Changes in endovascular trophoblast invasion and spiral artery remodelling at term in a transgenic preeclamptic rat model
Article Outline
- Abstract
- 1. Introduction
- 2. Materials & methods
- 3. Results
- 3.1. Endovascular trophoblast and associated remodelling in the whole mesometrial triangle
- 3.2. Endovascular trophoblast invasion and associated remodelling in three depth levels of the mesometrial triangle
- 3.3. CD31 immunostaining of endovascular trophoblast
- 3.4. Atherosis-like lesions in spiral arteries
- 3.5. Doppler results
- 4. Discussion
- Acknowledgement
- References
- Copyright
Abstract
As a follow-up to our previous study which revealed a surprisingly deeper endovascular trophoblast (ET) invasion on day 18 in a transgenic preeclamptic (PE) rat model (hAngiotensinogen ♀ × hRenin ♂) compared to non-PE controls, we examined further changes in ET invasion and associated spiral artery (SA) remodelling at term (day 21). PE transgenic rats and non-PE reversely mated (RM) transgenic rats were compared to normal SD rats (C). Sections were stained to visualize trophoblast, fibrinoid, vascular smooth muscle (VSM) and endothelium. SA were evaluated in three depth levels in the mesometrial triangle (MT) using the KS-400 image analysis system. In separate transgenic rats, Doppler ultrasound was performed in uterine arteries, and the resistance indices (RI) were calculated. Although for the whole MT differences in ET invasion were no longer significant between the PE and C, indicating a partial catching up in C rats, there was still significantly more ET in the deepest level in the PE group as compared to the C and RM groups. At the same time the SA walls in PE rats contained significantly more fibrinoid (versus RM and C) and VSM (versus C). In all SA cross-sections, re-endothelialisation was prominent, but significantly different between PE and C group. The Doppler results showed a significantly lower RI in the arcuate uterine artery of the PE group compared to the C group. There was no evidence of elimination of deeply invaded ET at term, previously considered as a possible mechanism for restriction of vascular remodelling in human PE. The differences in vascular remodelling, previously described on day 18 by histology and Doppler data, were maintained on day 21, but there was extensive endothelial repair in the three groups. Atherosis-like lesions were observed in the three groups, most frequently in the RM group, but were never associated with placental infarcts.
Keywords: Endovascular trophoblast, Trophoblast invasion, Pre-eclampsia, Rat, Spiral arteries, Vascular remodelling, Uteroplacental haemodynamics
1. Introduction
The pioneering work by Brosens and colleagues on the human placental bed revealed a link between preeclampsia (PE) and the defective ‘physiological change’ in the myometrial segments of spiral arteries (SA), which was hypothesized to result from impaired deep endovascular trophoblast (ET) invasion [1]. Deep ET invasion in the SA and associated vascular remodelling has also been described in the rat [2]. This species is therefore considered to be an attractive experimental model to study these events. Using the transgenic PE rat model developed by Bohlender and Dechend [3], [4], we initially hypothesized that in such animals shallow ET invasion and restricted vascular remodelling would occur [5]. Unexpectedly, on day 18 of their pregnancies the PE rats showed deeper and more intense ET invasion compared to C rats, although in the PE rats SA had undergone less vascular remodelling, as they showed more vascular smooth muscle (VSM) and more endothelium in parts of the vessel wall covered by ET.
Although one has tended to regard the defective remodelling in the human as the result of a primary invasion defect, the possibility that poor vascular remodelling may result from a maternal (over)reaction–via inflammatory or immune processes – towards an initially normal trophoblast invasion should also be considered [6]. Indeed, some human placental bed biopsies, invaded SA may show heavy infiltrations by inflammatory cells (illustrated in Pijnenborg et al., 2007) [7]. It has also been suggested that the lesion called “acute atherosis”, which has previously been claimed to be pathogenic for PE [8], might be caused by such an overreaction, since trophoblastic remnants can occasionally be found in the wall of such vessels [9]. Although the relevance of this lesion is still uncertain, vessels with acute atherosis have been shown to communicate with infarcted placental areas at term [1]. In PE rats on day 18 of pregnancy vascular lesions with features comparable to the atherosis lesions in human and further referred to as “atherosis-like lesions” were found in 2 of 7 implantation sites, but no communications with infarcted placental regions could be seen at that time.
