Implantation and extravillous trophoblast invasion: From rare archival specimens to modern biobanking

Abstract

Extravillous trophoblast invasion serves to attach the placenta to the uterus and to enable access to nutrients for the embryo throughout pregnancy – secretions of the uterine glands in the first trimester, maternal blood in the second and third trimester. For assessing extravillous trophoblast invasion, histology (in combination with immunohistochemistry) still plays a major role in placental research. This is especially true for the re-assessment of rare archival specimens from early human implantation sites or placenta in utero with the background of recent knowledge which may help to strengthen current hypotheses. This review summarizes the recently expanded picture of extravillous trophoblast invasion, gives an overview about fundamental archival specimens in placental research, presents new images of archival specimens, gives insights into the latest developments in the field of biobanking and provides insight into the current situation on sample usage in the absence of biobanks. Modern techniques allow expanding our hitherto believed concept of extravillous trophoblast invasion, which is not restricted to spiral arteries: Extravillous trophoblasts also invade into uterine glands and uterine veins and thereby connect all these luminal structures with the intervillous space. All biomedical research dramatically depends on the quality of the assessed biological samples. Hence, researchers should be aware that the time between collection of a sample from a body and the beginning of analysis (pre-analytical phase) may have more impact on the outcome of a study than previously assumed.

1. Introduction

Although there is a huge variety of highly sophisticated omics technologies available today, one can still learn a lot from histology and morphological assessment of tissues and organs. In placental research, histology in combination with immunohistochemistry and electron microscopy still plays a major role in identifying new players in the interaction between mother and fetus. Especially, histology of rare archival specimens from early human implantation sites or placenta in utero can be of great value.

At the same time, science has evolved and has led to the development of structured and systematic collections of samples to be used in research, biobanks. Today, biobanks have taken over a major role in biomedical research, offering the hope to enable higher sample quality, better linkage to clinical data and less garbage in – garbage out when analyzing human-derived samples.

Looking at placental development, extravillous trophoblasts originate from the distal side of trophoblast cell columns derived from cytotrophoblasts of anchoring villi. Within these cell columns proliferative but non-invasive cytotrophoblasts switch to an invasive but non-proliferative extravillous phenotype and start to enter maternal tissues (interstitial trophoblast) [1]. From a histological point of view one of the most important characteristics of extravillous trophoblasts is their expression of major histocompatibility complex, class I, G (HLA-G) [2–4]. Only HLA-G is a specific marker for extravillous trophoblasts, rather than the still commonly used cytokeratin 7. The latter also reacts with maternal glandular epithelial cells [5]; this may lead to the misidentification of maternal uterine glands and maternal vessels where in the latter the endothelium is replaced by extravillous trophoblasts. Researchers should take care when choosing antibodies against HLA-G, since not all antibodies recognize HLA-G1, which is the only isoform expressed at the cell surface [6]. The anti-HLA-G antibody we recommend to use is the clone 4H84; it recognizes all isoforms of HLA-G, including HLA-G1 (which is the full length, membrane bound isoform, expressed on the cell surface) [5].

This review is divided in three parts, the first part summarizes the latest developments in the field of extravillous trophoblast invasion towards uterine glands and vessels. The second part gives an overview about fundamental archival specimens in placental research for “placental rookies”. For well-established placental researchers some new images of archival specimens are discussed in the context of the latest literature. In the third part, this review aims at extending this view to the latest development in the field of biobanking and provides insight into the current situation on sample usage in the absence of biobanks.

2. Extended view on extravillous trophoblast invasion into uterine vessels and glands

It is commonly accepted, that extravillous trophoblasts start their invasive pathway by invading the decidual stroma (interstitial trophoblast) [7] and thereby anchor the placenta to the uterus. It is also long accepted that extravillous trophoblasts invade and transform uterine spiral arteries (endovascular trophoblast) [8,9]. This leads to plugging of such vessels during the first trimester of pregnancy and after release of the plugs enables flow of maternal blood towards the placenta starting with the beginning of the second trimester. Only recently, the morphological appearance of trophoblast plugs have been described in detail [10].

