Human amniotic membrane as newly identified source of amniotic fluid pulmonary surfactant

Pulmonary surfactant (PS) reduces surface tension at the air-liquid interface in the alveolar epithelium of the lung, which is required for breathing and for the pulmonary maturity of the developing foetus. However, the origin of PS had never been thoroughly investigated, although it was assumed to be secreted from the foetal developing lung. Human amniotic membrane (hAM), particularly its epithelial cell layer, composes the amniotic sac enclosing the amniotic fluid. In this study, we therefore aimed to investigate a potential contribution of the cellular components of the hAM to pulmonary surfactant found in amniotic fluid. We identified that cells within the native membrane contain lamellar bodies and express all four surfactant proteins as well as ABCA3. Lipidomic profiling by nanoESI – MS/MS revealed the presence of the essential lipid species as found in PS. Also, the biophysical activity of conditioned cell culture supernatant obtained from hAM was tested with captive bubble surfactometry. hAM supernatant showed the ability to reduce surface tension, similar to human PS obtained from bronchoalveolar lavage. This means that hAM produces the essential PS-associated components and can therefore contribute as second potential source of PS in amniotic fluid aside from the foetal lung.

Pulmonary surfactant (PS) is known to be produced and secreted by alveolar epithelial type II cells as a surface-active lipid-protein complex. Its function relies on minimizing surface tension at the air-liquid interface and therefore enabling the gas exchange in the lung by preventing the collapse of alveoli during expiration 1, 2 . The predominant component of PS are lipids (90%), mainly comprising dipalmitoylphosphatidylcholine (DPPC; 40%) and other phospholipids (40%). The latter ones include unsaturated PC species and around 15% of anionic phosphatidylglycerol (PG) and phosphatidylinositol (PI) 2, 3 . These lipids changed evolutionary to enhance surfactant activity 4 and to enable extreme minimization of surface tension 5 . Only a minor part of PS is composed of plasma proteins (5%) and surfactant-associated proteins (SP-A, B, C and D, 5%), which are playing an important role irrespective of their minor presence 6 .
In general, if the function of PS is disturbed or if it is present at insufficient amounts, the consequences can be fatal. Surfactant deficiency can occur in babies from immature lung development during foetal growth, leading to infant respiratory distress syndrome (IRDS). This occurs in approximately one per cent of all newborns and is the leading cause of death in preterm babies 7,8 . But also acquired PS failure, emerging from acute lung injury, can lead to acute respiratory distress syndrome (ARDS), chronic lung diseases and even lung failure, also in the adult 9 .
Therapeutic options to improve complications associated with these diseases are limited. One therapeutic strategy is the exogenous application of surfactant formulations, such as synthetically produced or animal-derived surfactants. This can increase the chances of survival 8,10,11 , although beneficial effects in ARDS are limited 12,13 .
During foetal development, PS does not form until the end of gestation, which is the reason why insufficient surfactant is a major problem in preterm neonates 4 . An indication of the foetal pulmonary maturity is the examination of amniotic fluid for levels of surfactant 14,15 . However, so far the origin of this surfactant had not been ultimately determined, though it was supposed to be secreted by the foetal developing lung 16 . The amniotic fluid is surrounded by the human amniotic membrane (hAM), the innermost of the foetal membranes. hAM is of Isolation and cultivation of human amnion-derived epithelial (hAE) and mesenchymal stromal (hAMS) cells. hAE cells were isolated according to a protocol similar to Stadler et al. 27 . Briefly, hAM (placental and reflected) was cut into small pieces (3-4 × 2 cm²) and placed into digestion solution. hAE cells were segregated from the basal membrane during 3 digestion steps, each comprising 10-20 min, using 0.05-0.1% trypsin/ EDTA (Sigma-Aldrich, USA). Cell suspensions of the second and third digestion step were filtered through a 100 µm cell strainer (BD, Austria). Then they were centrifuged at 400 g for 9 min, and either directly fixed for flow cytometry or seeded at a density of 3 × 10 4 cells/cm² onto gelatine-coated glass cover slips (Carl Roth, Germany) in 24-well culture plates containing cell culture medium for immunofluorescence staining. After reaching confluence of approximately 80%, the cells were analysed.