We decided to extend our observations to day 21, the day before parturition, led by the following working hypotheses: (1) in the PE group, the more deeply invaded trophoblastic cells on day 18 are subsequently eliminated; (2) vascular remodelling differs in the PE and the C group; (3) invaded vessels which had been remodelled on day 18 show maternal repair on day 21; (4) in the PE group there is an increased incidence of atherosis-like lesions which on day 21 are associated with placental infarction. As in our previous study [5], we calculated the percentages of SA cross-sections invaded by ET in the whole mesometrial triangle (MT) and studied the associated SA remodelling. Subsequently we evaluated the invasion and remodelling at three different depth levels in the MT. Our results for both pregnancy days will be compared in the discussion. Doppler ultrasound was performed on a separate group of transgenic rats on day 21 of pregnancy, providing a physiological context for our histological findings.
2. Materials & methods
All methods used in this study are identical to those used in our previous publication [5]. Therefore we include only a brief summary of these methods.
2.1. Rats
Female transgenic Sprague–Dawley (SD) rats harbouring the human angiotensinogen gene (hAogen) were mated with an SD male harbouring the human renin gene (hRen). Since these females become hypertensive and develop proteinuria during their pregnancy, this group is further referred to as the PE group. Similarly, females harbouring the hRen gene were mated with a male harbouring the hAogen gene. Such females do not show PE symptoms and are referred to as the reverse mating group (RM group). Non-transgenic SD females, mated with an SD male, were used in the control group (C group) [4]. The rats were kept at 24 +/− 2
°C, were fed ad libitum and had free access to water. Experiments were approved by and conducted according to guidelines of the Berlin animal care committee. For the histology part of this study, each group contained nine pregnant rats; for the Doppler part, five pregnant rats were used in each group.
2.2. Tissue preparations and immunohistochemistry
On day 21 of pregnancy (copulation plug on day 1) the females were sacrificed, fetuses were weighed and numbers of resorptions recorded. From each rat, all implantation sites (placenta with associated MT) were removed, fixed in JB-Fix fixative [10] and embedded into Paraplast Plus (Sherwood Medical Co, St Louis, USA). Sets of ten step-serial sections were cut from each implantation site parallel to the mesometrial–fetal axis. The periodic acid Schiff (PAS) staining was performed on one section of each set of parallel sections and a set containing a central maternal arterial channel was selected for further analysis. Parallel sections from this set were immunostained for α-actin as a marker for vascular smooth muscle (VSM) cells, for cytokeratin (using the broad spectrum anti-keratin reagent MNF 116, which according to the supplier reacts with intermediate and low-molecular-weight keratins) as a trophoblast marker and for the endothelial marker CD31. Details of the immunostaining procedures are summarized in Table 1.
Table 1. Procedures for α -actin, cytokeratin and CD31 immunostaining.
| α-Actin | Cytokeratin | CD-31 | |
|---|---|---|---|
| Endogenous peroxidase block | 0.5% H2O2 in 100% methanol 30min | – | – |
| Endogenous alkaline phosphatase block | – | HCl 0.2M 10min | HCl 0.2M 10min |
| Non-specific Ig binding Block | 2% BSA, 0.1% Tween-80, 1% non-fat dried milk TBS, 15min | 2% BSA, 0.1% Tween-80, 1% non-fat dried milk TBS, 15min | 2% BSA, 0.1% Tween-80, 1% non-fat dried milk TBS, 15min |
| Primary antibody | Clone 1A4, Dako, 1/200, 120min, RT | Clone MNF 116, Dako, 1/500, overnight, 4°C | Clone TDL-3A12, Pharmigen, 1/200, overnight, 4°C |
| Secondary antibody | PO-conjugated goat anti-mouse, 1/100, 30min | Unconjugated goat anti-mouse, 1/50, 30min | Unconjugated goat anti-mouse, 1/50, 30min |
| Detection of secondary antibody | _ | APAAP complex, 1/100, 30min | APAAP complex, 1/100, 30min |
| Colour reaction | DAB 10min | NBT/BCIP 80min | NBT/BCIP 60min |
| Counterstaining | Mayer's Hematoxylin | 0.1% α-amylase digestion, 10min, 37°C; PAS | Mayer's Hematoxylin |
| Mounting | Depex | Glycerin jelly | Glycerin jelly |
2.3. Quantitative analysis
For the quantitative analysis, exactly the same evaluation methods were used as described in our previous publication [5]. Briefly, the KS-400 image analysis system (Carl Zeiss, Zaventem, Belgium) connected to a Zeiss microscope (Axioskop 50) fitted with a colour camera (Axiocam MRc5) was used. The whole MT was photographed and the images were shade-corrected. The border of the MT was manually delineated on the α-actin immunostained section, and the MT was divided into three concentric depth levels of equal thickness for the evaluation of the depth of invasion (Fig. 1). The lumen of each SA cross-section was manually delineated and stretches of ET, fibrinoid, VSM and CD31 positive endothelium were traced separately over the lumen contour tracing. For every cross-section, the length of each traced line was measured and for every length the percentage was calculated in relation to (1) the total length of the SA lumen contour, (2) the SA lumen contour length corresponding to parts of the artery wall containing endovascular trophoblast (sub-ET) and (3) the SA lumen contour length of parts without endovascular trophoblast (extra-ET) (Fig. 2). Additionally, the numbers of total, invaded and non-invaded SA cross-sections, as well as the numbers of vessels with atherosis-like lesions, were counted in the whole MT as well as in each of the three different depth levels in all three groups.
