Twin-twin transfusion syndrome (TTTS) complicates up to fifteen percent of monochorionic twin pregnancies and carries a high mortality rate when untreated (WAPM consensus group on Twin-to-twin transfusion syndrome et al., 2011). Imbalances in intertwin blood flow from placental arteriovenous anastomoses result in hypoperfusion of the donor twin and hyperperfusion of the recipient twin, with the latter manifesting polyhydramnios, cardiomegaly, and hydrops fetalis (Wohlmuth et al., 2016). Selective fetoscopic laser photocoagulation (SFLP) of the vascular anastomoses aims to modify the natural course of the disease prior to the onset of irreversible damage, however, early detection of twins that will benefit from SFLP remains a significant challenge (Manning and Archer, 2016; Behrendt and Galan, 2016; Crombleholme et al., 2007). Current staging systems of TTTS, including the Quintero, Cincinnati, and Children’s Hospital of Philadelphia systems, rely on fetal ultrasound and echocardiography (ECHO) measurements as the primary metrics signifying disease progression and informing the decision to proceed with SFLP (Quintero et al., 1999; Harkness and Crombleholme, 2005; Rychik et al., 2007). Shortcomings of the current staging systems include complete dependence on ultrasound which is operator dependent and can be affected by fetal positioning, maternal habitus or timing of exam. These factors lead to an inherit limitation to the staging system reflected in the heterogenous outcomes after SFLP (Stirnemann et al., 2010; Espinoza et al., 2021; Hessami et al., 2021; Rossi and D'Addario, 2009).

The incorporation of objective biochemical and genetic tests has revolutionized the diagnosis of fetal genetic and anatomic anomalies (Kim et al., 2023; Allyse et al., 2015) and may serve to complement existing ultrasound-based staging systems for TTTS by introducing operator and time independent mechanisms to evaluate TTTS. Prior studies have identified differences in the amniotic fluid biochemical milieu of patients with TTTS, however, the invasive nature of amniocentesis has precluded their use in clinical practice (Coleman et al., 2015; Dunn et al., 2016; Hoffman et al., 2020; Takano et al., 2001; Van Mieghem et al., 2010). Non-invasive biomarkers for TTTS, such as those obtained via maternal venipuncture, are comparatively rare, with prior studies only demonstrating differences in angiogenic biomarkers in maternal plasma of TTTS patients (Fox et al., 2010; Yinon et al., 2014). The paucity of identified maternal plasma biomarkers for TTTS stems conceivably from the wide range of maternal factors that influence plasma-level expression and differences in transplacental trafficking of certain types of biomolecules (Manokhina et al., 2017).

microRNAs (miRNAs) are short, 20–22 nucleotide non-coding ribonucleic acids (RNA) involved in posttranscriptional messenger RNA regulation. They have been demonstrated to transfer from the placenta to maternal plasma in both physiologic and pathophysiologic states, suggesting their potential role as non-invasive biomarkers for placental diseases such as TTTS (Luo et al., 2009; Higashijima et al., 2013). Though differences in miRNA expression exist within the amniotic fluid of TTTS patients (Schuchardt et al., 2022; Willner et al., 2021), no differences in miRNA expression were observed in maternal plasma in a recent, albeit underpowered, study (Mackie et al., 2019).

In this study, we aimed to compare the expression of a targeted panel of miRNAs in the maternal plasma of patients with TTTS and normal monochorionic diamniotic (MCDA) twin pregnancies. Given prior evidence for transplacental transfer of miRNAs and established differences present in amniotic fluid, we hypothesized that expression of maternal plasma miRNAs would differ between TTTS and normal MCDA twin pregnancies and could serve in a predictive capacity for the identification of TTTS via non-invasive sampling.

In this study, we performed an exploration of potential non-invasive biomarkers for twin-twin transfusion syndrome, given the heterogeneity of ultrasound-based staging syndromes for prognostication. The findings of this study suggest that differential expression of circulating miRNA is present in the maternal plasma of patients with TTTS compared to uncomplicated MCDA twin pregnancies, and furthermore, that specific miRNAs in combination demonstrate strong predictive capacity to distinguish TTTS from uncomplicated MCDA twin pregnancies. Compared to amniotic fluid, the identification of circulating biomarkers for TTTS has been extremely limited, with prior studies only showing differences in circulating angiogenic factors in TTTS patients (Fox et al., 2010; Yinon et al., 2014). A prior discovery study examining circulating miRNA profiles in TTTS did not identify any differences (Mackie et al., 2019). However, the precedent for circulating miRNAs to serve as biomarkers for TTTS has remained given their established relevance for other diseases of the fetus and placenta, suggesting the potential for transplacental transfer of miRNAs into maternal circulation (Higashijima et al., 2013; Mackie et al., 2019; Subramanian et al., 2023; Gu et al., 2019). In the present study, we identified four miRNAs, miR-26a-5p, miR-320a-3p, miR-222-3p, and miR-145-5p, that are differentially expressed in patients with TTTS compared to MCDA controls.