The other cell type found in hAM, hAMS cells, were isolated by shaking in 0.1% collagenase (Biochrome, Germany) in DMEM (low glucose, Sigma-Aldrich, USA) for 2 hours in an incubator. After filtration through a 100 µm cell strainer, the cell suspension was centrifuged at 400 g for 9 min, the cells counted and afterwards fixed for flow cytometer analysis or cultured on glass cover slips until immunofluorescence staining.
Transmission electron microscopy. hAM explants directly taken after isolation were fixed at 4 °C for 2 hours in 2.5% glutaraldehyde in 0.1 mol/l sodium phosphate buffer, pH 7.4. The samples were then rinsed with the same buffer, postfixed in 2% osmium tetroxide in ddH 2 O, washed in ddH 2 O, dehydrated in a graded series of acetone, and embedded in Agar 100 resin. 70-nm sections were post-stained and examined with an FEI Morgagni 268D (FEI, Eindhoven, The Netherlands) operated at 80 kV. Images were acquired using an 11 megapixel Morada CCD camera (Olympus-SIS).
Scanning electron microscopy. hAM explants were washed in 1x PBS, then fixed in 2.5% glutaraldehyde in PBS (AppliChem, Germany) for 60 min at room temperature and afterwards dehydrated in an ascending ethanol series. After incubating in an ascending hexamethyldisilazane (HMDS, Sigma-Aldrich, USA) series, the membranes were air dried under the fume hood overnight. Then the biopsies were mounted onto double-sided sticky tape on top of an aluminium stub and sputter coated with Pd-Au using a Polaron SC7620 sputter coater (Quorum Technologies Ltd, GB). Afterwards they were analysed using a JEOL JSM-6510 scanning electron microscope (Jeol GmbH, Germany).
Nile red (NR) staining. NR specifically stains intracellular neutral lipids (450-500 nm excitation wavelength) and phospholipids (515-560 nm excitation wavelength). Stock solution of NR (1 mg/mL in EtOH absolute, Sigma-Aldrich, USA) was prepared and stored under light protection. Staining was performed on hAM explants prior fixed in 10% neutral buffered formalin (VWR, USA) for 30 min. Explants were incubated in NR staining solution for 15 min after dilution to a final concentration of 500 ng/mL and kept in PBS afterwards. Nuclei were counterstained with DAPI. Analysis was carried out with an Axio Observer.A1 microscope (Zeiss, Germany) immediately afterwards using the 450-500 nm and 515-560 band pass excitation filters to detect intracellular neutral and phospholipid droplets, respectively.
Immunohistological staining of hAM explants. Fixed hAM explants were rinsed in water and dehydrated in an ascending alcohol series. The tissue was embedded in paraffin and 4 µm thick sections were sliced in a rotary microtome. Sections were dried at 40 °C overnight, deparaffinised in xylene and rehydrated in a graded series of alcohol. Immunohistochemistry against pro-and mature surfactant protein B (SP-B) and prosurfactant protein C (pro-SP-C) was performed and sections were stained in Lab Vision Autostainer 360 (Thermo Fisher Scientific, UK). In the case of pro-SP-C, antibody retrieval was performed for 20 min at 95 °C using HIER T-EDTA pH 9.0 buffer (Zytomed, Germany) and for SP-B proteinase K (DAKO, Denmark) was used for 8 min. Afterwards, sections were washed in Tris-buffered saline and Tween (TBS, Thermo Fisher Scientific, UK). In SCIEnTIFIC RePoRtS | 7: 6406 | DOI:10.1038/s41598-017-06402-w order to deactivate endogenous peroxidase, sections were treated with UltraVision ™ Hydrogen Peroxide Block (Thermo Fisher Scientific, UK) for 10 min and rinsed with the same buffer again. Subsequently, sections were exposed to the primary antibody against SP-B (working dilution 1:200, Abcam, UK) and pro-SP-C (working dilution 1:500, Abcam, UK) for 1 hour at room temperature. Afterwards, sections were rinsed with TBS and incubated with EnVision ™ HRP anti-rabbit detection system (DAKO, Denmark) as a secondary antibody for 30 min at room temperature. Visualization was performed with ImmPACT ™ NovaRED ™ peroxidase substrate kit (Vector Laboratories, USA) and sections were counterstained with haematoxylin. Finally, slides were dehydrated and mounted with permanent mounting medium.
