Stereology

Stereology enables 3D interpretation of 2D sections of tissue. In essence, random systematic sampling and measurement of tissue cross-sections can provide unbiased and quantitative data of 3D structures such as volume, surface area, length and number. Undoubtedly, stereology has been a prevalent imaging technique to characterise placental structure, particularly in the rodent. Stereology has been used to assess placental morphology in humans, ^{Reference Higgins, Felle, Mooney, Bannigan and McAuliffe42–Reference Jackson, Mayhew and Boyd47} mice, ^{Reference Wyrwoll, Noble and Thomson18,Reference Nuzzo, Camm and Sferruzzi-Perri48–Reference Coan, Ferguson-Smith and Burton61} and rats. ^{Reference Napso, Hung, Davidge, Care and Sferruzzi-Perri62–Reference Hewitt, Mark and Waddell64} Via unbiased randomised sampling of 2D histological images, structures of interest can be quantitated and mathematically extrapolated to infer 3D structure. In the placenta, stereology is commonly used not only to evaluate information of placental vascular length, diameter, surface area, volume and density ^{Reference Wyrwoll, Noble and Thomson18,Reference Higgins, Felle, Mooney, Bannigan and McAuliffe42–Reference Mayhew, Ohadike, Baker, Crocker, Mitchell and Ong44,Reference Wyrwoll, Seckl and Holmes54,Reference Serman, Zunic and Vrsaljko63,Reference Hewitt, Mark and Waddell64} but also to estimate its theoretical diffusion capacity by measuring capillary interhaemal membrane thickness and volumes of placental components, that is, labyrinth zone, spongiotrophoblast and decidual volumes. ^{Reference Coan, Vaughan and Sekita53–Reference Coan, Angiolini, Sandovici, Burton, Constancia and Fowden55,Reference Coan, Ferguson-Smith and Burton59,Reference Coan, Ferguson-Smith and Burton61}

Stereology has been implemented to assess placental structure in human pregnancy complicated by FGR, pre-eclampsia and diabetes. ^{Reference Higgins, Felle, Mooney, Bannigan and McAuliffe42–Reference Mayhew, Ohadike, Baker, Crocker, Mitchell and Ong44} FGR placentas exhibited thicker trophoblastic epithelium, a reduction in peripheral villous surface areas and a reduction in volume of peripheral villi and intervillous space. ^{Reference Mayhew, Manwani, Ohadike, Wijesekara and Baker43} In contrast, there were no significant differences in the volume of intervillous space and all types of villi in pre-eclampsia, when compared to normal pregnancy. ^{Reference Mayhew, Manwani, Ohadike, Wijesekara and Baker43,Reference Mayhew, Ohadike, Baker, Crocker, Mitchell and Ong44} These changes support the notion that feto-placental vascular complexity is an important determinant of fetal growth. Higgins and colleagues utilised stereological analysis to examine human placentas from diabetic pregnancies which exhibited increased terminal villous volume, capillary volume and capillary length. ^{Reference Higgins, Felle, Mooney, Bannigan and McAuliffe42} Interestingly, maternal gestational hyperglycaemia affects capillary, but not stromal, development, and these changes may increase specifically the surface areas for maternal-fetal exchange and potentially represent a structural mechanism which accounts for the overgrowth of fetus and abnormally high birth weight. Alternatively, hyperglycaemia elicits excessive insulin secretion which promotes abnormal vascular proliferation, causing downstream abnormalities. Overall, these findings suggest differential condition-specific effects on placental integrity: FGR (but not pre-eclampsia) leads to reduction in structural size and complexity, while gestational diabetes leads to changes in terminal structures.

