Introduction

Fetal alcohol spectrum disorders (FASD) include fetal alcohol syndrome (FAS), partial FAS, alcohol-related neurodevelopment disorder (ARND) and neurobehavioral disorder1. The prevalence of FASD is conservatively estimated to affect 11.3 to 50.0 per 1000 children in the United States, although the prevalence may be as high as 98.5 per 1000 children2, and consequences include maladaptive behavioral characteristics, cognitive impairments, and physical abnormalities3. In addition, children with FASD often show sleep problems4, though limited work has been done to characterize these sleep disturbances and their consequences5. Given the impact of sleep disturbance on neurobehavioral functioning1 and health6, this is a significant, clinically relevant issue that is not well understood in the FASD population.

It is well recognized that the timing and duration of sleep are controlled by the circadian system, which keeps a roughly 24 h internal rhythm that entrains to environmental stimuli, and the sleep homeostat. The circadian system is a dynamic biological timekeeper system that controls daily timings of physiological, biochemical, and behavioral functions7. At the molecular level, it consists of various clock genes that participate in maintaining a 24-h internal timekeeping. The core clock genes BMAL1 (Brain and muscle arnt-like1) and CLOCK (from circadian locomotor output cycles kaput) rhythmically regulate expression of other core clock genes that encode PER (period)1/2/3 and CRY (Cryptochrome Circadian Regulator) 1/2 proteins. The PER/CRY complex subsequently translocates to the nucleus from cytoplasm to suppress transcription of BMAL1/CLOCK, thus completing a transcription-translation negative feedback loop that occurs over a 24-h period. Expression of these clock genes within the hypothalamic suprachiasmatic nucleus (SCN), the mammalian brain's circadian central pacemaker, and other peripheral pacemaker cells, integrate SCN temporal information into physiological processes. The molecular composition of the timing mechanisms in the SCN clock and peripheral oscillators is highly resembling, despite region-specific differences in particular gene products to achieve local physiological programs8. Therefore, clock genes serve as the basis of an intracellular timekeeping system, present throughout the body, which generates approximately 24-h rhythms in physiology and behavior. Many clock-controlled genes show circadian oscillation in their mRNA expression patterns, resulting in numerous physiological processes that exhibit circadian rhythmicity9.

In addition to core clock genes, several additional genes and feedback loops participate in fine tuning the rhythmic expression of clock genes and clock-regulated physiological functions. For example, BMAL1 expression is regulated by the secondary feedback loop, in which CLOCK and BMAL1 induce NR1D1 (Nuclear Receptor Subfamily 1 Group D Member 1 also known as REV-ERBα) and NR1D2 (REV-ERBβ) expression and then NR1D1/NR1D2 proteins repress BMAL1 expression10,11,12. The other negative regulators of the molecular clockwork, DEC1 and DEC2 (also known as Basic helix-loop-helix family member e41, or BHLHE41) are also shown to inhibit CLOCK-BMAL1-mediated transcription of E-box-containing promoters13. Additional motifs involved in rhythmic gene expression are D-box elements, which are activating transcription factors DBP (D-box binding protein), and the repressor NFIL3 (nuclear factor, interleukin-3 regulated). DBP rhythmically activates transcription of various genes through a DNA cis-element, D-box. The DBP-dependent transactivation is repressed by competitive binding of NFIL3 to the D-box. Collectively, D-box-mediated transcriptional regulation plays important roles in input and output in the circadian clock system14,15. Finally, neuronal PAS domain protein 2 (NPAS2) has been shown to bind with DNA at E-box sequences and form heterodimers with BMAL1 to control cellular rhythmicity16.

Available data indicate the relevance of circadian dysregulation associated with drug addiction, including alcohol17,18,19,20. In a study employing male patients with alcohol dependence, it has been shown that basal levels of circadian genes are significantly reduced in Peripheral blood mononuclear cells (PBMC) of alcohol-dependent patients21. However, the effect of prenatal ethanol on circadian clock gene expression has been mainly studied in experimental animals22,23,24,25. There are number of studies indicating that prenatal alcohol exposure affects circadian expression of hormones, body temperature, blood pressure, and regulation of the stress response22,26,27. Our laboratory has used a rat model of prenatal alcohol exposure to show altered circadian expression of clock genes, Per1 and Per2, in the arcuate nucleus and the SCN of male rats17,23. Thus, animal studies indicate that prenatal alcohol exposure may disrupt clock gene expression, but this has not been investigated in clinical populations.