2.1. Endoglandular trophoblast

In the last few years, the picture of trophoblast invasion has been massively expanded. Over the last six years it has been demonstrated that extravillous trophoblasts also invade into uterine glands [11,12]. Already in 2002 Burton and colleagues have demonstrated that secretion products of uterine glands fill the intervillous space of the placenta during the first trimester of pregnancy facilitating histiotrophic nutrition of the embryo prior to hemotrophic nutrition with the onset of maternal blood flow within the placenta [13–15]. In the same year Craven at al described extravillous trophoblasts surrounding uterine glands [16]. Thirty years earlier, Knoth and Larsen have already described close contact between invading trophoblast and uterine glands [17]. However, at that time it was totally unclear how glands are opened towards the intervillous space. It was only in 2010 when Moser et al. detected that trophoblast invasion into uterine glands (endoglandular trophoblast) leads to the connection of the glandular lumen to the intervillous space allowing secretion products to enter this space [11].

Additionally, these authors took a closer look into publicly available archival human implantation sites and also looked for archival material at their own site. Using this approach these authors identified a very close connection between an implanting embryo and subjacent uterine glands already at day 10 of pregnancy [12]. Fig. 1 shows images of an implantation site retrieved from the archive of the Institute of Cell Biology, Histology and Embryology at Medical University of Graz, Austria. These serial sections reveal how a uterine gland below the implantation site is widening and seems to open towards the very early placenta of that specimen (Fig. 1, images 13 and 14). So far, it was hypothesized that it is the extravillous trophoblast, developing in week 3 post conception, that invades into luminal structures. Looking at the archival material presented in Fig. 1, this view needs to be revisited: The evidence for a direct connection between the implanting blastocyst and uterine glands already at the time of implantation (Fig. 1, images 13 and 14) opens new ways for looking at early erosion of uterine luminal structures by the trophoblast.

Fig. 1. Connection between uterine glands and the early conceptus during implantation.

(1–22) Hematoxylin and eosin stained serial sections of a human implantation site at day 10 (archival specimen) in putative serial order. Images in (13) and (14) demonstrate a clear connection (arrows) between the early conceptus (EC) and the uterine gland beneath. Uterine glands are marked by asterisks. Lumina of uterine glands beneath the EC seem expanded and filled with fluid (probably a mixture of glandular secretion products and maternal blood).

2.2. Endovascular trophoblast stratified into endoarterial and endovenous trophoblast

Very recently another route of trophoblast invasion has been discovered. Although denied and disputed for decades, Moser et al. (2016) have identified, that using the correct markers (like EphB4 for venous endothelium and desmin as well as smooth muscle actin for the vascular smooth muscle layer) also invasion into uterine veins can be detected [18]. This of course is more than logical, as also veins need to be connected and opened towards the intervillous space. Only by opening and connecting these vascular structures, maternal blood flowing into the placenta via eroded spiral arteries can flow back into the maternal circulation. Two other groups have been able to prove this invasion into veins (personal communication, J. Pollheimer, Medical University of Vienna, Austria and He et al 2017 [19]). Besides that, even in the classic paper by Harris et al. 1966 the authors stated “large multinucleated giant cells are frequently encountered within the veins” and “In one specimen … the trophoblast had opened up a vein in the myometrium” [20]. Also Craven et al. [21] have proposed invasion of maternal decidual veins by trophoblast since they found trophoblasts in the lumen of such vessels. The concept of venous invasion does not question but rather broadens the concept of arterial trophoblast invasion, which is associated with extensive remodeling, transformation and plugging of the arteries.

This new route of invasion into uterine veins asks for a new definition of the terminology of extravillous trophoblast. So far, trophoblast cells within the wall or lumen of spiral arteries have been termed endovascular trophoblasts. With the invasion into uterine veins, this term is no longer specific enough. Hence, we propose a new definition of extravillous trophoblasts (Fig. 2): Those trophoblasts invading glandular structures are termed “endoglandular trophoblast”, those trophoblasts invading vascular structures (arteries and veins) are termed endovascular trophoblasts. The latter are further stratified into endoarterial trophoblasts (invading into and transforming spiral arteries) and endovenous trophoblasts (invading into uterine veins). An example for endovenous invasion is presented in Fig. 3. Beside that, we have recently collected strong evidence for extravillous trophoblast invasion into lymphatic vessels (unpublished data and also personal communication, J. Pollheimer, Medical University of Vienna, Austria and He et al. 2017 [19]).

Fig. 2.

Nomenclature of various subtypes of trophoblast cells depending on their location.

Fig. 3. Extravillous trophoblasts (EVTs) invade uterine veins (endovenous invasion).

