Periostin promotes extensive neovascularization in placenta accreta spectrum disorders via Notch signaling

Abstract Objective Extensive neovascularization, closely linked to massive intraoperative blood loss, is a pathological hallmark of placenta accreta spectrum (PAS) cases. This study aims to explore proteins related to neovascularization and elucidate their regulatory roles in PAS, enhancing our understanding of this condition. Methods The isobaric tags for relative and absolute quantitation technique were used to identify and quantify the differentially expressed proteins in placentas from PAS and healthy pregnant women. Immunofluorescence staining and western blot analysis were conducted to determine the protein expression and localization. Gain-of-function experiments were used to conduct cell proliferation and migration assays. In addition, the tube formation assay was performed to evaluate angiogenesis in vitro. The Notch inhibitor DAPT was used to determine the involvement of Notch signaling in angiogenesis in PAS. Results Periostin (POSTN) exhibited higher expression in PAS placentas than in normal placentas. Moreover, the overexpression of POSTN in endothelial cells promoted cell proliferation, mobility, and endothelial angiogenesis via the Notch signaling pathway in vitro. Conclusion Elevated POSTN expression in PAS is associated with increased angiogenesis, indicating its potential as a molecular marker for significant intraoperative blood loss.


Introduction
Placenta accreta spectrum (PAS) disorders, which are high-risk morbidity/mortality-associated conditions, refer to the abnormal attachment of the placenta to the uterine myometrium.Any attempt to manually remove a placenta accreta typically provokes heavy bleeding [1].Recent evidence suggests that women with PAS can experience a large blood loss ranging from 3,000 to 8,600 ml [2, 3].Consequently, massive obstetric hemorrhage may lead to an emergency postpartum hysterectomy, disseminated intravascular coagulation, and maternal death [4-6].
In addition, massive hemorrhage might be attributed to hypervascularity at the utero-placental and utero-bladder interfaces, as reflected by bridging vessels between the myometrium and the posterior bladder wall, indicating that vessels extend from the placenta across the myometrium and beyond the serosa [7].Meanwhile, in invasive placentation cases, placental bed hypervascularization observed at delivery suggests neovascularization at the scar area and radial and/or arcuate uterine vasculature vasodilatation in the accreta area [4].Therefore, extensive neovascularization is apparent in most PAS cases.Moreover, according to a recent study, researchers have associated the expression of various proangiogenic proteins with PAS development [8].However, the molecular mechanism by which neovascularization in PAS is regulated remains unclear.
The use of isobaric tags for relative and absolute quantitation (iTRAQ) is a well-established isobaric labeling approach for quantitative proteomics involving mass spectrometry.It can process eight samples with high precision in one experiment.So far, it has been widely used to quantify proteins, such as biomarkers and therapeutic targets, from different resources [9,10].In the present study, we determined the differentially expressed proteins in placentas donated by healthy and PAS pregnant women using iTRAQ to identify potential PAS-related proteins and to explore their roles in PAS pathogenesis.

Patient enrollment and sample collection
This study was approved by the hospital ethics committee (#No: 2023ER070-1).All recruited patients were informed and signed the consent forms before the study was initiated.This study recruited pregnant individuals with PAS and normal pregnant (NP) individuals without previous cesarean sections, all of whom were singleton pregnancies.PAS was diagnosed using the International Federation of Gynecology and Obstetrics guidelines [11].All cases with PAS were diagnosed antenatally based on ultrasonography and were verified during the operation.All PAS cases had placenta previa.The exclusion criteria included patients with any of the following: 1) gestational diabetes, hypertension, chronic kidney disease, infectious disease, or other maternal complications and comorbidities other than placenta accreta; 2) history of smoking or drinking; 3) fetal structural abnormalities or chromosomal abnormalities; 4) unwillingness to donate their placenta; 5) incomplete clinical records.Four PAS and three NP placental tissues were collected.The maternal and delivery characteristics (maternal age, body mass index (BMI), gravidity, parity, newborn weight, gestational age at delivery, placenta weight, and history of previous cesarean deliveries) between the PAS and NP individuals were compared to ensure unbiased group matching.Placental tissues from the PAS group were intruding into the uterine myometrium and were dissected while the corresponding parts of the placenta were isolated from the NP group.Tissue specimens were snap-frozen in liquid nitrogen and stored at −80 � C.