It has been established that biological fluid can be utilized for the investigation of endogenous circadian expression of clock genes28,29,30. Human saliva is an excellent source of genetic (DNA/RNA), protein, and metabolite which makes it an ideal biofluid for non-invasive clinical assessment of systemic diseases31. Saliva sampling is also less stressful and is thus ideal for sequential sampling. In fact, salivary cortisol level was measured in children with FASD and found to be higher across the day compared to children without FASD32. ARRB1 (β-arrestin1) has been identified to function as a circadian salivary biomarker which plays an important role in circadian regulation of diverse physiological functions10. However, there are no studies available wherein FASD saliva transcripts were analyzed in connection to circadian rhythmic pattern. Here we report for the first time that core-clock genes are expressed in a diurnal pattern in saliva, and that prenatal ethanol exposure significantly alters the circadian expression of these clock genes. We also identified the role of chromatin modifiers in the circadian expression of clock genes, uncovering possible molecular mechanisms triggering pathological states associated to circadian misalignment by prenatal exposure to ethanol.

Materials and methods

Human subjects

Children aged 6–10 years, with and without heavy prenatal alcohol exposure (alcohol-exposed [AE], control [CON]), were recruited from the FASD Research Registry, the San Diego State University Center for Behavioral Teratology and the University of California San Diego (UCSD) FASD Research Registry, and the MotherToBaby program. Subjects were enrolled as part of a larger study examining sleep and neurobehavioral functioning in individuals with heavy prenatal alcohol exposure. Two groups of children are recruited and assessed: children with histories of heavy prenatal alcohol exposure (AE; n = 23) and typically developing children (CONT; n = 26). Children of all sexes, races, and ethnicities between 6:0 and 10:11 years of age was selected based on the inclusion and exclusion criteria described previously4,33.

Two saliva KITs (Zymo Research saliva/sputum collection kit; R1210) were sent to each participant. Parents were instructed to collect saliva samples at 2 time points during the daytime (ZT3-11) for each child. One child in the control group gave only one saliva sample. All methods were carried out in accordance with relevant guidelines and regulations. All experimental protocols underwent Institutional Review Board (IRB) approval from University of California San Diego/San Diego State University. Informed consents were obtained from all subjects and/or their legal guardians.

Saliva sample collection

The saliva samples from control and AE children were collected at different time points. Saliva/sputum collection kits (R1210) were used to collect saliva samples in specified containers containing DNA/RNA shield solution which protected the DNA/RNA present in the samples. The samples and shield solutions were mixed at a ratio of 1:1 according to the manufacturer’s instructions. A total 51 control and 46 AE samples were collected. Time of collection was recorded, and kits with saliva samples were sent to Rutgers, New Jersey.

RNA isolation

The RNA isolation was done according to the protocol of Zymo Research kit (D7005). In brief, to the 1:1 mixture of saliva and shield solution Proteinase K and Proteinase K Digestion Buffer were added. The whole mixture was then incubated at room temperature (20–30 °C) for 30 min. After incubation, samples were vortexed and centrifuged at maximum speed for 2 min to pellet debris. The 400 ml of the cleared supernatant was transferred to the nuclease-free tube. The DNA/RNA lysis buffer was added at the ratio of 1:1 and mixed well. The samples were then transferred into the Zymo-Spin™M IC-XM Column1 with the collection tube and centrifuged to flow-through which contained the RNA. The flow-through was then mixed with one volume of 95% to 100% ethanol. After mixing, the samples were transferred into the Zymo-Spin™ IC Column' with the collection tube and centrifuged. The flow-through was discarded and the column was washed with 400 µl of DNA/RNA prep buffer followed by 2 washings with DNA/RNA wash buffer. After washing, 15 µl of DNase/RNase-free water was added directly to the column matrix and centrifuged to collect the RNA and used immediately for cDNA preparation or stored frozen.

Conventional cDNA synthesis high-capacity cDNA reverse transcription kit

The following strategy was taken according to the previously published protocol34. For analyzing the expression of pan-bacterial rRNA (16S), and human rRNA (18S), total salivary RNA was converted into complementary DNA (cDNA) via reverse transcription using the High-capacity cDNA reverse transcription kit (Applied Biosystem, Life Technologies). The RNA concentration was measured in NanoDrop and total 500 ng of RNA was used as input for all the samples. After reverse transcription, cDNA was diluted and used for qRT-PCR for the detection of 16S and 18S rRNAs.