Flow cytometric analysis of hAM-derived cells. hAE and hAMS cells were fixed immediately after isolation with 1% formaldehyde in PBS for 10 min. After two washing steps in 1% BSA (Sigma Aldrich, USA) in PBS, cells were either prior permeabilised for 20 min in Permeabilization Buffer (Affymetrix, USA) for intracellular staining or directly incubated with the respective primary antibodies for 45 min. Cells were washed and incubated with polyclonal Alexa Fluor 488-conjugated goat anti-rabbit or Alexa Fluor RPE-conjugated goat anti-mouse IgG secondary antibodies (1:750, Thermo Fisher Scientific, USA) for 30 min. Isotype controls were incubated with rabbit and mouse IgG. After washing, cells were analysed on a flow cytometer (Cytoflex, Beckman Coulter, USA). SP-A, Oct-4 and pan-CK were obtained from Santa Cruz Biotechnology (USA) and were used in a dilution of 1:100, all the other antibodies (SP-B, pro-SP-C, ABCA3, SSEA-4) were purchased from Abcam (UK) and were diluted 1:250. Every incubation step was performed on ice.

Quantification of SP-D by ELISA.
Conditioned cell culture supernatant of hAM biopsy punches and isolated hAE cells was collected within the first week of culture (hAM) or after reaching 80% confluence (hAE cells). After three to four days without medium change, the supernatant was frozen and kept at −80 °C until analysis. Subsequent to centrifugation to remove cell debris, SP-D was measured by ELISA (Hycult Biotechnology, The Netherlands) following the manufacturer's instructions. As a positive control, bronchoalveolar lavage fluid (BALF) from male pigs was collected similar to Gushima et al. 28 . Briefly, 15 mL of sterile PBS was infused into each lung and gently aspirated. Afterwards the fluid was sterile filtered and centrifuged at 200 g for 10 min to remove debris and mucus clumps. BALF was stored at −80 °C until analysis and diluted 1:5 in PBS for ELISA.

Extraction of lipids for mass spectrometric (MS) analysis.
Cell culture supernatant of hAM explants was withdrawn and stored as described above. Extraction of lipids from hAM-derived supernatant and porcine BALF was carried out by liquid-liquid extraction using methyl-tert-butyl ether (MTBE) according to Matyash et al. 29 . For this, 400 µL of sample was mixed with 320 µL of methanol (VWR, USA) and 1 mL of MTBE (Honeywell Specialty Chemicals Seelze GmbH, Germany) and shaken for 1 h at 24 °C and 180 g. 250 µL of MS-grade water (Merck Millipore, USA) was added afterwards to induce phase separation. Following 10 min shaking at 24 °C and 180 g, centrifugation at 1500 g for 1 min was applied and the top organic layer carefully collected. Samples were then stored at −20 °C until MS analysis.
Qualitative lipid profiling using shotgun MS. The MTBE extracts derived from hAM supernatants were measured using a Triversa Nanomate (Advion, USA) coupled to a LTQ Velos Orbitrap mass spectrometer (Thermo Fisher Scientific, USA). 8 μL of sample were mixed with 80 μL of measurement mix (70% (v/v) methanol, 30% (v/v) chloroform, 7.5 mM ammonium acetate); chloroform and ammonium acetate were purchased from Sigma-Aldrich, USA. Triversa Nanomate settings included 0.3 psi as backpressure and 1.4 kV spray voltage with a pick-up volume of 5 µL. The mass spectrometric analysis was conducted in the negative ion mode with an m/z window from 350-1,000 and normalised collision energy applied for fragmentation of 35. Data-dependent acquisition was chosen along with an isolation width for the precursor ion of 1 m/z and a dynamic exclusion m/z width of previously fragmented peaks of 0.5 m/z. Only singly charged ions were allowed for fragmentation, which was accomplished by collision induced dissociation (CID). Porcine BALF was used as a positive control and for comparison. Data analysis of the acquired raw files was carried out applying the free software LipidXplorer (Version 1.2.7).