Serial sections of invaded first trimester decidua with immunohistochemical double staining (gestational age 7 weeks). Image in (a) is immuno-double stained for von Willebrand factor (VWF) (blue, serves as marker for vascular endothelial cells) and major histocompatibility complex, class I, G (HLA-G) (brown, serves as marker for EVTs). Image in (b) is immuno-double stained for EphB4 (blue, serves as marker for venous endothelial cells) and HLA-G (brown). (a) Interstitial trophoblasts (examples marked by asterisks) are situated within the decidua, the uterine vessel (circle) is nearly completely surrounded by EVTs. (b) The neighboring section shows that the endothelium of the vessel (circle) is Eph4 positive (blue) and thus identifies the vessel as uterine vein. Arrows point to EVTs within the venous lumen and within the endothelium. AB, air bubble; no nuclear counterstain.

3. Fundamental archival human implantation sites

3.1. Publicly available archival human implantation sites

As can be seen from the achievements above, revisiting archival material with current knowledge may well enable detection of new, so far undetected features. Unfortunately, today only a few images from a handful of human early implantation sites are available for the scientific (and public) community. As an example we describe the image collections of the well-known collections that are publicly available at the Centre for Trophoblast Research (http://www.trophoblast.cam.ac.uk/Resources). These image collections include:

• -The Boyd Collection, comprising histological sections (no blocks) of placenta-in-situ specimens and isolated placental specimens. As is outlined on the website of the Boyd Collection: “The placenta-in-situ specimens often demonstrate low-lying placentas, and may have resulted from hysterectomy for antepartum hemorrhage. The isolated placental samples are likely to have been obtained from spontaneous miscarriages.” The gestational age of sections of in situ specimens directly available as scanned images range between 8.5 weeks and 32.5 weeks.
• -Images of implantation, provided by Allen Enders. This collection of images comprises sets of images from the implantation of the human, the armadillo, the baboon, the guinea pig, the macaque, the marmoset, the mustelid, the rabbit and the rat. There are also sets of images of equine development. The images showing human implantation are mostly derived from the Carnegie Collection that was created by Hertig et al. [22]. Some images are from Ender's personal collection.

Interestingly, in the set of images of human implantation, the description of Enders for an early lacunar stage (image 14, #8171) contains the following sentence: “Trophoblast appears to be invading a gland in the lower right.” Some images later (image 54) Ender describes the following: “There is an extensive trophoblastic shell and well-formed anchoring villi. Note the broadly opened venule and its communication through the shell to intervillous spaces. ” Hence, it seems as if the description of endoglandular and endovenous invasion has already been done and can be attributed to Allen Enders.

Another well-known source for images of archival material of early human implantation and placentation is the Virtual Human Embryo (http://www.ehd.org/virtual-human-embryo). At this site a team led by Raymond F. Gasser created thousands of restored, digitized, and labeled serial sections from the world's largest collection of preserved human embryos, the Carnegie Collection. They used the serial sections to create animations, fly-throughs, and 3D reconstructions.

Similar projects are the Kyoto Human Embryo Visualization Project (http://bird.cac.med.kyoto-u.ac.jp/index_e.html) under the leadership of Kohei Shiota or the “Blechschmidt collection” at the University Medical Centre Göttingen (https://embryology.med.unsw.edu.au/embryology/index.php/Blechschmidt_Collection).

Both of these collections include embryonal rather than placental samples beginning from Carnegie stage 7 (represents day 16 post conception). Based on the samples of the Carnegie collection very recently an interactive three dimensional digital atlas and quantitative database of human development including detailed analyses of organ development was presented [23].

In PubMed titles of publications of a large number of historical articles about human placentation can be found; however, unfortunately even the abstracts are hardly commonly available. Access to the full texts is even more restricted. One of the best preserved human implantation sites is “The Peter's embryo” which was described by several authors. This embryo is estimated to be thirteen days old, was first described by Peters in 1899 [24], later extensively studied by Hertig et al. 1959 [25] and is also part of the classics “The Human Placenta” by Boyd and Hamilton [26]. Boyd and Hamilton also describe the syncytial surface of the Barnes embryo – another remarkable early human implantation site, estimated to be eleven days old – as “adjacent to an endometrial gland which is showing degenerative changes”. They also described that “in some instances the glandular epithelium has disappeared at the point of contact with the trophoblast which thus comes to plug the glandular lumen”. In the light of the recent studies mentioned above on endoglandular invasion this sounds up-to date more than ever.

