Published online before print July 18, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Biomedical Sciences, University of Guelph, Guelph, Ontario, Canada; and
Department of Anatomy and Cell Biology, Queen's University, Kingston, Ontario, Canada
2 Correspondence: Department of Biomedical Sciences, University of Guelph, Guelph, ON, Canada, N1G2W1. E-mail: ctayade{at}uoguelph.ca
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and with elevations in IFN-
and TNF- and suggested endometrial lymphocytes played a pivotal, proangiogenic role in fetal survival. Here, we address more comprehensively porcine endometrial angiogenesis by comparing transcription between endometrial endothelium and lymphocytes during early (gd20) and midgestation (gd50) losses and by incorporation of histopathology and protein immunolocalization of VEGF, placenta growth factor (PlGF), VEGF receptor I (VEGFRI), and VEGFRII. In healthy sites, endometrial lymphocytes transcribed more VEGF at gd50 than gd20, and transcripts were more abundant in lymphocytes than in endothelium or trophoblast. Arterial endothelial cells showed the most abundant transcription of PlGF. With fetal arrest, maternal transcripts for VEGF but not PlGF dropped, and fetal transcripts remained relatively stable. Maternal and fetal HIF-1 transcription declined. Lymphocytes preferentially transcribed VEGFRI over VEGFRII, and endometrial arterial endothelium and trophoblast preferentially transcribed VEGFRII. IFN- and TNF- transcripts were present in gd20 and gd50 healthy- and arresting-implantation sites. gd20 arrest was associated with greater transcription of IFN- than TNF- in maternal and fetal tissues. At gd50, this was reversed. Endometrial, vascular pathology was evident only at gd50. These data suggest the critical importance for lymphocyte-driven endometrial angiogenesis, which extends to midgestation.
Key Words: endothelium • pregnancy • proinflammatory cytokines • trophoblast • vascular endothelial growth factor
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In all mammals studied, including pigs, maternal endometrium becomes transiently enriched in lymphocytes, particularly NK cells [6
7
8
9
10
], early in gestation. Uterine (u)NK cells have been best studied in species with invasive, hemochorial placentation [7
, 8
]. In these species, uNK cells are recruited during endometrial decidualization, expand during early gestation, and decline in numbers from midgestation [11
]. In mice, uNK cells trigger pregnancy-induced structural enlargement of maternal uterine arteries by mechanisms, which include IFN- production [12
]. In mice and humans, uNK cells synthesize hypoxia-inducible factor 1- (HIF-1)-regulated vascular endothelial growth factor (VEGF) and its related protein placenta growth factor (PlGF) [13
14
15
16
], a splice variant in the VEGF family, which binds with high affinity to VEGF receptor I (VEGFRI) and serves a decoy function by releasing VEGF from this receptor [17
]. PlGF does not bind to VEGFRII [17
].
Porcine endometrium does not decidualize during pregnancy [1
2
3
4
5
6
], and uNK cell recruitment is triggered by an attachment (implantation) of noninvasive conceptuses between gd12 and gd15 [6
, 10
]. In pigs, uNK cells are detected first at gd12 [6
] and endometrial angiogenesis onsets, 
gd15 [18
, 19
]. Compared with trophoblast, porcine endometrial lymphocytes transcribe relatively high levels of VEGF and HIF-1 between gd15 and gd23 [20
] in healthy but not arresting, conceptus-attachment sites. During this interval, porcine trophoblast produces IFN-
and - [21
], which function during blastocyst attachment and in initiation of endometrial angiogenesis [22
]. Little is known about porcine uNK cells after gd35.
The first aim of this investigation was to determine the proangiogeneic potential of midgestation (gd50) porcine uterine lymphocytes and to quantify differences in the potential for lymphocyte-promoted angiogenesis between gd20 and gd50 and between healthy and arresting implantation sites. Transcription of VEGF, HIF-1, PlGF, and VEGFRI and -RII was analyzed. Transcript levels in lymphocytes were compared with those in laser capture-microdissected (LCM) endometrial arterial endothelium, endometrial biopsies, and trophoblasts to assess the relative importance of lymphocyte-mediated promotion of implantation-site angiogenesis. Immunohistochemistry was used to verify and localize gene translation.
