Quantitative assessment of placental morphology may identify specific causes of stillbirth

To address the first hypothesis we obtained placental tissue from cases of stillbirth, defined as the birth of an infant with no signs of life after 24 weeks gestation. Parents gave permission for the use of samples for research at the time of consent for post-mortem examination. A favourable ethical opinion was given by the Greater Manchester South Research Ethics Committee (09/H1012/11) and approval given from the Research and Innovation Division of Central Manchester University Hospitals NHS Foundation Trust to conduct the study. Cases of stillbirth were classified using the ReCoDe system [15] by a multidisciplinary meeting following a full panel of investigations including: post-mortem, histopathological examination of the placenta, chromosomal analysis and maternal biochemical, haematological, immunological and serological tests. We obtained samples from the following classifications of stillbirth: cord accident (n = 8), diabetes (n = 5), FGR (n = 10), hypertension (n = 8), infection (n = 8) and from stillbirths of an unknown cause (n = 10). For comparison, matched placental samples were used from appropriately-grown live born infants and preterm births (26-36 weeks) (demographics shown in Table 1). Samples from live births were collected following written informed consent as part of the Maternal and Fetal Research Centre (MFHRC) Biobank (08/H1010/55). To address the second hypothesis, placental samples were taken from a further cohort of stillbirths attributed to FGR (n = 13) and from live births with FGR (n = 13) with the same ethical approvals as described above. FGR was defined as a customised birthweight <5th centile (demographics shown in Table 2). To address the final hypothesis, placental tissue was obtained from appropriately grown (n = 7) and FGR (n = 5) live births following written informed consent as part of the MFHRC biobank previously described. For all cases, maternal and infant demographic information was recorded from medical case notes and post-mortem reports (for stillbirths). An estimate of the duration of in utero retention was made according to Genests’ descriptions of findings at post-mortem and from histopathological examination of the placenta [21‐23].

Tissue from live births was obtained within 30 min of delivery. For assessment of placental morphology biopsies of villous tissue were dissected from the centre, middle and edge of the placenta. Tissue was fixed in 4 % neutral buffered formalin for 24 h before being wax embedded. For stillbirth samples three blocks of placental tissue not obtained from specific lesions were obtained for each placenta.

Unless otherwise stated, all reagents were supplied by Sigma-Aldrich Chemical Company (Poole, UK). To examine the acute impact of continued maternal blood flow in the absence of fetal blood flow single-sided (maternal) ex vivo human placental lobule perfusion was adapted from the dual-sided perfusion model [24]. Perfusion was performed on placentas from normal pregnancy (n = 7) and placentas from pregnancies complicated by FGR (n = 5). An intact peripheral lobule was selected, devoid of post-partum tears, deep decidual damage and marginal membrane separations. Prior to perfusion two villous biopsies were sampled from neighbouring lobules taken 5 cm apart, and fixed immediately in 4 % neutral buffered formalin forming “pre-perfusion samples”. The fetal artery and vein on the chorionic surface, serving the villous trees within the lobule designated for perfusion, were each ligated using sutures (Mersilk 3/0, Ethicon, supplied by NuCare, UK) to confine a static fetal blood pool within the fetal vasculature of the associated cotyledons. The maternal surface was cannulated using five 10 cm lengths of polythene tubing (Smiths Medical, UK) arising from a perfusion manifold (Harvard Apparatus, UK). The distal ends of the cannulae were cut into apices and inserted through the decidual surface of the lobule with an even spatial distribution. The perfusate was modified Earle's bicarbonate buffer (EBB 117 mM NaCl, 10.7 mM KCl, 5.6 mM D-glucose, 3.6 mM CaCl, 1.8 mM NaH₂PO₄, 13.6 mM NaHCO₃, 0.04 mM L-arginine, 0.8 mM MgSO₄, 3.5 % (w/v) dextran, 0.1 % (w/v) bovine serum albumin, 5000 IU/L Heparin sodium) equilibrated with 95 % O₂ / 5 % CO₂ to pH 7.4 and warmed to 37 °C, delivered by a roller pump (Watson Marlow, UK) at 14 ml/min. Lobule preparations were only considered acceptable for experimentation when maternal-side perfusion was established within 30 min of delivery. Open-circuit perfusion was for 6 h, and then the physiological buffer was switched to a 4 % neutral buffered formalin at T = 6 h for a 10 min maternal-side perfusion fixation period. Following this, the lobule was excised and two further full thickness (vertical and horizontal) biopsies slices were taken as the “post-perfusion samples”. These wide tissue sections where then immersion fixed in 4 % neutral buffered formalin for 24 h before being wax embedded. Placental structure in these biopsies was examined as described above.

Placental cell turnover, structure and vascularity were assessed using antibodies specific for Ki67 (Dako, Ely, Cambridgeshire, UK; 0.16 μg/ml), cytokeratin 7 (Dako; 0.9 μg/ml) and CD31 (Dako; 0.16 μg/ml). The number of leukocytes was assessed by an antibody specific for CD45 (Dako; 0.4 μg/ml). Negative controls were performed using non-immune mouse IgG (Dako) at matching concentrations to the primary antibody. Immunohistochemistry was performed as previously described with antigen retrieval performed by microwaving the sections for 10 min in 0.01 M sodium citrate buffer [19, 20].

Quantification of syncytial nuclear aggregates (SNAs, also known as syncytial knots) was conducted on sections stained with haematoxylin and eosin as previously described [19, 20]. Dewaxed and rehydrated sections were stained with Harris’s haematoxylin for 10 min before differentiation in acid-alcohol. Slides were stained with eosin for 2 min, rinsed in cold tap water, and dehydrated and mounted as described above.

For all analyses images were captured using an Olympus BX41 light microscope (Southend-on-Sea, UK) and QIcam Fast 1394 (QImaging, BC, Canada) and Image Pro Plus 6.0 and 7.0 (Media Cybernetics Inc., MD, USA). During image acquisition and analysis the presumed cause of stillbirth was concealed from the observer. Between images the microscope was taken out of focus to prevent selection bias, if the randomly selected image was not mostly of terminal villi another area was identified. Five random images of terminal villi were taken of each section, giving a total of 15 images per placenta for each component evaluated.

The number of SNAs were counted and total villous area measured using image analysis software, expressed as the number of SNAs per mm² of villous tissue as previously described [25]. Proliferative index was the number of Ki67 positive nuclei as a proportion of total nuclei as previously described [19, 20]. Vascularity was expressed as the number of capillaries per terminal villus and the percentage of avascular villi (defined as a villus with no evidence of CD-31 immunostaining or morphological evidence of vessels) [19, 20]. Trophoblast area was expressed as the proportion of villous area positive for CK-7 immunostaining. The number of leukocytes was assessed by the number of CD45 positive cells per 1,000 nuclei.

For comparison of different causes of stillbirth data from each variable was compared to the median level in healthy controls using Wilcoxon signed rank test. Cases of FGR that were live-born were compared to those who were stillborn using Mann-Whitney U test. Data from pre- and post-perfusion samples were compared using Wilcoxon matched-pairs test. Demographic variables were compared using Kruskal-Wallis test with Dunn’s post-hoc test for multiple comparisons and Mann-Whitney U test for single comparisons. For all statistical tests a p-value of 0.05 was considered to be statistically significant. All statistical analyses were carried out using GraphPad PRISM (Version 6, La Jolla, CA).