Cocaine use disorder effects on blood oxytocin levels and OXTR DNA methylation

Substance use disorders have been associated with alterations in the oxytocinergic system, but few studies have investigated both the peptide and epigenetic mechanisms potentially implicated in the regulation of oxytocin receptor. In this study, we compared plasma oxytocin and blood DNA methylation in the OXTR gene between people with and without cocaine use disorder (CUD). We measured the oxytocin levels of 51 people with CUD during acute abstinence and of 30 healthy controls using an enzyme immunoassay. The levels of DNA methylation in four CpG sites at exon III of the OXTR gene were evaluated in a subsample using pyrosequencing. The Addiction Severity Index was used to assess clinical characteristics. We found higher oxytocin levels in men with CUD (56.5 pg/mL; 95% CI: 48.2 – 64.7) than in control men (33.6 pg/mL; 95% CI: 20.7 – 46.5), while no differences between women with and without CUD were detected. With a moderate effect size, the interaction effect between group and sex remained significant when controlling for height, weight and age data. A positive correlation in the CUD sample was found between oxytocin levels and days of psychological suffering prior to treatment enrollment. No group differences were observed regarding DNA methylation data. This suggests that CUD is associated with higher peripheral oxytocin levels in men during acute abstinence. This finding may be considered in future studies that aim at using exogenous oxytocin as a potential treatment for cocaine addiction.


Introduction
Substance use disorder is characterized by the compulsion to seek and consume substances, loss of control when attempting to limit use, the emergence of negative emotional states when access to drugs is hindered, and a clinical presentation of chronic relapse [1][2][3].Despite the significant impact of substance use disorders on individuals and society, treatment approaches for this condition are still limited.Thus, further studies are needed to elucidate the biological hypotheses that aid in understanding the pathophysiology of these disorders and in developing new therapeutic targets.
There is a growing body of evidence that recognizes the oxytocinergic system as a promising new target for the treatment of addictive disorders [4].The nonapeptide oxytocin is a member of the family of peptides produced in the neurohypophysis [5], and it plays an important role in the encephalon [6].This peptide is thought to be implicated in an array of social effects [7].Oxytocin and its receptors are expressed widely throughout the body, in both the central and peripheral nervous systems, which reflects the diversity of endocrine and paracrine activities attributed to them so far [4].Based on extensive preclinical evidence from animal model studies, a series of clinical trials has tested the administration of exogenous oxytocin in humans, mostly via intranasal route.Despite its chemical instability and rapid pharmacokinetics, exogenous oxytocin achieves direct transport from the nasal passages to the brain via intranasal administration, crossing the blood-brain barrier and resulting in a rapid onset of therapeutic effects [8].The findings are still inconsistent, but some studies have reported a decrease in craving and depression symptoms during acute withdrawal compared to placebo in methamphetamine use disorder [9].In alcohol use disorder, some studies have observed an improved ability to recognize negative emotions, representing a decrease in social cognition impairments associated with this condition [10].A preliminary study has also pointed to increased chances of sustained abstinence in cocaine users following intranasal oxytocin administration [11].Several neurobiological hypotheses have emerged to clarify the potential benefits of exogenous oxytocin in the clinical management of addiction, with one of the main ones covering the interaction between the oxytocinergic and dopaminergic systems in the mesolimbic brain reward system [8].
The endogenous oxytocinergic system, however, has not received as much attention in human studies, especially concerning its interaction with psychoactive substances in both genders [12].Some evidence suggests that acute or chronic drug exposure may alter the signaling of the oxytocinergic system in human samples [13].For example, higher blood concentrations of oxytocin have been identified in patients with alcohol use disorder in the acute withdrawal period compared to healthy controls [14].Similarly, an association between peripheral oxytocin levels and the severity of cravings was observed in patients with heroin use disorder [15].In addition, preclinical studies in rodents have shown that chronic exposure to nicotine, methamphetamine, cocaine, and morphine commonly result in an upregulation of the expression of the OXTR gene in various brain regions [16].
The OXTR gene encodes oxytocin receptors, which are G proteincoupled receptors that act both centrally and peripherally.The OXTR gene comprises 17 kb and contains three introns and four exons.Exons I and II correspond to the 5′ non-coding region, while exons III and IV encode the amino acids of the receptor.Within the gene, a CpG island extends from the first to the third exon, and changes in DNA methylation levels of this CpG island are suggested to affect OXTR expression [17].However, there is still little evidence on whether chronic consumption of psychostimulants can alter such epigenetic regulation process.
Considering the still-inconsistent results of clinical trials on oxytocin as a treatment for substance use disorders, the inconsistent findings regarding oxytocin levels as potential biomarkers associated with addictive disorders, and the lack of studies investigating both peptide levels and DNA methylation in regulatory regions of the OXTR gene body, the present study sought to compare such outcomes between people with and without cocaine use disorder (CUD).Specifically, we compared plasma levels of oxytocin and levels of DNA methylation at CpG sites within the third exon of the OXTR gene.We also sought to investigate possible associations with clinically relevant features.