The early elegant stereological assessments by Coan and colleagues established the developmental dynamics of C57BL/6J mouse placental morphology. ^{Reference Coan, Ferguson-Smith and Burton59,Reference Coan, Ferguson-Smith and Burton61} Maternal blood spaces were shown to expand rapidly between E14.5 and 16.5, whereas fetal capillary development continuously increased in volume and surface area right up to late gestation. ^{Reference Coan, Ferguson-Smith and Burton61} These findings highlight the importance for feto-placental capillary expansion to meet increased fetal demand for nutrients in late gestation when fetal growth is rapid. These initial assessments have been extended to the comparison of different mouse strains, revealing differences in feto-placental structure including growth of feto-placental capillaries and interhaemal membrane thickness, ^{Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52} an important consideration when selecting experimental models. In well-established mouse models of developmental origins of health and disease (DOHaD) such as maternal protein restriction ^{Reference Rutland, Latunde-Dada, Thorpe, Plant, Langley-Evans and Leach58} and maternal undernutrition, ^{Reference Coan, Vaughan and Sekita53} stereology has demonstrated changes in placental vascular morphology. Moreover, stereological investigations of knockout-mouse models have given important insights into the dynamic nature of the placental function, especially into the co-ordination of maternal and fetal signals to in part compensate for an adverse environment in utero. ^{Reference Nuzzo, Camm and Sferruzzi-Perri48–Reference Sferruzzi-Perri, Lopez-Tello, Fowden and Constancia50,Reference Coan, Fowden, Constancia, Ferguson-Smith, Burton and Sibley57,Reference Sibley, Coan and Ferguson-Smith60,Reference Napso, Hung, Davidge, Care and Sferruzzi-Perri62} Placental structural development in rats has also been examined by stereology and revealed greater feto-placental complexity in comparison to mice. ^{Reference Serman, Zunic and Vrsaljko63,Reference Hewitt, Mark and Waddell64} In a rat model of glucocorticoid (‘stress’ hormone) excess in pregnancy, the reduction in fetal weight is accompanied by markedly reduced placental weight and feto-placental vascular volume, length and diameter, although not maternal blood spaces. ^{Reference Hewitt, Mark and Waddell64} Similar stereological outcomes have been obtained in other mouse studies of glucocorticoid excess ^{Reference Vaughan, Sferruzzi-Perri, Coan and Fowden51} including 11β-hydroxysteroid dehydrogenase 2 knockout (11β-HSD2^-/-) mice which also exhibit reduced cardiac function. ^{Reference Wyrwoll, Noble and Thomson18,Reference Wyrwoll, Seckl and Holmes54} Interestingly, administration of pravastatin to stimulate placental vascular growth factor A and enhance feto-placental angiogenesis in 11β-HSD2^-/- fetuses also ameliorated fetal cardiac function, ^{Reference Wyrwoll, Noble and Thomson18} highlighting the likely direct influence of feto-placental vascular structure on fetal organ development. ^{Reference Wyrwoll, Noble and Thomson18,Reference Wyrwoll, Seckl and Holmes54} However, none of the studies mentioned have shown a conclusive role of feto-placental vasculature directly informing adult cardiovascular health.

Stereology is a well-established and low-cost method, albeit laborious, with the advantage of encompassing the entire placental vasculature. Stereology has also yielded important insights into placental vascular morphology in human, mouse and rat alike. However, stereology approaches necessitate dismembering the structural integrity and connectivity of placental vascular networks. Thus, this approach cannot be used to evaluate placental vascular tree topology. To overcome these limitations, high-resolution contrast-enhanced micro-CT (discussed below) may well replace stereology in years to come.

