TABLE 2.
Summary of studies exploring the relationship between DNAm and clinical improvement.
| Studies | Genome Region analyzed | Sample size | Drugs (duration) | Tissue | Platform | Main results |
| Schizophrenia | ||||||
| Tang et al., 2014 | HTR1A Promoter region | 82 SCZ | Co-medications (10 weeks) | Blood | Bisulfite conversion + PCR | • DNAm at CpG 13 showed a positive correlation with changes in total PANSS scores (p = 0.006) and changes in negative factor (p < 0.001) • DNAm at CpG 13 was positively correlated with baseline negative factor (p = 0.019) |
| Shi et al., 2017 | CYP3A4, CYP2D6, ABCB1, HTR2A, DRD2 Upstream of promoter region | 288 SCZ | Risperidone* (4 weeks) | Blood | MassARRAY Analyzer 4 | RES vs. NRES: • Decrease of DNAm at CpG_193 (p = 0.0008, q = 0.012), CpG_242:244:250 (p = 0.000051, q = 0.00076) and CpG_284 (p = 0.0023, q = 0.034) in CYP2D6 • Increase of DNAm at CpG_−367:-372:−374 (p = 0.0018, q = 0.028), CpG_−222(p = 0.012, q > 0.05) and CpG_−243 (p = 0.02, q > 0.05) in CYP3A4 • Decrease of DNAm at CpG_−36 (p = 0.000091, q = 0.0014), CpG_−258 (p = 0.000082, q = 0.0013), CpG_−296 (p = 0.000091, q = 0.0014) in CYP3A4 |
| Kinoshita et al., 2017 | Whole genome | 21 SCZ | Clozapine (1 year) | Blood | HM450K | • Increases in DNAm of the CREBBP (cg05151055, p = 2.7 × 10-7, q < 0.05) gene was significantly correlated with clinical improvements in treatment-resistant SCZ. |
| Nour El Huda et al., 2018 | COMT CpG island of the 5′ upstream | 138 SCZ 132 HC | Co-medications (NA) | Blood | Bisulfite conversion + PCR | • DNAm of COMT was negatively correlated with excitement (p < 0.001) and depressed (p = 0.001) scores |
| Miura et al., 2018 | ANKK1 -162C to + 260C of the 5′ region | 34 SCZ | Aripiprazole*(6 weeks) | Blood | Bisulfite conversion + PCR | • RES vs. NRES: no differences in DNAm levels at overall CpG sites; hypermethylation at CpG site 387 (p = 0.017) • DNAm at CpG sites 387 was negatively correlated with changes in total PANSS (p = 0.031), positive (p = 0.037) and negative scores (p = 0.039) |
| Bipolar disorder | ||||||
| Burghardt et al., 2018 | AKT1, AKT2, AKT3 Promoter region | 30 BD | AAPs or MSs at least 3 months | FSM | PCR + MS-HRM | • AAPs patients: AKT2 DNAm and HOMA-IR were positively correlated (p = 0.3) • MSs patients: AKT2 DNAm and HOMA-IR was negatively correlated (p = 0.1) |
| Major depressive disorder | ||||||
| Kang et al., 2013 | 5-HTT Promoter region | 108 MDD | Ads (12 weeks) | Blood | Bisulfite conversion + PCR | • CpG site 2 and the average DNAm of 5-HTT were negatively correlated with changes of HAM-D scores (p < 0.05) • CpG sites 1 was negatively correlated with changes of HAMA scores (p < 0.05); • average DNAm was negatively correlated with changes of SOFAS scores (p < 0.05) |
| Powell et al., 2013 | IL11 CpG island | 113 MDD | Escitalopram or Nortriptyline (12 weeks) | Blood | Bisulfite conversion + PCR | • DNAm of CpG unit 5 was negatively correlated with changes of MADRS scores (p = 0.005, q = 0.055); • Escitalopram group: CpG unit 4 was positively correlated with changes of MADRS scores (p = 0.005, q = 0.055); • Nortriptyline group: CpG unit 4 was negatively correlated with changes of MADRS scores (p = 0.005, q = 0.055) |
| Tadić et al., 2014 | BDNF Promoter region | 39 MDD | Monoaminergic drugs (6 weeks) | Leus | NA | • RES vs. NRES: higher DNAm at CpG-87 of BDNF (p = 0.003, q = 0.03) |
| Okada et al., 2014 | 5-HTT CpG island | 50 MDD 50 HC | ADs (6 weeks) | Blood | MCS | • DNAm of CpG 76 was positively correlated with total HAM-D scores (p = 0.03); • DNAm of CpG 3 was positively correlated with improvement ratio (HAM-D, p = 0.02) |
| Domschke et al., 2014 | 5-HTT Upstream of exon 1A | 94 MDD | Escitalopram*(6 weeks) | Blood | Bisulfite sequencing | • CpG 1 (p = 0.048), CpG 2 (p = 0.002, q < = 0.05), CpG 4 (p = 0.029) and average DNAm (p = 0.005, q < = 0.05) of 5-HTT were positively correlated with changes of HAMD-21 scores |