Our second aim was to evaluate relative transcription of proinflammatory cytokines. Elevated IFN- and TNF- are associated with negative pregnancy outcomes [23
]. In spontaneously aborting women, an IFN-/TNF- synergy is thought to activate procoagulants in endothelial cells, reducing blood flow to implantation sites and promoting embryonic or fetal arrest and subsequent death [24
]. We reported previously localized increases in transcription of IFN-, TNF-, IL-1ß, and IL-1R in porcine endometrial biopsies collected from gd23 arresting-attachment sites [25
]. To understand the cell types involved in this response and whether endometrial responses in midgestation failure are similar to those in peri-implantation failure, IFN- and TNF- transcripts were quantified at gd20 and gd50 in endometrial biopsies, lymphocytes, and arterial endothelium and in trophoblast collected from healthy- and arresting-attachment sites. Angiogenic and proinflammatory cytokine gene expression patterns differed by gestation length, cell type, and conceptus health status. Extremely localized differences occurred in lymphocytes associating with adjacent-arresting versus healthy-attachment sites. These data support the hypothesis that endometrium actively contributes to loss of fetal viability and that endometrial lymphocyte-promoted angiogenesis contributes to fetal health to at least midgestation.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
30 mg) before the next conceptus was removed. Samples were collected from an individual attachment site and not pooled. Trophoblast was dissected, and at gd50, each fetus in each litter was weighed and measured. Molecular and histological analyses were conducted independently on three healthy and three arresting conceptuses from each gilt at gd20 and gd50. When possible, these three paired samples per gilt were chosen as adjacent littermates. Samples were placed into 600 µl lysis buffer or into embedding medium (Cryomatrix, ThermoShandon, Pittsburgh, PA, USA) for cryosectioning and LCM or 4% paraformaldehyde for routine histopathology or immunohistology.
RNA isolation from conceptus-attachment sites and trophoblasts
Tissues were disrupted using Kontes pestles (Fisher Scientific, Ottawa, ON, Canada), and total RNA was extracted using Qiagen's RNeasy mini kit (Qiagen, Mississauga, ON, Canada). The cleared tissue lysate was mixed with 700 µl 70% ethanol and loaded on an RNeasy mini column and centrifuged at 8000 g for 30 s. After washing the column with wash buffer, RNA was eluted using 50 µl nuclease-free water by centrifugation at 8000 g for 1 min. After assessing the concentration and purity, RNA was frozen at –80°C until required.
LCM, RNA purification, and amplification
Snap-frozen endometrial biopsies were cut as 7 µm-thick sections (CM3050S, Leica, Richmond Hill, ON, Canada), mounted onto positively charged slides (Superfrost/Plus, Fisher Scientific), and stained using a modified, rapid H&E protocol described previously [25
] to identify endometrial lymphocytes and endothelium. All solutions, including the stains, were supplemented with 0.5 unit/µl RNase inhibitor (Promega, Madison, WI, USA). After air-drying (5 min), individual endometrial lymphocytes or endothelial cells were dissected onto high-sensitivity LCM caps using laser pulse settings of 55 mW for 0.7 ms at 7.5 µm (Pix Cell IIe, Arcturus Engineering, Mountain View, CA, USA). Five hundred captured endometrial lymphocytes or endothelial cells were pooled for each sample. RNA was extracted using Picopure RNA isolation kit (Arcturus Engineering) as per the manufacturer's instructions. The cell extract was purified using an RNA purification column. The RNA was eluted in 11 µl nuclease-free water and stored at –80°C. RNA amplification was carried out as described previously [25
, 26
] using the MessageAmp II antisense (a)RNA kit (Ambion, Austin, TX, USA) as per the manufacturer's instructions. cDNA was synthesized and eluted in 20 µl nuclease-free water and was used for in vitro transcription to synthesize aRNA as per kit instructions. The amplified aRNA was eluted in 100 µl nuclease-free water and stored at –80°C.
