Amniotic Fluid Exosome Proteomic Profile Exhibits Unique Pathways of Term and Preterm Labor

Abstract

Our objective was to determine the amniotic fluid–derived exosomal proteomic profile in patients who had spontaneous preterm birth (PTB) or preterm premature rupture of membranes (pPROM) compared with those who delivered at term. A cross-sectional study of a retrospective cohort was used to quantify and determine the protein content of exosomes present in amniotic fluid, in PTB or pPROM, and normal term labor (TL) or term not in labor (TNIL) pregnancies. Exosomes were isolated by differential centrifugation and quantified using nanocrystals (Qdot) coupled to CD63 and placental alkaline phosphatase (PLAP) by fluorescence nanoparticle tracking analysis. The exosomal proteomic profile was identified by liquid chromatography–tandem mass spectrometry, and a small ion library was constructed to quantify the proteomic data by Sequential Window Acquisition of All Theoretical analysis. Ingenuity Pathway Analysis determined canonical pathways and biofunctions associated with dysregulated proteins. Amniotic fluid exosomes have similar shape and quantity regardless of the conditions; however, the PLAP/CD63 ratios for TL, PTB, and pPROM were significantly higher (∼3.8-, ∼4.4-, and ∼3.5-fold, respectively) compared with TNIL. The PLAP/CD63 ratio was also significantly higher (∼1.3-fold) in PTB compared with pPROM. Biological functions primarily indicated nonspecific inflammatory response regardless of condition, but unique profiles were also identified in cases (PTB and pPROM) compared with term. Amniotic fluid exosomes provide information specific to normal and abnormal parturition. Inflammatory marker enrichment and its uniqueness in term and preterm pregnancies support the value of exosomes in determining underlying physiology associated with term and preterm parturition.

Uterine tissue–specific amniotic fluid exosomes from term and preterm pregnancies are enriched in nonspecific inflammatory proteins, indicative of underlying (patho)physiology.

Preterm birth (birth before the 37th completed week of pregnancy) is a major socioeconomic problem that affects 11% of all pregnancies worldwide (1). The most common (45%) phenotype of preterm birth (PTB) occurs spontaneously (2). To address the problem of PTB, a clear understanding of the hormonal signals that induce normal labor and delivery at term is needed. Conventional theories regarding the signaling mechanisms for the initiation of parturition are primarily linked to fetomaternal endocrine and immune changes in the intrauterine cavity, correlating with fetal growth and development (3–8). Nonetheless, these signals and their precise mechanism of initiating parturition are still unclear. Understanding these signals and their mechanisms in normal term pregnancies can provide insight into the pathologic activation of signals that can cause spontaneous PTB.

We have recently proposed a paracrine signaling mechanism mediated in part by fetal tissue (amniochorion and placenta) stress (e.g., physiologic or pathophysiologic oxidative stress and senescence), which generates sterile inflammation within the maternal-fetal interface (fetal membranes, decidua, and myometrium) (9). As reported by many groups, inflammation is one of the key mediators of parturition-associated changes in both fetal and maternal tissues (10–13). Fetal tissue senescence has been correlated with fetal growth and organ maturation (14). The inflammatory mediators produced by senescent fetal tissues are propagated to maternal gestational tissues, to signal fetal readiness for parturition. This is mechanistically affected by increasing the inflammatory load in various gestational tissues. Interestingly, our data suggest the novel concept that senescent tissues package and propagate their contents via exosomes (15–17). Exosomes are bioactive, cell-derived intercellular signaling vesicles (30 to 150 nm) secreted by fusion of multivesicular bodies with the plasma membrane that contain a cargo of proteins, lipids, and nucleic acids. Exosomal content (protein, DNA, and all forms of RNAs) is a representation of the physiologic state of the cell of origin, making it a good vector of paracrine signaling (15, 18). We have noted that fetal cell–derived exosomes can cause labor-associated functional changes and inflammatory changes in maternal decidual and myometrial cells (19). Thus, exosomes function as communication channels between fetomaternal uterine compartments with specific roles at various target sites.

