**1. Introduction**

Proper and strictly controlled invasion of extravillous trophoblasts is mandatory for placental development, enabling the normal growth of a fetus in the maternal uterus. The trophoblast cell line develops at the time of blastocyst formation and divides into two main cell populations, (1) the villous trophoblast with villous cytotrophoblasts and the multinucleated syncytiotrophoblast, forming the outer cover of all placental villi; and (2) the extravillous trophoblast that invades into the maternal uterine tissues, reaching down to the inner third of the myometrium.

Extravillous trophoblasts start their journey at trophoblast cell columns that develop at the tips of anchoring villi that attach to the uterine wall. Within these cell columns the trophoblast cells in direct contact to the villous basement membrane proliferate and build the source for all extravillous trophoblasts. Their daughter cells leave the cell cycle and are pushed toward maternal tissues by the proliferative pressure of the cells at the basement membrane. After a transitional phase, the daughter cells start their active migration and invade into the uterine connective tissues. This is why these cells have been termed "interstitial trophoblasts" [1].

Traditionally, the visualization of extravillous trophoblasts has been achieved by using antibodies against cytokeratin isoforms, such as cytokeratins 8 and 18 [2], and mostly cytokeratin 7 [3–5]. Although trophoblast staining for cytokeratin was always referred to as highly specific, especially in the placenta of a first-trimester placenta, other fetal and maternal cells display immunostaining for these cytoskeletal proteins, including epithelial cells of the embryonic amnion or epithelial cells of maternal uterine glands [6]. With the identification of the highly specific expression of the major histocompatibility complex protein HLA-G on extravillous trophoblasts [7–9] followed by the development of suitable antibodies that specifically bind to only this type of HLA proteins, a new era of identification of extravillous trophoblasts began [6,10].

As will be described below, alterations of trophoblast invasion have been associated with pregnancy pathologies, including preeclampsia, intra-uterine growth restriction (IUGR), spontaneous abortion, and placenta accrete/increta/percreta [11]. So far, scientists tried to link all the above pathologies with trophoblast invasion in total or invasion into uterine arteries only. In specific cases, hypotheses finding associations between pathology and trophoblast invasion were developed but failed testing due to conceptual challenges [12]. Similarly, the placental expression and release of growth factors such as sFlt-1 and/or PlGF were associated with specific pregnancy pathologies, but also, here, the direct link between placental growth factor expression and pathology development could not be established [12,13].

So far, the new routes of trophoblast invasion have not been investigated regarding their impact on placental and thus fetal development. Hence, there is a grea<sup>t</sup> knowledge gap that needs to be filled.

### **2. Historical Thinking of Trophoblast Invasion**

Early descriptions of uterine spiral arteries during pregnancy date back to 1774, when William Hunter described "convoluted arteries that passed between the womb and the placenta" [14]. At that time, no one thought of placental cells invading into the maternal uterus. About a century later, Friedländer (1870) was the first to describe "endovascular cells" in these spiral arteries, without mentioning any information on the source of these cells. This observation of Friedländer was described in a book that was published another 50 years later, by Grosser (1927) [15]. This author was the one who first imagined that the endovascular cells in spiral arteries during pregnancy are not necessarily derived from the maternal decidua but rather could well be of trophoblastic/placental/fetal origin.

It was not until the 1950s and 1960s that the aspect of arterial transformation was revisited, and major observations on spiral artery transformation were published by Harris and Ramsey [16,17], as well as Boyd and Hamilton [18,19]. Both groups described perivascular trophoblast in the decidual stroma surrounding arteries, mural trophoblasts in the walls of these arteries, and intraluminal trophoblasts residing in the vessel lumen. Already at this time, it was speculated that the trophoblast cells within the lumen of spiral arteries may well be washed out into the intervillous space of the placenta.

Interestingly, both groups missed any other route of trophoblast invasion into other luminal structures of the decidua. One of the reasons for this may be the fact that, at that time, an identification of cells by using specific probes, e.g., antibodies, was not possible. Moreover, only few specimens were available at the time; and hence, knowledge on changes of spiral arteries until delivery, even during normal pregnancy, was very sparse.