Adjustment of human RNA input for cDNA synthesis

The adjustment of input was done according to the previously published protocol34. As NanoDrop spectrophotometer only provides total RNA values which consists of the mixture of human and bacterial RNA, we calculated the ratio of the detected raw Ct-values of human rRNA (18S) to pan-bacterial rRNA (16S). The rRNA is representative for all human and bacterial RNA specimens. The ratio was used to calculate fold change value using the following formula: Fold change = 2^(Ct18S rRNA − Ct16S rRNA). After that, the total amount of RNA loaded for each sample (500 ng) was divided with fold change value which gave the amount of human RNA in individual sample. Finally, normalization was done based on the amount of human RNA (5 ng) and samples were used to synthesized cDNA using oligo d(T) primers (Supplementary Fig. 1).

Modified cDNA synthesis—SuperScript III First-Strand Synthesis System

To perform gene expression analysis of human origin only, the eukaryotic messenger RNA (mRNA) was reverse transcribed to cDNA by using the SuperScript® III First-Strand Synthesis System with Oligo (dT) primers34. As a result, pan-bacterial RNA and other human RNA species like tRNA, rRNA were excluded from cDNA conversion. According to the kit description, the amount of starting material can vary from 1 pg to 5 μg of RNA. The amount of 5 ng of human RNA per reaction was set to synthesize cDNA. Using the concentration from repeated NanoDrop™ measurements and the calculated 18S/16S ratio in each sample, the corresponding amount of measured total RNA input was calculated. The reverse transcription was performed according to the manufacturer’s instructions.

RT PCR (qPCR)

The c-DNA was prepared from the RNA of each sample and 1:3 dilution was done using nuclease free water. The real-time qPCR was carried out in 384 well plates. Each well of 384 well plates contain a reaction mixture of 10 µl consisting of 1:3 dilution of c-DNA, primers (Supplementary Table 1) in Power SYBR Green PCR Master Mix and comparative CT (ΔΔCT) standard run method was carried out in Quant Studio 7 real time PCR system. The expression of genes was expressed in terms of relative quantification using the formula 2−(ΔΔCT)35.

Statistics and reproducibility

We did adjust our time of collection to Zeitgeber Time (ZT; ZT0 is the equivalent of sunrise at 06:00). All data are presented as ZT time. Graph Pad Prism (version 10) software was used for statistical analysis of data. We have presented data in both spline graph of individual data at various time points and the line graph of mean + SEM values of each group at various time points. Cosinor analysis was used to determine rhythmicity and approximate acrophase values. The significance difference between AE and control group was assessed using unpaired t-test. When two variables (time and treatment or sex and treatment) were compared, 2-way ANOVA followed by Tukey's multiple comparison test were used, with p < 0.05 considered significant. In the cosinor model, data of period X units of time is modeled as Y = M + A* cos (2πt/X-Ø)36. In this equation, Y represents the modeled outcome in the units of the circadian variable of interest. M represents the mesor, Ø represents the acrophase, and A represents the curve amplitude37.

Results

Demograpic data of study subjects are presented in Table 1. Children of all sexes, races, and ethnicities between 6:0 and 10:11 years of age (mean age 8.6 years) with histories of heavy prenatal alcohol exposure (AE; n = 23) and typically developing children (CONT; n = 26) were assesed. These two subject groups did not significantly differ on age, sex, race, ethnicity, handedness or parental income (Table 1).