Participants
This is a case-control study carried out with a convenience sample.Participants with CUD (n = 51) were recruited from detoxification units for alcohol and drug abuse in Porto Alegre, Brazil.Participants included in the study met the following criteria: (1) aged between 18 and 45 years; (2) diagnosed with CUD based on the Structured Clinical Interview for the DSM-5 [SCID] [18]; physiologically dependent on snorted or smoked cocaine (crack) [19]; and (3) having tested positive for cocaine on a urine screening test during the first three days of treatment (indicating less than 1 week since last cocaine use prior to seeking treatment).All participants with CUD were in a controlled abstinence environment without access to drugs, including alcohol and cigarettes.The SCID-I was used to identify mental disorders and confirm the diagnosis of CUD [18].The substance use and abuse disorders module of the SCID-I was modified for this study.We considered both disorders (substance abuse and substance dependence) as a single disorder, which fits the DSM-5 criteria.
Healthy control participants (n = 30) were selected by convenience sampling.Eligibility of controls involved the absence of any past or current Axis I psychiatric disorders, no severe or unstable clinical illnesses, no neurological disorders, no use of any substances for at least 30 days prior to the study (self-report), and no positive results for drugs on a urine screening by the time of the assessment.The research protocols were previously reviewed and authorized by the Ethics Committee of the Pontifícia Universidade Católica do Rio Grande do Sul (PUCRS) (protocol n. 10/05214).

Demographic and clinical information
The Addiction Severity Index (ASI) published by McLellan et al. [20], is a semi-structured interview and clinical/research instrument to assess various dimensions of substance use behavior patterns.We collected data on years of regular use (at least three days per week over 12 months) of cannabis, tobacco, alcohol, and cocaine.In Brazil this instrument was validated by Kessler et al. [21].Based on a series of collected data, the ASI-6 also generates severity scores for psychiatric problems.The questions used to generate this score are: days in psychiatric treatment prior detoxification (last month, quantitative variable), days with psychological suffering prior treatment (last month, quantitative variable), days unable to do activities due psychological problems (last month, quantitative variable), rating of trouble/bothered by psychological problems (quantitative variable), rating of importance of treatment for psychological problems (quantitative variable), rating of need for psychological-trauma related treatment (quantitative variable), depressive symptoms (last month, binary variable), anxiety symptoms (last month, binary variable), impulsivity (last month, binary variable), suicide thoughts (last month, binary variable), hallucinations (last month, binary variable), aggression (last month, binary variable).Scores range from 30 to 80 points.In addition, during the interview we inquired questions related to sociodemographic (age and ethnicity).Height and weight data were also evaluated.

Oxytocin measurement
All blood samples were collected in BD Vacutainer® tubes with ethylenediaminetetraacetic acid (EDTA) and centrifuged at 2000g for 10 min at 4 • C for separation of the plasma fraction.The recovered plasma was aliquoted, added aprotinin as a protease inhibitor, and stored at − 80 • C until further use.Oxytocin extraction protocol was performed by adding an equal volume of 0.1% of trifluoroacetic acid in water (TFA-H20) to the sample.Samples were then centrifuged at 17000g for 15 min at 4 • C to clarify and save the supernatant.Supernatant were then added to columns of solid-phase extraction (C18 Sep-Pak®) were fitted into stopcocks and connected to a vacuum manifold (Vaccum Manifold system -Waters®).With the stopcocks opened and the vacuum extraction Manifold turned on, the sample was loaded onto the column at a flow rate of 1 ml/min.The vacuum pressure was gradually increased to maintain a constant flow rate throughout the loading process.After sample addition, the column was washed with acetonitrile and with 10-20 mL of 0.1% TFA-H20.Oxytocin was eluted slowly (gravity-fed) by applying 3 mL of a solution comprised of 95% acetonitrile/5% of 0.1% TFA-H2O into microtubes.The eluted solutions were evaporated under 12 PSI of nitrogen gas for 60 minutes and reconstituted with assay buffer, then stored at − 20 • C. Oxytocin levels were then measured using the commercially available via enzyme-linked immunosorbent assay (ELISA-Enzo Life Sciences®) with high sensitivity and low levels of cross reaction of oxytocin with vasopressin.All measurements were performed, according to the manufacturer's protocol, with plasma samples diluted at a ratio of 1:2.After analyzing the absorbance data and cross-referencing them with the standard curve data using a 4PL polynomial curve fit (all plates with R values above 0.99).