Recent observational data from several laboratories have reported the usefulness of exosomes as a potential bioindicator of cellular physiology (20–25). Exosomes have been isolated and characterized in various biological fluids to understand their origin during pregnancy (26, 27), and test their biomarker potential in identifying pregnancy complications (28–30). Interestingly, no studies have isolated or reported amniotic fluid (AF) exosome characteristics and their changes associated with PTB or normal term birth. AF predominantly reflects fetal metabolic changes, although its contents can also have maternal contributions (31). Metabolomic analysis of AF has been reported to reflect fetal organ–associated functional changes during normal and preterm parturition (32–35). This study tested the hypothesis that AF exosomes, mostly released by fetal cells, and their cargo reflect the overall physiologic status of the fetus and uterine tissues during pregnancy and parturition. In this study, exosomes were isolated from AF obtained from a well-characterized cohort of spontaneous PTB without rupture of membranes, preterm premature rupture of the membranes (pPROM), and normal term birth, and their concentration and proteomic content were determined. Herein we report the differences between the AF exosome protein profile in PTB, pPROM, and normal term birth.

Materials and Methods

Study group and samples

Patient recruitment and sample collection for this study were approved by the Tristar Nashville institutional review board and Western institutional review board for the Perinatal Research Center, Nashville, Tennessee, for a parent study that examined racial disparity associated with biomarkers of PTB. Informed written consents were obtained from patients prior to enrollment, and reuse of those samples for PTB-related projects was approved by the institutional review board at The University of Texas Medical Branch (Galveston, TX). Patients were recruited at the Centennial Medical Center (Nashville, TN) between 2003 and 2008 as a part of Nashville Birth Cohort to study racial disparity in biomarkers of PTB.

For this study, our cohort of PTB, pPROM, term labor (TL), and term not in labor (TNIL) consisted of 45 participants. Mothers between the ages of 18 and 40 years were recruited upon admission for term or preterm deliveries. Gestational age was determined by last menstrual period and corroborated by ultrasound dating. AF samples were collected from women who underwent (1) PTB (n = 13), defined as having two contractions per 10 minutes followed by delivery at <37^0/7 weeks’ gestation; (2) pPROM (n = 8), defined as rupture of membranes (confirmed by tests such as AF pooling, “ferning,” nitrazine, and Amnisure positivity) followed by PTB before onset of labor at <37^0/7 weeks’ gestation; (3) normal TL and delivery (n = 11), defined as having two contractions per 10 minutes followed by delivery at ≥37^0/7 weeks’ gestation with no medical or obstetrical complications during pregnancy; and (4) normal TNIL (n = 13), undergoing elective cesarean deliveries between 37 and 42 weeks’ gestation. Participants with multiple gestations, preeclampsia, placental previa, fetal anomalies, gestational diabetes, polyhydramnios and oligohydramnios, intrauterine growth restrictions, and other complications such as surgeries during pregnancies were excluded.

For vaginal deliveries, AF samples were collected during labor immediately before artificial rupture of the membranes by transvaginal amniocentesis of intact membranes using a 22-gauge needle through the dilated cervical opening. Samples were also collected by placement of an intrauterine pressure catheter in a few cases where indicated for clinical reasons. This procedure avoided contamination of AF from vaginal and cervical fluids. In cases undergoing cesarean delivery, samples were collected by transabdominal amniocentesis. All pPROM samples collected for this study were obtained by transabdominal amniocentesis under ultrasound guidance prior to cesarean sections. AF samples were centrifuged immediately for 10 minutes at 2000g to remove cellular and particulate matter, and supernatant aliquots were processed rapidly and stored in the dark at −80°C in filled tubes to minimize artifactual oxidation until analysis.

Isolation of exosomes from AF

Exosomes were isolated from AF as previously described (27, 28, 36, 37). In brief, AF samples were diluted with an equal volume of phosphate-buffered saline (PBS) (pH 7.4) and centrifuged at 2000g for 30 minutes at 4°C (Sorvall, high-speed microcentrifuge, fixed rotor; Thermo Fisher Scientific, Asheville, NC). The 2000g supernatant fluid was then centrifuged at 12,000g for 45 minutes at 4°C (Sorvall, high-speed microcentrifuge, fixed rotor; Thermo Fisher Scientific, Ashville, NC). The resultant supernatant fluid (2 mL) was transferred to an ultracentrifuge tube (10 mL; Beckman) and centrifuged at 100,000g for 2 hours (Sorvall, T-8100, fixed ultracentrifuge rotor; Thermo Fisher Scientific). The 100,000g pellet was suspended in PBS (10 mL) and filtered through a 0.22-μm filter (Steritop; Millipore, Billerica, MA) and then centrifuged at 100,000g for 2 hours. The pellet containing the enriched small vesicles was resuspended in 500 μL PBS and placed onto a size exclusion chromatography column (Sepharose CL-2B resin; Sigma-Aldrich, St. Louis, MO), and the exosomes were obtained by elution with PBS. Vesicles were concentrated using a 100,000 Nominal Molecular Weight Limit Amicon Ultra-15 Centrifugal Filter Unit (Merck Millipore, Billerica, MA) by centrifugation at 4000g for 10 minutes (4°C). Exosomes were characterized by size distribution, abundance of proteins associated with exosomes, and morphology in accordance with the recommendation of the International Society of Extracellular Vesicles (38), using nanoparticle tracking analysis (NTA) [Research Resource Identifier (RRID): cSCR_014239], western blot analysis, and electron microscopy, respectively.