Protein preparation, quality control, and mass spectrometry characterization
Protein samples (10 lg) harvested from the human placenta (n ¼ 7) were solubilized, separated by sodium dodecyl sulfate-polyacrylamide electrophoresis (SDS-PAGE, 12%), and subjected to iTRAQ analysis conducted by the Beijing Genomics Institute (China), as previously reported [12].Subsequently, the separated proteins were visualized with Coomassie blue G-250 and excised for proteolysis with trypsin.Then, the generated peptides were labeled using the iTRAQ labeling 6-plex kit (Applied Biosystems, USA).Next, the labeled peptides were separated and collected with a liquid chromatography 20AB liquid-phase system (Shimadzu, Japan).Afterward, the peptides were separated by ultra-high-performance liquid chromatography using a Thermo Ultimate 3000 instrument coupled to a nanoelectrospray ionization system and a Q-Exactive HF mass spectrometer (Thermo Fisher Scientific, USA) for data-dependent acquisition mode detection.
The proteins were identified and quantified by the MASCOT search engine (http://www.matrixscience.com) against Homo sapiens.The following parameters were selected: the mass tolerances were set as 20 Da (ppm) for intact peptide masses and 0.05 Da for fragmented ions.One missed cleavage was permitted for trypsin digests.Oxidation (M), deamidation (NQ), and iTRAQ8plex (Y) were selected to represent variable modifications, and carbamidomethyl (C), iTRAQ8plex (N-term), and iTRAQ8plex (K) were set to represent fixed modifications.All unique peptides were allowed for protein quantitation.The quantitative analysis of the labeled peptides was conducted using IQuant software, as reported previously [13].Q value < 0.05 and fold change > 1.2 were set as the significant thresholds for differential expression.

Western blot analysis
Western blot analysis was performed to evaluate the expression of POSTN, as reported previously [14].In brief, the tissues were solubilized in radioimmunoprecipitation assay lysis buffer with protease inhibitors, and the protein concentration was determined using the bicinchoninic acid method.Identical total tissue lysates were separated by SDS-PAGE and subsequently transferred onto polyvinylidene difluoride membranes.The membranes were blocked with 5% goat serum (1 h, room temperature) before adding the primary antibodies for overnight incubation.The primary antibodies used included anti-POSTN (1:5,000; ProteinTech, China), anti-Notch1 (1:500; ABclonal, USA), anti-Hairy/Enhancer of Split 1 (HES1) (1:500, ABclonal, USA), anti-Hairy/Enhancer of Split 1 related with YRPW motif 1 (Hey1) (1:1,000; Abcam, UK), and anti-GAPDH (1:5,000; ProteinTech, China).The following day, the membranes were incubated with horseradish peroxidase-conjugated secondary antibody, developed with an enhanced chemiluminescent kit, and the protein bands were analyzed using ImageJ software (Version 1.53, National Institutes of Health, USA).

Cell culture and treatment
Human umbilical vein endothelial cells (HUVECs) were obtained from the American Type Culture Collection and cultured in RPMI-1640 (Corning, USA) with 10% fetal bovine serum (Gibco, USA) in humidified air at 37 � C with 5% CO 2 .

Cell viability assay
Cell viability was determined with a Cell Counting Kit 8 (CCK-8) (Beyotime, China), according to the manufacturer's instructions.In brief, HUVECs with or without POSTN overexpression were seeded into a 96well plate (5 � 10 3 /well) and cultured for 72 h.At the indicated time points, cells were incubated with CCK-8 reagent for 1 h, and the absorbance was determined with a microplate reader at 450 nm.The optical density value was used to reflect the relative cell viability.

Scratch assay
HUVECs with or without POSTN overexpression were cultured until confluence, and scratches were generated with a 10-lL pipette tip.When the detached cells were washed off, images were acquired at the beginning (0 h) and every 24 h afterward until the end of the experiment at four random fields using an optical microscope (�100).Wound closure was reflected by the closure rate (0 h scratch width − 24/48 h scratch width)/0 h scratch width � 100%.

Tube formation assay
A tube formation assay was conducted, as reported previously [15].In brief, 1.5 � 10 4 HUVECs transfected with pCDH-POSTN or empty vector were subcultured onto a l-slide coated with growth factor-reduced Matrigel (BD Bioscience, China) in triplicate.The images were taken at 4-6 h after the cells were seeded [16,17].Twenty-four hours later, DAPT (10 mM, HY-13027, MedChemExpress, USA) was used to determine the involvement of Notch signaling in angiogenesis.The number of tubes, tube length, and branch points from each group were analyzed using Image J software (National Institutes of Health, USA) to reflect the angiogenic potency.