Electron microscopy

Electron microscopy is a 2D imaging technique which includes transmission electron microscopy (TEM) and scanning electron microscopy (SEM). TEM creates high-resolution and high-magnification images of fine structural characteristics by passing electron beams through tissues. However, TEM destroys the connectivity of the placental vascular networks due to the necessity of tissue-sectioning during sample preparation. SEM scans the surface of a tissue specimen and also generates a 2D image of a 3D surface via interaction of high energy electron beam with a prepared specimen. However, SEM destroys surrounding tissues due to the requirements of sample preparation and the high electron beams. In 1955, Wislocki and Dempsey implemented electron microscopy to investigate placental morphology in rats, ^{Reference Wislocki and Dempsey65} pigs ^{Reference Dempsey, Wislocki and Amoroso66} and humans. ^{Reference Wislocki and Dempsey67} Thereafter, electron microscopy has been used by other investigators to delineate fine detail in the placental vasculature of monkeys, ^{Reference Lee and Yeh68} minks, ^{Reference Krebs, Winther, Dantzer and Leiser69,Reference Pfarrer, Winther, Leiser and Dantzer70} horses, ^{Reference Macdonald, Chavatte and Fowden71} rodents ^{Reference Adamson, Lu and Whiteley40,Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52,Reference Coan, Ferguson-Smith and Burton59,Reference Detmar, Rennie and Whiteley72–Reference Cahill, Rennie and Hoggarth74} and humans. ^{Reference Lee and Yeh68,Reference Bobinski, Pielesz and Waksmanska75–Reference Cronqvist, Tannetta and Morgelin86} Demir and colleagues have used TEM to investigate the development of feto-placental villous trees during the early period of ectopic human pregnancy compared to normal pregnancy. ^{Reference Demir, Demir, Ustunel, Erbengi, Trak and Kaufmann80} A decade later, Coan and colleagues investigated the mouse placental exchange region, which included characterising cytoplasmic bridges attached to several areas of the interhaemal membrane. ^{Reference Coan, Ferguson-Smith and Burton59} This was the first study that combined TEM and SEM to determine ultrastructural characteristics of mouse placental development. ^{Reference Coan, Ferguson-Smith and Burton59} Additionally, TEM has contributed to the understanding of placental cellular mechanisms in gas exchange surrounding capillaries, ^{Reference Coan, Ferguson-Smith and Burton59,Reference Sooranna, Oteng-Ntim, Meah, Ryder and Bajoria85} as well as the interior of feto-placental endothelial cells. ^{Reference Cronqvist, Tannetta and Morgelin86} Moreover, TEM has been used to inspect the maldevelopment of the trophoblast and placental terminal villous compartment in human preterm growth-restricted ^{Reference Krebs, Macara, Leiser, Bowman, Greer and Kingdom77} and pre-eclamptic pregnancies, ^{Reference Abdel Salam, Alam, Ahmed and Al-Sherbeny87} illustrating irregular trophoblast surface and elongated villi in preterm intrauterine growth restriction, ^{Reference Krebs, Macara, Leiser, Bowman, Greer and Kingdom77} and excessive villous arborisation with attenuated villi in pre-eclampsia. ^{Reference Abdel Salam, Alam, Ahmed and Al-Sherbeny87}

Vascular corrosion casting of feto-placental vasculature can be visualised using SEM. ^{Reference Adamson, Lu and Whiteley40,Reference Whiteley, Pfarrer and Adamson73,Reference Leiser, Krebs, Ebert and Dantzer84} Briefly, a vascular cast is generated with a polymer (i.e. Batson’s casting compound) and after polymerisation, the surrounding tissue is removed with a corrosive agent to reveal the surface of the vascular cast. ^{Reference Adamson, Lu and Whiteley40,Reference Whiteley, Pfarrer and Adamson73,Reference Leiser, Krebs, Ebert and Dantzer84,Reference Sricharoenvej, Tongpob, Lanlua, Piyawinijwong, Roongruangchai and Phoungpetchara88,Reference Bamroongwong, Somana, Rojananeungnit, Chunhabundit and Rattanachaikunsopon89} SEM has been utilised to reveal the intricate connectivity of feto-placental vasculature of healthy pregnancies ^{Reference Detmar, Rennie and Whiteley72,Reference Burton78,Reference Burton79} and complicated pregnancies such as FGR, whereby pronounced variations in placental surface were observed. ^{Reference Detmar, Rennie and Whiteley72,Reference Bobinski, Pielesz and Waksmanska75} SEM has also been used to provide anatomical comparison between humans and rodents, as well as a deeper understanding of how placental structure associates with fetal outcomes. ^{Reference Adamson, Lu and Whiteley40,Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52,Reference Detmar, Rennie and Whiteley72–Reference Krebs, Macara, Leiser, Bowman, Greer and Kingdom77} For example, SEM has revealed that the human placentas of FGR pregnancies exhibit abnormalities in their terminal villi compartments, including reductions in numbers of capillary loops, branches and coiled vessels. ^{Reference Krebs, Macara, Leiser, Bowman, Greer and Kingdom77} SEM has been applied to image the transition from arterioles to capillaries in the mouse feto-placental tree and revealed that expansion of the feto-placental vasculature in late gestation is strain dependent. ^{Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52} Moreover, SEM revealed that chronic hypoxia during pregnancy in mice caused expansion of the feto-placental capillary bed and thinning of the interhaemal membrane, ^{Reference Cahill, Rennie and Hoggarth74} possibly as a compensatory adaptation to ensure maternal-fetal exchange.