Fig. 1
Histological section of a day 21 implantation site including the placenta and its associated mesometrial triangle. A maternal arterial channel is visible in the central part of the placenta. The mesometrial triangle is divided into three depth levels (1, 2 and 3).
Fig. 2
Diagram illustrating the measurements and calculations performed on individual spiral artery cross-sections. Stretches of endovascular trophoblast (ET, light blue) are traced on the lumen contour (yellow), delineating sub-ET (sum of light blue traced lengths) and extra-ET (non-traced) parts of the lumen contour. Similarly stretches of fibrinoid (magenta), vascular smooth muscle (green) and endothelium (dark blue) are projected on the lumen contour, and can thus be related either to the total lumen contour or to the presence or absence of endovascular trophoblast.
Because we needed the full implantation sites for this analysis, i.e. with placentas still attached, placental weights could not be determined. Instead we measured the area of the placental compartments (labyrinth and trophospongium) in the studied set of parallel sections after delineating the different contours on the PAS stained section as previously described [5], which should be sufficiently representative considering the central position of the selected sets of parallel sections containing a central maternal arterial channel.
2.4. Doppler studies
Doppler analyses were performed on a separate group of transgenic rats. The pregnant rats were imaged transcutaneously on day 21 using the ultrasound biomicroscope (UBM) and the transducer (Model Vevo 660, VisualSonics Inc., Toronto, Canada). Doppler waveforms were obtained from the uterine and the arcuate artery. Peak systolic velocity (PSV) and end diastolic velocity (EDV) were measured and averaged from three consecutive cardiac cycles that were not affected by maternal movements. The resistance index (RI) was calculated using the formula RI=(PSV-EDV)/PSV [11].
2.5. Statistics
For each group, the total number of resorptions was expressed as percentage of the total number of implantation sites present in that group, irrespective of the number of mothers. Similarly, for each group, numbers of cross-sections with atherosis-like lesions and invaded and non-invaded SA cross-sections were expressed as percentages of the total number of SA cross-sections examined in that group, irrespective of the number of implantation sites. The foregoing data were evaluated by Fisher's exact test. Fetal weights were analysed with 1-way ANOVA followed by the Bonferroni post test, and are represented as means and standard errors of the means, calculated per group. Data on litter sizes, histological features and Doppler waveforms were analysed using the Mann–Whitney test and are represented as medians and ranges, per group, irrespective of number of implantation sites. P values<0.05 were considered as statistically significant.
3. Results
There were no significant differences in litter sizes (expressed as median and ranges) (C: 11 [8–14], RM: 8 [7–12], PE: 12 [7–16]) and numbers of resorptions (expressed as % of total implantations) (C: 8%, RM: 6%, PE: 3%) between any of the three groups. The fetal weights (expressed as means±SEM) in the PE group (3.11g±0.09) were significantly lower than in the C group (3.91g±0.08) (P < 0.0001) and the RM group (3.73g + 0.08) (P < 0.0001). Total placental areas (expressed as medians {ranges}) were 9.4 {7.4–12.1}, 8.6 {7.6–14.5} and 7.6 { 6.2–10.1}mm2 in the C, RM and PE groups; trophospongial areas were 2.3 {1.9–3.3}, 2.5 {1.4–4.0} and 1.5 {1.3–3.7} mm2 in the C, RM and PE groups; and labyrinthine areas were 7.0 {4.9–8.7}, 6.4 {5.8–10.4} and 5.9 {4.8–7.2} mm2 in the C, RM and PE groups respectively. The total placental area (P < 0.05) and the trophospongial area (P < 0.05) were significantly smaller in the PE group than in the C group.
3.1. Endovascular trophoblast and associated remodelling in the whole mesometrial triangle
At this stage of pregnancy, it was difficult to distinguish between interstitial trophoblasts, which closely approached the arterial lumina, and the ET (previously intraluminal) which were now often covered again by endothelium (see below) and thus became intramural. The different morphological characteristics used to define interstitial trophoblast and ET on day 18 were applied to the day 21 biopsies, although the differences between the two types of trophoblasts were not always as clear as on day 18. Interstitial trophoblasts are usually rounded cells with empty cytoplasmic areas caused by leaching of glycogen, while the ET have a dense cytoplasm, are still rounded when they are intraluminal, but tend to be flattened when they have become intramural. Moreover the associated fibrinoid, which had been so obvious underneath the intraluminal trophoblast on day 18, often disappeared once the ET became intramural. For the quantification of invasion and associated vascular remodelling, we did not distinguish between intraluminal and intramural trophoblasts, which were both always designated as endovascular (Fig. 3).