NTA

NTA measurements were performed using a NanoSight NS500 instrument (NanoSight NTA 3.0 Nanoparticle Tracking and Analysis Release Version Build 0064; Malvern, Worcestershire, United Kingdom) as previously described (27). Exosomes were diluted to 100 µg/mL protein, and five videos were processed and analyzed. A minimum of 200 completed tracks per video were collected for each analyzed sample. NTA postacquisition settings were optimized and kept constant between samples, and each video was then analyzed to give the mean, mode, and median particle size together with an estimated number of particles per milliliter of AF. A spreadsheet (Excel; Microsoft Corp., Redmond, WA) was automatically generated, recording the concentration of each particle size. The 100-nm polystyrene latex microspheres (Malvern NTA 4088) were routinely analyzed to confirm instrument performance.

Fluorescence NTA

Qdots (Qdot nanocrystals) were conjugated to anti-CD63 (RRID: AB_648179; catalog no. sc-15363; Santa Cruz Biotechnology), placental alkaline phosphatase (PLAP) (RRID: AB_557555; catalog no. MA1-20245; Thermo Fisher Scientific), or IgG1 isotype control antibody (RRID: AB_672089; catalog no. sc-34665; Santa Cruz Biotechnology) with a SiteClick Qdot 605 Antibody Conjugation Kit (Life Technologies), executed according to the manufacturer’s instructions as previously described (29, 39). Exosomes were diluted in PBS and incubated with FcR blocking reagent (10 µL, 10 minutes at 4°C) (MACS Miltenyi Biotec), followed by incubation with anti–CD63-Qdot605, PLAP-Qdot605, or IgG1-Qdot605 (10 µL, 1:100) for 30 minutes in the dark at room temperature. Samples (i.e., exosomes alone, exosomes plus IgG1-Qdot605, exosomes plus anti–CD63-Qdot605, and exosomes plus anti–PLAP-Qdot605 and background controls; FcR blocking reagent plus IgG1-Qdot605; FcR blocking reagent plus anti–CD63-Qdot605; and FcR blocking reagent plus anti–PLAP-Qdot605) were then diluted to 500 µL with PBS and analyzed using the NanoSight NS500 instrument and NTA software. In fluorescence mode (i.e., camera level 9, shutter speed 11.25 ms, and slider gain 250), five 60-second videos were captured for each sample and analyzed.

Western blot analysis and transmission electron microscopy

Exosome proteins separated by polyacrylamide gel electrophoresis were transferred onto a polyvinylidene difluoride membrane using the Trans-Blot Turbo Transfer System (BioRad Laboratories, Hercules, CA). After transfer, the blot was blocked with Odyssey Blocking Buffer at room temperature for 1 hour. The blocking buffer was removed and the antibodies [i.e., CD63 (RRID: AB_648179; catalog no. sc-15363; Santa Cruz Biotechnology), Flotilin-1 (RRID: AB_2106567; catalog no. sc-25506; Santa Cruz Biotechnology), and TSG101 (RRID: AB_10974262; catalog no. ab125011; Abcam)], placenta-derived exosomes (PLAP) (RRID: AB_10678257; catalog no. ab96588; Abcam), and a negative control for Grp94 (RRID: AB_2722657; catalog no. 20292; Cell Signaling Technology) were added and incubated overnight at 4°C. The blots were then washed for 5 minutes in Tris-buffered saline supplemented with 0.1% volume-to-volume ratio (v/v) Tween 20 for a total of three times. Anti-rabbit conjugated with DyLight 800 (RRID: AB_10697505; catalog no. 5151S; Cell Signaling Technology) secondary antibody diluted 1:15,000 in Odyssey Blocking Buffer was added. Washes of 5 minutes each of Tris-buffered saline supplemented with 0.1% v/v Tween 20 and an additional wash of Tris-buffered saline for 5 minutes removed any residual Tween-20. The blots were scanned using a LI-COR Biosciences Odyssey IR Imaging System and data quantified using Image Studio software version 4.0. For electron microscopy analysis, exosome pellets were fixed in 3% (weight-to-volume ratio) glutaraldehyde and analyzed under an FEI Tecnai 12 transmission electron microscope (FEI, Hillsboro, OR) as we previously described (28).