This became obvious when the first studies dealt with alterations of trophoblast invasion into arteries in pregnancy pathologies. In one of these early studies the authors stated the following: "The examination of the spiral arteries in pregnancy associated with hypertension has not been easy because of two factors. The first was the difficulty of obtaining suitable material and the second the occurrence in the same spiral arteries of extensive morphological changes due to pregnancy itself" [20]. Thus, at the time when scientists and clinicians became aware of structural alterations of invaded spiral arteries in pathological cases, they realized that there was not enough knowledge on how a normal placental bed with invaded structures looks like.

The combination of the following facts may be the reason why, even today, the knowledge on trophoblast invasion is very restricted:

(1) In the very beginning, only very few groups looked into changes of uterine vessels, in particular focusing on spiral arteries.

(2) These groups visualized arteries and the entrance of blood into the intervillous space of the placenta by dye injection and thus missed the veins.

(3) These groups visualized arteries and the infiltrating trophoblasts mostly in monkeys and only a few human cases, and thus may have missed the differences between monkeys and humans.

(4) These groups had no tools in hands to specifically identify invading trophoblasts.

(5) Maybe due to the combination of the above facts, trophoblast invasion into other luminal structures never came into focus.

(6) Scientists following these initial studies simply used this knowledge as a basis and did not scrutinize the real variety of structures invaded by extravillous trophoblast.

#### **3. Looking into Invaded Uterine Structures from the Embryo's Nutritional Point of View**

Looking from the side of the embryo in terms of nutritional support from the mother, shortly after implantation, there is the need to increase nutritional support due to the massively increasing volumes of embryo and placenta in the absence of any supporting blood vessel. Within the endometrium of the human uterus, this is best performed by eroding uterine glands in direct vicinity to the placenta, allowing direct contact of the syncytiotrophoblast to the glandular secretion products. Hence, it looks as if histiotrophic nutrition of the embryo already starts a few days after implantation [21].

In the collection of images of Allen Enders at the Centre for Trophoblast Research in Cambridge, there are images from case 8020, which is considered the earliest specimen in the Carnegie collection, probably of day one after initiation of implantation. In one of the images from case 8020, the margin of the trophoblastic plate is displayed (Figure 1A). Here, the initially invading syncytiotrophoblast has already invaded uterine glands underlying the embryo. This is the earliest description of glandular invasion by trophoblast. In a later-stage case of the early lacunar stage (Figure 1B), invasion into a uterine gland can be seen again [22].

**Figure 1.** (**A**) Image #7 of case 8020: Margin of the trophoblastic plate. Allen Enders explained: "Syncytial trophoblast with small nuclei has invaded the underlying endometrial gland. It is not known whether the small nuclear syncytium is synctiotrophoblast or is partially a heterokaryon involving fusion of trophoblast and uterine cells." The black arrow points to invasion into a uterine gland. (**B**) Image #13 of case 8171: Early lacunar stage (stage 5B). Allen Enders' explained: "Note that the appearance of endometrial glands is similar to that seen in one of the stage 5A sites." He further explained (under image #14 of case 8171): "Note continuity of a capillary with a lacuna that anastomoses with other lacunae. Trophoblast appears to be invading a gland in the upper right." The black arrow points to invasion into a uterine gland, while the blue arrow points to invasion into a uterine blood vessel. Image are provided by courtesy of Allen C. Enders and the Carnegie Collection.

Following this early invasion by the initially invasive syncytiotrophoblast, the extravillous trophoblast population takes over and further invades into uterine glands, resulting in opening these luminal structures toward the developing intervillous space of the placenta [21]. As soon as the intervillous space of the placenta is established, the glandular secretion products flow into this space and are transferred from the placenta to the embryo [23]. At the same time, the remaining secretion products and the respective fluids need to be drained back into the maternal system. Hence, erosion and connection of uterine veins to the intervillous space of the placenta needs to take place next (Figure 2A) [24–26]. Other images of the Enders collection show the junctional zone of trophoblast invasion at the secondary villus stage. Here, invasion into veins and glands can be found, while arteries next to these two luminal structures do not show any signs of invasion [22].