The identification of circulating miRNAs capable of differentiating TTTS from controls is significant and reveals a potential biomarker source with translational relevance. In addition, we found the top differentially expressed miRNAs, miR-26a-5p, miR-320a-3p, and miR-222-3p, are highly predictive of the presence of TTTS in combination. While the clinical utility of biomarkers in TTTS would be enhanced by their ability to discriminate based on disease severity, we did not find differences in expression according to several metrics of severity: cardiomyopathy, fetal demise, Quintero stage or Cincinnati stage. However, it is likely that this study was underpowered to detect true differences between subgroups as smaller patient subsets were assessed. An interesting future study would be evaluation of plasma miRNA expression in those patients with initially normal or equivocal ultrasound-based signs who progressed to TTTS on subsequent ultrasounds. First-trimester ultrasound screening to predict future development of TTTS utilizing features such as fetal nuchal translucency thickness and crown rump length discordance had significant false positive and false negative rates in prior studies (Mackie et al., 2017; Kagan et al., 2007; El Kateb et al., 2007; Sebire et al., 2000). Therefore, these patients could benefit from the identification of biochemical differences early in gestation that may predict TTTS development, aiding in risk stratification and/or earlier treatment. Importantly, we evaluated whether miRNA expression changes in response to treatment, and in paired patient samples of maternal plasma prior to and after SFLP, the expression of miR-26a-5p, miR-320a-3p, miR-222-3p and miR-145-5p corrected to levels closer to that of controls after treatment. Though outside of the scope of the present study, it is of considerable interest to assess whether plasma miRNA normalization after SFLP may predict post-procedure prognosis, as this study suggests that treatment of TTTS physiology may induce changes at the level of miRNA expression.

The miRNAs included in the panel were selected due to their relevance in other fetal and placental disease states. All four miRNAs that were significantly dysregulated are known to be dysregulated in other gestational diseases including preeclampsia; specifically, miR-320a-3p and miR-222-3p are dysregulated in trophoblast cells of patients with preeclampsia (Han et al., 2017; Xie et al., 2019; Yang et al., 2022). miR-26a-5p, which regulates uterine epithelial remodeling and immune signaling during implantation, is also dysregulated in pregnancy complications such as placenta previa and IUGR, and modulates inflammation and oxidative stress via NF-κB and cytokine pathways in other disease contexts, supporting its potential role as a pregnancy-relevant immune regulator (Hromadnikova et al., 2015; Myszczynski et al., 2025; Hromadnikova et al., 2024; Bian et al., 2024). The source of the circulating miRNAs is not known, and the differential expression could represent changes at the fetal, placental, or maternal level given that bidirectional miRNA trafficking occurs at the maternal-fetal interface (Xu et al., 2021; Morales-Prieto et al., 2020; Chang et al., 2017). In this study, we aimed to understand the potential interrelatedness of these differentially expressed miRNAs that could guide further research of downstream mediators of disease. To this aim, gene-miRNA network analysis revealed 18 genes that were simultaneously regulated by two miRNAs and three genes regulated by three miRNAs. Each of these three genes, KRT8, NIN, and FKBP3 are expressed in the human placenta, though their potential relevance to TTTS pathophysiology has not yet been elucidated (Gerli et al., 2024). Interestingly, a meta-analysis of single-cell RNA sequencing studies within placental samples identified KRT8 as the sole gene that is universally expressed within all trophoblast cell subtypes (Derisoud et al., 2024). Furthermore, KRT8 knockout mouse models produced embryonic lethality due to impairment in trophoblast giant cell layer formation (Jaquemar et al., 2003). The present study raises new questions of whether differences in miRNA expression could affect regulation of structural genes such as KRT8 implicated in placental development, an interesting avenue for further study. Additionally, it raises the question of whether in utero miRNA delivery could be a potential future treatment strategy aimed at placental modulation, as has been demonstrated in other fetal pathologies such as congenital diaphragmatic hernia (Ullrich et al., 2023; Khoshgoo et al., 2019). Interestingly, gene set enrichment analysis revealed the overexpression of target genes within several pathways of established and potential significance in TTTS pathophysiology including progesterone regulation, carbohydrate catabolism, and non-membrane spanning protein tyrosine kinase activity (Dunn et al., 2016; Hoffman et al., 2020; Parchem et al., 2023; Kajiwara et al., 2022). The mechanisms governing differences in steroid hormone expression and carbohydrate metabolism in TTTS are incompletely understood, and it is interesting to speculate whether miRNAs could be implicated in regulation of these critical fetal and placental developmental pathways (Dunn et al., 2016; Hoffman et al., 2020).

The strengths of this study include the use of a non-invasive biomarker source as most biomarker research for TTTS has been performed in amniotic fluid samples. Further strengths were the selection of uncomplicated MCDA twins as controls, multiple testing correction, and collection of samples before and after SFLP. However, this study has several limitations. First, though this study aimed to understand whether miRNAs could predict the presence of TTTS via random forest modeling, the study was retrospective and limited by sample size. Therefore, conclusions cannot be drawn on whether these miRNAs could prospectively predict disease development. Further, it is not known whether the changes in miRNA expression are unique to TTTS or could be present in other gestational twin pathologies. Several miRNAs of potential relevance to TTTS were not included in the panel due to limitations of the size of custom microarray cards, and therefore this study is not a comprehensive evaluation of all miRNAs that may be differentially expressed in maternal plasma. Assessment of a more comprehensive miRNA profile with additional validation experiments in well-powered patient cohorts is a goal of future study. Additionally, lack of an endogenous control is a limitation of the study as the intended endogenous control was not present in the samples. Although endogenous controls have not been established for this population, we hypothesize that this is due to a technical error in manufacturing of the custom plate. Finally, this study was limited in its ability to compare miRNA expression according to TTTS stage. The ability of maternal plasma miRNAs to complement existing staging systems requires prospective investigation.

We conclude that differential expression of miRNAs is present in maternal plasma of patients with TTTS compared to MCDA controls. miR-26a-5p, miR-320a-3p, miR-222-3p and miR-145-5p are significantly dysregulated in patients with TTTS. miR-26a-5p, miR-320a-3p, and miR-222-3p demonstrate strong predictive capacity for TTTS in a random forest model.