DNA methylation analysis
DNA methylation analysis was carried out in a subsample (CUD n = 31; Control n = 29).Mononuclear cell lysis and DNA isolation and purification were carried out using the Gentra Puregene® Blood kit (Qiagen, Hilden, Germany) according to the manufacturer's standard protocol.Genomic DNA was quantified using the NanoDrop® Lite spectrophotometer (Thermo Fisher Scientific, Wilmington, USA).We used two PyroMark CpG assays (OXTR1, cat.No. PM0001682, Qiagen) that allows for the analysis of the AGGCGGTA-TAGTAGGTCGGGTTCGTAGAAGCGGA sequence (Chr: 3p25, bp: 8809467-8809658).This sequence contains four CpG sites (proximal to the 3′ end of the exon).The PCR primer for OXTR1 was as follows: forward 5′-TTTAGGGATATGAGTAGTAGTAGTAGGTAGG-3′ and reverse 5′-ACCCCTCTCTTCTTCTTCTTCATAAAACACC-3′, which generated a 193 bp amplicon.Pyrosequencing was performed using PyroMark® Gold Q24 reagents (Qiagen) according to the manufacturer's instructions.Data were analyzed using the PyroMark® CpG software version 2.0.6 (Qiagen), which identifies the cytosine methylation (%) at each CpG unit within a given sequence.The PyroMark® CpG software generates a quality score for each CpG site throughout the sequencing run, which allowed us to exclude low-quality samples from the pyrosequencing analysis.

Statistical analysis
All variables were tested for normality through the Shapiro-Wilk distribution test.All outcome variables were found to have a normal distribution.Sociodemographic and clinical characteristics were compared between the CUD group and control group using independent t-test for continuous variables or, when categorical, the chi-square (χ2) test.Data regarding years of regular use of licit (alcohol and tobacco) and illicit (cocaine and marijuana) substances were adjusted for age, generating a ratio of years of regular substance use by age.Outcome variables were compared between groups through two-way ANOVA (group and sex), with Bonferroni post-hoc analysis for a paired comparison between the four conditions.Analysis of covariance (ANCOVA) were applied including potential clinical variables with group differences as covariates.Effect size was estimated through Partial Eta Squared (η p 2 ) for ANOVA and ANCOVA.Pearson correlations were performed to assess for possible associations between biological variables and ASI-6 variables.The significance threshold adopted for all analyses was p < 0.05.All statistical analyses were carried out using the Statistical Software R Core Team, version 4.0.4(Vienna, Austria) [22].

Sociodemographic and clinical characteristics of the groups
The clinical and sociodemographic characteristics of the groups are summarized in Table 1.Significant differences between the groups were evidenced with respect to age, height and weight.There were no significant differences between the groups regarding ethnicity.In relation to the ratio of years of regular use by age, we observed significant differences beyond the pattern of cocaine use, in particular because men with CUD had more years of alcohol and cannabis use compared to control men.With regard to the ASI-6, both men and women with CUD had a significantly higher psychiatric-problems score than controls.In particular, the rating of importance of treatment for psychological problems were higher in both men and women with CUD than controls, while the rating of need for psychological-trauma related treatment was higher in women with CUD than men with CUD.No significant differences were found regarding days reporting psychological suffering, days in psychiatric treatment, and days unable to do daily activities due to mental health problems, in the last month prior to detoxification enrollment.

Plasma oxytocin levels and methylation levels in the OXTR gene
We observed a significant group × sex interaction effect (F = 9.Control men and control women presented significant post-hoc differences (p = 0.025), with women showing higher levels of oxytocin.
Regarding OXTR methylation levels, we did not observe significant group, sex, or interaction effects at any of the four sites analyzed (all p values > 0.05, Fig. 1B).
Given that groups differed in weight, height, and age, we performed an ANCOVA model for oxytocin levels including these as covariates, with group and sex as fixed factors.The interaction effect between group and sex remained significant in this analysis (F = 9.1; p = 0.003; ηp2 = 0.11 -moderate effect size).

Correlations between biological and clinical variables
We assessed correlations (Table 2) between oxytocin levels, age, height, weight, ratio of years of regular substance use by age, ASI-6 score for psychiatric problems, and quantitative variables included in the ASI-6 score (variables in Table 1).This analysis was restricted to the CUD  sample.The only significant correlations found were between oxytocin levels and height (r = 0.324; p = 0.025), and between oxytocin and days reporting psychological and emotional suffering in the last month prior to treatment enrollment (r = 0.294; p = 0.045), indicating that higher oxytocin levels were associated with more psychological and emotional suffering prior to treatment.Including the whole sample (controls), we tested correlations between oxytocin with age, height and weight, and no significant associations were found.In addition, no correlation between oxytocin and OXTR methylation levels were found.