3.2. Locally available archival human implantation and placentation sites

At the Institute of Cell Biology, Histology and Embryology of Medical University of Graz, Austria, serial sections of an implantation site are used for teaching medical students since decades. From these sections the penetration of single trophoblasts into uterine glands has been demonstrated [12]. Collecting and assessing these serial sections enabled us to show that there is a direct connection between early conceptus and the lumen of uterine glands at the time of implantation (Fig. 1). The underlying gland seems expanded and filled with fluid – probably a mixture of glandular secretion products and maternal blood – and due to the direct connection the early conceptus also seems surrounded with this fluid. The same picture of blood filled expanded glands beneath the implantation site has been described as early as 1899 by Peters et al. [24]. Researchers at that time were discussing about possible connections between the implanting blastocyst and uterine glands, but were not able to find corresponding morphological evidence; while more than 100 years later the respective evidence has been described.

The investigation of later stages of placentation and hence trophoblast invasion is hindered by the fact that samples from first trimester placenta often contain solely villi and/or non-invaded decidua parietalis rather than at least parts of the placental bed. If the specimen contains parts of the decidua basalis, these pieces are usually present as little pieces with unknown orientation. Also, parts from the deeper placental bed containing myometrium are extremely rare.

Hysterectomy specimens with placenta in utero provide the ideal material for the assessment of structure and orientation of vessels and glands and progression of trophoblast invasion. As listed above, a few of these specimens are publicly available. At the same time, the main focus of the assessment of such specimens was on invasion and transformation of spiral arteries, rather than on invasion of uterine glands and veins.

At the Institute of Cell Biology, Histology and Embryology of Medical University of Graz, Austria, an archival specimen with a first trimester placenta in utero was evaluated using immunohistochemistry (Fig. 4). Staining sections of the placenta in utero with antibodies against HLA-G and cytokeratin 7 reveals the pattern of extravillous trophoblast distribution around placenta and embryo. At that time of gestation, eroded uterine glands can be visualized especially at the margin of the placenta (Fig. 4b and c). This may be due to the fact that the developing placenta expands rapidly and laterally [21,27]. New glands are thereby approached and opened constantly at the edges of the placenta. Their secretion products reach the intervillous space and thereby secure a continuous supply with histiotrophic nutrients until the beginning of the second trimester when hemotrophic supply starts. There is neither information of the origin nor the gestational age nor the processing of this archival placenta in utero. However, it is most likely from the first trimester. The placenta in utero is embedded in paraffin and due to the yellow color the specimen it was very likely fixed in picric acid (Bouin's solution) rather than in formalin. Factors like these need to be taken into account when assessing archival material, they may influence the results of staining protocols or other techniques like microdissection with subsequent isolation of nucleic acids, and in situ hybridization. However, it was recently demonstrated that it is possible to isolate DNA from archival specimens, even after long-term preservation in formalin or Bouin's solution [28].

Fig. 4. Eroded glands especially at the edge of the developing placenta.

Serial sections of an archival placenta in utero from hysterectomy, image in (a) is immunostained for HLA-G (brown), images in (b, c) are immuno-double-stained for HLA-G (brown, serves as marker for extravillous trophoblasts) and Ck7 (blue, serves as marker for uterine glands). (a) Overview: shows the embryo in the center of the uterus, the uterus with a thick layer of myometrium underneath the decidua, the non-invaded decidua parietalis, the junctional zone between mother and fetus (decidua basalis, strongly invaded by extravillous trophoblasts), the villous part of the placenta with the intervillous space inbetween, and cell islands. (b, c) Details from inset in (a). Inset marks the zone of the placental edge and thus the transitional zone between invaded decidua basalis, non-invaded decidua parietalis and the uterine cavity. (b) Uterine glands (UG) are invaded by endoglandular trophoblasts particularly at the edge of the expanding placenta. The most upper gland appears nearly completely eroded, endoglandular trophoblast (green arrow) replace the glandular epithelium. (c) Higher magnification of uterine glands invaded by endoglandular trophoblasts (green arrows) at the edge of the placenta. Nuclei were counterstained with Hemalaun (a) or no nuclear counterstain (b, c).

