Principal findings of this study: This is the first study of PLAP-EVs and proteomics of maternal plasma EVs for the identification of women with Hb Bart’s fetuses. We demonstrated that: 1) EV particle size in Bart fetuses with hydropic features was significantly smaller than that in normal pregnancy and disease control groups; 2) Bart fetuses with hydropic features had higher PLAP exosome number than that in normal pregnancy; although, PLAP exosome level was highest in the placental associated complications groups; 3) Bart fetuses without hydropic features tended to have higher PLAP-EV level compared to normal pregnancy group; 4) hydropic fetuses due to non-Bart’s causes had similar PLAP level compared to normal pregnancy group; 5) EV proteomic analysis revealed aberrant immune response, pro-inflammatory cytokine regulation mediated by TLRs-MyD88-IRAK4-MAPK axis and vascular injuries as part of pathophysiology of placental edema and hypoxia in Bart’s hydrops.
Prenatal identification program for Hb Bart’s fetalis fetuses
Previous studies have shown that there are morphological and functional changes in the responses to persistent fetal anemia[55, 56, 57, 58, 59, 60, 61, 62, 63]. Prior to the fetal hydropics stage, hemodynamic function shows a normal and good compensatory adaptation, by increasing turnover blood volume, or cardiac work and distribution of blood flow into essential organs such as brain and spleen as detected by cardiomegaly, increased middle cerebral artery peak systolic velocity (MCA-PSV) or splenic-PSV[5]. Subsequently, at the high-output hydropic stage, there are fluid accumulations in the body space such as pleural/pericardial effusion or ascites, and these features are considered late ultrasonographic signs usually detected in the second or third trimester. Lastly, persistent fetal anemia results in low cardiac output heart failure as demonstrated by low cardiac output, poor contractility and a markedly increased preload, which usually occur in the last trimester[5]. Therefore, the determination of biomarkers according to disease pathophysiology prior to hydropic state is the key.
Currently, sonographic and maternal serum biomarkers have been extensively evaluated as screening tools to identify Hb Bart’s fetuses in pregnancy at risk for this disease. Common sonographic markers include the measurement of cardiothoracic diameter and MCA-PSV[5, 51]. Thus far, only cardiothoracic diameter (≥0.50) can accurately predict fetal Hb Bart’s disease as early as the late first trimester (12–15 weeks of gestation), with a detection rate of 75-100% at 90-100% specificity[5, 51]. The detection rate of MCA-PSV (≥1.5 MoM) was approximately 64%-85% at 16-22 weeks of gestation, at 98-100% specificity[5, 51]. Other sonographic signs including nuchal translucency, placental thickness, liver length, splenic circumference had limited values for early identification of Hb Bart’s fetuses[5, 51]. Recently, Harn-A-Morn et al. demonstrated that serial ultrasound screening (using cardiothoracic diameter and MCA-PSV) for the identification of Hb Bart’s disease during pregnancy, beginning in the first trimester and continuing every 2–4 weeks until 24 weeks, has a sensitivity of 100%, at 10.9% false positive rate, in detecting pre-hydropic signs[11]. The mean gestational age at Hb Bart diagnosis was 15.5 ± 2.6 weeks of gestation. Therefore, our national standard prenatal care program implements such sonographic markers for the routine screening of fetal Hb Bart’s disease in all pregnancies at risk for Hb Bart’s fetuses. The main limitations of serial ultrasound are that it requires specific equipment, and it is operator-dependent, which needs additional training. Several maternal serum biomarkers such as free beta-human chorionic gonadotropin (β-hCG), inhibin-A, pregnancy-associated plasma protein-A (PAPP-A), alpha-fetoprotein (MAFP), unconjugated estriol (uE3) or angiogenic factors (soluble fms-like tyrosine kinase or placental growth factor) have poor predictive values for the detection of Hb Bart’s fetuses[5, 51].
Placenta-derived EVs in Hb Bart’s fetuses
In this study, we isolated EVs using the combinatory methods of stepwise centrifugation, ultrafiltration and qEV size exclusion chromatography[64, 65, 66, 67] and then validated the presence of EVs in the isolates using NTA, TEM and Western blot analysis (Figures 1-3) following the ISEV guidelines[52]. Then, placental-specific protein PLAP containing EVs (PLAP-EVs) were measured by ELISA, while other EV proteins were detected by SWATH proteomics compared between Bart’s hydrops group and other comparators as detailed in the methods section. Accordingly, we would like to highlight that this is the first study reporting PLAP-EVs in Bart’s hydrops fetalis.