Quantitative analysis of lipids with supercritical fluid chromatography coupled with tandem mass spectrometry (SFC-MS/MS). The SFC analyses were carried out using the Agilent 1260 SFC System directly coupled to the electrospray ionisation (ESI) source followed by a 6490 triple quadrupole mass spectrometer (Agilent Technologies, USA). The separation was performed on a 150 × 3 mm Luna HILIC separation column (Phenomenex, UK) employing a flow rate of 2 mL/min. Gradient mode was used for elution of the phospholipids. As mobile phase A supercritical carbon dioxide (Messer, Austria) and as mobile phase B methanol with 0.1% formic acid and 0.05% ammonium formate (both from Sigma-Aldrich, USA) were used. The gradient was started at 5% mobile phase B and increased to 40% B in 6 min. Mobile phase constitution was kept constant at 40% B for 1 min and reduced to 5% at 7.1 min. The chromatographic run was finished after a total measurement time of 9 min. The backpressure was set to 150 bar and the column temperature to 40 °C. For the targeted mass spectrometric analysis of DPPC the multiple reaction monitoring (MRM) transitions and the collision energies were optimized during the experiments. For the final analysis as quantifier the transition from m/z 734.4 to 184.1 occurring from the head group was used and as qualifiers m/z 734.4 to 86.1 (N,N,N-trimethyletheneaminium), m/z 734.4 to 478.32 (cleavage from the SN2 and H 2 O) and m/z 734.4 to 551.3 (cleavage from the SN1). For all transitions, the collision energies of 40 eV were used. Positive ionization mode was applied and the ESI spray voltage was 4000 V, gas flow 11 L/min, sheath gas temperature 400 °C, sheath gas flow 12 L/min, gas temperature 250 °C, and nebulizer gas pressure 35 psi. For quantification, an external calibration was performed.
Biophysical activity assay: Captive Bubble Surfactometer (CBS). Supernatants of placental and reflected hAM explants were collected as described above to isolate PS-like complexes. Samples were concentrated by consecutive ultracentrifugation at 120,000 g until getting a phospholipid concentration of 10 mg/mL, quantified using a commercial kit (Spinreact, Spain). Then, surface activity of the different surfactant samples were tested using the CBS 30 .
The assay is based on introducing a small amount of surfactant sample (typically 0.5 µl) below the surface of an air bubble floating against an agarose cap located in a chamber containing a buffer solution (Tris 5 mM, NaCl 150 mM, 10% sucrose, pH 7) (Supplementary Fig. S1). The agarose cap is attached to a piston, which allows performing compression-expansion cycles into the chamber. The resulting changes in surface tension can be deduced from the changes in bubble shape ( Supplementary Fig. S1). First, the adsorption of a surfactant added to a clean air-liquid interface is recorded for 5 min. Then, the chamber is closed and the bubble is subjected to a quick expansion to register post-expansion adsorption for 5 min. Finally, the bubble is exposed to consecutive slow (quasi-static cycles) or fast (dynamic cycles) compression-expansion cycles (at 20 cycles/min) mimicking respiratory dynamics in the lung. The different parts of the CBS experiment are recorded with a camera focusing the bubble throughout the experiment. Surface tension values are calculated from bubble dimensions with a Data Analyses Software, according to a model published elsewhere 31 . PS isolated from BALF of porcine lungs (native surfactant, NS) was used as a reference of an optimal surface active agent.
Evaluation and statistical analysis. ELISA results and flow cytometric analysis as well as DPPC quantification were expressed as mean and S.D.
Data availability. The data generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Results
Amniotic epithelium ultrastructure differs between placental and reflected region. The ultrastructure of the epithelial layer of hAM from term placenta separated into placental and reflected region was analysed by transmission (TEM) and scanning electron microscopy (SEM) (Fig. 1). hAE cells from both subregions showed cell membrane protrusions covering the apical surface resembling microvilli, which are generally associated with secretion and typically found on PS-secreting pulmonary epithelial cells 32 . In general, the majority of cells from the placental region exhibited a columnar phenotype and a lower cell frequency on the membrane surface. In comparison, cells from the reflected area showed a dense and regular morphology, cuboidal appearance and a considerable amount of secretory vesicles inside the cytoplasm, indicating a strong secretory nature of the cells.