Even though electron microscopy has obvious advantages, especially visualising the feto-placental vasculature (particularly capillary and terminal villi) at extremely high resolutions and high magnifications, it only provides 2D images and also a very restricted field of view of a small segment over the entire vascular network. Moreover, each examination requires demanding sample preparation which destroys the connectivity of the placental vascular networks and the surrounding tissue. Other 3D imaging modalities such as confocal microscopy, light-sheet microscopy and micro-CT may overcome these limitations of electron microscopy.

Confocal microscopy

The development of 3D imaging techniques has altered the scope to study microvasculature with greater accuracy and precision. Confocal microscopy relies on point illumination; the technique utilises a confocal pinhole and a photomultiplier detector to exclude out-of-focus light surrounding the focal plane. This generates a series of 2D optical sections with enhanced contrast, precision and definition. These 2D optical sections can then be combined and reconstructed into a 3D representation which results in high-contrast and high-resolution images. ^{Reference Liu and Chiang90}

Confocal microscopy has been applied to human placentas to visualise the architecture of chorionic villous vascularisation in the first trimester, ^{Reference Lisman, van den Hoff, Boer, Bleker, van Groningen and Exalto91} second trimester ^{Reference Merz, Schwenk, Shah, Necaise and Salafia92} and at term. ^{Reference Palaiologou, Etter and Goggin81,Reference Cronqvist, Tannetta and Morgelin86,Reference Plitman Mayo, Abbas, Charnock-Jones, Burton and Marom93,Reference Perazzolo, Lewis and Sengers94} It has also been used to examine the spatial arrangement of the capillary bed. ^{Reference Jirkovska, Janacek, Kalab and Kubinova95,Reference Plitman Mayo, Charnock-Jones, Burton and Oyen96} Confocal has been utilised to investigate the distribution of villous capillaries in human diabetic placentas, including the surface and volume ratios between the fetal capillaries and the villus. It has also been implemented to provide a structural basis for computational modelling of oxygen flux in human placenta. ^{Reference Plitman Mayo, Charnock-Jones, Burton and Oyen96,Reference Jirkovska, Kucera and Kalab97} Recently, Sargent detailed methods of confocal microscopy in rhesus macaques placentas. ^{Reference Sargent, Roberts, Gaffney and Frias98}

Although confocal microscopy can be used to visualise the architecture of chorionic villous vascularisation, a major limitation is the maximum depth of focus (~1 mm³), ^{Reference Merz, Schwenk, Shah, Necaise and Salafia92} which means that a full-specimen coverage of the placental vasculature is currently impossible using this technique. Nonetheless, 3D confocal microscopy is still a useful approach for quantifying fine details of placental structure.

Micro-computed tomography

In recent decades, X-ray micro-CT has become a valuable and increasingly popular technique for non-destructive ex vivo imaging and has been used extensively to visualise placental vasculature in 3D, ^{Reference Junaid, Bradley, Lewis, Aplin and Johnstone6–Reference Langheinrich, Vorman and Seidenstucker8,Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52,Reference Detmar, Rennie and Whiteley72,Reference Cahill, Rennie and Hoggarth74,Reference De Clercq, Persoons and Napso99–Reference Langheinrich, Wienhard, Vormann, Hau, Bohle and Zygmunt111} as well as other organs including brain, ^{Reference Vasquez, Gao and Su112–Reference Chugh, Lerch and Yu114} heart, ^{Reference Vasquez, Gao and Su112,Reference Jorgensen, Demirkaya and Ritman115} lung, ^{Reference Yang, Yu, Rennie, Sled and Henkelman105,Reference Vasquez, Gao and Su112,Reference Kampschulte, Schneider and Litzlbauer116} liver ^{Reference Op Den Buijs, Bajzer and Ritman117–Reference Kline, Zamir and Ritman119} and kidney. ^{Reference Vasquez, Gao and Su112,Reference Jorgensen, Demirkaya and Ritman115,Reference Nordsletten, Blackett, Bentley, Ritman and Smith120–Reference Marxen, Thornton and Chiarot123} While micro-CT scanning is clearly suitable for imaging bone, blood vessels with low-inherent contrast properties cannot be directly visualised. To overcome this issue, a radio-opaque compound, such as Microfil® (Flow Tech Inc., Carver, MA, USA), Batson’s No.17 (Polysciences Inc., Warrington, PA, USA), BriteVu® (Scarlet Imaging, Murray, UT, USA) or barium sulphate, is perfused into blood vessels to heighten the contrast of the vessel lumen against surrounding tissues. This delineates the vascular structures for qualitative and quantitative analysis. ^{Reference Junaid, Bradley, Lewis, Aplin and Johnstone6,Reference Junaid, Brownbill, Chalmers, Johnstone and Aplin7,Reference Bentley, Ortiz, Ritman and Romero121} In recent years, this approach has been successfully applied to examine placental vascular integrity in mice, ^{Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52,Reference Detmar, Rennie and Whiteley72,Reference Cahill, Rennie and Hoggarth74,Reference De Clercq, Persoons and Napso99–Reference Conroy, Silver and Zhong106} rats ^{Reference Bappoo, Kelsey and Parker107,Reference Tongpob, Xia, Wyrwoll and Mehnert108} and humans. ^{Reference Junaid, Bradley, Lewis, Aplin and Johnstone6–Reference Langheinrich, Vorman and Seidenstucker8,Reference Pratt, Hutchinson and Melbourne109–Reference Langheinrich, Wienhard, Vormann, Hau, Bohle and Zygmunt111} Briefly, ex vivo micro-CT of vasculature consists of four main steps: (i) generation of vascular casts using a radio-opaque compound; (ii) ex vivo micro-CT scanning and 3D image reconstruction; (iii) vascular network segmentation and (iv) visualisation and characterisation of vascular network. ^{Reference Tongpob, Xia, Wyrwoll and Mehnert108} Importantly, the field of view using micro-CT is sufficiently large to capture the entire span of feto-placental vasculature in rodents and humans. Moreover, casting approaches for micro-CT can be modified to distinguish arterial and venous networks within the placenta, and the non-destructive nature of the imaging means the same sample can be preserved for follow-up histological assessments.

