Due to collaboration of public Primary Schools in the Center of Italy, we recruited 63 families with 6–10 years old children. In accordance with the Declaration of Helsinki, the study was approved before its start by the Ethical Committee of the Department of Dynamic and Clinical Psychology at Sapienza, University of Rome (protocol number 27/2016). Informed consent procedures included orally informing the children (using age-adequate approaches) and illustrating to parents the aims and scope of the study and all procedures and measures; the parents gave their written informed consent for their child to participate in this study. In addition, we declare that collected biological materials were used solely to the purpose of this study.

After receiving the consent of the primary school headmaster, a group of psychologists, specifically trained for the purposes of the study, reached children and parents at school premises. Participants (both parents and children) were assessed through buccal swabs (Isohelix Swab Pack). Buccal cell sampling is a feasible, non-invasive method that yields reproducible results also in DNA methylation studies (49). Subjects were aware of not having to eat (including chewing gum, candy, etc.), drink (except water), smoke, and brush their teeth for at least 1 h before sampling. Epithelial cell samples were carefully collected through the buccal swabs. The biological samplings were transported, slightly chilled by normal ice (+4°C), to the laboratories of the co-author, Esterina Pascale, for further processing.

After buccal swabs were gathered, mothers and fathers filled out independently self-report and report form questionnaires (described below). The order of administration of these measures was randomly selected. The following tools were chosen because they are very widely used and proved to be able to capture a wide range of difficulties that can be experienced by adults and children in general population (50, 51).

Parents were administered the Symptom Check-List-90 item-Revised (SCL-90-R), a 90-item self-report questionnaire. It measures psychological symptoms and psychological distress in adult from general and clinical population (52). The SCL-90-R is rated on a Likert scale of 0 (not at all) to 4 (extremely) and asks participants to report if they have suffered in the past week from symptom such as: Headaches (Somatization scale), Trouble remembering things (Obsessive-Compulsivity scale), Feeling critical of others (Interpersonal Sensitivity scale), Blaming oneself for things (Depression scale), Feeling fearful (Anxiety scale), Feeling easily annoyed or irritated (Hostility), Feeling afraid to will faint in public (Phobic Anxiety), Feeling watched or talked about by others (Paranoid Ideation scale), and The idea that something is wrong with one’s mind (Psychoticism scale). Besides these nine primary scales, the questionnaire provides a Global Severity Index (GSI), used to determine severity and degree of psychological distress. The SCL-90-R showed good internal coherence (α = 0.72–0.96) in this study (Italian validated version, 51).

Parents also filled out the Italian version of the Child Behavior CheckList/6–18 (51, 53), which is one of the most widely used instruments to assess child and adolescent psychopathology both in epidemiological and clinical samples. The CBCL/6–18 is a 113-item informant-report questionnaire, which asks parents (independently) to rate specific emotional–behavioral problems of their child during the past 6 months. Items are rated on a 3-point Likert scale, ranging from “0” (not true) to “2” (very true or often true), and they are grouped into eight empirically based syndrome scales: anxious/depressed, withdrawn/depressed, somatic complaints, social problems, thought problems, attention problems, rule-breaking behavior, and aggressive behavior. These subscales are, in turn, combine in three broad-band scales: internalizing problems scale is comprised of items from the anxious/depressed, withdrawn-depressed, and somatic complaints scores; externalizing problems combines rule-breaking and aggressive behavior. There also is a total problems score, which is comprised the scores of all the problem items. For this study, the T scores of the internalizing and externalizing problems, and a total problems T score from both mother and father reports were used (53). Mean test–retest reliabilities of r = 0.88 have been reported for the school-age forms (53). In this study, statistical analyses were performed on raw scores. Furthermore, as suggested by several studies (54–56), data were obtained by mothers and fathers (independently). In fact, international literature has emphasized that parents may be discordant in the observation of their children.

Buccal cell DNA isolations were performed using the Buccal-Prep Plus DNA isolation kit (Isohelix) according to the manufacturer’s instructions. The yield of DNA is usually between 3 and 10 µg. The 3′-UTR repeated sequence of DAT was amplified by the polymerase chain reaction (PCR) as it has been described previously (38, 57).

DNAs from the buccal swabs were further processed for assessing amount of methylation in the DAT 5′-UTR sequence (notably, not the transcription promoter region). Amount of methylation was determined in six specific CpG residues [termed M1, M2, M3, M5, M6, and M7; see Figure 1, reproduced with permission of Springer, license n. 4198250303549, from Adriani et al. (38)]. Notably, M1–M3 represent a CGGCGGCGG motif, while M5/M6 represents a CGCG motif. The following primers (5′–3′) were used to amplify the gene for DAT: Fwd, AGCTACCATGCCCTA TGTGG; Rev, ATCAGCACTCCAAACCCAAC. Bisulfite-treated DNA was amplified by PyroMark PCR Kit (Qiagen, Hilden, Germany) in accordance with the manufacturer’s protocol. PCR conditions were as follows: 95°C for 15 min, followed by 45 cycles of 94°C for 30 s, 56°C for 30 s, 72°C for 30 s, and, finally, 72°C for 10 min. PCR products were verified by agarose electrophoresis. Pyrosequencing methylation analysis was conducted using the PyroMark Q24 (Qiagen, Hilden, Germany). The level of methylation was analyzed using the PyroMark Q24 Software (Qiagen, Hilden, Germany), which calculates the methylation percentage [mC/(mC + C)] for each CpG site, allowing quantitative comparisons (mC is methylated cytosine and C is unmethylated cytosine).

