Tiny Blood MicroRNA Shows Inverse Link to Oxytocin in Men With Hypersexual Disorder

Deep in the human genome, more than 2,654 mature microRNAs act as molecular volume dials, tuning the output of genes by binding to the 3′-untranslated regions of messenger RNA and triggering its degradation or blocking its translation. These small non-coding RNAs regulate more than half of all protein-coding genes, and a single microRNA can govern several hundred different targets, making them uniquely powerful coordinators of cellular pathways. Now, a team of Swedish researchers has reported the first direct human examination of whether one specific microRNA, hsa-miR-4456, covaries with the neuroendocrine hormone oxytocin when both are measured in the same individuals — a paired observation that connects previously separate epigenetic, expression, pathway-prediction, and hormonal findings in the study of compulsive sexual behavior.

The new study, published in Frontiers in Neuroscience, was led by Adrian E. Desai Boström of Karolinska Institutet and Umeå University, together with colleagues including Andreas Chatzittofis, Peter Andersson, and senior author Jussi Jokinen. It is a secondary cross-sectional analysis of baseline, pretreatment measurements drawn from clinically characterized men with hypersexual disorder and healthy male controls, with samples collected in Stockholm during 2013–2014 under approved ethical protocols. Hypersexual disorder was proposed for DSM-5, while the overlapping construct of compulsive sexual behavior disorder was subsequently included in ICD-11 as an impulse-control disorder; the participants in this historical cohort were assessed using the proposed HD criteria.

The scientific backdrop is a sequence of convergent observations from an overlapping research program. Earlier hypothesis-free work in 60 men with hypersexual disorder and 33 healthy controls identified higher DNA methylation at a locus associated with hsa-miR-4456, and subsequent expression analysis indicated lower peripheral blood-derived miR-4456 in the disorder. In-silico target prediction then implicated neuronal pathways, including oxytocin signaling — providing biological motivation but not evidence of direct regulation. A subsequent study from the same program found higher baseline plasma oxytocin in men with hypersexual disorder, and that oxytocin decreased in a treated subgroup. What had never been examined was whether miR-4456 expression and oxytocin levels quantitatively covary within the same individuals.

To answer that question, the team measured miR-4456 by RT-qPCR in EDTA blood-derived material and plasma oxytocin by radioimmunoassay at the Clinical Neurochemistry Laboratory in Mölndal, Sweden. The submitted pooled robust regression included 74 paired measurements; revised group-adjusted and interaction models included 73 complete cases, comprising 51 men with hypersexual disorder and 22 controls. Because Shapiro–Wilk tests indicated departures from normality, miR-4456 expression was rank-based inverse-normal transformed using Blom’s method, while oxytocin remained on its original picomole-per-liter scale. Each regression coefficient therefore represents the difference in plasma oxytocin per one-unit increase in Blom-transformed miR-4456 expression.

The headline result was striking. In the submitted pooled robust model, higher miR-4456 was associated with lower plasma oxytocin, with a coefficient of −4.039 pmol/L (95% CI −6.149 to −1.930; p < 0.001). But when the researchers adjusted for diagnostic group, the estimate remained negative yet was substantially attenuated to −1.522 (95% CI −3.376 to 0.333; p = 0.106), with the confidence interval now spanning zero. The control-specific estimate was −0.613 (p = 0.696), the hypersexual-disorder estimate was −2.223 (p = 0.057), and the formal miR-4456-by-group interaction was not significant (p = 0.414). In other words, the marked between-group difference in plasma oxytocin accounts for part of the pooled relationship, and the data do not establish a group-independent association.