In mice, the innovative work of Rennie and colleagues using Microfil® and micro-CT has added considerable depth of understanding in the structural changes of feto-placental vasculature. ^{Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52,Reference Rennie, Cahill, Adamson and Sled100–Reference Rennie, Detmar and Whiteley102} The developmental course of mouse feto-placental vasculature has been captured, ^{Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52,Reference Rennie, Cahill, Adamson and Sled100–Reference Rennie, Detmar and Whiteley102} and comparison between mouse feto-placental arterial and venous trees has been drawn. ^{Reference Rennie, Cahill, Adamson and Sled100} Again, mouse strain differences have been revealed using this technique, such as differences in the expansion of the feto-placental arterial and capillary vasculature in late gestation. ^{Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52} Experimental models have also been characterised using this technique. As such, endothelial nitric oxide synthase deficiency reduced feto-placental vascular diameter and lengths in peripheral vessels. ^{Reference Rennie, Rahman, Whiteley, Sled and Adamson101} Also, maternal chronic hypoxic mice showed a significant reduction in the absolute volume of feto-placental arterial vasculature and the total number of vessel branches (−44% and −30%, respectively), when compared to controls. ^{Reference Cahill, Rennie and Hoggarth74} Moreover, environmental pollutant exposure polycyclic aromatic hydrocarbons in C57Bl/6J mice reduced the volume and surface area of the feto-placental arterial vasculature, ^{Reference Detmar, Rennie and Whiteley72} as well as resulting in significantly increased feto-placental arterial vascular tortuosity and reduced intra-placental arteriole vessels. ^{Reference Rennie, Detmar and Whiteley102} These alterations can cause reduced placental blood flow, leading to ineffective maternal-fetal exchange as well as potentially impaired fetal development and outcomes. Further investigations are required however to clarify these associations, as well as the consequences for adult health.