PLAP is a plasma membrane enzyme isoform of alkaline phosphatase specifically produced by the syncytiotrophoblast[68]. PLAP-EVs are specific for pregnancy and are not found in the circulation of non-pregnant women[69, 70]. Herein, we are reporting that the PLAP-EVs was significantly higher in Bart’s fetuses with hydropic features compared to the couple at risk with normal pregnancy outcome (Figure 4). The increment of PLAP-EVs has demonstrated only in Bart’ fetuses’ group since women with hydropic fetuses due to non-Bart’s causes had similar levels of PLAP-EVs as those with normal pregnancy. However, PLAP-EVs in Bart’s fetuses with hydropic feature were lower than pregnant women with placental-associated complications (e.g., preeclampsia and fetal growth restriction). This result suggested that PLAP-EV levels may depend on the spectrum of placental pathologies. Further studies in a larger cohort are warranted to investigate the usability of PLAP-EVs in various pregnancy complications.
There are very few studies about EVs in patients with thalassemia and none of these studies evaluated the role of EVs in pregnant women or Bart disease[71, 72, 73]. EVs obtained from patients with β-thalassemia/HbE induced platelet activation, platelet aggregation and platelet-neutrophil aggregation, which partly contributes to the hypercoagulable state[74]. In addition, such microparticles have an effect on endothelial cell activation that leads to the increased adhesion of leukocytes to endothelial cells resulting in thrombosis and vascular dysfunction especially in patients with splenectomized β-thalassemia/HbE disease[71].
Within the maternal circulation, different EV populations are released from several cell types including erythrocytes[75], endothelial cells[76], lymphocytes, dendritic cells, and placenta during gestation. The roles of EVs are thought to be involved in cell-to-cell communication between the placenta and maternal immune system[77]. Placenta-derived EVs are thought to promote maternal immune tolerance towards the fetal allograft as they can suppress maternal T-cell signaling[69]. The local immune privilege at the feto-maternal interface has been attributed to the expression of placental-derived EV-associated functional Fas ligand (FasL), programmed death ligand 1 (PD-L1), and TNF-related apoptosis inducing ligand (TRAIL)[69, 78]. In addition, NK cell activity has been shown to be down-regulated, and this is mediated by the expression of NKG2D receptor ligands, UL-16 binding proteins (ULBP) and MHC class I chain-related (MIC) proteins on placenta-derived EVs. The NK cell activity down-regulation leads to maternal cytotoxic activity suppression, thereby, promoting fetal allograft survival[79]. In normal pregnancies, placenta-derived EVs have been shown to modulate the function of maternal endothelium to promote trophoblast migration, angiogenesis and spiral artery remodeling[80]. Yet, under inflammatory or pathological conditions (i.e., obesity, hypoxia and high blood sugar), high number of circulating EVs can induce the release of pro-inflammatory cytokines from endothelial cells which lead to perturbation of physiologic function[80, 81, 82]. Altogether, these suggest that the content and effects of EVs depend on the physiological or pathological of the pregnant women.
Evidence of placental hypoxia in Hb Bart’s fetuses includes: 1) the placenta in Hb Bart’s hydrops demonstrated an increased number of immature intermediate villi, which persist despite advancing in gestational age, and this finding is referred to as “generalized delayed villous maturation”[49, 83]; 2) the presence of villous edema as well as numerous and cytotrophoblastic cells that cover the villous stroma[83]; and 3) increased number of stromal cells at the periphery beneath the trophoblast layer, also called “peripheral villous stromal hypercellularity (PVSH)”[83], a sign of placental adaptation to chronic hypoxia. The immature intermediate villi have a large diameter, and their predominance in the Hb Bart’s placenta leads to the narrowing of the intervillous space, impeding fetoplacental oxygen and nutrition transfer. Contractions of the myofibroblastic cells in PVSH lead to the reduction the villous size, thereby widening the intervillous space for an increase in the maternal blood flow[49, 83]. In addition, morphometric studies have shown a branching vascular pattern, which is associated with placental hypoxia. This change is thought to be responsible to the marked placental enlargement, which compromised the blood flow, from the uterine distention and the generally diminished intervillous space due to the numerous intermediate types of villi. Lastly, the dramatic reduction capacity of Hb Bart’s to extract oxygen from the intervillous space also contributes to placental hypoxia. Altogether, the presence of placental hypoxia can stimulate the release of EVs from the placenta into the maternal circulation. The observations that PLAP-EV levels were highest in women with placental associated complication group were consistent with previous studies[82, 84, 85, 86].
We hypothesized that the degree of hemodynamic changes or placental hypoxia is less in women with Bart’s fetuses without hydropic changes as demonstrated by a non-significant trend of high PLAP-EVs. Interestingly, women with hydropic fetuses due to non-Bart’s causes had a comparable PLAP-EV level as the normal pregnancy. Causes of hydrops in these fetuses are due to non-immune causes: cardiac disease (n=1), Down’s syndrome (n=1), and unknown non-immune causes (n=3). These cases had no fetal anemia as shown by normal MCV-PSV. Pathophysiology of hydrops fetalis is not completely understood; however, mainly due to the developmental defects in the microcirculation and lymphatic system, decreased ventricular filling or increased central venous pressure resulting from increased right heart pressure or obstructed lymphatic drainage[87]. The presence and degree of placental hypoxia in these fetuses is currently unknown.