A preliminary screening of data showed few data missing for each psychometric instrument (3% for each instrument). Missing data were corrected using multiple imputation in SPSS software (version 23.0). To verify the possible differences in the state of methylation in subjects with 10-repeat allele polymorphism, preliminary descriptive analyses and a T-test for independent samples were carried out. Furthermore, bivariate Pearson correlations were conducted to verify aims 1 and 2 and regression analysis were conducted to verify aims 3 and 4 (see above in this manuscript). Also, we tested for the impact of confounding variables, such as smoking, alcohol use, concurrent medical illness, and traumatic experiences. All presented results were adjusted for these confounding variables.

For the aims of this study, the following cases were excluded from the sample: families with children with mental and/or physical disability (N = 4); families in which one or both parents could not understand the Italian language (N = 9); families who did not complete all the psychometric tools (N = 7); families in which one or more members were following a psychoactive pharmacological or psychological treatment (N = 10); families in which parents were not the biological parents of the child (N = 3); and families who refused to participate in the study (N = 9). Participants included 21 children (12 females and 9 males with age ranging from 6 to 10 years; average age 7.43 years and SD, 1.53), their mothers (average age 42.33 years and SD 5.21), and fathers (average age 44.91 years and SD 5.26). Most of the families recruited for the study (85.7%) had a middle socioeconomic status (59), and a large majority (95.2%) comprised intact family groups. Ninety percent of the families were Caucasian, and 80.9% relied on more than one income. Confounding variables (such as alcohol use, smoking, drugs of abuse, current medical illness, traumatic experiences, and social-economic status) were assessed through an anamnestic questionnaire, specifically created for this study.

Figure 2 shows the distribution of dopamine transporter gene among mothers, fathers and children.

Furthermore, Table 1 shows average scores and SDs of mothers and fathers’ psychopathological risk of parents, according to the subscales of the SCL-90-R, and the scores of the children’s emotional–behavioral functioning at questionnaires filled out by mothers and fathers.

Preliminarily, we wanted to verify possible differences in the state of methylation of DAT1 based on the presence of the 10-repeat allele on DAT1 gene; thus, an independent samples T test has been conducted. Independent variables were a two-level genotype groups, based on the number of 10-repeat allele copies (group 9/10, with one allele “10” and group 10/10, with two alleles “10”), while dependent variables were the levels of methylation of DAT1 5′-UTR at all considered sites [M1, M2, M3, M5, M6, and M7; see Ref. (38), for the rationale of this choice]. Results showed that there were no differences between groups on levels of methylation, neither in mothers (at all CpG sites p > 0.05), fathers (at all CpG sites p > 0.05), and child (at all CpG sites p > 0.05).

Bivariate Pearson correlation analyses were performed separately for mothers, fathers and children. In mothers and fathers, we verified the possible presence of correlations between the methylation levels at the six-selected CpG sites (of DAT1 5′-UTR) and psychopathological symptoms (correlations run between all SCL-90-R subscales and at all loci of DAT methylation, for mothers and fathers). Tables 2 and 3 shows significant correlations.

For children, we verified the possible presence of correlations between the methylation levels at the six selected CpG sites (of DAT1 5′-UTR) and their own (parent-reported) emotional–behavioral functioning (correlations run between all CBCL/6–18 subscales, as reported by mothers and fathers, and at all children’s loci of DAT methylation). Table 4 shows significant correlations.

We carried out bivariate Pearson correlations to verify the presence of correlations between parents and children’s DAT methylation (at all considered CpG sites). Results showed no significant correlations between mothers and children’s DAT methylation. Instead, significant correlations were found between fathers’ methylation at CpG M1 and children’s methylation at CpG M3 (r = 0.45; p < 0.05; Figure 2), fathers’ methylation at M5 and children’s methylation at M3 (r = 0.53; p < 0.05; Figure 3), and fathers’ methylation at M1 and children’s methylation at M6 (r = 0.56; p < 0.01). Figure 3 shows scatter plots of correlations between fathers’ methylation vs. children’s methylation at different CpG positions.