In humans, Pratt and colleagues have evaluated different methodologies in placental preparation and imaging. ^{Reference Pratt, Hutchinson and Melbourne109} They attempted to optimise the application of contrast agent, perfusion pressure, perfusion location and perfusion vessel in tissue preparation. ^{Reference Pratt, Hutchinson and Melbourne109} As such, they found the quality of the feto-placental cast from Microfil® more superior than barium sulphate as well as the contrast injection through umbilical artery more effective than via the umbilical vein or chorionic vessels. ^{Reference Pratt, Hutchinson and Melbourne109} Aughwane and colleagues recently investigated the degree of heterogeneity in human feto-placental vasculature using multiscale micro-CT and histological analysis. ^{Reference Aughwane, Schaaf and Hutchinson110} They found that the vascular density varied throughout placentas, and the patterns of vascular branching differed across a number of normal placentas. ^{Reference Aughwane, Schaaf and Hutchinson110} Langheinrich and colleagues have also analysed the vascular density and surface area of human feto-placental vasculature using micro-CT with barium sulphate perfusion. ^{Reference Langheinrich, Wienhard, Vormann, Hau, Bohle and Zygmunt111} Furthermore, they quantitated feto-placental vasculature in FGR and found arterial vascular volume in FGR was significantly lower than healthy controls. ^{Reference Langheinrich, Vorman and Seidenstucker8} Junaid and colleagues combined two techniques [corrosion casting (Batson’s No.17) and micro-CT] to image the whole human placental vascular morphology in FGR. ^{Reference Junaid, Bradley, Lewis, Aplin and Johnstone6,Reference Junaid, Brownbill, Chalmers, Johnstone and Aplin7} Notably, this study was the first to evaluate the whole human placenta. They detected an alteration of chorionic arterial branching patterns and a reduction in capillary network density in FGR placentas, ^{Reference Junaid, Brownbill, Chalmers, Johnstone and Aplin7} as well as shorter median vessel length density in arteries but longer median vessel length density in veins. ^{Reference Junaid, Bradley, Lewis, Aplin and Johnstone6}

We have recently established micro-CT methodology in our laboratory to characterise mouse and rat feto-placental vascular structures. ^{Reference Bappoo, Kelsey and Parker107,Reference Tongpob, Xia, Wyrwoll and Mehnert108} The qualitative assessments of mouse and rat feto-placental vascular structures at the end of gestation revealed striking species differences. Compared to mouse, rat feto-placental vasculature exhibits far greater branching complexity, vascular tree span and tree depth (Fig. 1). While differences in vascular structure are readily apparent qualitatively, obtaining quantitative data has proven more challenging. Despite the excellent work of others, quantitation of these vascular trees has been limited by costly software needed for such analyses. Indeed, Aughwane and colleagues recently also highlighted these methodological issues and barriers in the analysis of human feto-placental vasculature. ^{Reference Aughwane, Schaaf and Hutchinson110} To assist with this, we have recently developed an end-to-end methodology for 3D quantitative characterisation of feto-placental arterial vascularity. ^{Reference Tongpob, Xia, Wyrwoll and Mehnert108} The methodology enables quantification of numerous parameters including tortuosity, vessel diameter, number and order of vessel segments. ^{Reference Tongpob, Xia, Wyrwoll and Mehnert108}

Fig. 1. Visualisation of a feto-placental arterial network at the end of gestation in a mouse (left) and a rat (right). Colour rendering represents different levels of the feto-placental vascular tree. Top row, superior view; bottom row, lateral view; scale bar, 2000 µm.

Beyond assessing the broad expanse of the feto-placental vascular tree, higher resolution scans with a narrower field of view can reveal the capillary structure of the placenta. We have recently assessed this in a rat placenta at the end of gestation, and the capillary tuft can be distinguished in exquisite detail (Fig. 2). Thus, micro-CT approaches may be able to replace electron microscopy and enable investigators a non-destructive way to image capillary networks in their experimental models. Moreover, contrast-enhanced micro-CT (work in progress in our laboratory) can provide potential novel opportunities to visualise greater detailed structures in placental vascular architecture. Iodine ^{Reference Zhou, Cahill, Wong, Seed, Macgowan and Sled124} and zirconium-substituted Keggin polyoxometalate ^{Reference De Clercq, Persoons and Napso99} have also been used on mouse placenta and gross structures can be clearly visualised. However, further work still needs to be developed to explore the full potential of contrast-enhanced micro-CT techniques for assessing placental vasculature.

Fig. 2. High-resolution visualisation of the rat feto-placental capillary tuft at the end of gestation.

Micro-CT is a quantitative 3D imaging technique that is advantageous for investigating 3D placental vascular networks. Evidently, the combination of the contrast agent casting technique and the micro-CT scanning can generate striking images of feto-placental vascular geometry and connectivity in 3D. However, the methodology does involve technical challenges: the casting technique can be difficult, visualisation of data requires considerable computing power and quantitation can be challenging.