Proteomics profile of plasma EVs in Hb Bart’s fetuses
We performed proteomic analysis of the plasma EVs of pregnant women who had Hb Bart’s fetuses, demonstrating 16 proteins in the plasma EVs were different from normal pregnancy and women with placental associated complication group (Figure 5). Fibrinogen complex, fibrin clot formation and integrin signaling were functions and processes related to the significantly different proteins. Among differentially expressed proteins, hnRNPA2B1 was highest in Bart fetuses with hydropic features compared to placental associated complications (10 times higher) or normal pregnancy group (2 times higher).
hnRNPA2B1 (A2B1), a member of the hnRNPABs subfamily, consists of two structural homologous proteins (hnRNPA2 and hnRNPB1) characterized by a tight sequence correlation and conserved domain structure, with B1 having 12 additional amino acids at N terminus compared with A2. A2B1 is an RNA-binding protein that affects the localization, shearing, stability, translation and other biochemical functions of RNA[88]. The hnRNPs are RNA binding proteins and they complex with heterogeneous nuclear RNA (hnRNA). These proteins are associated with pre-mRNAs in the nucleus and appear to influence pre-mRNA processing and other aspects of mRNA metabolism and transport. hnRNPs is expressed in several tissues such as placenta, ovary, musculoskeletal, endocrine organs and gastrointestinal tracts[89]. In pregnancy, these proteins were differentially expressed in the placenta of women with gestational diabetes mellitus and may play a role in regulating the occurrence and development of gestational diabetes[89]. It is possible that the presence of placental hypoxia in Bart fetuses triggers the release of hnRNPA2B1 since early gestation and this is specific to Hb Bart fetuses. However, a larger trial is required to further validate these findings.
Clinical and research implications
Ultrasound parameters and some soluble maternal blood biomarkers have been proposed to be used for the prediction of Hb Bart’s fetuses. However, we still lack sensitive methods for effective early screening. In addition, ultrasound parameters require standardization and proper training. Biomarkers from plasma EVs could improve the prediction of Hb Bart’s fetuses. Indeed, this study revealed that the PLAP-EV level in maternal circulation is higher in Bart’s fetuses and 16 EV proteins as well as hnRNPA2B protein in maternal plasma exosome have different concentrations in cases of Hb Bart’s fetuses compared to normal pregnancy and those with placental associated complications. Measuring PLAP-EVs together with differentially expressed EV proteins, might provide independent information to improve the ability to identify affected fetuses with Hb Bart disease.
We envision that soluble markers secreted from the placenta can be assessed non-invasively in ongoing pregnancies through a “liquid biopsy”. A term “liquid biopsy” refers to a test performed on any biofluid specimens such as plasma, urine amniotic fluid or cervico-vaginal fluid[90, 91, 92] in which the result can be used to infer pathologic changes in distant tissues (e.g., cancer) or other pathologic processes (e.g., atherosclerosis)[93, 94, 95]. We believe that the non-invasive liquid biopsy could provide information in relation to placental health. In this report, the liquid biopsy based on an examination of placenta-derived EVs in maternal blood for the identification of Hb Bart’s fetuses is feasible.
For the strengths and limitations, this is the first study to evaluate PLAP-EVs and proteomics profile of maternal plasma EVs in women with Bart’s fetuses. The strengths of this study are that we have included several groups such as Hb Bart’s group with or without hydropic features, hydropic fetuses due to other causes, or placental associated complication group. The latter group is considered the disease comparator group since compelling evidence suggests abnormal profile of placenta-derived EVs in women with placental associated complications. In addition, we confirmed by definite invasive prenatal diagnosis test that all couple at risks with normal pregnancy group had non-Bart’s fetuses and delivered without maternal or neonatal complications. The separation of EVs in the current study was confirmed by published methods including NTA, TEM and EV markers by Western blotting analysis following the international standard. The main limitation of our study is related to a small sample size. Future research is needed to confirm our findings in a larger number of populations.
In the conclusions, EV particle size is smaller, while placental-derived EV level is significantly higher in women with Hb Bart’s fetalis fetuses compared to normal pregnancy. In addition, several EV proteins were differential expressed in women with Bart’s fetuses and these proteins may play roles in innate immunity, TLRs-MyD88-IRAK4-MAPK axis, and vascular injury as the consequences of placental hypoxia in Bart’s hydrops. The result of this study support in future investigations to determine whether placenta-derived EVs in maternal circulation can be used as a liquid biopsy or non-invasive prenatal test for the early identification of Bart’s fetuses. This will ultimately reduce the rate of unnecessary invasive prenatal diagnosis testing.