Distinct expression of PS-associated components in hAM explants.
To characterize hAM for intracellular lipid vesicles, explants from both amniotic regions were stained for lipid droplets with NR immediately after isolation. As shown in Fig. 2a, epithelial cells from the reflected area stained with great intensity also in the red spectrum suggesting a high content of phospholipid vesicles. hAE cells of the placental region were also homogenously stained, but at both wavelength ranges to a lower degree than cells of reflected amnion. The next step was to identify specific PS protein components within the cells of hAM. For this, histological sections of hAM explants were stained for pro-SP-C and SP-B (Fig. 2b). The majority of epithelial cells of the reflected and the placental amnion exhibited a distinct intracellular staining for both surfactant proteins. Furthermore, the membrane of the reflected area demonstrated SP-B positive parts distributed in the extracellular matrix. In contrast, fewer (pro-SP-C) or hardly any (SP-B) of the hAMS cells expressed PS proteins in both hAM subregions ( Supplementary Fig. S2). hAE cells are main source of PS proteins. In order to examine the cellular origin of PS proteins in hAM, hAE and hAMS cells were isolated from the two different subregions of the membrane and analysed for the production of surfactant proteins and ATP binding cassette subfamily A member 3 (ABCA3, Fig. 3a). This is a lamellar body membrane protein required for the formation of PS, which is expressed in alveolar epithelial type II cells. As summarized in Table 1, hAE cells analysed immediately after isolation via flow cytometry showed a high expression of three of the four surfactant proteins (SP-A, SP-B and pro SP-C) as well as ABCA3, whereas only a minor fraction of hAMS cells was positive for PS proteins (Supplementary Table S1). Fewer epithelial cells were positive for SP-D and almost none of the mesenchymal cells. In general, a higher percentage of hAE cells derived from reflected hAM expressed the surfactant-associated proteins, whereas cells from placental amnion showed higher percentages of cells positive for Oct-4 (8.1% ± 7.8 vs. 3.3% ± 2.8), a marker expressed in pluripotent stem cells. Almost all of the hAM-derived epithelial cells of both subregions expressed stage-specific embryonic antigen-4 (SSEA-4), in contrast to the mesenchymal cells, of which approximately half of the population was positively stained. The expression of SPs was further confirmed via immunofluorescent staining of cultured hAM-derived cells (Fig. 3b, Supplementary Fig. S3).
SCIEnTIFIC RePoRtS | 7: 6406 | DOI:10.1038/s41598-017-06402-w SP-D is actively secreted by hAM. To confirm the potential of hAM to actively secrete PS-associated proteins, the conditioned supernatant of both isolated placental and reflected hAM explants was examined for SP-D by ELISA and compared with porcine BALF (Fig. 4a). SP-D was detectable in the supernatant of both subregions, but at considerably higher concentrations in reflected hAM-derived medium. Notably, SP-D concentration decreased moderately over time ( Supplementary Fig. S4a), but was still detectable after a three-week culture period and was also secreted from isolated and cultured hAE cells (Supplementary Fig. S4b). The expression of surfactant proteins was sustained in hAE cells over several passages (Supplementary Fig. S4c). Moreover, ultrastructural analysis of these cells exhibited lamellar granular vesicles, strongly resembling lamellar bodies, which are crucial in storage and secretion of PS in the lung 33 (Fig. 4b). , also referred to as DPPC, is the most important phospholipid species for biophysical function of PS 5 . While a plethora of other signals associated with lipids were detected on the MS1 level, these four could also be identified on the MS2 level with satisfactory confidence (Fig. 5).

Quantitative DPPC analysis of hAM-derived supernatants. DPPC as important player for surfactant
functionality was examined on a quantitative level. Therefore, a targeted SFC-MS/MS method was developed for the exact quantitative analysis of DPPC. While DPPC could be found in all supernatants derived from hAM explants at concentrations ranging from 7.5 to 10.3 µg/mL, no differences in DPPC levels were detected between supernatants from the two distinct amniotic subregions (Fig. 6).