Sequential Window Acquisition of All Theoretical mass spectrometry ion library generation

To generate the ion library used in the Sequential Window Acquisition of All Theoretical (SWATH) mass spectra analysis, exosome-derived samples from AF were reduced, alkylated, and trypsinized using an in-gel digestion method. In brief, exosome samples were first mixed with Bolt LDS sample buffer (Thermo Fisher), sonicated for 5 minutes, and heated at 72°C for 10 minutes. Samples were then separated based on molecular weight on a Bolt Bis-Tris Plus polyacrylamide gel (Thermo Fisher) at 200 V until full separation. For each pooled biological fluid sample, a total of 10 gel fractions were excised, resulting in a total of 120 samples. Tryptic peptides were separated using an Eksigent NanoLC system coupled with a ReproSil-Pur Basic-C18-HD, 5-µm column over a 90-minute gradient ranging from 2% to 35% [buffer A: 0.1% formic acid (v/v); buffer B: 100% acetonitrile, 0.1% (v/v) formic acid]. The resulting peptide samples were processed in an information-dependent acquisition on an AB Sciex 5600 TripleTOF mass spectrometer with the top 20 precursor ions automatically selected for fragmentation (40).

SWATH

For SWATH acquisition, the 5600 Triple TOF was operated in a looped product ion mode. Using an isolation width of 26 m/z, a set of 32 overlapping windows (1 m/z overlap) was constructed covering the mass range of 400 to 1200 m/z. For individual patient samples, 10 µg of exosomes was digested using the Filter Aided Sample Preparation as previously described (40).

Data processing

To generate the ion libraries, the mass spectra were processed using the ProteinPilot version 4.5b software (Sciex, Framingham, MA) and the Paragon Algorithm to search against a human SwissProt database. A global false discovery rate (FDR) of 1% was used as the threshold for the number of proteins for import. All identified proteins were submitted to PeptiteAtlas with the identifier PASS01131 (ftp://PASS01131:QN4683m@ftp.peptideatlas.org/). For SWATH processing, the SWATH Acquisition Microapp (version 2.0) within PeakView (RRID: SCR_015786; version 2.2) was used. Within the Microapp, a setting of three peptides per protein, four transitions per peptide, peptide confidence threshold corresponding to 1% global FDR, and FDR threshold of 1% was used. The retention time was then manually realigned with a minimum of five peptides with constantly high signal intensities and distributed along the time axis. The resulting peak area for each protein after SWATH processing was exported to MakerView (version 1.3.1; Sciex), and the resulting data were normalized using the most likely ratio method.

Ingenuity Pathway Analysis of identified proteins

Pathway enrichment analyses were performed with Ingenuity Pathway Analysis (RRID: SCR_008653; Qiagen, Hilden, Germany). Ingenuity Pathway Analysis was performed to identify canonical pathways, diseases and functions, and protein networks. Heatmap analysis was used to demonstrate the expression patterns of biological functions based on z scores. Significantly enriched pathways for the proteins and pathways were identified with the criterion P value <0.05.

Results

Patient cohort data

Maternal age was between 24 and 29 years and not different between case groups (PTB and pPROM) and control groups (TNIL and TL) (P = 0.3), and each group had a similar amount of nulliparous women and smokers (P = 0.1 and P = 0.7, respectively). Race, marital status, and level of high school and college education (determined socioeconomic status) were similar across all groups. As per definition of cases, women in the PTB and pPROM groups delivered babies at an earlier gestational age and with a lower birth weight than women in the TNIL and TL groups (P < 0.001 and P < 0.001, respectively; Table 1).

Table 1.