Finally, around mid-first trimester spiral arteries are the next target of the invading extravillous trophoblasts. While glands and veins have thin walls and only need to be eroded and connected to the placenta (Figure 2A), the arterial walls need to be prepared prior to invasion into them. Finally, the spiral arteries are invaded as well, and their lumen is plugged until the beginning of the second trimester (Figure 2B) [27,28].

Hence, during the first trimester, a plasma flow from the plugged arteries, plus a flow of glandular secretion products, enters the intervillous space of the placenta, which is drained back into the maternal system by the utero-placental veins (Figure 2B). This allows the nutritional support of the embryo during the first trimester of pregnancy with substances in maternal plasma, plus the secretion products of the uterine glands. This has been termed histiotrophic nutrition by glands rather than vessels [23].

While the histiotrophic nutrition seems to be sufficient in the first trimester of pregnancy, with the massive growth of the fetus later in pregnancy, a different nutritional support is needed. With the dissolution of the plugs from the arteries and the establishment of the flow of maternal blood into the placenta at the beginning of the second trimester, the nutritional supply of the fetus changes from a histiotrophic to a hemotrophic nutrition (Figure 2C) [23,29]. At the same time, the number and the input of uterine glands diminish, while, of course, the veins remain, to drain back maternal blood into the maternal circulation. Due to the lack of normal placental-bed specimens of the time around mid-gestation, it is not clear when the glandular connection to the placenta disappears. It seems as if this occurs around week 20 of pregnancy, but this still needs further elucidation.

**Figure 2.** Schematic representation of the routes of trophoblast invasion during normal pregnancy. (**A**) Very early in pregnancy, prior to six weeks of gestation, invasion of the early invading syncytiotrophoblast during implantation, as well as invasion of early extravillous trophoblasts, results in opening uterine glands and veins toward the intervillous space of the placenta. Endoglandular trophoblasts open uterine glands, to enable the flow of "uterine milk" toward the placenta. This is followed by invasion of endovenous trophoblasts into uterine veins, to enable backflow of fluids into the maternal system, including villous material and endoglandular trophoblasts (shown in vein). The arrows in gland and vein represent the material transported in these structures (green arrow: glandular secretion products). (**B**) Later, during the first trimester, endoarterial trophoblasts invade into uterine spiral arteries, transform their walls, and plug their lumen, to hinder flow of maternal blood into the placenta. At that stage, only blood plasma is seeping through the plugs (indicated by the dashed red arrow). During this stage of pregnancy, the backflow via utero–placental veins comprises glandular secretion products, plus plasma from the spiral arteries (green arrow plus dashed red arrow), including villous material plus endoglandular and endoarterial trophoblasts (shown in vein). (**C**) At the beginning of the second trimester, the arterial plugs disintegrate, and the flow of maternal blood into the placenta is finally established. So far, it is not clear at which time point the glandular input diminishes and disappears, but in the second half of pregnancy, respective glands can hardly be found. Hence, this schematic drawing only shows arteries and veins (red arrows: maternal blood). Now, the venous backflow contains villous material, as well as endoarterial trophoblasts (shown in vein). A, artery; G, gland; V, vein; GA, gestational age; ST, syncytiotrophoblast; vCT, villous cytotrophoblast; EVT, extravillous trophoblast.

#### **4. New Routes of Trophoblast Invasion**

All the above considerations are only conceivable due to the recent progress in the identification of new routes of invasion of extravillous trophoblasts (Figure 2) [30]. Interestingly, the identification of these new types of cellular pathways can only be performed by using the original tissue organization. Any dissolution of the tissue would have destroyed the possibility to identify these routes.

Of course, this is a purely descriptive approach, which needs to be followed by the elucidation of the functional di fferences of the cells in the di fferent routes and pathways. It needs to be clarified whether the cells already "know" from the beginning where to go and which luminal structure they will go for or whether they just invade uterine tissues and reach a luminal structure simply by chance.