Statistical analysis
The experimental results are expressed as the mean-± standard deviation from at least three independent experiments.The continuous variables are presented as the median (interquartile range), while categorical variables are presented as numbers and percentages.Statistical analyses were conducted using SPSS 25.0 software (SPSS, Inc., USA).The categorical and nonnormally distributed continuous variables were analyzed with the Fisher's exact test and the Mann-Whitney U test, respectively.The three groups of variables were analyzed with one-way analysis of variance with Tukey's multiple-comparison test.A p-value of less than 0.05 was considered statistically significant.

Baseline and delivery characteristics of the PAS and control patients
The enrolled patients' maternal age, BMI, gravidity, parity, placenta weight, and history of previous cesarean delivery status as well as the newborn weight were compared between the two groups, and no significant differences were observed other than the history of previous cesarean delivery (Table 1).Moreover, in all PAS cases, Doppler ultrasonography indicated numerous newly formed vessels and the corresponding subplacental or uterovesical hypervascularity observed in the surgery (Supplementary Figure ).All four PAS subjects received conservative management, and their clinical characteristics, including intraoperative blood loss, are shown in Table 2.

Differentially expressed proteins in the PAS and control patients
The iTRAQ method is a highly accurate and repeatable technique used in proteomic studies to screen and quantify differentially expressed proteins [10].Thus, we leveraged this technology to interrogate the proteome of placental tissues from PAS and healthy individuals.The labeled peptides in human placentas from the PAS and NP groups were analyzed.As a result, we detected 37,980 unique peptides and identified 6015 proteins (false discovery rate ¼ 1%) containing 682 differently expressed proteins (fold change � 1.2, and q-value < 0.05) (Figure 1A).Subsequently, a fine-tuned selection criteria (unique peptide number � 2, fold change � 1.3, and q-value < 0.05) was applied, and 81 differently expressed proteins were identified, including 46 upregulated and 35 downregulated proteins (Figure 1B, full list of identified proteins is shown in Supplementary Table S1).

Localization and validation of POSTN in PAS tissue
POSTN is a well-documented protein closely related to angiogenesis [18].More importantly, a recent study has suggested that abnormally expressed POSTN may be associated with dysregulated decidualization, leading  to PAS [19].Interestingly, POSTN was one of the most differentially expressed proteins identified in our study (Supplementary Table S1), indicating its potential significance in PAS initiation/progression.Therefore, we explored the expression of POSTN in the placental samples and found that POSTN expression in the PAS placentas was greater than in the normal placentas (Figure 2A).Also, the immunofluorescence staining  results indicate that it is colocalized with CD31 (Figure 2B), a well-established endothelial cell marker [20] in the vasculature of the placenta.Additionally, POSTN has been demonstrated to regulate the angiogenic functions in HUVECs in vitro and in vivo [21].These findings suggest that POSTN plays an important role in PAS and may exert its effects by regulating endothelial cell functions.

POSTN promoted endothelial cell invasion, migration, and proliferation in vitro
To determine how POSTN is involved in endothelial cell biology during PAS development, we generated POSTN-overexpressing HUVECs and conducted invasion, migration, and proliferation assays.RT-qPCR (Figure 3A) and western blot (Figure 3B) confirmed that POSTN was successfully upregulated in HUVECs transfected with pCDH-POSTN.Moreover, we noticed that POSTN overexpression significantly promoted HUVECs proliferation (Figure 3C).Furthermore, we compared the migratory potency of HUVECs with or without POSTN overexpression and found that POSTN significantly facilitated the HUVECs mobilities as reflected by faster wound closure (Figure 3D).Of note, proliferation and migration play essential roles in endothelial cells during angiogenesis [15].Thus, these findings further highlighted the importance of POSTN in endothelial biology during the PAS progression.

POSTN enhanced the formation of vasculature
To validate our speculation that POSTN regulates PAS by controlling vasculature formation, a tube formation assay was conducted, and the results showed that POSTN overexpression significantly promoted tube formation, as reflected by the elevated tube amount, length, and branch points in the pCDH-POSTN group (Figure 3E).

Overexpression of POSTN activated the Notch signaling pathway in HUVECs
It has been documented that POSTN regulates Notch signaling [22], which is pivotal for vasculature formation and maturation [23].In our study, we noticed that overexpression of POSTN in HUVECs resulted in the significant upregulation of Notch1, HES1, and Hey1 (Figure 4).These results suggest that POSTN might participate in the activation of Notch signaling in HUVECs during neovascularization.

DAPT rescued the vasculature formation in HUVECs with POSTN overexpression
To further test our speculation that POSTN regulates HUVEC neovascularization through Notch signaling, a Notch inhibitor, DAPT, was used.DAPT rescued the POSTN-promoted vasculature formation (Figure 5A-F).
In addition, our results demonstrated that POSTN overexpression significantly elevated the number of tubes, tube length, and branch points compared with the control group in which the cells were transfected with the empty vector.However, the introduction of DAPT significantly attenuated the promotive effects of POSTN on neovascularization (Figure 5G-I).These results suggest that POSTN promotes blood vessel formation via the Notch signaling pathway.