3.3. Transition from archival material to modern tissue sampling

All the above images and data show the power of revisiting archival tissues and their sections. However, this can only be one part of a project and needs to be broadened using other technologies. Today's methods of collecting human samples for any type of analysis have dramatically changed during the last decade or so and the main focus of sample collection and handling is no longer to store as many samples as possible but to best maintain sample quality.

For any scientific analysis, the time between collection of a (human) sample and analysis is critical. This time is called pre-analytical phase. It starts at the time of collection and includes the collection procedure, the handling of samples (intermediate storage, transport, cutting, aliquoting etc.), the fixation of samples (freezing, formalin, etc.), the storage of samples at various temperatures and finally the distribution to the scientist who will perform the analysis.

4. Aspects of modern biobanking

As can be seen from the list above, the number of options and chances for a suboptimal treatment of samples is extremely high. Typical post-sampling variables during the pre-analytical phase include but are not limited to the time between collection and fixation, the temperature during transport, the storage temperature and duration, the number of freeze-thaw cycles and the extraction method for biomolecules. This list needs to be extended to the variables at the time of analysis, since taking the same set of samples but changing the assay may lead to a completely different set of results [29].

The time between collection of a sample from the body (e.g. tissues during surgery) and fixation, i.e. the pre-analytical phase, has a profound effect on a variety of biomolecules. Proteins for example are differently affected and show acetylation, methylation and phosphorylation differences that directly influence the pathways these proteins are actively involved in, such as regulation of glycolysis, translation and transcription as well as regulation of the cytoskeleton [30]. Variables of the pre-analytical phase will further extend the effects on biomolecules and reduce the chance for reproducibility.

These problems are not only present in scientific publications, but occur on a daily basis in routine care in hospitals. As was stated by Lippi et al. 2011: “Pre-analytical errors still account for nearly 60%-70% of all problems occurring in laboratory diagnostics, most of them attributable to mishandling procedures during collection, handling, preparing or storing the specimens” [31].

This scenario can easily be extended by adding another level of complexity to it. It is not only the handling of samples that may change the results of an assay; it may simply be the labeling of a sample that changes outcome. It has been estimated that nearly 100,000 people die every year in hospitals in the USA because of preventable medical errors (M. Rasanen, “http://www.leicabiosystems.com/pathologyleaders/specimen-tracking-helping-prevent-misdiagnosis/”). And one of the most common (and preventable) reasons for misdiagnosis is labeling of tissue samples. Why is this? Because today's routines in hospitals as well as research labs include the following typical error-prone procedures: Ambiguous hand written labels that cannot be clearly read and are misidentified, labels that have not been tested thoroughly and fall off during storage, as well as errors that occur while typing numbers of codes rather than retrieving them from the hospital LIMS (laboratory information and management system).

Irrespective of whether such error-prone procedures and variables occur in research or clinical routine, it has a major impact on the development of new research hypotheses, new therapies and biomarkers. It is clear now that the obstacle number one for biomarker validation is the missing access to high quality biospecimens [32]. Biospecimen quality in this context includes the direct quality of the samples, but also the limited linkage to clinical data, the poorly developed standardization for classification, coordination and distribution of samples and finally the neglect of standards for the pre-analytical phase. The latter has been identified today to be highly important for the biobanking field [33] and is now leading to the harmonization of biospecimen management practices and the development of evidence-based standards [34]. Only recently, an example for application of Standard Operation Procedures for sample collection for scientific research has been published [35]. Standardized sampling procedures like this are key for successful collections of samples for future research. Otherwise the percentage of irreproducible results will further increase [36].

5. Conclusion

Archival material of human implantation sites and placentas in situ from hysterectomy specimens are still an invaluable treasure for placental research, especially when re-assessing it with the background of the recent state of knowledge. The hypothesis that there is a direct connection between the implanting blastocyst and uterine glands was stated more than hundred years ago and only now there is proof and evidence for this. Modern techniques allow expanding our hitherto believed concept of extravillous trophoblast invasion, which is not restricted to spiral arteries. Extravillous trophoblasts also invade uterine glands, veins (and lymphatic vessels) and thereby connect all these luminal structures with the intervillous space. All biomedical research depends dramatically on the quality of the assessed biological samples. Hence, all researchers should be aware that the so-called pre-analytical phase (the time between collection of a sample from a body and the beginning of analysis) may have more impact on the outcome of a study than previously assumed.