Discussion
In the present study, we compared plasma oxytocin levels and blood DNA methylation in regulatory regions of the OXTR gene body, specifically in its exon III region, between men and women with and without CUD.We observed an interaction effect of both group and sex with respect to oxytocin levels, whereas multiple comparison analyses indicated higher levels of the circulating peptide in men with CUD compared to healthy controls, but did not indicate differences between women with and without the diagnosis.In addition, we identified an association between oxytocin levels and psychological suffering during the last month before treatment enrollment.However, we observed no group or sex effects with respect to OXTR gene methylation.
The findings on increased peripheral oxytocin in men with substance use disorder compared to controls are similar to those found by Gerra et al. [23].However, the participants in Gerra's study were in withdrawal from heroin use.In a study with male patients in heroin withdrawal on a residential regimen, and using the same methodology (ELISA) as the present study, Nikolaou et al. [24] also found elevated oxytocin levels and higher levels of withdrawal stress associated with early discharge from treatment.In Nikolaou's study, patients with higher serum oxytocin levels had a worse prognosis at 24 months.Interestingly, another study also showed a correlation between higher serum oxytocin levels and severity of alcohol consumption in male patients [14].In addition, higher plasma oxytocin levels after consecutive administration of cocaine for five days were observed in baboons [25].In this sense, the increase in circulating oxytocin after excessive consumption of drugs of abuse such as cocaine may point not only to the effects of the substances themselves on the oxytocinergic system but also to the neuroendocrine stress response associated with withdrawal and acute abstinence.As Jong et al. [26] had already evidenced, different types of stress can produce increased release of circulating oxytocin.However, when it comes to the differences in CUD-related effects on plasma oxytocin between men and women observed in our study, more data is needed to identify associated hypotheses.It is known, however, that oxytocin secretion is influenced by steroid hormones such as estrogen [27], which could be a factor behind the sex differences observed in our analyses.Other studies in neuropsychiatry have already reported higher levels of oxytocin in healthy female controls when compared to healthy male controls [28], similar to what was observed in the our investigation.
We observed no group or sex effects with respect to OXTR methylation levels.In this sense, it seems that there are no associations between CUD and changes in methylation patterns of the sites evaluated in exon III of the OXTR gene, neither in men nor in women in our cohort.Other mechanisms of epigenetic regulation may be related to the differences found in the levels of the peptide itself.This lack of group effect is partially corroborated by data from an epigenome-wide association study (EWAS) which found no relationship between OXTR gene methylation and substance use disorders, including CUD [29].
The present study has some limitations.We conducted a case-control study with a small convenience sample.The CUD participants (cases) were on a controlled treatment regimen, in abstinence, and mostly using neuroleptics.However, prior investigations involving patients with schizophrenia who were prescribed daily neuroleptic doses did not show any significant correlations between factors such as chlorpromazine

Table 2
Correlation analyses in the CUD group.dosage and plasma oxytocin levels [30,31].Also, mixed findings exist regarding the correlation between peripheral oxytocin levels and oxytocin levels in the brain [32,33], suggesting caution when interpreting the plasma findings from our study in the context of the central nervous system.In addition, this was a cross-sectional study.Thus, the findings of this study raise questions that should be explored further in future longitudinal studies.

Conclusion
It is understood that the oxytocinergic system exhibits great plasticity in response to a variety of environmental and pharmacological stimuli.Our study showed that CUD may be associated with a considerable increase in plasma oxytocin levels, particularly in males.This finding is supported by the literature on other addictive disorders that point to higher peripheral concentrations of the peptide in males and to associations with clinically relevant variables for the treatment of these patients.This suggests that, in the future, peripheral oxytocin may be further investigated as a biomarker in addiction.This finding thus has implications for studies that are seeking to investigate intranasal administration of oxytocin as a potential therapeutic approach for addictive disorder.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.Oxytocin and OXTR DNA methylation data.Legend: A -Oxytocin levels; B -DNA methylation levels of the four CpG dinucleotides assessed; CUDcocaine use disorder; Significant group per sex interaction is depicted as *.Data presented as mean and standard deviation.
cocaine use disorder; ASI-6 -Addiction Severity Index; a lifetime use ratio (years of regular use divided by age); b days in psychiatric treatment prior detoxification; c days reporting psychological and emotional suffering prior to treatment; d days unable to do activities due psychological problems prior to treatment; e rating of trouble/bothered by psychological problems; f rating of importance of treatment for psychological problems; g rating of need for psychological-trauma related treatment; Significant correlations are highlighted with * for p-values < 0,05 and ** for p-values < 0,01.M.S. Souza et al.

Table 1
Clinical characteristics of the sample.