Table 1 Demographic information by groups.
Full size table

We compared the diurnal changes in the expression levels of core clock genes (BMAL1, CLOCK, PER1,2,3, CRY1,2) in saliva samples of control and AE subjects. Our study revealed that in control subjects, expression patterns (as determined by spline graph of individual data at various time points and the line graph of mean + SEM values of each group at various time points) of primary circadian activator genes BMAL1 (Fig. 1A,B) and CLOCK (Fig. 1D,E) fluctuated throughout the day and showed lower expression levels in the afternoon. Cosinor analysis of expression data identified a circadian nature of the rhythm (Fig Y; Zero-amplitude test p values: BMAL1, 0.008, CLOCK, 0.0001; Acrophase times for BMAL1, ZT20.46.83h; CLOCK, ZT19.49h). In subjects with AE, the gene expression patterns of these two activators BMAL1 and CLOCK were elevated and circadian expression patterns were altered (Zero-amplitude test p values: BMAL1, 0.0.311; CLOCK, 0.0906; acrophase times: BMAL1,ZT4.42 0.4h; CLOCK, ZT7.27). The mean levels of Bmal1 and Clock genes at various time points (Fig. 1B,F and supplementary Table 2) and overall mean levels (Fig. 1C) were higher (p < 0.001) in AE group than those in the control group. The overall mean expression of BMAL1 did not show any sex difference in AE effects [F (1, 93) = 1.972, p = 0.1636], while CLOCK1 expression did show significant sex differences, with females showing reduced expression compared to males [F(1, 93) = 10.60, p = 0.0016] (supplemental Fig. 2B,C). In the case of primary repressor genes of circadian clock in CONT group, we found that all PER genes (PER1, Fig. 1G,H: PER2, Fig. 1J,K; PER3, Fig. 1M,N) and CRY genes (CRY1, Fig. 1P,Q: CRY2, Fig. 1S,T) showed rhythmic and elevated expression during the late morning to early afternoon in CONT children. Cosinor analysis of expression data identified the circadian nature of the rhythm (Fig. 1Y; Zero-amplitude test p values for PER1, 0.00048; PER2, 0.00001; PER3, 0.00014; CRY1, 0.00001: CRY2, 0.00001) with acrophase times as follows in control subjects (PER1, ZT07.26; PER2, ZT8.36; PER3, ZT7.55; CRY1, ZT7.83: CRY2, ZT7.80). However, in subjects with AE, all PER and CRY gene expression were reduced and circadian oscillations were disrupted (Cosinor p values; PER1, 0.09267; PER2, < 0.09403; PER3, 0.0.2263; CRY1, 0.49685: CRY2, 0.41161) and acrophase times were modified (PER1, ZT10.84; PER2, ZT11.68; PER3, ZT14.93; CRY1, ZT21.54: CRY2, ZT17.34). The mean levels of PER1-3 and CRY1-2 genes at various time points (Fig. 1H,K,N,P,S and supplementary Table 2) and overall mean levels (Fig. 1I,L,O,R,U) were lower (p < 0.001) in AE groups than those in the control groups. The overall mean levels of of these genes did not show any sex differences [PER1: F (1, 93) = 0.1922, p = 0.6621; PER2: F (1, 93) = 0.06520, p = 0.7990; PER3: F (1, 93) = 0.3453, p = 0.5582; CRY1: F (1, 93) = 0.06445, p = 0.8001; CRY2: F (1, 93) = 0.04588, p = 0.8309] (supplemental Fig. 2D–H).

Fig. 1
figure 1

Effects of prenatal alcohol exposure on circadian expression of peripheral circadian markers (A–C) and core clock genes (D–X). (A,D,G,J,M,P,S,T) Showing changes in the rhythmic expression of core clock genes in salivary samples collected at two time points (morning and afternoon) from each AE and control subjects throughout the day (ZT3 to ZT11) and plotted as spline curves (B,E,H,K,N,Q,T,W) Showing changes in the expression of these clock genes when plotted as mean ± SEM at each hour during ZT 3:00 to ZT 11:00 in AE (ZT3, n = 3; ZT4, n = 6; ZT5, n = 3; ZT6, n = 4; ZT7, n = 9; ZT8, n = 2; ZT9, n = 10; ZT, n = 3: to ZT11, n = 6) and CONT subjects (ZT3, n = 3; ZT4, n = 7; ZT5, n = 6; ZT6, n = 3; ZT7, n = 5; ZT8, n = 6; ZT9, n = 5; ZT, n = 9: to ZT11, n = 7). Data were analyzed using Two-way ANOVA and the interaction between the treatment and time were identified [BMAL1: F (8, 79) = 3.218, p = 0.0032; CLOCK: F (8, 79) = 2.306, p = 0.0282; PER1: F (8, 79) = 3.446, p = 0.0019; PER2: F (8, 79) = 5.536, p < 0.0001; PER3: F (8, 79) = 3.440, p = 0.0019; CRY1: F (8, 79) = 9.740, p < 0.0001; CRY2: F (8, 79) = 10.62, p < 0.0001; ARR1B: F (8, 83) = 16.60, p < 0.0001]. (C,F,I,L,O,R,U,X) Overall Mean ± SEM levels of gene expression in two groups were analyzed using one-way ANOVA. ***p < 0.001. (Y) Cosinor analysis was used to determine the rhythmicity (Zero-amplitude test p value < 0.05 identified rhythmicity and the acrophase value shows peak time of the rhythm) of these clock genes. (Z) Schematic diagram showing postulated mechanisms by which alcohol exposure (AE) alters circadian oscillation by modifying expression of core clock genes in AE children. The daily circadian oscillation which is maintained by a coordinated transcription-translation feedback loop between the core clock activators (BMAL and CLOCK) and core clock repressors (PER and CRY) is affected by AE. Down arrow indicates inhibition.