Clinical and Demographic Information

Variable	TNIL (n = 13)	TL (n = 11)	PTB (n = 13)	pPROM (n = 8)	P Value
Age, y	28.6 ± 4	27 ± 5.6	26.5 ± 5.5	24 ± 6.8	0.3
Gestational age, wk	38.6 [38.2–39.3]	39 [37.3–39.5]	36 [35.1–36.3]	34 [32.2–34.5]	<0.001
Birth weight, g	3950 [3225–3724]	3322 [2863–3402]	2948 [2291–3058]	2346 [2182–2723]	<0.001
Gravidity	2 [2–3]	2 [1–4]	2 [1–4.5]	2 [1.25–2]	0.31
Nulliparity	0 (0)	4 (36)	5 (38)	3 (38)	0.1
History of PTD	1 (8)	0 (0)	5 (38)	0 (0)	0.02
Smoking	1 (8)	0 (0)	1 (8)	1 (12.5)	0.7
GBS					0.04
Negative	8 (62)	11 (100)	7 (54)	4 (57)
Positive	4 (31)	0 (0)	1 (8)	1 (14)
Unknown	1 (8)	0 (0)	5 (38)	2 (29)
Race					0.32
White	12 (92)	8 (73)	11 (85)	8 (100)
African American	1 (8)	3 (27)	2 (15)	0 (0)
Marital status					0.28
Single	2 (15)	5 (45)	6 (46)	4 (50)
Married	11 (85)	5 (45)	7 (54)	4 (50)
Divorced	0 (0)	1 (9)	0 (0)	0 (0)
High school education, y	12 [12–12]	12 [12–12]	12 [12–12]	12 [12–12]	0.48
College education, y	2 [0–4]	0 [0–2.5]	0 [0–2]	0 [0–2.25]	0.48

Data are reported as mean ± standard deviation, median [interquartile range], or number (percent).

Abbreviation: GBS, group B Streptococcus.

Exosome isolation and characterization

The characteristics of exosomes isolated and purified using a well-established and validated method (38, 41) are presented in Fig. 1. NTA identified particles with a diameter of 35 to 125 nm (Fig. 1A–1C). Transmission electron microscopy identified vesicles with a cup shape morphology, which is characteristic of an exosome (Fig. 1D). Furthermore, exosomes were positive for exosomal protein marker CD63 (42) and PLAP (Fig. 1E). No significant differences between the NTA characteristics of exosomes isolated from PTB, pPROM, TL, and TNIL were identified. Interestingly, a positive correlation between the protein concentration and the number of vesicles was identified, demonstrating the purity of the enrichment of exosomes.

Figure 1.

Characterization of exosomes from AF. Exosomes were isolated from AF obtained from (1) TNIL (n = 13), (2) TL (n = 11), (3) spontaneous PTB (n = 13), and (4) pPROM (n = 8). The size distribution of the particles in the samples was determined using nanoparticle tracking analysis (see “Materials and Methods”). (A) Mean. (B) Mode. (C) A representation of the size distribution. (D) Representative electron micrograph of exosomes (arrows). (E) Representative protein profile staining with SimplyBlue SafeStain and representative western blot for PLAP and exosome enriched marker CD63. (D) Scale bar = 100 nm. Std., standard.

Quantification of exosomes

Exosomes were quantified using nanocrystals (Qdot) coupled to CD63 to evaluate the total number of exosomes or coupled to PLAP antibodies to establish the number of exosomes from placenta present in AF. The total number of exosomes (expressed as total exosomes × 10¹¹ vesicles per milliliter of fluid) present in AF was 2.2 ± 1.8, 1.8 ± 1.7, 2.4 ± 2.6, and 2.9 ± 2.3 for PTB, pPROM, TL, and TNIL, respectively (Fig. 2A). No significant differences [one-way analysis of variance (ANOVA), P = 0.59] in the total number of exosomes in AF were observed. Conversely, the number of PLAP-positive (PLAP^+ve) exosomes (expressed as × 10¹⁰ vesicles per milliliter of fluid) present in AF was 5.8 ± 4.5, 3.4 ± 3.6, 5.2 ± 4.8, and 1.7 ± 1.3 for PTB, pPROM, TL, and TNIL, respectively (Fig. 2B). A significant difference (one-way ANOVA, *P = 0.03) in the number of PLAP^+ve exosomes present in AF was identified. Post hoc test Bonferroni multiple-comparisons analysis showed that the levels of PLAP^+ve exosomes were significantly higher (∼3.6-fold) in PTB compared with TNIL. Finally, we evaluated whether different stages of parturition can modify the contribution of placental exosomes, measured as the PLAP/CD63 ratio and expressed as a percentage, to the total exosomes present in AF (Fig. 2C). Interestingly, the contribution of placental exosomes was significantly different between all the groups analyzed (i.e., PTB, pPROM, TL, and TNIL; ANOVA, P < 0.0001). Bonferroni multiple-comparisons analysis showed that the values in PTB, pPROM, and TL were significantly higher (∼4.4-, ∼3.5-, and ∼3.8-fold, respectively) compared with TNIL. The PLAP/CD63 ratio was significantly higher (∼1.3-fold) in PTB compared with pPROM.