Discussion
PAS is a well-established cause of postpartum hemorrhage that can seriously threaten life.It is characterized by increased subplacental and intervillous circulation and neovascularization in the peritoneum [7].Hence, gaining a deeper insight into the pathogenesis of PAS would aid in its diagnosis and management.
We explored the differentially expressed proteins in placentas from the PAS and NP groups using human samples.Among the upregulated proteins identified by iTRAQ, POSTN attracted our attention.POSTN is broadly expressed in various tissues, such as the heart, skin, tumors, and vasculature, with the highest content in the aorta, lower digestive tract, placenta, uterus, thyroid tissue, and breast [24][25][26][27].Moreover, POSTN is a versatile protein that participates in a variety of cellular activities, including angiogenesis [18].Although, as mentioned above, PAS is closely associated with elevated neovascularization, the relationship between POSTN and PAS has not been previously reported to the best of our knowledge.Therefore, we evaluated the expression of POSTN in the placentas of PAS and NP women.We found that POSTN was upregulated in the PAS group compared with the NP group.More importantly, our findings indicate that POSTN was colocalized with CD31, suggesting that POSTN is expressed in endothelial cells.These data prompted us to speculate that POSTN may play a role in PAS by regulating angiogenesis in the placenta.To test this hypothesis, we modulated the POSTN expression in endothelial cells and evaluated angiogenesis in vitro.The results showed that endothelial cells with elevated POSTN expression exhibited enhanced angiogenic potential compared with control cells with normal   POSTN expression.These findings are in alignment with previous studies.For instance, POSTN was recently found to promote the growth and branching of the vasculature in tumors [18,21].
Notably, the proliferation and migration functions of endothelial cells play a critical role in angiogenesis.It has been suggested that endothelial cells and endothelial progenitor cells need to travel to the site of blood vessel formation to carry out their functions [15,28].Therefore, we suspected that POSTN may affect angiogenesis by regulating these functions of endothelial cells.So, we conducted proliferation and migration assays using HUVECs with or without POSTN overexpression, and the findings demonstrated that upregulated POSTN significantly promoted endothelial cell proliferation and motility, echoed by previous studies [21].Thus, these data imply that POSTN might be an important regulatory factor involved in the angiogenesis of PAS.Consistent with our findings, a recent study on endometriosis performed RNA sequencing using tissue samples from patients with endometriosis and healthy participants.Their sequencing data showed that POSTN was upregulated in the samples from patients with endometriosis.Furthermore, they reported that changes in its gene expression would result in impaired decidualization [19], which could ultimately lead to PAS.Our findings were supported by their sequencing data and, in turn, validated their findings.
These discoveries have heightened our interest in exploring the underlying mechanisms by which POSTN regulates angiogenesis.Researchers have suggested a close correlation between Notch signaling and cell proliferation and migration, which prompted us to ask whether POSTN modulates angiogenesis through this pathway.Interestingly, our data revealed the involvement of Notch1 in POSTN-mediated neovascularization.We noticed that POSTN activated the Notch signaling pathway and concomitantly promoted angiogenesis in HUVECs.However, the POSTN-mediated promotive effect on neovascularization was attenuated by a Notch inhibitor, DAPT.These results were supported, at least partially, by previous findings showing that silencing POSTN resulted in reduced Notch signaling activity [29,30] in vitro and in vivo.As POSTN has been implicated in Notch signaling [22] and has been demonstrated to bind to Notch1 through interactions with the epidermal growth factor-like domains [29], it is reasonable to speculate that POSTN may bind to Notch1 to maintain its stability in HUVECs, ultimately leading to its increased expression at the cell surface.However, further studies are needed to validate this hypothesis.
It is worthwhile to mention that within the vascular system, Notch signaling is composed of multiple receptors (Notch1, Notch3, and Notch4) that modulate downstream targets (HES and Hey) [31].While in-vitro studies have reported that Notch4 inhibits angiogenesis, results from in-vivo studies have demonstrated that combined inhibition of Notch1 and Notch4 exaggerated the vascularization defects observed in Notch1-knockout mice [32], indicating a potential proangiogenic role of Notch4.In addition, sustained activation of Notch4 in endothelial cells has been shown to upregulate delta like canonical Notch ligand 4, a well-documented Notch1 and Notch4 ligand [33].Thus, both Notch1 and Notch4 may contribute to angiogenesis modulation in response to stimuli-mediated Notch signaling activation.Intriguingly, our data mining of the sequencing results revealed upregulation of Notch4 in the PAS group (data not shown), suggesting that both receptors may contribute to the pathogenesis of PAS; however, whether these two receptors function independently or synergistically needs further investigation.
The upregulation of POSTN in PAS patients prompted us to consider the broader implications beyond our studied cohort.First, POSTN as a biomarker could be utilized for the early diagnosis and screening of PAS, aiding in the early recognition of potential issues during pregnancy.Second, by assessing POSTN levels (for instance, building nomogram curves of serum POSTN for different gestational ages), the degree of risk for severe bleeding in PAS patients could be evaluated, assisting in tailoring more targeted treatment plans.Additionally, predicting the efficacy of different treatment methods and serving as an effective tool during follow-up and monitoring are also potential applications of POSTN detection.However, further research and validation are essential to ascertain the effectiveness of these applications.
This study stands as one of the limited endeavors delving into the proteomics of placental accreta.We aimed to unravel the molecular mechanism behind increased angiogenesis in placenta accreta.It is important to acknowledge that our study has its limitations, notably the small sample size and its exploratory nature as a pilot study.Therefore, a more expansive sample size will be essential for validating and substantiating our findings in the future.