| Domschke et al., 2015 | MAO-A Promoter/exon1/intron1 regions | 94 MDD | Escitalopram* (6 weeks) | Blood | Bisulfite sequencing | • Females: DNAm of CpG 1 (p = 0.04) and CpG 5 (p = 0.009) were positively correlated with changes of HAMD-21 scores; while average DNAm across all CpGs showed no association with treatment response • Males: no significant associations detected |
| Kahl et al., 2016 | GLUT1/4 Promoter region | 37 MDD | ADs (6 weeks) | Blood | Bisulfite sequencing | • REMs vs. NREMs: hypomethylation of the average DNAm of GLUT1 (p < 0.001); • No significant differences were observed for GLUT4 DNAm. |
| Iga et al., 2016 | 5-HTT Promoter region relate to TSS | 28 MDD; 29 HC | ADs (8 weeks) | Leus | Bisulfite conversion + PCR | • DNAm of CpG 3 (p = 0.003) and CpG 5 (p = 0.004) were negatively correlated with baseline HAMD-17 scores • DNAm of CpG 2 (p = 0.04) was negatively correlated with clinical improvement as assessed with HAMD-17 |
| Gassó et al., 2017 | HTR1B Promoter region | 57 MDD | Fluoxetine*(12 weeks) | Blood | Pyrosequencing | • A negative correlation was found between the average DNAm level and GAF/CGAS changes (p = 0.004) |
| Lieb et al., 2018 | BDNF Exon IV or P11 promoter | 561 MDD | Escitalopram*(8 weeks) | Blood | Bisulfite conversion + PCR | • DNAm status at CpG-87 of promoter exon IV (p = 0.029) was positively correlated with remission of the depressive symptomatology, especially in severe MDD patients (n = 199, p = 0.031) |
| Wang et al., 2018a | HTR1A/B promoter region | 85 MDD | Escitalopram* (8 weeks) | Blood | Illumina Hiseq | REM VS. NREM: • Hypermethylation of CpG 668 (HTR1A, p = 0.025) and CpG 1401 (HTR1B, p = 0.033); • Higher DNAm changes after treatment in CpG 2793 (p = 0.015), CpG 2834 (p = 0.002), CpG 2927 (p = 0.023) and CpG 2937 (p = 0.003) of HTR1A and CpG-100 (p = 0.021) and CpG 1401 (p = 0.029) of HTR1B, compared with NREMs |
| Wang et al., 2018b | BDNF Promoter region | 85 MDD | Escitalopram* (8 weeks) | Blood | Illumina Hiseq | • REM vs. NREM: hypermethylation of amplicon BDNF_1 (p = 0.006), BDNF_3 (p = 0.016), BDNF_4 (p = 0.029) and BDNF_5 (p = 0.036); • BDNF_1 (p = 0.004), BDNF_3 (p = 0.013), BDNF_4 (p = 0.034), BDNF_5 (p = 0.027) and average (p = 0.038) DNAm of BDNF was positively correlated with HAMD-17 changes |
| Wagner et al., 2019 | BDNF Exon IV and P11 promoter | 110 MDD | Escitalopram*(8 weeks) | Blood | Bisulfite conversion + PCR | • The BDNF exon IV promoter and P11 gene methylation did not predict a normalization of executive dysfunctions |
| Hsieh et al., 2019 | BDNF Exon IX promoter | 36 MDD | SSRIs (4 weeks) | Blood | Pyrosequencing | • RES vs. NRES: hypermethylation of CpG 24 (p = 0.029) and CpG 324 (p = 0.031) of BDNF promoter region |
| Draganov et al., 2019 | IL6, IL6R, IL1-β CpG island relative to 5′ regulatory region | 153 MDD | SSRIs (NA) | Blood | Bisulfite-pyrosequencing | • RES vs. NRES: hypermethylation of CpG IL6R_4 (p = 0.05) |
| Takeuchi et al., 2017 | Whole genome | 20 MDD | Paroxetine*(6 weeks) | Blood | HM450K | • BR vs. WR: 218 sites were nominally significant (p < 0.05), and 2 sites (PPFIA4 and HS3ST1) were significantly hypomethylated after FDR correction (q < 0.05). |
| Ju et al., 2019 | Whole genome | 177 MDD | Escitalopram (8 weeks) | Blood | EPIC | • RES vs. NRES: 303 sites were nominally significant (p < 0.05, Δβ ≥ ± 2%) |
| Bipolar disorder & Major depressive disorder | ||||||
| Carlberg et al., 2014 | BDNF Exon I promoter | 207 MDD; 59 BD; 278 HC | ADs (NA) | Blood | Bisulfite conversion + PCR | • No significant correlations were found for DNAm levels and the total BDI sum score (available from n = 81 MDD) |
Maudsley Staging Method, MSM; Responder, RES; Non-responder, NRES; Best-responder, BR; Worse-responder, WR; Methylation-sensitive High-resolution Melting, MS-HRM; Antipsychotics, APs; Atypical Antipsychotics, AAPs; Typical Antipsychotics, TAPs; Classic Antipsychotics, CAPs; Mood Stabilizers, MSs; Valproate, VPA; Lithium, Li; Fasting skeletal muscle, FSM. *Monotherapy was permitted for co-medications with other psycho-pharmacological agents or medications for insomnia.