National Institute on Alcohol Abuse and Alcoholism [grant number R01AA029758].
Data availability
The relative expression levels of m6A regulatory genes are displayed in Supplementary Table 2a–j. The RNA-seq FASTQ files for 192 human postmortem brain tissue samples are publicly available in the NCBI Gene Expression Omnibus (GEO) under accession number GSE181982.
References
- 1. Meyer K D, Jaffrey S R. Rethinking m(6)A readers, writers, and erasers. Annu Rev Cell Dev Biol. 2017; 33: 319–42. 10.1146/annurev-cellbio-100616-060758 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Shi H, Wei J, He C. Where, when, and how: context-dependent functions of RNA methylation writers, readers, and erasers. Mol Cell. 2019; 74: 640–50. 10.1016/j.molcel.2019.04.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Jiang X, Liu B, Nie Z et al. The role of m6A modification in the biological functions and diseases. Signal Transduct Target Ther. 2021; 6:74. 10.1038/s41392-020-00450-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Fu Y, Dominissini D, Rechavi G et al. Gene expression regulation mediated through reversible m(6)A RNA methylation. Nat Rev Genet. 2014; 15: 293–306. 10.1038/nrg3724 [DOI] [PubMed] [Google Scholar]
- 5. Dominissini D, Moshitch-Moshkovitz S, Schwartz S et al. Topology of the human and mouse m6A RNA methylomes revealed by m6A-seq. Nature. 2012; 485: 201–206. 10.1038/nature11112 [DOI] [PubMed] [Google Scholar]
- 6. Alarcon C R, Goodarzi H, Lee H et al. HNRNPA2B1 is a mediator of m(6)A-dependent nuclear RNA processing events. Cell. 2015; 162: 1299–308. 10.1016/j.cell.2015.08.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mathoux J, Henshall D C, Brennan G P. Regulatory mechanisms of the RNA modification m(6)A and significance in brain function in health and disease. Front Cell Neurosci. 2021; 15: 671932. 10.3389/fncel.2021.671932 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Woldemichael B T, Bohacek J, Gapp K et al. Epigenetics of memory and plasticity. Prog Mol Biol Transl Sci. 2014; 122: 305–40. [DOI] [PubMed] [Google Scholar]
- 9. Dabrowski K R, Daws S E. Morphine-driven m6A epitranscriptomic neuroadaptations in primary cortical cultures. Mol Neurobiol. 2024; 61: 10684–704. 10.1007/s12035-024-04219-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Ou C, Zhang K, Mu Y et al. YTHDF1 in periaqueductal gray inhibitory neurons contributes to morphine withdrawal responses in mice. BMC Med. 2024; 22: 406. 10.1186/s12916-024-03634-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Xue A, Huang Y, Li M et al. Comprehensive analysis of differential m6A RNA methylomes in the Hippocampus of Cocaine-Conditioned Mice. Mol Neurobiol. 2021; 58: 3759–68. 10.1007/s12035-021-02363-4 [DOI] [PubMed] [Google Scholar]
- 12. Shi Z, Wen K, Zou Z et al. YTHDF1 mediates translational control by m6A mRNA methylation in adaptation to environmental challenges. Biorxiv. 2024; 2024.08.07.607063. [Google Scholar]
- 13. Hess M E, Hess S, Meyer K D et al. The fat mass and obesity associated gene (Fto) regulates activity of the dopaminergic midbrain circuitry. Nat Neurosci. 2013; 16: 1042–48. 10.1038/nn.3449 [DOI] [PubMed] [Google Scholar]
- 14. Liu Y, Zhang H. RNA m6A modification changes in postmortem nucleus accumbens of subjects with alcohol use disorder: a pilot study. Genes. 2022; 13: 958. 10.3390/genes13060958 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Pan X S, Li B W, Wang L L et al. Kupffer cell pyroptosis mediated by METTL3 contributes to the progression of alcoholic steatohepatitis. FASEB J. 2023; 37: e22965. 10.1096/fj.202300059RR [DOI] [PubMed] [Google Scholar]