Full size image

We determined the expression pattern of ARRB1 which is considered a molecular marker of the peripheral circadian rhythm10. ARRB1 levels in saliva of control subjects fluctuated throughout the day in a rhythmic fashion and showed lower expression levels in the afternoon, as determined by spline graph of individual data at various time points (Fig. 1V) and the line graph of mean + SEM values of each group at various time points (Fig. 1W). Cosinor analysis of expression data identified the circadian nature of the rhythm (Fig. 1Y; Zero-amplitude test p value < 0.0001) with an acrophase value of ZT19.80. However, ARRB1 levels in saliva of subjects with AE were markedly reduced and fluctuated without a clear circadian pattern (Zero-amplitude test p value 0.5283) with an acrophase time of ZT9.36. The overall mean level of ARRB1 was lower (p < 0.0001) in AE group compared to control group (Fig. 1X). Also, there were no significant sex differences [F (1, 93) = 0.4623, p = 0.4983] in the mean expression level of ARRB1 in AE and control groups (supplemental Fig. 2A).These data suggest that AE affects the daily circadian oscillation, which is maintained by a coordinated transcription-translation feedback loop between the core clock activators (BMAL and CLOCK) and core clock repressors (PER and CRY) (Fig. 1Y).

We also determined the salivary expression patterns of an activator NPAS2, and several repressors (NFL3, NR1D1, DEC1, DEC2, DBP) of transcription/translation of core circadian control genes, as they are involved in a negative feedback loop responsible for the control of circadian rhythm38. Our study revealed that in control subjects, expression patterns (as determined by spline graph of individual data at various time points and the line graph of mean + SEM values of each group at various time points) of NAPS2 (Fig. 2A,B) and NFL3 (Fig. 2D,E) fluctuated throughout the day. The NPAS2 gene expression pattern showed circadian rhythm (Cosinor p = 0.0021) with the estimated acrophase time is at ZT23.5 in the control subjects and it remained rhythmic (Cosinor p < 0.006) with the estimated acrophase time is at ZT1.70in subjects with AE (Fig. 2S). Themean levels of NPAS2 genes at various time points (Fig. 2B, and supplementary Table 2) or overall mean levels (Fig. 2C) were higher (p < 0.001) in AE group than those in the control group, without showing any sex differences in the AE response [F (1, 87) = 3.469, p = 0.0659] (supplemental Fig. 3A). The levels of NFL3, NR1D1, and DEC2 in saliva showed no rhythmicity in control (Fig. 2S; Cosinor p = ; NFL3—0.1733, NR1D1—0.1463, and DEC2—0.8245) or AE subjects (Fig. 2S; Cosinor p = ; NFL3—0.863, NR1D1—0.1673, and DEC2—0.4544), but had acrophase modification (Fig. 2S; control—NFL3—ZT21.64, NR1D1—ZT7.24, and DEC2—ZT8.32; AE- NFL3—ZT8.45, NR1D1—ZT15.58, and DEC2—ZT1.16) and mean level elevation (p < 0.001) at various time points (Fig. 2B; supplementary Table 2) or overall mean levels (Fig. 2C). Other repressors of core circadian control genes (DEC1 and DBP) were expressed in a circadian manner in both control (Cosinor p; DEC1-0.0001, DBP-0.0212) and AE subjects (Cosinor p; DEC1-0.0012, DBP-0.0329) in saliva samples of control subjects; however, acyclic expression in some but not all of these genes were seen (Cosinor p; NR1D1-0.1562, DEC1-0.0004, DEC2-0.445; DBP-0.0001) in AE subjects (Fig. 2G,J,M,P). The estimated acrophase times of these genes also differed between Control (NR1D1-ZT 7.22, DEC1-7.45h, DEC2-ZT8.139h, DBP-ZT7.78h,) and AE subjects (NR1D1-ZT 16,23h, DEC1-ZT 4.937, DEC2 ZT 1.201, DBP-ZT 5.5). The mean levels of DEC1, DEC2, DBP genes but not NR1D1 genes at various time points (Fig. 2H,K,N,Q; supplementary Table 2) or overall mean levels of DEC1, DEC2, DBP genes but not NR1D1 were also lower (p < 0.0001) in the AE group compared to those in the control group (Fig. 2I,L,O,R). The responses of these clock regulatory genes did not show any sex differences [NPAS2: F (1, 87) = 3.469, p = 0.0659; NFL3: F (1, 89) = 0.009085, p = 0.9243; NR1D1: F (1, 85) = 8.675, p = 0.0042; DEC1: F (1, 93) = 3.280, p = 0.0733; DEC2: F (1, 92) = 1.039, p = 0.3018; DBP: F (1, 93) = 5.681, p = 0.0192] (Supplementary Fig. 3B–F). These data suggest that AE effects may involve an alteration in the CLOCK/BMAL1 complex, which is regulated through epigenetic modifications via various activators and repressors at the CRY and PER promoters to influence their circadian expression (Fig. 2S).