Linear regression analysis has been performed using mothers and fathers’ SCL-90-R GSI as predictors, and children’s methylation of DAT (in all CpG sites) as dependent variables. Results showed that mothers’ GSI predicted children’s levels of methylation particularly at CpG M2 (R² = 0.33; β = 0.57; t = 3.07; p < 0.01) and at CpG M6 (R² = 0.25; β = 0.5; t = 2.53; p < 0.01), whereas its predictive power on methylation at CpG M3 was weaker (R² = 0.21; β = 0.46; t = 2.26; p < 0.05). Furthermore, fathers’ GSI predicted children’s levels of methylation particularly at CpG sites: M2 (R² = 0.45; β = 0.67; t = 3.98; p < 0.01), M5 (R² = 0.43; β = 0.66; t = 3.82; p < 0.01), and M6 (R² = 0.39; β = 0.62; t = 3.46; p < 0.01), whereas its predictive power on methylation at M1 (R² = 0.23; β = 0.48; t = 2.37; p < 0.05), M3 (R² = 0.3; β = 0.54; t = 2.82; p < 0.05), and M7 (R² = 0.25; β = 0.5; t = 2.5; p < 0.05) was weaker. Figure 4 shows the relationships between the investigated variables, with the relative regression indices and arrows in bold indicating stronger predictive power.

As Figure 4 shows, mothers and fathers’ psychopathological risk predicted children’s methylation of DAT at specific CpG sites. As seen in the previous analysis (see Table 4), methylation of DAT at CpG M2 and M6 loci were correlated with some of the subscales of CBCL 6–18 (somatic complains, internalizing problems, attention problems, withdraw), concerning the children themselves.

Therefore, we verified whether the children’s scores on these sub-dimensions were predicted by children’s methylation.

At this aim, we carried out linear regression analysis using methylation of DAT as predictors, while the subscales of CBCL/6–18 questionnaires compiled by mothers and fathers were used as dependent variables. Children’s methylation of DAT at CpG M2 predicted scores of somatic complaint (R² = 0.2; β = 0.45; t = 2.18; p < 0.05) and internalizing problems reported by mothers (R² = 0.22; β = 0.47; t = 2.31; p < 0.05), and attention problems reported by fathers (R² = 0.26; β = 0.51; t = 2.6; p < 0.05). Furthermore, children’s methylation of DAT at CpG M6 predicted scores of withdraw reported by mothers (R² = 0.22; β = 0.46; t = 2.29; p < 0.05). Figure 5 shows the relationships between the investigated variables, with the relative regression indices and arrows in bold indicating stronger predictive power.

This study has some limitations. First, parental psychopathology and children’s emotional–behavioral functioning were assessed, respectively, through self-report and report form measures. Further studies should evaluate these variables through clinical observations and/or clinical interviews, to obtain more robust data on psychological profiles of the subjects (76, 77). Nonetheless, one merit of the present study was to consider the point of view of both mothers and fathers on the psychological functioning of their child, maintaining the differences that have been demonstrated in previous literature (54–56) and that concur to a more accurate representation of offspring mental health.

Second, our sample was relatively small. It must be stressed, however, that no other study has, to our knowledge, focused on parents and children in this developmental stage (school-age) by assessing both genotype and methylation, with a comparable attention to psychopathological symptoms. Although rarely performed, the study of psychopathological symptoms in community samples is supported by emerging literature underlining that up to a fifth of normative populations may show psychiatric problems (78). We acknowledge the fact that regression analyses suggesting a causal link between variables in cross-sectional studies cannot be considered conclusive in their results, due to limited effectiveness of this methodology outside longitudinal research. Nevertheless, some authors (79) have suggested the usefulness of cross-sectional studies even for assessing causal links between variables, and other authors have suggested that results from relatively small samples are still informative for the programming of assessment protocols and intervention plans (80), paving the way to more advanced investigations.

Our preliminary results suggest that epigenetic processes (such as DNA methylation) dynamically regulate our genomes responding to early environmental experiences (such as exposure to parental psychopathological symptoms in the first years of life), but they do so not just during the early stages of development but even in school-age children (81). Furthermore, we found that DNA methylation correlated with sub-clinical psychopathological symptoms, while previous literature focused only on psychiatric conditions, such as Mood disorders (82), Post-traumatic stress disorder (83), and Schizophrenia (84–86). Besides, this study may have important clinical implication for the prevention and treatment of emotional–behavioral difficulties in children, as it adds to previous knowledge about the role of genetic and environmental factors in predicting psychopathological symptoms within non-clinical populations. This approach can be useful for the identification of early indicators of maladaptive psychological functioning, observed in children well before they crystalize in adolescence and adulthood. Moreover, such studies (that assess individual genetic, biological, and psychological characteristics) can be essential for understanding the differential proneness of vulnerable versus resilient individuals for the onset of symptoms, secondary to inadequate environments and/or relationships (in this case, parental psychopathological symptoms). In fact, it has been demonstrated that children and parents with altered phenotype in dopaminergic system may be less emotionally attuned, due to a reduced sensitivity to reward and reinforcement (87).