Figure 2.

(A) No significant differences were found in total exosomes per mL of AF when comparing TNIL (n = 13), TL (n = 11), spontaneous PTB (n = 13), and pPROM (n = 8). (B) PLAP-positive exosomes per mL of AF are significantly higher (∼3.6-fold) in PTB compared with TNIL. (C) PLAP/CD63 ratios for PTB, pPROM, and TL were significantly higher (∼4.4-, ∼3.5-, and ∼3.8-fold) compared with TNIL, respectively. The PLAP/CD63 ratio was significantly higher (∼1.3-fold) in PTB compared with pPROM. *P < 0.05.

Differentially expressed proteins quantified by SWATH–mass spectrometry analysis in exosomes from AF

The protein profile in exosomes isolated from AF obtained from PTB, pPROM, TL, and TNIL was compared by SWATH–mass spectrometry to identify differentially expressed proteins that can be used to differentiate between distinct signals of parturition. We generated an ion library using the same amount of protein per group and identified 501 proteins with FDR 1% (Supplemental Tables 1 and 2). Interestingly, 18 of the top 20 proteins associated with exosomes (see exocarta) were identified in the library, such as CD63, TSG101, CD9, ALIX, and CD81, suggesting the purity of our preparations. Moreover, an abundance of PLAP in exosomes from AF was also identified, suggesting placental and other reproductive tissue origin of these exosomes.

The overall protein profile for each sample was subjected to principal component analysis and the proteins contributing to the differences between samples shown as eigenvectors (Fig. 3). Using these protein profiles, the samples were separated into four distinct groups: PTB, pPROM, TL, and TNIL. Subsequently, statistical analysis between each pair of groups was performed, and significant proteins (adjusted P < 0.05) with a log₂ fold change ±0.6 were displayed as a volcano plot (Fig. 4) and listed in Supplemental Tables 3–6. Of note, the log₂ fold change values of pregnancy zone protein were 1.73, 4.46, and 6.18 for pPROM vs PTB, PTB vs TL, and pPROM vs TL, respectively. The log₂ fold change values of Ras-related protein Rab-10 and Ras-related C3 botulinum toxin substrate 2 were −0.73 and −2.01, respectively, for PPROM vs PTB, and the log₂ fold change value of Ras-related protein Ral-A was −0.91 for PTB vs TL.

Figure 3.

Principal component analysis for each group was plotted (left) and the proteins contributing to the differences between samples are shown as eigenvectors (right). Spontaneous PTB (n = 13), pPROM (n = 8), TL (n = 11), and TNIL (n = 13).

Figure 4.

Volcano plots for pPROM vs PTB, PTB vs TL, pPROM vs TL, and TNIL vs TL. Significant proteins (adjusted P < 0.05) with a log₂ fold change ±0.6 are represented by green dots. Differentially expressed proteins that did not reach statistical significance are represented by gray dots. Spontaneous PTB (n = 13), pPROM (n = 8), TL (n = 11), and TNIL (n = 13).