cDNA synthesis, cloning, and sequencing
Total RNA from the endometrial biopsies and trophoblast and aRNA from the LCM-isolated endometrial lymphocytes and endothelium were reverse-transcribed using the First Strand cDNA synthesis kit (Amersham, Bioscience, Piscataway, NJ, USA) as per the manufacturer's instructions. The resulting cDNA was stored at –20°C. PCR-amplified products of VEGF, PlGF, VEGFRI, VEGFRII, HIF-1, IFN-, and TNF- were cloned using the topoisomerase-TA cloning kit (Invitrogen Life Technologies, Burlington, ON, Canada) as per the manufacturer's instructions. Plasmid DNA was purified using Genelute plasmid DNA purification kit (Sigma-Aldrich, St. Louis, MO, USA). Sequencing was performed at the Laboratory Services Division, University of Guelph. The resultant sequences were analyzed by the BLASTN portal of the National Center for Biotechnology Information and were deposited to GenBank (Table 1
).
|
View this table: [in a new window] |
Table 1. Porcine Partial Coding Sequences Deposited to GenBank
|
|
View this table: [in a new window] |
Table 2. Primers Used for Real-Time Quantitative PCR
|
Statistical analysis
Statistical analyses were performed by the nonparametric Friedman test using SAS software (SAS 8.2, SAS Institute, Cary, NC, USA) for comparison among groups. Variances between attachment sites within the animals or between animals were handled by sum of ranks using the nonparametric Friedman test. Post hoc analysis for planned comparisons between different paired groups was done using the Wilcoxon signed rank test. A value of P < 0.05 was considered significant.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (76K): [in a new window] |
Figure 1. An example showing healthy (A, a) and arresting (A, b) conceptuses (littermates). At gd20, vascular differences were the predominant discriminating feature. At gd50, fetal size and weight were additional identification criteria used. (B) Means (± SD) for length and weight of 38 healthy and 12 arresting gd50 conceptuses in this study. This rate of conceptus loss is consistent with our previous observations [25
] and published studies. No significant histopathological alterations were found at gd20. At gd50, histological analyses for elastin (C, a) revealed uniformly distributed elastin in arterial walls at healthy-attachment sites (solid arrowheads, C, b). In arresting sites, arterial elastin was absent or fragmented. An abnormally high amount of perivascular collagen was present at arresting sites (C, d) and appeared to induce vascular lumen stenosis. (C, a–d, original magnifications, x200.)
|
 View larger version (26K): [in a new window] |
Figure 2. Quantitative analysis of VEGF (A) and PlGF (B) transcription in endometrial tissue, endometrial lymphocytes (ENDO LY), endothelium, and trophoblasts obtained from the attachment sites of three healthy (H) and three arresting (A) conceptuses from each gilt at gd20 and gd50 (n=4/gd time-point). Gene expression is analyzed by RelQuant software (Roche Diagnostics) and is expressed as a normalized ratio with the housekeeping gene ß-actin. VEGF was expressed abundantly in all of the endometrial microdomains associated with healthy conceptuses compared with their arresting counterparts (P<0.01). PlGF was expressed abundantly in endothelium compared with endometrial biopsies and endometrial lymphocytes, and lymphocytes associated with arresting, conceptus-attachment sites transcribed significantly more PlGF at gd20 and gd50 (P<0.001; B).
|
Expression of PlGF in endometrial biopsies, lymphocytes, endothelium, and trophoblasts
Endometrial transcription of PlGF was not elevated above that in the virgin uterus in gd20-healthy sites. Endometrium from gd20-arresting sites expressed significantly less PlGF (P<0.05) than either. PlGF transcription was elevated significantly at gd50, but there was no significant difference in PlGF transcription between healthy and arresting sites (Fig. 2B)
. Endometrial lymphocytes from all pregnant uteri had elevated PlGF transcription compared with lymphocytes from virgin uteri. Lymphocyte transcription at gd50 was significantly greater than at gd20 in cells from healthy and arresting sites. At gd20, lymphocytes associated with arresting, conceptus-attachment sites had significantly greater (P<0.05) PlGF transcription than lymphocytes from healthy sites, but the difference was not significant at gd50.