Fig. 3
Parallel cross-sections of an invaded spiral artery immunostained for cytokeratin (A) and the endothelial marker CD31 (B). The intraluminal endovascular trophoblast (arrow) is associated with fibrinoid (A), the intramural endovascular trophoblast is flattened (A) (arrowheads). Reendothelialization has occurred in the contour part not covered by intraluminal trophoblast (between bars) (B).
A total of 141, 141 and 114 SA cross-sections were examined in the MT of the 9C, 9 RM and 9 PE rats respectively (Table 2). In these three groups 81% (C), 59% (RM) and 95% (PE) of all SA cross-sections were invaded by ET, with a significantly higher percentage in the PE group compared to the RM group (P < 0.0001) and a lower percentage in the RM group than in the C group (P < 0.0001), but no difference between PE and C. When expressed as percentages of the total SA contour length (see Materials and Methods), there were significantly more ET in both the PE and the C groups compared to the RM group (P < 0.0001), but there was no difference between the PE and C groups (Table 2). Moreover, in several arterial cross-sections intraluminal endovascular trophoblasts with underlying fibrinoid were present within the lumen contour (Fig. 3). This was especially obvious in the PE group, with intraluminal ET being present in 64.8% of the invaded SA cross-sections, which is significantly higher than in the C (20.5%) (P < 0.0001) and in the RM group (22.7%) (P<0.01). There were also more intraluminal trophoblasts present in the invaded SA cross-sections in the RM group compared to the C group (P < 0.01).
Table 2. Presence of endovascular trophoblast, fibrinoid, vascular smooth muscle and CD31 positive endothelium in spiral artery cross-sections in the complete mesometrial triangle.d
| Control n = 141 | Reverse Mating n = 141 | Preeclampsia n = 114 | |
|---|---|---|---|
| % ET in SA contour | 49.4 {0–100} | 11.0 {0–100}a< | 41.5 {0–100}c> |
| % Fibrinoid in SA contour | 0 {0–100} | 0 {0–100} | 17.7 {0–100}b/c> |
| % Fibrinoid in sub-ET | 0 {0–100} | 0 {0–100} | 39.1 {0–100}b/c> |
| % Fibrinoid in extra-ET | 0 {0–100} | 0 {0–43} | 0 {0–100}b/c> |
| % VSM in SA contour | 27.0 {0–100} | 33.8 {0–100} | 40.3 {0–100}b> |
| % VSM in sub-ET | 13.6 {0–100} | 0 {0–100}a< | 27.1 {0–100}b/c> |
| % VSM in extra-ET | 22.0 {0–100} | 31.3 {0–100} | 39.1 {0–100}b> |
| % Endothelium in SA contour | 100 {0–100} | 100 {0–100} | 100 {0–100}b< |
| % Endothelium in sub-ET | 100 {0–100} | 74.6 {0–100}a< | 100 {0–100}c> |
| % Endothelium in extra-ET | 100 {0–100} | 100 {0–100} | 100 {0–100} |
aRM vs C, P < 0.01. |
bPE vs C, P < 0.05. |
cPE vs RM, P < 0.0001. |
dCalculated per SA cross-sections as % of the total SA contour length, % of the contour length containing trophoblast (sub-ET), and the % of the contour length outside the trophoblast-containing parts of the SA (extra-ET), respectively. “Absence of a sub-ET or extra-ET part resulted in a zero value for the respective %.” |
The amount of fibrinoid in the arterial wall, expressed as a percentage of the total lumen contour, the lumen contour parts containing endovascular trophoblast (sub-ET) and the lumen contour parts without endovascular trophoblast (extra-ET), was significantly higher in the PE group compared to the RM and control group (P < 0.0001) (Table 2).
The length of the VSM, expressed as a percentage of the total lumen contour, was significantly higher in the PE group than in the C group (P < 0.01). Results for VSM in the sub-ET contour parts showed more VSM in the PE group compared to the C (P < 0.01) and the RM group (P < 0.0001). The RM group showed less VSM in the sub-ET contour parts compared to the C group (P < 0.01). Also in the extra-ET contour parts, the PE group had more VSM than the C group (P < 0.05) (Table 2).
In the PE group, significantly less endothelium was covering the total lumen contour when compared to the C group (P < 0.05), but for the sub-ET and extra-ET lumen contour parts the differences were not significant. In the RM group there is significantly less endothelium in the sub-ET lumen contour parts compared to both the PE and the C groups (P < 0.0001) (Table 2).