Ingenuity Pathway Analysis

Ingenuity Pathway Analysis was performed on differentially expressed proteins between pPROM vs PTB, pPROM vs TL, PTB vs TL, TNIL vs TL, PTB vs TNIL, and pPROM vs TNIL. The top 10 canonical pathways for pPROM vs PTB, PTB vs TL, pPROM vs TL, and TNIL vs TL (P < 0.05) are represented in Fig. 5. Immunoglobulins and acute phase proteins, which consisted of but were not limited to inflammatory cytokines, complement factors, and coagulation factors, were overexpressed in all comparisons, suggesting that AF exosome cargo proteins represent nonspecific inflammation. This is not unexpected in any of these conditions because inflammation is an underlying pathophysiologic process of parturition irrespective of gestational age, and we determined that AF exosomes carry that signature of the cell. Activation (positive z score) of the liver X receptor/retinoid X receptor (LXR/RXR) and production of nitric oxide and reactive oxygen species in macrophages pathways were found in pPROM AF exosomes compared with PTB and TL. Canonical pathways between PTB and TL had lesser differences, with a smaller positive z score for LXR/RXR and production of nitric oxide and reactive oxygen species in macrophage pathways and a z score of 0 for the complement system pathway. Uniquely, TNIL AF exosomes revealed a downregulation of the complement system (downregulation of complement activation, coagulation cascade, and intrinsic prothrombin activation) compared with TL, suggesting that activation of the complement cascade in the intrauterine compartments as a part of the inflammatory process is associated with the initiation of labor. Fig. 6a represents the top 10 biological functions for pPROM vs PTB, PTB vs TL, pPROM vs TL, and TNIL vs TL (P < 0.05). Consistently, in all comparisons, the top three biological functions are cellular movement, cell-to-cell signaling, and interaction and inflammatory response. Cell death and survival is also a top 10 biological function for all comparisons. A heatmap was created for the biologic functions of differentially expressed proteins between pPROM vs PTB, pPROM vs TL, PTB vs TL, TNIL vs TL, PTB vs TNIL, and pPROM vs TNIL (Fig. 6b). Interestingly, downregulation of the synthesis of eicosanoids was noted in all comparisons aside from TNIL vs TL.

Figure 5.

Ingenuity Pathway Analysis–determined top 10 canonical pathways for pPROM vs PTB, PTB vs TL, pPROM vs TL, and TNIL vs TL (P < 0.05). pPROM vs PTB and TL, respectively, revealed positive z scores for liver X receptor/retinoid X receptor (LXR/RXR) activation, complement system, and production of nitric oxide and reactive oxygen species in macrophage pathways. The PTB vs TL z scores were positive to a lesser extent for LXR/RXR activation and production of nitric oxide and reactive oxygen species in macrophage pathways, and the complement system pathway had a z score of 0. This reflects more similarities between PTB and TL than pPROM. Spontaneous PTB (n = 13), pPROM (n = 8), TL (n = 11), and TNIL (n = 13).

Figure 6.

(a) Ingenuity Pathway Analysis–determined top 10 significantly different (P < 0.05) biological functions. In all comparisons, the top three biological functions are cellular movement, cell-to-cell signaling, and interaction and inflammatory response. (b) Heatmap represents upregulation (blue, positive z score) or downregulation (red, negative z score) of biological functions for all comparisons. Again, cellular movement, cell-to-cell signaling, and interaction and inflammatory response are represented. Of note, downregulation of the synthesis of eicosanoids was noted in all comparisons aside from TNIL vs TL. Spontaneous PTB (n = 13), pPROM (n = 8), TL (n = 11), and TNIL (n = 13).

Discussion

In pregnancy, exosomes are involved in embryogenesis, placental development, and maintenance of pregnancy as mediators of paracrine communication (43–45). Although several pregnancy-associated maternal biologic fluid exosome characterization studies have been reported, to our knowledge, no studies on AF exosomes to determine if their quantity, cargo, and character (representing biological pathways) differ between term and preterm labor have been performed. This study examined the AF exosomes associated with PTB, pPROM, TL, and TNIL. Comparisons of the AF exosomal protein signature between these four different groups of samples revealed unique differences between groups. Although biobanked samples were used for this report, we did not observe any variation in size, shape, or characters and any major differences in the number of proteins in exosomes from various groups.

AF collected at the time of PTB, pPROM, TL, and TNIL gives us samples representative of the fetal environment during abnormal parturition, normal parturition, and nonlabor, respectively. Interestingly, the total number of exosomes does not change in relation to these conditions. This finding is consistent with prior reports in preeclamptic and PTB pregnancies (29). When comparing preterm samples, the PLAP/CD63 ratio was higher in AF exosomes in PTB compared with pPROM. Although PLAP^+ve exosomes are generally referred to as placenta specific, we disagree with this argument, because PLAP^+ve exosomes are released by amnion, chorion, maternal decidua, and myometrium (unpublished report). It is speculative that involvement of PLAP^+ve exosomes may be indicative of a specific role of these exosomes in PTB, in contrast to pPROM. In addition, PTB and TL samples showed no differences in the PLAP/CD63 ratio, suggesting similarities between labor and delivery at both preterm and term.