Fig. 2
figure 2

Effects of prenatal alcohol exposure on circadian expression of an activator NPAS2 (A–C), and repressors (NFL3, D–F; NR1D1, G-I; DEC1, J–L; DEC2, M–O; DBP, P-R) of transcription/translation of core circadian control genes via its role in a negative feedback loop responsible for the control of circadian rhythm. (A,D,G,J,M,P) Showing changes in the rhythmic expression of these genes from each AE and control subjects throughout the day and plotted at the time of collections during ZT 3:00 to ZT 11:00 using spline curves. (B,E,H,K,N,Q) Showing changes in the expression of these clock genes when plotted as mean ± SEM at each hour during ZT 3:00 to ZT 11:00 (for n values see Figure legend 1). Data were analyzed using Two-way ANOVA and the interaction between the treatment and time were identified [NPAS2: F (8, 79) = 0.6095, P = 0.7673; NFIL3: F (8, 79) = 2.063, P = 0.0494; NR1D1: F (8, 79) = 2.795, p = 0.0089; DEC1: F (8, 79) = 2.197, p = 0.0363; DEC2: F (8, 79) = 1.815, p < 0.0864; DBP: F (8, 79) = 5.500, p = 0.0001]. (C,F,I,L,O,R) Overall Mean ± SEM levels of gene expression in two groups were analyzed using one-way ANOVA. ****p < 0.0001. (S) Cosinor analysis was used to determine the rhythmicity (Zero-amplitude test p value < 0.05 identified rhythmicity and the acrophase value shows peak time of the rhythm) of these clock genes. (T) Schematic diagram showing how prenatal alcohol exposure (AE) alters circadian oscillation by modifying expression of core clock genes in children with FASD. The daily circadian oscillation is maintained by a coordinated transcription-translation feedback loop between the core clock activators (BMAL and CLOCK or BMAL and NAPS2) and core clock repressors (PER and CRY). The other negative regulators of the molecular clockwork, Dec proteins, are also shown to inhibit CLOCK/BMAL1-mediated transcription of E-box-containing promoters. Other feedback loops also participate in maintaining circadian rhythmicity consist of NR1D and ROR proteins, which compete to bind RORE to alter the expression of various circadian clock genes (CCGs), including BMAL. Additional regulators controlling this circuitry include activator DBP and the repressor NFIL3 through the D-box element. Prenatal alcohol exposure affects (stimulates and inhibits} various sites of this circadian feedback loop to disrupt the daily circadian oscillation.

Full size image

Prenatal alcohol exposure is known to impact the body’s epigenome to modulate the global transcriptome39,40. There are reports that epigenetic modulators interact with the CLOCK:BMAL1 complex and impose an oscillatory pattern to the epigenetic mark at the promoter of PER/CRY gene to maintain circadian rhythmicity15. Expression patterns of MLL1 and P300 genes, which are known to interact with CLOCK:BMAL1 and are recruited in the CCG promoters in a circadian fashion41, are rhythmic. This isrevealed by the finding that in control subjects, expression patterns (as determined by spline graph of individual data at various time points and the line graph of mean + SEM values of each group at various time points) of MLL1 (Fig. 3A,B) and P300 (Fig. 3D,E) fluctuated throughout the day. The MLL1 gene expression pattern showed circadian rhythm (Cosinor p; MLL1, 0.0482; P300, 0.0003) and had estimated acrophase times in late night (MLL1, ZT 19.895; P300, ZT 19.830) in control subjects, like BMAL1 and CLOCK (Fig. 2). In AE subjects, the expression pattern of these epigenetic modifiers was arhythmic (Cosinor p; MLL1, 0.362; P300, 0.524), and estimated acrophase times were advanced (MLL1,ZT10.662; P300, ZT20.149) (Fig. 3B,E,S). The mean levels of MLL1 gene and P300 gene at various time points (Fig. 3B,E; supplementary Fig. 2) or overall mean levels were lower (p < 0.05) in AE group than those in the control group (Fig. 3C,F). Expression patterns of other epigenetic modulators (SIRT1, EZH2, HDAC3, JAR1D1) known to favor a repressive state of chromatin and be involved in repression of PER/CRY gene transcription42,43 were also measured. In control subjects, expression patterns of these repressors (Fig. 3G,J,M,P) were rhythmic (Cosinor p; SIRT1, 0001; EZH2, 0.0002; HDAC3,0.0001; JAR1D1, 0.0001) and had estimated acrophase times in late afternoon (SIRT1, ZT10.6; EZH2, ZT9.819; HDAC3, ZT10.398; JAR1D1, ZT10.126), whereas in AE subjects, expression patterns these genes were arhythmic (Cosinor p; = SIRT1, 0.899; EZH2, 0.264; HDAC3, 0.276; JAR1D1, 0.665), and their estimated acrophase times were altered (SIRT1, ZT21.561; EZH2, ZT5.113; HDAC3, ZT21.266; JAR1D1, ZT5.788). The mean levels of all CLOCK regulatory epigenetic genes at various time points (Fig. 3H,K,N,Q; supplementary Table 2) and the overall mean levels of these genes were lower (p < 0.0001) in AE group than those in the control group (Fig. 3I,L,O,R). We did not find any sex differences in the response to AE in these epigenetic genes [MLL1: F (1, 93) = 0.1136, p = 0.736; NFL3: F (1, 89) = 0.009085, p = 0.9243; P300: F (1, 93) = 0.02040, p = 0.5876; SRT1: F (1, 93) = 0.4435, p = 0.5071); EZH2: F (1, 93) = 0.2548, p = 0.614; HDAC3: F (1, 93) = 0.4185, p = 0.5193; JAR1D1: F (1, 93) = 1.513, p = 0.2218] (Supplementary Fig. 4A–F). These data suggest that AE may also mediate alterations in the CLOCK/BMAL1 complex-regulated epigenetic machinery which involves activators and repressors at the CRY and PER promoters to influence their circadian rhythmic expression (Fig. 3S).