Comparing invaded with non-invaded SA cross-sections, there was significantly less VSM in the extra-ET parts of invaded than in the vessel wall of non-invaded cross-sections in all three groups (P < 0.01). There was also significantly less endothelium covering the extra-ET lumen contour parts of the invaded than in the vessel wall of the non-invaded SA cross-sections in the C group (P < 0.05) and the RM group (P < 0.001) (Table 3). In contrast to the C group, in the PE group there was no significant difference in percentage of endothelium between the sub- and extra-ET parts of the invaded cross-sections.
Table 3. Presence of fibrinoid, vascular smooth muscle and CD31 positive endothelium in invaded and non-invaded spiral artery cross-sections in the complete mesometrial triangle.d
| Sub-ET invaded SA | Extra-ET invaded SA | Non-invaded SA | |
|---|---|---|---|
| CONTROL | n = 114 | n = 114 | n = 27 |
| Fibrinoid | 0 {0–100}a/b> | 0 {0–100} | 0 {0–0} |
| VSM | 19.3 {0–100}b< | 15.5 {0–100}c< | 37.5 {0–100} |
| Endothelium | 100 {0–100}a> | 100 {0–100}c< | 100 {0–100} |
| REVERSE MATING | n = 84 | n = 84 | n = 58 |
| Fibrinoid | 0 {0–100}a/b> | 0 {0–42.7}c> | 0 {0–0} |
| VSM | 25.9 {0–100}b< | 14.7 {0–100}c< | 44.3 {0–100} |
| Endothelium | 100 {0–100}b< | 100 {0–100}c< | 100 {0–100} |
| PREECLAMPSIA | n = 108 | n = 108 | n = 6 |
| Fibrinoid | 46.2 {0–100}a/b> | 0 {0–100} | 0 {0–0} |
| VSM | 29.2 {0–100}b< | 37.1 {0–100}c< | 100 {34.1–100} |
| Endothelium | 100 {0–100} | 100 {0–100} | 100 {100–100} |
aSub-ET invaded SA vs Extra-ET invaded SA, P < 0.001. |
bSub-ET invaded SA vs Non-invaded SA, P < 0.05. |
cExtra-ET invaded SA vs Non-invaded SA, P < 0.01. |
dCalculated per SA cross-section as % of the contour length containing endovascular trophoblast (Sub-ET), of invaded spiral artery parts without endovascular trophoblast (Extra-ET) and as % of the total SA contour length of non-invaded SA, respectively. |
3.2. Endovascular trophoblast invasion and associated remodelling in three depth levels of the mesometrial triangle
The number of SA cross-sections in depth level 1 remained low because of the straight course of the arteries, joining to maternal arterial channels. Depth level 1 has also the smallest surface area, while the cross-sections of the SA joining the maternal arterial channels are more dilated and have a thicker wall than in deeper levels. Depth levels 2 and 3 show increasing numbers of SA cross-sections because of the increased coiling of the vessels.
A comparison of the overall ET invasion between the three groups showed that 66% of all invaded SA in the PE group were situated in depth level 3, which is significantly more than 33% in the C group (P < 0.0001) and 25% in the RM group (P < 0.0001).
A clear trend to a more intense and deeper ET invasion in the PE group was revealed by the percentages of invaded SA calculated for the three different depth levels separately: in depth level 1, 100% of all SA were invaded in all three groups; in depth level 2, 91%, 84% and 90% of all SA were invaded in the C, RM and PE group respectively, and in depth level 3, 64%, 30% and 96% of the SA were invaded in the C, RM and PE group respectively (P < 0.0001 between PE and C, and between PE and RM, P < 0.001 between RM and C). This was also reflected by the percentage of total lumen contour length containing ET. In depth level 1, there was no significant difference between the three groups. In depth level 2, there was more ET present in the PE group (P < 0.05) and the C group (P=0.01) compared to the RM group. In depth level 3 the highest % of ET was present in the PE group compared to the C group (P < 0.001) and the RM group (P < 0.0001), the latter showing less ET in the total lumen contour compared to the C group (P < 0.001) (Table 4).