- 16. Feng Y, Guo S, Zhao Y et al. DNA 5mC and RNA m(6)A modification successively facilitates the initiation and perpetuation stages of HSC activation in liver fibrosis progression. Cell Death Differ. 2023; 30: 1211–20. 10.1038/s41418-023-01130-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Lim Y, Beane-Ebel J E, Tanaka Y et al. Exploration of alcohol use disorder-associated brain miRNA-mRNA regulatory networks. Transl Psych. 2021; 11: 504. 10.1038/s41398-021-01635-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Matsumura Y, Stiles K M, Reid J et al. Gene therapy correction of aldehyde dehydrogenase 2 deficiency. Mol Ther—Methods Clin Dev. 2019; 15: 72–82. 10.1016/j.omtm.2019.08.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Koo J S, Zhan Q, Zhang H. Acetaldehyde-driven mRNA methylation and expression changes in ethanol-metabolizing enzyme genes. Epigenetics. 2025; 20: 2493865. 10.1080/15592294.2025.2493865 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Lindros K O, Stowell A, Pikkarainen P et al. Elevated blood acetaldehyde in alcoholics with accelerated ethanol elimination. Pharmacol Biochem Behav. 1980; 13 Suppl 1 119–24. 10.1016/S0091-3057(80)80019-0 [DOI] [PubMed] [Google Scholar]
- 21. Liu Y, Koo J S, Zhang H. Chronic intermittent ethanol exposure-induced m6A modifications around mRNA stop codons of opioid receptor genes. Epigenetics. 2024; 19: 2294515. 10.1080/15592294.2023.2294515 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. McClintick J N, Thapa K, Liu Y et al. Effects of chronic intermittent ethanol exposure and withdrawal on neuroblastoma cell transcriptome. Alcohol. 2020; 85: 119–126. 10.1016/j.alcohol.2019.12.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Karkhanis A N, Rose J H, Huggins K N et al. Chronic intermittent ethanol exposure reduces presynaptic dopamine neurotransmission in the mouse nucleus accumbens. Drug Alcohol Depend. 2015; 150: 24–30. 10.1016/j.drugalcdep.2015.01.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Korsten M A, Matsuzaki S, Feinman L et al. High blood acetaldehyde levels after ethanol administration. Difference between alcoholic and nonalcoholic subjects. N Engl J Med. 1975; 292: 386–89. 10.1056/NEJM197502202920802 [DOI] [PubMed] [Google Scholar]
- 25. Eriksson C J, Peachey J E. Lack of difference in blood acetaldehyde of alcoholics and controls after ethanol ingestion. Pharmacol Biochem Behav. 1980; 13 Suppl 1 101–105. 10.1016/S0091-3057(80)80016-5 [DOI] [PubMed] [Google Scholar]
- 26. Lindros K O. Human blood acetaldehyde levels: with improved methods, a clearer picture emerges. Alcohol: Clin Exp Res. 1983; 7:70–75. 10.1111/j.1530-0277.1983.tb05414.x [DOI] [PubMed] [Google Scholar]
- 27. Bell R L, Rodd Z A, Engleman E A et al. Scheduled access alcohol drinking by alcohol-preferring (P) and high-alcohol-drinking (HAD) rats: modeling adolescent and adult binge-like drinking. Alcohol. 2014; 48: 225–34. 10.1016/j.alcohol.2013.10.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Melendez R I. Intermittent (every-other-day) drinking induces rapid escalation of ethanol intake and preference in adolescent and adult C57BL/6J mice. Alcohol: Clin Exp Res. 2011; 35: 652–58. 10.1111/j.1530-0277.2010.01383.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Jury N J, Pollack G A, Ward M J et al. Chronic ethanol during adolescence impacts corticolimbic dendritic spines and behavior. Alcohol: Clin Exp Res. 2017; 41: 1298–308. 10.1111/acer.13422 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Pan Y, Kaiguo M, Razak Z et al. Chronic alcohol exposure induced gene expression changes in the zebrafish brain. Behav Brain Res. 2011; 216: 66–76. 10.1016/j.bbr.2010.07.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Law C W, Chen Y, Shi W et al. voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol. 2014; 15: R29. 10.1186/gb-2014-15-2-r29 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Ritchie M E, Phipson B, Wu D et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015; 43: e47. 10.1093/nar/gkv007 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The relative expression levels of m6A regulatory genes are displayed in Supplementary Table 2a–j. The RNA-seq FASTQ files for 192 human postmortem brain tissue samples are publicly available in the NCBI Gene Expression Omnibus (GEO) under accession number GSE181982.