Fig. 3
figure 3

Effects of prenatal alcohol exposure on circadian expression of epigenetic modulators (transcription activators-MLL1, A–C; P300, D–F; transcription inhibitor-SIRT1, G–I; EZH2, J–L; HDAC3, M–O; JAR1D1, P–R) known to interact with CLOCK:BMAL1 and impose oscillatory patterns to the epigenetic mark at the promoter of PER/CRY gene to maintain circadian rhythmicity. (A,D,G,J,M,P) Showing changes in the rhythmic expression of epigenetic modulators of core clock genes in salivary samples collected at two time points (morning and afternoon) from each AE and control subjects throughout the day and plotted at the time of collections during ZT 3:00 to ZT 11:00. (B,E,H,K,N,Q) Showing changes in the expression of these clock genes when plotted as mean ± SEM at each hour during ZT 3:00 to ZT 11:00 (for n values see Figure legend 1). Data were analyzed using Two-way ANOVA and the interaction between the treatment and time were identified [MLL1: F (8, 83) = 5.711, P = 0.0001; P300: F (8, 83) = 3.316, P = 0.0025; SIRT1: F (8, 83) = 4.739, p = 0.0001; EZH2: F (8, 83) = 4.211, p = 0.0003; HDAC3: F (8, 83) = 4.398, p < 0.0002; JAR1D1: F (8, 83) = 4.525, p = 0.0001]. (C,F,I,L,O,R) Overall Mean ± SEM levels of gene expression in two treatment groups were analyzed using one-way ANOVA. *p < 0.5; ***p < 0.001; ****p < 0.0001. (S) Cosinor analysis was used to determine the rhythmicity (Zero-amplitude test p value < 0.05 identified rhythmicity and the acrophase value shows peak time of the rhythm) of these clock genes. (T) Schematic diagram showing postulated mechanisms by which alcohol exposure (AE) affects epigenetic modifications and alters circadian clock gene (CCG) transcriptions during the circadian cycle in children with FASD. The effects may involve alterations in the Clock/Bmal1 complex-regulated epigenetic modifications by various activators and repressors at the Cry and Per (other clock regulatory genes) promoters to influence their circadian expression. Down arrow indicates inhibition.

Full size image

Discussion

Data presented here are consistent with the previous finding that CLOCK genes are expressed in a circadian manner in oral mucosa in healthy diurnally active volunteers44. Peak time of expression for PER genes and BMAL1 in our study follows what might be predicted by the animal literature and human postmortem brain tissues. PER genes peaks during early afternoon around ZT 07:00–08:00 while BMAL1 expression peaks in late night around ZT 18:00–19:0023,45,46,47. The peak times for PER1 and CRY1 by the cosinor curve fit are similar to those found in human oral mucosa44, and in the SCN and the peripheral tissues in rodents48,49. Furthermore, the ARRB1 gene, known to be a candidate biomarker of peripheral circadian rhythm10, also showed rhythmic expression in saliva in the control subjects. Our study also revealed a loss of rhythmicity in all core clock gene transcripts, as well as ARRB1, in saliva samples of subjects with AE, suggesting that alcohol exposure during the prenatal period adversely affects the circadian expression of core CLOCK genes.