Table 4. Presence of endovascular trophoblast in spiral artery cross-sections in three depth levels of the mesometrial triangle.d
| Control n = 141 | Reverse Mating n = 141 | Preeclampsia n = 114 | |
|---|---|---|---|
| Depth Level 1 | 100 {50.5–100} | 100 {14.7–100} | 100 {39.1–100} |
| (n=13) | (n=11) | (n=9) | |
| Depth Level 2 | 72.2 {0–100} | 31 {0–100}a | 64 {0–100}c |
| (n=69) | (n=62) | (n=31) | |
| Depth Level 3 | 14.8 {0–100} | 0 {0–100}a | 31.3 {0–100}b/c |
| (n=59) | (n=68) | (n=74) |
aRM vs C, P < 0.01. |
bPE vs C, P < 0.001. |
cPE vs RM, P < 0.05. |
dCalculated per SA cross-sections as % of the total SA lumen contour length. |
Regarding the vascular remodelling features at the different depth levels, we found no difference in fibrinoid deposition and amount of VSM as percentage of total lumen contours between the three groups at depth level 1, but here significantly less endothelium was present in the total lumen contours in the PE than in the C group (P < 0.05). At depth level 2, there was more fibrinoid in the PE group compared to the C (P < 0.01) and the RM group (P < 0.05). No difference was found in amount of VSM and endothelium between the three groups. At this depth level, endothelial covering attained almost 100% of the lumen contours of the SA in all three groups. At depth level 3, the same difference was found for the % fibrinoid as in level 2, i.e. a higher % fibrinoid in the PE group compared to the C (P < 0.0001) and the RM group (P < 0.0001). There was more VSM present in the PE group compared to the C group (P < 0.05). At this depth level there was almost 100% endothelial covering of the lumen contours in all three groups (data not shown).
3.3. CD31 immunostaining of endovascular trophoblast
On day 18, we noted positive staining of CD31 in some of the ET although showing a lower intensity than the typical staining of endothelial cells. On day 21, such CD31 staining was less pronounced than on day 18 and was limited to intraluminal endovascular trophoblasts in all groups. CD31 staining in intraluminal trophoblasts showed positive cytoplasmic dots, thus being considerably different from the very intense labeling of the endothelial cell membrane (Fig. 4).
Fig. 4
Parallel cross-sections of a partially invaded spiral artery immunostained for cytokeratin (A) and the endothelial marker CD31 (B). Endovascular trophoblasts (arrow) show CD31 positivity at a lower staining intensity than the endothelial cells (arrowhead). Notice the different cell size of trophoblasts and endothelial cells.
3.4. Atherosis-like lesions in spiral arteries
Atherosis-like lesions were earlier described as areas of focal necrosis or arteriolosclerosis by Dechend and colleagues [4]. In pregnancy day 18 of the transgenic PE rat model, such lesions were mostly present in SA of the RM group. The same results were found for day 21. Atherosis-like lesions were present in 25/141 SA cross-sections in seven out of nine placentas of the C group, in 46/141 SA cross-sections in eight out of nine placentas of the RM group and in 11/114 SA cross-sections in five out of nine placentas of the PE group. There was no significant difference in the incidence of these vascular lesions between the PE and the C group. However, significantly more atherotic lesions were found in the RM group compared to the C group (P < 0.01) and the PE group (P < 0.0001). Atherosis-like lesions were not associated with the presence of infarcted areas in the placenta.
3.5. Doppler results
Resistance index measurements in the uterine artery did not show any significant difference between the three groups. In the arcuate artery, the resistance index of the PE group (0.4582 {0.3524–0.4770}) was significantly lower compared to the C group (0.5527 {0.5000–0.7301}) (P < 0. 05).
4. Discussion
In a previous publication we reported a more extensive ET invasion in the SA of this transgenic PE rat model on day 18 of pregnancy [5]. This observation differs from earlier findings in human PE, where the myometrial SA segments showed a deficient remodelling which was thought to be the consequence of impaired ET invasion [1]. The aim of the present study was to examine the fate of the deeply invaded ET (observed on day 18) and their effect on vascular remodelling in the rat at term. Our first working hypothesis was that the more deeply invaded ET would have been eliminated at term. This hypothesis was refuted by our observations on day 21, which showed, as on day 18, a higher % of ET covering the lumen contours in the deepest depth level of the MT of the PE group compared to the other two groups (Table 4). However, when considering the whole MT rather than depth level 3 only, these values were no longer significantly different between the PE and C group (Table 2), suggesting a (partial) catching up of invasion in the C group later in gestation. This finding might suggest that endovascular invasion of the mesometrial triangle may be accelerated and/or may have an earlier start in the PE group. In an extra series of day 15 pregnant rats we did already find more invaded spiral arteries in the PE than in the C group (unpublished observations), but the exact time when endovascular invasion of the mesometrial triangle is started in PE and C rats could not yet be determined. If confirmed, a possibly earlier onset of invasion in PE rats may have resulted in a lead in invasion depth on day 18, which was partially, but not completely overtaken by the C group on day 21 as seen in the results in depth level 3 (Table 4)
On day 21, it was difficult to distinguish between ET and interstitial trophoblast, since most of the ET were embedded into the vessel walls and thus had become intramural, the fibrinoid deposits were less obvious and the arterial wall was covered by new endothelium. The difficulty in distinguishing different trophoblastic cell types brought in an element of uncertainty which we tried to overcome by a rigid application of the morphological criteria described in the Results. Admittedly, there is no certainty that all the trophoblasts immediately below the restored endothelium are really derived from endovascular trophoblasts, and this may therefore confound our quantitation of endovascular invasion. Nevertheless, it is clear that in the PE group more intraluminal ET (not covered by endothelium) were still lining the vessel walls, which show, as expected, a higher percentage of fibrinoid. This higher number of cross-sections with intraluminal trophoblasts in the PE group may indicate a still ongoing endovascular migration at that time. Additionally, ET in PE rats may be less able to embed into the arterial walls and may thus be prevented to participate in full trophoblast-associated remodelling. In contrast to what has been observed in the human, rat ET is not completely embedded within the fibrinoid material, except in depth level 1 at the junction with a transplacental maternal channel. Instead, this fibrinoid forms a thick basal lamina supporting seemingly free-lying cells which apparently disappears when the trophoblast becomes intramural. It is not inconceivable that intraluminal trophoblasts are able to pass the deposited fibrinoid layer, which may explain the not-infrequent finding of fibrinoid stretches in extra-ET arterial cross-section parts.