Conclusions
In conclusion, our findings suggest that upregulated POSTN expression contributes to neovascularization in THE JOURNAL OF MATERNAL-FETAL & NEONATAL MEDICINE PAS placentas, potentially involving the Notch signaling pathway.Therefore, POSTN might be a possible molecular marker for excessive angiogenesis and massive intraoperative blood loss in PAS.

Figure 1 .
Figure 1.(A) Volcano plots of the differentially expressed proteins identified in placenta accreta spectrum (PAS) and control placentas.The plot depicts the volcano plot of log2 fold change versus -log10 q-value.A Q-value < 0.05 and fold change > 1.2 were the thresholds for significantly differential expression.Green dots represent downregulated proteins, gray dots represent expression-intact proteins, and red dots represent upregulated proteins.(B) Heat map exhibiting the differentially expressed proteins identified in PAS and control placentas.Red-and blue-colored bars indicate up-and downregulation, respectively.

Figure 2 .
Figure 2. (A) Representative images of the western blot analysis detecting periostin (POSTN) expression in the placentas from the placenta accreta spectrum (PAS) and normal pregnancy (NP) women.(B) Results of the immunofluorescence staining detecting POSTN and CD31 in PAS and normal placentas.POSTN is colocalized with CD31 in the blood vessels of the placenta.Red: CD31; Green: POSTN; blue: DAPI.

Figure 3 .
Figure 3. Results of the reverse transcription-quantitative polymerase chain reaction (A) and a representative image of the western blot analysis (B) validating the expression of periostin (POSTN) in human umbilical vein endothelial cells (HUVECs).(C-D) Representative images and analyzed results demonstrating that overexpression of POSTN promoted the proliferation (C) and migration (D) of HUVECs.(E) Representative images and analyzed results of the tube formation assay.The left panels are microscopic images, and the right panels are Image J-analyzed images (the cyan, green, yellow, red, and blue colors were used to contour the tubular structures, branches, segments, nodes, and isolated structures).analyzed data comparing the indicated parameters between the two groups.(n ¼ 3, � p < 0.05, �� p < 0.01, POSTN vs. vector).

Figure 4 .
Figure 4. Representative images of the western blot analysis detecting the expression of Notch1, Hairy/enhancer of Split 1 (HES1), and hairy/enhancer of Split 1 related with YRPW motif 1 (Hey1) in human umbilical vein endothelial cells with or without periostin overexpression.

Figure 5 .
Figure 5. Representative images and analyzed results of the tube formation assay.(A-F).the microscopic images (A, C, and E) and Image J-analyzed images (B, D, and F) of the tube formation assay using human umbilical vein endothelial cells under the indicated conditions.(the cyan, Green, yellow, red, and blue colors were used to contour the tubular structures, branches, segments, nodes, and isolated structures).(G-I).analyzed data comparing the indicated parameters among the three groups.(n ¼ 3, � p < 0.05, �� p < 0.01, ��� p < 0.001, vs. vector; # p < 0.05 vs. pCDH-POSTN).

Table 1 .
Baseline and delivery characteristics of the placenta accreta spectrum and control patients.
a Mann-Whitney U-test b Fisher's exact test.

Table 2 .
Clinical characteristics of the subjects with placenta accreta spectrum disorders.