PS-like material secreted by hAM shows limited functional activity. Lung surfactant is synthesized
to reduce surface tension along respiratory cycles allowing respiratory dynamics and avoiding alveoli collapse. Because of that, biophysical properties of tissue culture supernatants derived from placental and reflected hAM explants were tested using CBS. Supernatants of both amniotic subregions did not reduce surface tension below 50 mN/m once injected to an air-liquid interface (Fig. 7a). This behaviour was not improved after subjecting the system to a quick expansion (Fig. 7b). However, compression of the surface films reduced surface tension down to around 20 mN/m, which required more than 60% area reduction, either under slow (quasistatic, Fig. 7c) or fast (dynamic "breathing-like", Fig. 7d) compression-expansion cycles. Surface activity did not improve to yield lower activity after repetitive compression-expansion cycling of the films. No obvious differences in surface activity were observed between the supernatants of placental and reflected hAM.

Discussion
hAM, as the innermost of the foetal membranes, encloses the growing foetus and the amniotic fluid, in which PS, consisting mainly of lipids and specific proteins, can be found as an indication of the foetal pulmonary maturity 14,15 . As origin of this amniotic fluid-derived PS, the developing foetal lung has been reported but not been   Table 1. Flow cytometric analysis of hAE cells immediately after isolation: Percentages of hAE cells isolated from the placental and reflected hAM region, positive for the respective marker. Illustrated are mean ± S.D. of given biological replicates. exclusively defined so far. We have therefore investigated hAM and cells thereof as potential additional source of PS, as the presence of the two hydrophilic proteins SP-A and SP-D in hAM was already described 34 . For these proteins expressed in hAM, an immunoregulatory and anti-pathogenic function has been proposed 35 , however no direct relation to the amniotic fluid PS. We identified that cells within the native hAM of term placenta express all four surfactant-associated proteins in both amniotic subregions, which have crucial functions regarding surfactant: Aside from its biophysical properties lowering surface tension, surfactant also has an important function in pulmonary host defence mediated by SP-A and SP-D. These two hydrophilic proteins are involved in clearance of bacteria, fungi, dying cells and viruses through binding, thus initiating phagocytosis; also, they stimulate chemotaxis and have an anti-inflammatory effect 36,37 . SP-B and SP-C are hydrophobic proteins required for the proper biophysical function and consequently indispensable for breathing 4,38,39 . Lack of functional SP-B, such as for genetic reasons, leads to respiratory failure unless lung transplantation is performed 40,41 . SP-C deficiency is not as critical, but leads to a decreased stabilization of surfactant and affected patients usually develop severe and chronic lung diseases 42 .   43,44 . While SP-A and SP-D are also present in other organs 45,46 , SP-C is found exclusively in surfactant-producing cells 47 , which is a clear indication of a relation between hAM and the surfactant found in amniotic fluid. Moreover, we identified lamellar body-like structures, which are crucial for PS storage and secretion 48 and the expression of ABCA3, required for the lamellar body biogenesis 49 , in amnion-derived cells. Also, secretory capability was demonstrated via ELISA, for which SP-D could be identified in hAM-derived cell culture supernatant over several weeks. These results show that foetus-derived PS proteins do not only adsorb on hAM, but demonstrates that they are actively produced and secreted by hAM cells.
The main part of PS consists of lipids; therefore, we analysed hAM-derived supernatant for lipids and compared it with porcine BALF. We identified four lipids in supernatant of both placental and reflected hAM explants, hence representing mostly the same lipid class constitution as BALF. The lipids included phosphatidylcholine, ethanolamine plasmalogen, sphingomyelin and phosphatidylserine, with the exception of phosphatidylglycerol, which was not identified in the hAM-derived surfactant-like material. Yet the phospholipid class, especially DPPC, is responsible for the extreme reduction of surface tension at the air-water interface 5 , a lipid that was definitely identified as major component in our hAM-derived supernatants. However, proper PS function is also conferred by their protein components 5 rendering natural PS usually more efficient compared to most of the synthetically-produced ones 50 .