Post hoc sensitivity analyses added nuance without converting the finding into confirmation. An ordinary least-squares model with heteroskedasticity-consistent standard errors gave −0.881 (p = 0.493). A group-stratified non-parametric bootstrap of 10,000 resamples yielded a 95% percentile interval of −3.373 to 0.613, with 92.0% of resampled slopes below zero. The group-adjusted partial Spearman correlation was −0.201, with a stratified permutation p value of 0.089. Across 73 leave-one-out refittings of the group-adjusted robust model, every estimate remained negative, ranging from −1.856 to −1.260 — indicating that no single retained participant determined or reversed the direction of the estimate, though these overlapping samples should not be read as 73 independent replications.

The authors are explicit about what the study does and does not show. The cross-sectional peripheral measurements do not establish directionality, causation, or central nervous system regulation. The analysis was not preregistered, reused partially overlapping source cohorts, and included only men. Clinical covariates for the exact paired set were unavailable, leaving potential confounding by medication, symptoms, trauma, lifestyle, or metabolic factors. Four miR-4456 observations had been excluded in the source study using visual boxplot inspection and could not be restored from the filtered archive. The miRNA and oxytocin source studies documented different fasting conditions, and the retained records do not establish that both analytes came from the same blood draw or aliquot.

Preanalytical limitations further constrain interpretation. The source miRNA material was affected by erythrocyte hemolysis and included plasma and corpuscular-blood components, which is why the measurement is described as peripheral EDTA blood-derived rather than a validated cell-free plasma miRNA. Cohort-specific U6 stability, spike-in recovery, platelet contamination, and complete batch and freeze–thaw information were unavailable. Oxytocin assay variability — with intra- and interassay coefficients of variation ranging from 8% to 15% depending on concentration — and uncertain correspondence between peripheral and central concentrations further limit precision and mechanistic claims, a challenge long recognized in human oxytocin measurement research.

Why does this matter for the broader field? MicroRNAs are emerging as important regulators of normal and pathological behavior, with potential roles as master regulators of cellular processes and as candidate biomarkers in conditions such as major depressive disorder, where several circulating miRNAs have been proposed for diagnosis and treatment assignment. A recent animal study implicated septal miR-132 downstream of oxytocin receptor signaling in social fear extinction, hinting that microRNA–oxytocin crosstalk is biologically plausible. Circulating microRNAs may even act in a hormone-like fashion, secreted from living cells within lipid structures or protein complexes and capable of intercellular crosstalk that could influence adjacent or distant neuronal populations.

The paired observation defines a focused next step. Independent studies should obtain paired miR-4456 and oxytocin measurements under harmonized collection conditions, with validated plasma-miRNA normalization, hemolysis and platelet quality control, complete clinical covariates, and sufficient power for within-group and interaction estimates. Mechanistic studies can then test whether experimental manipulation of miR-4456 changes validated oxytocin-pathway targets, oxytocin synthesis or release, or related signaling in relevant cells and animal models. Such work is necessary before miR-4456 can be described as a regulator or therapeutic target — but the present paired measurement provides a specific, experimentally testable rationale for undertaking it, turning a scattered set of indirect findings into a single coherent hypothesis about how gene-regulatory molecules and neuroendocrine signaling may intersect in human behavior.

Subject of Research: Neuroscience

Article Title: Association between peripheral hsa-miR-4456 expression and plasma oxytocin in men with hypersexual disorder and healthy controls: a secondary cross-sectional analysis

Article References: Desai Boström, A. E., Chatzittofis, A., Andersson, P., Savard, J., Görts Öberg, K., Arver, S., Flanagan, J. N., & Jokinen, J. (2026). Association between peripheral hsa-miR-4456 expression and plasma oxytocin in men with hypersexual disorder and healthy controls: a secondary cross-sectional analysis. Frontiers in Neuroscience, 20, Article 1946302. https://doi.org/10.3389/fnins.2026.1946302

Image Credits: AI Generated

DOI: 10.3389/fnins.2026.1946302

Keywords: microRNA, miR-4456, oxytocin, hypersexual disorder, neuroendocrine science, biomarkers, epigenetics, Association, between, peripheral, hsa-miR-4456, expression