Light-sheet microscopy

Light-sheet microscopy has the potential to be a powerful tool for investigating feto-placental vasculature at a macro and micro scale, using either contrast agents or antibodies. However, to our knowledge, little work has been published on this with regard to the placenta. Light-sheet microscopes are traditionally built in-house, require significant computing power for visualisation of the images and optical clearing of the tissue (for a comprehensive review, see ^{Reference Power and Huisken125,Reference Reynaud, Peychl, Huisken and Tomancak126} ). Optical clearance can be challenging for a structure such as the mouse labyrinth zone which is dense with blood vessels. Light-sheet microscopes also traditionally have a relatively small field of view. However, there have been recent exciting developments for tissue clearance and the use of light-sheet microscopy in whole mice, ^{Reference Cai, Pan and Ghasemigharagoz127} mouse eyes, ^{Reference Henning, Osadnik and Malkemper128} mouse hearts, ^{Reference Ding, Bailey and Messerschmidt129} mouse colons, ^{Reference Li, Hui, Hu, Ma, Yang and Tian130} mouse uterine walls, ^{Reference Kagami, Shinmyo, Ono, Kawasaki and Fujiwara131} human brains ^{Reference Mano, Albanese and Dodt132} and human fetal urogenital organs, ^{Reference Isaacson, McCreedy and Calvert133} as well as comparison between different tissue types (i.e. bone, muscle and brain). ^{Reference Buytaert, Goyens, De Greef, Aerts and Dirckx134} With these recent developments, there comes demands for new ways to visualise and quantitate these images. This is a rapidly emerging field and we anticipate that this method will add substantially to our understanding of placental vascular development.

Ultrasonography

None of the imaging modalities discussed thus far are able to assess the feto-placental unit in real time. Indeed, a key challenge for placental research is the ability to accurately monitor placental blood flow in real time, and in particular feto-placental blood flow in vivo with sufficient resolution to yield predictive and diagnostic data. Conventional 2D ultrasonography has been widely used as a primary imaging modality for the assessment of human placental structure and haemodynamics during pregnancy. ^{Reference Abramowicz and Sheiner135–Reference Kingdom, Audette, Hobson, Windrim and Morgen137} Utero-placental vasculature can be identified towards the end of the first trimester, ^{Reference Collins, Birks, Stevenson, Papageorghiou, Noble and Impey138} and there have been recent striking developments in 3D ultrasonography to obtain human placental volumes and overlaying power Doppler ultrasound of the major vasculature. ^{Reference Mathewlynn and Collins139} Moreover, 3D ultrasound has been used to reveal placental volume and vascularity. ^{Reference Mulcahy, Mone and McParland140–Reference Looney, Stevenson and Nicolaides144} These approaches provide not only the opportunity for a screening test for growth restriction but also novel insight into how placental vasculature informs future health outcomes.

With regard to experimental models, preclinical micro-ultrasound has provided useful insight into placental haemodynamics in models commonly used to understand early life origins of adult disease. ^{Reference Wyrwoll, Noble and Thomson18,Reference Rennie, Detmar, Whiteley, Jurisicova, Adamson and Sled52,Reference Rennie, Rahman, Whiteley, Sled and Adamson101,Reference Rennie, Detmar and Whiteley102,Reference Lin, Mauroy, James, Tawhai and Clark145–Reference Lopez-Tello, Barbero and Gonzalez-Bulnes152} For example, using preclinical high-resolution ultrasound, there is evidence for perturbed umbilical cord blood flow and velocity in 11β-HSD2^-/- mice. ^{Reference Wyrwoll, Noble and Thomson18} However, this was ameliorated by pravastatin administration and also improved fetal cardiac function. ^{Reference Wyrwoll, Noble and Thomson18} Current resolution of the conventional 2D ultrasound is not adequate for evaluating 3D vascular complexity and fine branching structure. Therefore, the technique cannot investigate the entire extent to which vascular networks are altered in FGR and the downstream effects on fetal outcomes. However, colour Doppler and photoacoustics are emerging as valuable tools for understanding placental haemodynamic changes and structure in real time. At E14.5 in mouse pregnancy, the utero-placental and feto-placental blood supply can be visualised including chorionic plate vessels and intra-placental arterioles. ^{Reference Foster, Hossack and Adamson153} Moreover, these assessments can be taken in conjunction with Doppler measures of fetal heart and other aspects of the fetal circulation including the aorta, inferior vena cava, ductus arteriosus, ductus venosus, inferior vena cava, carotid and pulmonary arteries. ^{Reference Foster, Hossack and Adamson153} Recently, photoacoustic technology has been utilised to image the human placenta in normal and pathologic pregnancy ^{Reference Maneas, Aughwane and Huynh154,Reference Yamaleyeva, Brosnihan, Smith and Sun155} including maternal hypoxia and hyperoxygenation, ^{Reference Arthuis, Novell and Raes156} and pre-eclampsia. ^{Reference Lawrence, Huda and Bayer157} These approaches will provide invaluable insight into how placental and fetal haemodynamic parameters in early life associate and inform adult cardiovascular outcomes.

