Introduction
The hystricognath rodents (Suborder Hystricomorpha, Infraorder Hystricognathi [
1]) appeared in the Eocene and underwent an extensive radiation in the Miocene. At this time they were able to capitalize upon the emergence of grasslands for which they were well adapted in a number of ways [
2]. They differ from other rodents in giving birth to precocial young. The newborn is well developed with open eyes and a full coat of hair [
3]. This reproductive strategy requires a lower rate of energy consumption and is well suited to an herbivorous diet [
2]. The hystricognath placenta has a number of distinctive features [
4,
5]. The exchange area or labyrinth is lobulated, an adaptation that allows an increase in the total exchange area and helps to support the larger fetus at the end of gestation [
6]. The lobules are separated by interlobular trophoblast that is the counterpart of the spongy layer found in the placenta of other rodents. Beneath this is a structure known as the subplacenta that is unique to the hystricognath rodents.
In this paper we review the structure of the subplacenta and examine possible functional correlates. The analysis is based on a description of the placenta of the red-rumped agouti (
Dasyprocta leporina). This is a medium sized rodent, larger than a guinea pig and with longer legs. It is found throughout the forest, where it lives mainly on fallen fruits and nuts. Agoutis usually sit erect to eat, holding the food in their hands. They bury excess nuts and fruits for use when food is scarce. Because they bury them singly (scatter hoarding) rather than many in a cache, they are important seed dispersers for a number of tree species [
7,
8].
The functions that have been ascribed to the subplacenta include hormone production [
9,
10]. Our findings are consistent with this interpretation, but suggest that hormone secretion is directed towards the fetal circulation. Trophoblast giant cells occur close to the margin of the subplacenta and we included these in our analysis. We note that, unlike the trophoblast of the subplacenta, the giant cells often occur in close proximity to maternal blood vessels.
Materials and methods
The study was based on 9 placentae from 6 red-rumped agoutis collected from early in midgestation to near term of pregnancy (Table
1). The research was authorized by the Brazilian Institute of Environment and Renewable Natural Resources (IBAMA). The experimental protocol was approved by the bioethics committee of the School of Veterinary Medicine, University of São Paulo. The samples were collected at an agouti breeding facility at São José do Rio Preto, São Paulo. Pregnant females were submitted to hemihysterectomy. Details of anaesthesia and surgical procedures are given elsewhere [
11].
Table 1
Measurements of fetus, umbilical cord and placenta in the specimens of red-rumped agouti used for the present study.
1 | Early to mid-gestation | Left | --- | --- | --- | 15 |
| | Left | 2 | 15 | 2 | 13 |
2 | Early to mid-gestation | Right | 2 | 15.5 | 2 | 12 |
3 | Midgestation | Right | 10 | 42 | 9.5 | 37 |
| | Left | 10 | 42 | 9 | 36 |
4 | Midgestation | Left | 10 | 40 | 9 | 34 |
5 | Midgestation | Right | 10.5 | 41 | 9.5 | 36 |
6 | Term gestation | Right | 16 | 100 | 13 | 78 |
| | Left | 12.5 | 76 | 9 | 50 |
Placental fragments were fixed in 10% formalin in 0.1 M phosphate buffer and processed by standard histological procedures for embedding in paraplast, and then sectioned at 5 μm (automatic microtome, Model RM2155, Leica, Germany). Sections were stained with haematoxylin and eosin, Masson's trichrome stain or by the periodic acid Schiff (PAS) reaction with and without pretreatment with 1% amylase (Sigma, St Louis, Missouri, U.S.A.) at 37°C for 30 min. The latter sections were counterstained with haematoxylin.