Native surfactant has the ability to extremely reduce surface tension at the air-liquid interface at the end of expiration in the lung 2, 3 . We tested the biophysical activity of our hAM-derived surfactant-like material with a captive bubble surfactometer. Although we could not reach the level of reduction of surface tension during initial and post-expansion adsorption as native surfactant 5 , we could demonstrate a certain functional activity of the hAM-derived surfactant-like material during quasistatic and dynamic cycling. It resembled the biophysical behaviour obtained with human non-bronchoscopic BALF 51 . This indicates at least some functional activity of the surfactant components secreted by hAM. One explanation for the lower biophysical activity of hAM-derived supernatant could be associated with the exposure of secreted surfactant to a medium enriched with FBS. Serum proteins, such as albumin, are potent PS inhibitors 52, 53 , which could have limited the functionality of the PS secreted by hAM. Moreover, the stimulation of secretion with β-adrenergic agents or secretagoges, such as ATP (adenosine triphosphate) and PMA (phorbol 12-mirystate 13-acetate) might be essential to induce substantial levels of PS secretion from hAM, as the PS exocytosis is tightly regulated 54 . This could also be the reason why we could not detect any significant differences in biophysical activity and DPPC concentration between the supernatants from the two subregions of hAM. Yet, a distinctly higher amount of lipid droplets was found in the cytoplasm of cells of reflected hAM compared to those of the placental region. Stimulation of hAM could increase secretion of lipid vesicles and consequently disclose a higher surfactant lipid concentration in reflected amnion-derived supernatant. This might then potentially lead to an improved PS function, for which a high phospholipid, most of all DPPC, concentration is required. Such experiments, including a direct extraction of PS from homogenized or secretion-stimulated hAM, could be part of future studies. Recently a difference in mitochondrial activity 55 as well as in gene expression 26 between the two subregions of hAM was demonstrated. Within this study, we identified an overall higher percentage of cells expressing the pluripotency marker Oct-4 in placental hAM-derived cells compared to those of the reflected area. Although not really relevant in regard to PS, this result might be interesting for differentiation and stem cell studies.
Cells isolated from hAM have previously been considered as a potential cell therapy for human lung diseases. For both hAE as well as mesenchymal decidual cells, the development of an alveolar epithelial cell phenotype, secreting PS proteins upon in vitro differentiation, has been reported 21, 54 . Moodley et al. did however not find these characteristics in undifferentiated hAE cells 21 . Moreover, the cells have proven promising in several pulmonary pathologic conditions in vivo without prior differentiation. They reduced inflammation and fibrosis and partially restored the physiological lung function after bleomycin-induced lung injury in mice 21, 56 SCIEnTIFIC RePoRtS | 7: 6406 | DOI:10.1038/s41598-017-06402-w or after inflammation-or ventilation induced lung injury in sheep 57,58 and resulted in an engraftment into the lung in some cases 21,58 . The reason for these beneficial effects in those studies could at least partially be hAM's ability to produce surfactant-like material. In vitro differentiation of hAM or hAM-derived cells similar to Moodley et al. 21 towards an alveolar epithelial cell type might be helpful and increase surfactant production. Both, isolated amniotic cells 59 , as well as the whole membrane 60,61 , have already been transplanted in human patients for diverse therapeutic reasons, without eliciting major adverse immune reactions. Therefore, the application of hAM and hAM-derived cells is considered to be safe, despite an allogeneic usage 62 . Formulations derived from human amnion, such as a PS-enriched secretome, may represent a novel therapeutic approach for surfactant-related diseases. These could include respiratory distress syndrome, asthma or surfactant metabolism dysfunction, as potential therapeutic targets. However, substantial improvements, including an increase of PS concentration and a potential enhancement of the biophysical activity of hAM-derived PS, would be a prerequisite.
We showed that hAM-derived cells from term placenta already produce and secrete components similar to PS secreted by alveolar epithelial type II cells in the lung. All in all, this study demonstrates that hAM from term placenta has the potential to serve at least as a second source of PS in amniotic fluid aside from foetal lung. The functional relevance, such as a possible role of hAM in foetal lung maturation or in pulmonary host defence of the growing fetus, remains however to be determined.