Numerous differentially expressed proteins were discovered when the four groups were compared. These differentially expressed proteins are representative of inflammation and immune response. Specifically, pregnancy zone protein, which has been found to be upregulated in pregnancy and other inflammatory states (46), was significantly higher in pPROM and PTB than in TL. Interestingly, pregnancy zone protein was higher in pPROM than in PTB. This may be indicative of the extent of inflammation involved in the respective conditions. Ras-related proteins were also differentially expressed between PTB, pPROM, and TL. In fetal membranes, Ras-GTPase activation has been suggested to be a result of oxidative stress and DNA damage, leading to inflammation and PTB (47). Our findings are not surprising; as stated earlier, parturition either at term or preterm is an inflammatory process, and both fetomaternal cells can contribute to this process. Therefore, exosomes reflect the precise physiologic state of gestational tissues.

We would like to highlight some unique canonical pathways and biologic functions represented in exosomes. Specifically, pPROM AF exosomes revealed an activation of LXR/RXR and production of nitric oxide and reactive oxygen species in macrophage pathways compared with PTB and TL. These pathways are involved in inflammation and oxidative stress, and they are known findings in pregnancies complicated by preeclampsia and pPROM (14, 47–50). Approximately 80% of pPROM pregnancies are associated with inflammation (sterile or infectious) of the uterine cavity, and a large number of them have histologic chorioamnionitis (51). Both of these conditions are known to cause leukocyte inflammatory activation or produce reactive oxygen radicals. Similarly, the complement system is a component of the innate immune response and is a key mediator of the inflammatory response. Dysregulation of complement has been implicated in adverse pregnancy outcomes such as preeclampsia and PTB (52). Complement activation and subsequent collapse of the balanced uterine immune system is a feature of TL. This is reflected in the AF exosome profile in our study; TNIL AF exosomes demonstrated a downregulation of the complement system compared with TL. In addition, eicosanoids identified in exosomes have been proposed to play a role in PTB (53).

Fetal membrane cell aging and cell death are well-documented biological processes associated with PTB and pPROM (10, 47, 50, 54–56). Therefore, it is logical that four of the top 10 biological functions in AF exosomes from pPROM, PTB, TL, and TNIL also showed cellular movement, cell-to-cell signaling and interaction, inflammatory response, and cell death and survival. Thus, the proteomic signature of AF exosomes shows common and unique biologic functions between each group.

In summary, we show that AF exosomes uniquely represent specific conditions, and their proteomic contents allowed us to cluster various pregnancy phenotypes (Fig. 3). Although the sample sizes used were limited due to the stringency of inclusion criteria and availability of AF samples (a major limitation of this study), we were able to produce a profile that is unique and reflective of underlying physiologic and pathophysiologic processes during normal and abnormal pregnancies. However, a stratified analysis to delineate pathways based on various risks associated with PTB and pPROM (e.g., group B Streptococcus colonization, chorioamnionitis, and length of latency with pPROM) was not attempted due to a limited sample size. Similarly, fetal sex may also contribute to biological variability in exosome cargo, and we have not addressed this in our study.

Conclusions

Intrauterine tissues generate various types of signals to facilitate communication between the fetus and mother. Endocrine signals are well known and well characterized. Unknown are paracrine signalers, and herein we project exosomes in AF as a potential source of communication reflecting uterine status. Generalized inflammation in preterm and term parturition and unique biologic functions associated with specific groups seen in AF exosomes reflect how these extracellular vesicles may function as either a biomarker or a carrier of specific communication. Although assessment of functional properties of these exosomes is beyond the scope of this report, we speculate that AF exosome cargo likely reflects their potential functional role at a target site, determining pregnancy outcome. The findings from our study clearly demonstrate that the quantity of uterine-derived exosomes and protein content of exosomes derived from AF provide information specific to normal and abnormal parturition. Future research could focus on the source of these exosomes, as well as their destination and functional significance. Finally, it is also possible that this new information will provide the foundation for the development of novel therapies for PTB in the future.

Glossary

Abbreviations:

AF

amniotic fluid

ANOVA

analysis of variance

FDR

false discovery rate

LXR/RXR

liver X receptor/retinoid X receptor

NTA

nanoparticle tracking analysis

PBS

phosphate-buffered saline

PLAP

placental alkaline phosphatase

PLAP^+ve

placental alkaline phosphatase–positive

pPROM

preterm premature rupture of membranes

PTB

preterm birth

RRID

Research Resource Identifier

SWATH

Sequential Window Acquisition of All Theoretical

TL

term labor

TNIL

term not in labor

v/v

volume-to-volume ratio