Immunohistochemistry was performed for cytokeratin (to identify epithelial cells and trophoblasts) and vimentin (to identify mesenchymal cells and stromal decidua). For this purpose, 5 μm sections were transferred to poly-L-lysine coated slides. Endogenous peroxidase was blocked with 0.1% hydrogen peroxide. To improve antigen retrieval, the sections were then treated in a microwave oven in 0.1 M Tris-HCl buffer, pH 7.4. To prevent non-specific binding, slides were blocked with 2% milk powder in phosphate-buffered saline for 20 min. They were incubated overnight with primary antibody at 4°C; either a rabbit polyclonal antibody against cytokeratin (1:500; PU071-UP, Biogenex, San Ramon, California, U.S.A.) or a goat polyclonal antibody against vimentin (1:500; SC -1226, Santa Cruz Biotechnology, Santa Cruz, California, U.S.A.). Immunostaining was then performed using a secondary antibody from a kit (LSAB-HRP Peroxidase, Dako, Carpinteria, California, U.S.A.) with diaminobenzidine in Tris-HCl buffer, pH 8.2, as the chromogen. The sections were counterstained with Harris's haematoxylin.
The description of the ultrastructure of agouti subplacenta is based on material from the middle of gestation. For transmission electron microscopy, small samples were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, washed in buffer and post-fixed in phosphate-buffered osmium tetroxide, pH 7.4 (Polysciences, Warrington, PA, USA). They were then dehydrated, washed with propylene oxide and embedded in Spurr's resin (Spurr's Kit, Electron Microscopy Sciences, Fort Washington, PA, U.S.A.). Sections were made at 60 nm and stained with 2% uranyl acetate (5 minutes) and 0.5% lead citrate (10 minutes). The ultrastructural observations were made with a transmission electron microscope (JEOL 1010, Peabody, MA, U.S.A).
To study the microvasculature of the subplacenta, an umbilical artery was injected with Mercox™ CL-2R (Okenshoji Co., Ltd, Tokyo, Japan) as previously described [
11]. Tissues were digested by immersion of the preparation in several changes of 20% NaOH solution at 50–60°C. The casts were rinsed thoroughly in distilled water and dried in an oven at 37°C. They were then refrigerated in 20% gelatin. For scanning electron microscopy, pieces of the casts were rinsed in distilled water to remove the gelatin, dried, and mounted on stubs with conductive carbon cement (Neubauer, Münster, Germany). They were then coated with gold using a sputter coater (Model K550, Emitech Products Inc., Houston TX, USA) and examined in a scanning electron microscope (Model 435 VP, Leo Electron Microscopy Ltd, Cambridge, UK).
As a further aid to understanding vessel distribution, some placentae were injected with coloured latex (Neoprene 650, DuPont, Brazil; Latex Stain, Suvinil, Glassurit do Brazil S/A, São Bernardo do Campo, S.P., Brazil). Different colours were injected in a uterine vein, a uterine artery, and the umbilical vein. The placentae were fixed in 10% formalin in 0.1 M phosphate buffer.
Discussion
The subplacenta is a unique structure that distinguishes hystricognath rodents from all other mammals [
4,
5]. The basic arrangement of cytotrophoblast and syncytiotrophoblast, supported by lamellae of connective tissue, is well conserved. The ultrastructure of these elements is also similar across species, including the guinea pig [
9,
10,
12], chinchilla [
13], cane rat [
14], paca [
15], rock cavy [
16] and agouti (this study). The function of the subplacenta is poorly understood.
In the guinea pig, fetal trophoblast invades the walls of the uterine arteries and this is associated with extensive remodelling of the vessel wall [
17,
18]. Recently, it was proposed that the source of the invasive trophoblast was the cytotrophoblast layer of the subplacenta [
19]. There is experimental evidence from the guinea pig and degu to support this hypothesis [
20,
21]. It does not, however, address the function of the syncytiotrophoblast. An earlier proposal that the subplacenta might play a role in maternal to fetal transfer of high molecular weight molecules [
22] receives no support from studies of its ultrastructure [
12,
13].
The placenta is a source of steroid hormones. In the guinea pig ovarian progesterone is not required after day 20 of gestation, implying that placental progesterone is adequate for pregnancy maintenance after this time [
23]. It has been shown for a variety of hystricognath rodents that the hormone circulates bound to progesterone-binding protein [
24]. The principal site of placental progesterone synthesis is the interlobar syncytiotrophoblast [
25]. This is supported by the presence there of smooth endoplasmic reticulum and mitochondria containing tubular cristae [
26]. In contrast Wolfer and Kaufmann [
12] argued against a role for the subplacenta in steroid synthesis. Attention has therefore been focussed on the subplacenta as a source of glycoprotein.