Magnetic resonance imaging

The technology of MRI has advanced over the past few years to enable assessment of placental structure, blood flow and oxygenation. MRI is a non-invasive, in vivo 3D imaging modality that has been used for fetal and placental investigations and is emerging as a possible clinical tool. For instance, using MRI in obstetrics to assess human placental abnormalities ^{Reference Otake, Kanazawa, Takahashi, Matsubara and Sugimoto158,Reference Allen and Leyendecker159} and to predict fetal volume and birth weight. ^{Reference Damodaram, Story and Eixarch160–Reference Damodaram, Story and Eixarch162} MRI has the advantage of capturing a large field of view, and thus full-specimen coverage, and, in the case of rodent models, multiple placentas. ^{Reference Avni, Neeman and Garbow163} However, in comparison to ultrasound and photoacoustics, there are significant limitations with translation of MRI studies to a clinical setting. These include machine accessibility outside of large clinical research centres, logistics for obese pregnant women undergoing an MRI and feasibility of using MRI to assess fetal oxygenation in high-risk pregnancies. The low resolution of MRI is also not suitable for detailed analysis of placental microvasculature and microcirculation.

Experimentally, MRI is used to reveal placental and fetal development across gestation. Blood oxygen level-dependent MRI (BOLD-MRI) can be used to evaluate changes in feto-placental oxygenation detecting changes in deoxyhaemoglobin concentration during a respiration challenge (i.e. T₂*). For example, BOLD-MRI has been used to measure changes in human placental oxygenation during maternal hyperoxia in normal pregnancies ^{Reference Sorensen, Peters, Frund, Lingman, Christiansen and Uldbjerg164,Reference Sorensen, Sinding and Peters165} and FGR pregnancies. ^{Reference Ingram, Morris, Naish, Myers and Johnstone166} Moreover, in rat models of FGR, there is a lower feto-placental response to maternal hyperoxygenation. ^{Reference Aimot-Macron, Salomon and Deloison167,Reference Chalouhi, Alison and Deloison168} Furthermore, dynamic contrast-enhanced MRI (DCE-MRI) with contrast agent intravenous injection (i.e. gadolinium chelate) has been used to quantify placental function such as blood volume, blood flow, oxygenation and perfusion rate in different placental compartments. ^{Reference Kording, Forkert and Sedlacik169} Recently, DCE-MRI has been used to investigate placental perfusion in hypoxic-induced FGR rodents, ^{Reference Alison, Quibel and Balvay170–Reference Deloison, Siauve and Aimot172} pre-eclamptic rats (induced by continuous delivery of L-nitro-arginine methyl ester to chronically inhibit nitric oxide synthase) ^{Reference Lemery Magnin, Fitoussi and Siauve173} and prenatal alcohol exposure in monkeys. ^{Reference Lo, Schabel and Roberts151} Placental blood flow and perfusion rate were reduced significantly in the animal models of hypoxia and ethanol exposure ^{Reference Lo, Schabel and Roberts151,Reference Alison, Quibel and Balvay170,Reference Tomlinson, Garbow, Anderson, Engelbach, Nelson and Sadovsky171} but not in a model of pre-eclampsia. ^{Reference Lemery Magnin, Fitoussi and Siauve173} A combination of DCE-MRI and fluorescence imaging has also been used to delineate and quantify the volume of maternal and fetal compartments in mouse placenta. ^{Reference Solomon, Avni and Hadas174} Therefore, both BOLD-MRI and DCE-MRI can be used for in vivo 3D assessment of placental and fetal function, blood flow and oxygenation. While these approaches are highly useful in an experimental context, it remains to be seen if MRI will transition to being standard use clinically.