Concerning vascular remodelling (hypothesis 2), we found significantly more VSM (expressed as percentages of their lengths versus the total lumen contour, sub-ET as well as extra-ET, respectively) in the PE group compared to the C group. A similar finding was previously reported for day 18 [5], but only in sub-ET vascular regions. The absence of a significant difference for total lumen contour and extra-ET may be explained by a confounding effect of a considerably lower % of non-invaded artery cross-sections in PE compared to C [5]. These observations might indicate that on day 21 there is also less VSM breakdown in the extra-ET regions in PE than in C rats. This seems to contradict earlier ideas of an exclusive VSM remodelling role by invading ET in the human [1], [6]. Indeed in mice VSM breakdown in SA is effected by uterine NK cells [12], [13], and also in the Wistar rat it was shown that the VSM had undergone a marked thinning, but not complete breakdown, in deciduomata, i.e. in the absence of any trophoblast [2]. Even in the human it was recently confirmed that the early steps in VSM remodelling in the decidua occur prior to trophoblast invasion, following infiltrations by leukocytes which include uterine NK cells (uNK) [14]. We therefore postulate that the uNK cells in the rat, representing another cell type involved in VSM breakdown, may be deficient in the transgenic rats. Meanwhile, the fact that in all three groups studied the extra-ET parts of the invaded SA cross-sections did show less VSM than the non-invaded cross-sections, indicates at least some relationship of VSM breakdown with the presence of ET.
In all groups re-endothelialization of the SA was observed, confirming previously reported findings in normal pregnant Wistar rats on day 21 [2]. This is a phenomenon also described in the human [1]. In the present study, we hypothesized that invaded vessels which had been remodelled on day 18 would show maternal repair on day 21 (hypothesis 3). However, only at depth level 1 there was a significant difference in re-endothelialization between the C and the transgenic groups, albeit in an opposite sense than hypothesized, i.e. more endothelium being present in the C group compared to the RM and PE groups.
The Doppler results were in agreement with our finding of a more intense ET invasion in the PE group compared to the C and RM groups. The PE rats showed a lower resistance index in the distal part of the uterine artery. These results seem to be contradictory to the persistence of VSM in PE, which may however be explained by the fact that the measured percentages of VSM may be insufficient to result in a significant vascular narrowing of the arteries. Moreover, the remodelling defects may be (partially) compensated by the more intense ET invasion. Furthermore, a lower arterial resistance in the maternal blood supply of the placenta in the PE than in the C group is difficult to reconcile with the smaller-sized placentas and lower fetal weights in the PE compared to the C group on day 21. It could be postulated that the smaller-sized placenta, which is mainly due to a smaller trophospongium, results from a depletion of invasive trophoblastic cells which are, at least in the mouse, derived from this placental compartment [15].
Hypothesis four suggested an increased incidence of atherosis-like lesions in the PE group which would be associated with zones of placental infarction. Such lesions were indeed seen in some SA cross-sections but were never associated with infarcted placental regions as reported for the human [1], which refutes our fourth working hypothesis. However, when considering the anatomy of the afferent vascular system of the mouse and the rat placenta [16] such associations are not likely to occur in those animals. Indeed the trophoblast-lined arterial channels, which traverse the placenta, are formed by convergence of two or three SA. Atherosis-like lesions in one artery is therefore not likely to have a major impact on the maternal blood supply to the placenta, since individual SA are not linked to particular regions of placental tissue. This minimal impact is reflected by the fact that there are no differences in litter sizes and resorptions between any of the three groups. Strangely enough, the RM group in the present rat model showed the highest incidence of atherosis-like lesions while these animals did not develop PE-like symptoms and had normal litter sizes.