This study also identified significant changes in overall transcript levels of core CLOCK genes in subjects with symptoms of FASD. The overall mean levels of PERs and CRYs were lower, whereas BMAL1 and CLOCK genes were higher in the AE group, suggesting that AE may act on the circadian positive feedback loop involving BMAL1 and CLOCK heterodimers effects on transcriptions of PER and CRY genes. Thus, AE affects the daily circadian oscillation maintained by a coordinated transcription-translation feedback loop between the core clock activators (BMAL and CLOCK) and core clock repressors (PER and CRY) (Fig. 1Y).

Alcohol exposures during the prenatal period also affects the expression patterns of various regulators of core CLOCK genes. In the core negative feedback loop, NAPS2 acts as an activator while NFL3, NR1D1, DEC1, DEC2, and DBP serve as repressors of transcription/translation of core circadian control genes and its circadian rhythm. These proteins are also shown to inhibit CLOCK-BMAL1-mediated transcription of E-box-containing promoters. Additional regulators controlling this circuitry include activator DBP and the repressor NFIL3 through the D-box element. Prenatal alcohol exposure affects both the circadian patterns and the expression levels of these repressor genes (Fig. 2S). These data suggest that AE may act on multiple components of the circadian feedback loop to impair typical circadian oscillation.

This study also determined the role of epigenetic modulation of circadian genes in FASD. It is known that the CLOCK:BMAL1 heterodimer activates the transcription of several CCGs through their binding to the E-Box elements in their promoters and the recruitment of coactivating complexes such as P300 that acetylates the lysine residues of the histones allowing euchromatin formation41. The MLL1 methyltransferase responsible for methylation in H3K4 (mark related to transcriptional activation) is associated with CLOCK and is recruited in the CCG promoters in a circadian manner to promote transcription41. In addition to the repression mediated by PER, CRY, histone deacetylation mediated by histone deacetylases like SIRT1 and HDAC3 play a role to favor a repressive state of chromatin. Thus, AE effects may involve alteration in the CLOCKC/BMAL-1 complex-regulated epigenetic modifications by various activators and repressors at the CRY and PER (and other clock regulatory genes) promoters to influence their circadian expression. Additional research is necessary for understanding how AE affects the core clock genes and the regulators of core clock genes.

One limitation of this study is that the data presented here illustrate dysregulation of clock genes when examining expression during the light cycle. However, to fully understand the impact of AE on circadian expression of clock genes, future studies need to be conducted to examine across the 24-h period. Moreover, to determine the relationship between circadian dysregulation and behavioral outcome in children with FASD, parent-reported sleep questionnaires and measures of behavior and cognition need be compared to circadian regulation. Importantly, individuals with AE in this study have reported disruptions in sleep and increased night-to-night sleep variability4, consistent with alterations in circadian control. However, correlations between such outcomes and circadian dysregulation will require measurements of individual circadian control over the 24-h period.

In summary we found that the daily circadian oscillation which is maintained by a coordinated transcription-translation feedback loop between the core clock activators (BMAL and CLOCK) and core clock repressors (PER and CRY) is affected by AE. The daily circadian oscillation is maintained by a coordinated transcription-translation feedback loop between the core clock activators (BMAL and CLOCK or BMAL and NAPS2) and core clock repressors (PER and CRY). The other negative regulators of the molecular clockwork, DEC proteins, are also shown to inhibit CLOCK-BMAL1-mediated transcription of E-box-containing promoters. Also, NR1D and ROR proteins which compete to bind RORE to alter the expression of various circadian clock genes (CCGs), including BMAL. Additional regulators controlling this circuitry include activator DBP and HLF and the repressor NFIL3 through the D-box element. Prenatal alcohol exposure affects various sites of this circadian feedback loop to disrupt the daily circadian oscillation. The effects may involve alterations in CLOCK/BMAL1 complex regulated epigenetic modifications by various activators and repressors at the Cry and Per (other clock regulatory genes) promoters to influence their circadian expression. Further studies are needed to clearly define the mechanism by which AE disrupt circadian mechanisms.

Importantly, there is a strong association between disruption of circadian rhythms and psychiatric illnesses including sleep disorders, anxiety behavior and hyperreactive stress responses often observed in children with FASD27,50,51,52,53,54. However, the relationship between the pathophysiology of FASD and circadian rhythm disruption remains poorly defined and needs future investigation. Thus, circadian clock dysfunction may represent both a marker for pathology and a crucial opportunity for early intervention in FASD. Hence, restoration of circadian rhythms in preclinical FASD may provide an opportunity for early intervention to alter the developmental trajectory and quality of life.