Davies et al. [
9,
10] remarked upon the strongly positive PAS reaction of the guinea pig subplacenta. This is in part due to the presence of glycogen, but there is a diastase-resistant component, as demonstrated here for the agouti. It was suggested that this was glycoprotein and that it was secreted into the lacunae [
9,
10]. Heap and Illingworth [
27] proposed that it might be the progesterone-binding protein, as this is a carbohydrate-rich transport protein. The association of the subplacenta with a transport protein likewise unique to hystricognath rodents constituted an attractive hypothesis. However, the cellular location of progesterone-binding protein was later shown to be the interlobar and marginal trophoblast of the main placenta [
28].
Davies et al. [
9,
10] had implied that the glycoprotein might be chorionic gonadotropin (CG), but the evidence for this is equivocal. Indeed, recent work suggests that CG is restricted to primates and equids [
29,
30]. There is some evidence for production of luteinising hormone, possibly CG, in guinea pig placenta [
31], but no detectable signal for the corresponding messenger RNA was seen in Northern blot assays of placental RNA [
32]. In primates, CG is secreted directly into the intervillous space from the maternal-facing syncytiotrophoblast, whilst in equids CG is secreted to maternal blood from the endometrial cup cells. Thus, if there is a guinea pig CG, it is more likely to be produced by the syncytiotrophoblast of the main placenta than the subplacenta as it can then be secreted to the maternal blood channels. As we have shown here, the blood supply to the subplacenta is derived from the fetal circulation. A longitudinal study in the guinea pig showed that maternal vessels were present only early in development and had disappeared by day 20 of gestation [
12].
The subplacental syncytiotrophoblast contains electron dense inclusions that likely are secretion granules as argued by Wolfer and Kaufmann [
12]. Their product is released into the elaborate network of lacunae. These do not have access to the maternal circulation, but King and Tibbitts [
13] suggested that products secreted to the extracellular spaces might be able to reach the fetal capillaries. As we show here, the fetal vessels within the subplacenta pursue a tortuous course with dilatations and constrictions as in an endocrine gland.
Secretion of most hormones produced by the human placenta is unidirectional: they are released to the maternal circulation and affect maternal physiology. However, some placental products do reach the fetal circulation. In sheep they include prostaglandin E
2 [
33], adenosine [
34] and steroids [
35]. As an example, fetal breathing movements occur throughout most of gestation, but are influenced by prostaglandin E
2 from the placenta, which tends to suppress them. The removal of the placenta and the placental prostaglandin E
2 is critical for the initiation of continuous breathing at birth [
33]. Thus, if the subplacenta is an endocrine organ, the function of its hormones may be to influence fetal physiology.
The origin of the trophoblastic giant cells was explored by Mossman [
36]. They tend to occur in clusters surrounded by amorphous extracellular material, as described here for the agouti and elsewhere for the chinchilla [
13]. Electron dense granules were seen both within the cells and in the extracellular matrix. This is consistent with a secretory function as suggested in a study of chinchilla giant cells by Tibbitts and Birge [
37]. The giant cell clusters sometimes occur close to maternal vessels. Thus, in contrast to those of the subplacenta, their secretory products may have access to the maternal circulation. On the other hand, many giant cells are vacuolated; this has been considered to be a degenerative change [
37].
In conclusion, the function of the subplacenta remains elusive. The syncytiotrophoblast secretes PAS-positive material, probably glycoprotein, to an extensive system of lacunae. These lacunae seem to be connected to the intercellular spaces of the cytotrophoblast layer, which lie directly beneath fetal capillaries. Indeed, the subplacenta is plentifully supplied with blood vessels from the umbilical circulation, but lacks direct access to the maternal circulation. Perhaps the function of the subplacenta is to secrete growth factors, hormones or cytokines to the fetal circulation. As in the sheep, they might function to shut down physiological functions that are not needed in fetal life. Placental signals disappear once the cord is severed, allowing immediate resumption of suppressed functions in the neonate. Like sheep, hystricognath rodents give birth to precocial young and we suggest that placental control of fetal function is an important feature of this reproductive strategy.