The RM group was initially planned to serve as an extra control, but after completing both the previous day 18 [5] and the present day 21 study, it has become clear that this group should rather be considered as a different transgenic combination. Although the RM group is clinically similar to the C group in lacking PE symptoms, it is clear that the data on trophoblast invasion and vascular remodelling for this group are aberrant from both the C and the PE data. Differences in the expression level and location of the extra human genes in the RM and the PE group may be important [17]. In the RM group the female is carrying the human renin gene, perhaps this leads to the human angiotensinogen in the trophoblast of the RM group being rapidly processed, while in the PE group the mother lacks the human renin to process the surplus of maternally expressed angiotensinogen. However, we do not know whether and how differences in the processing of human angiotensinogen in these animals may have an impact on trophoblast invasion. Also the more frequent finding of atherosis-like lesions in the RM group compared to the PE and the C group is intriguing, suggesting an impaired placentation without a manifest pathology.
We conclude that in this transgenic PE rat model the deeper and more intense ET invasion seen on day 18 of pregnancy is still progressing on day 21, although there is a partial catching up of invasion in the C group. Remodelling of the vascular smooth muscle is still lagging behind in PE rats on day 21, but no difference in endothelial repair could be found in the deeper invaded SA. The association of ET invasion and lower vascular resistance with less placental and fetal growth in the transgenic PE rats is not yet understood. Further investigation of possible effects of the extra human angiotensinogen and human renin genes in this model is also required. Finally, our observations make clear that impaired invasion as such is not the only pathophysiological event that may trigger the syndrome of preeclampsia. This fits very well with the growing awareness that the disease, as defined by a set of clinical features, may have multiple causes, and that the precise aetiology is still unknown. Using rat models may help to obtain a better understanding of possible interactions between trophoblast and maternal tissues, but may not exactly mirror the situation of human pregnancy.
Acknowledgement
The present study was supported by grant G.0547.06N of FWO Vlaanderen to R.P. and M.H.
References
- . The role of the spiral arteries in the pathogenesis of preeclampsia. Obstet Gynecol Annu. 1972;1:177–191
- . Endovascular trophoblast invasion and associated structural changes in uterine spiral arteries of the pregnant rat. Placenta. 2005;26:574–584
- . Rats transgenic for human renin and human angiotensinogen as a model for gestational hypertension. J Am Soc Nephrol. 2000;11:2056–2061
- Agonistic autoantibodies to the AT1 receptor in a transgenic rat model of preeclampsia. Hypertension. 2005;45:742–746
- Endovascular trophoblast invasion, spiral artery remodelling and uteroplacental haemodynamics in a transgenic rat model of pre-eclampsia. Placenta. 2008;29:614–623
- . The uterine spiral arteries in human pregnancy: facts and controversies. Placenta. 2006;27:939–958
- . Trophoblast invasion in pre-eclampsia and other pregnancy disorders. In: Lyall F, Belfort M editor. Pre-eclampsia, etiology and clinical practice. Cambridge: Cambridge University Press; 2007;p. 3–19
- . The making of the placental bed. Eur J Obstet Gynecol Reprod Biol. 1984;18:255–266
- . A study of placental bed spiral arteries and trophoblast invasion in normal and severe pre-eclamptic pregnancies. Br J Obstet Gynaecol. 1994;101:669–674
- . A simple technique for preservation of fixation-sensitive antigens in paraffin-embedded tissues. J Histochem Cytochem. 1994;42:1127–1134
- Uterine vascular function in a transgenic preeclampsia rat model. Hypertension. 2008;51:547–553
- Ultrastructural studies of implantation sites from mice deficient in uterine natural killer cells. Placenta. 2000;21:693–702
- Angiogenic growth factor messenger ribonucleic acids in uterine natural killer cells. J Clin Endocrinol Metab. 2001;86:1823–1834
- . Evidence for immune cell involvement in decidual spiral arteriole remodeling in early human pregnancy. Am J Pathol. 2009;174:1959–1971
- . Diverse subtupes and developmental origins of trophoblast giant cells in the mouse placenta. Dev Biol. 2007;304:567–578
- Interactions between trophoblast cells and the maternal and fetal circulation in the mouse placenta. Dev Biol. 2002;250:358–373
- Characterization of a transgenic rat model for preeclampsia – new insights. Hypertens Pregnancy. 2008;27:512
PII: S0143-4004(10)00037-8
doi:10.1016/j.placenta.2010.01.011
© 2010 Elsevier Ltd. All rights reserved.