PlGF expression in endothelial cells from all pregnant uteri was greater than from NP uteri. In healthy-attachment sites, endothelium was the most abundant site of PlGF transcription with equivalent expression at gd20 and gd50. Endothelial cells from arresting sites had less PlGF transcription than cells from gd-matched, healthy sites [gd20 (P<0.01); gd50 (P<0.05)]. PlGF transcription also occurred in trophoblast. At gd20, no significant difference was found between trophoblast from healthy and arresting littermates. At gd50, trophoblast from arresting sites had fewer PlGF transcripts than trophoblast from healthy sites (P<0.05).
Expression of VEGFRI in endometrial biopsies, lymphocytes, endothelium, and trophoblasts
VEGFRI transcription was equally elevated in gd20 and gd50 endometrial biopsies from healthy sites compared with biopsies from NP uteri. VEGFRI expression was lower in endometrium associated with arresting conceptuses at gd20 and gd50 in comparison with biopsies from healthy sites (P<0.01; Fig. 3A
). VEGFR1 transcription was elevated in endometrial lymphocytes from pregnant compared with NP uteri and was higher in lymphocytes from healthy versus arresting sites at gd20 (P<0.05) and gd50 (P<0.01). Endometrial lymphocytes expressed significantly more VEGFRI than endothelium or trophoblasts.
 View larger version (27K): [in a new window] |
Figure 3. Expression of VEGFRI (A) and VEGFRII (B) in endometrial tissue, lymphocytes, endothelium, and trophoblasts associated with three healthy and three arresting conceptuses from each gilt at gd20 and gd50 (n=4/gd time-point). Endometrial lymphocytes transcribed more VEGFRI than trophoblasts, and trophoblasts had more abundant transcription of VEGFRII. Gene expression is relative to ß-actin.
|
Expression of VEGFRII in endometrial biopsies, lymphocytes, endothelium, and trophoblasts
In endometrium biopsies from healthy-attachment sites, VEGFRII transcription was greater (P<0.05) at gd20 and gd50 than in virgin endometrial biopsies, and there was more transcript abundance at gd20. Endometrial transcription of VEGFRII in arresting sites was also high at gd20 but dropped significantly at gd50 (P<0.05 vs. gd50-healthy). VEGFRII transcription was elevated in endometrial lymphocytes by pregnancy. At gd20 VEGFRII, transcription was greater in lymphocytes from healthy versus arresting sites (P<0.05). There was no significant difference in VEGFRII expression between lymphocytes from gd50-healthy versus gd50-arresting sites (Fig. 3B)
. Endometrial lymphocytes displayed greater transcription of VEGFRI than VEGFRII during successful and failing pregnancy, suggesting VEGFRI is the more dominant receptor used by porcine endometrial lymphocytes.
Transcription of VEGFRII in endothelial cells from NP uteri and gd20 from healthy sites was equal and greater than in endothelium from healthy gd50 sites (P<0.01). At gd50, VEGFRII expression was approximately fourfold greater in endothelial cells from healthy versus sites arresting (P<0.05), but at gd20, VEGFRII transcription was slightly higher in endothelium from arresting sites compared with healthy. The expressions of trophoblast of VEGFRII exceeded that measured in any maternal sample studied and were greater than its expression of VEGFRI. VEGFRII expression was equivalent in healthy gd20 and gd50 trophoblasts and significantly lower in gd-matched, arresting trophoblast [gd20 (P<0.001); gd50 (P<0.05)]. Preferential transcription of VEFGRII versus -RI by trophoblast suggests the responses of trophoblast to VEGF and PlGF would differ from the responses of colocalized, VEGFR1-dominant, maternal cells.
Localization of VEGF, PlGF, and VEGFRI and -RII in porcine implantation sites
Translated VEGF (Fig. 4
), PlGF (Fig. 5
), VEGFRI (Fig. 6
), and VEGFRII (Fig. 7
) were detected in healthy and arresting gd20 and gd50 conceptus-attachment sites. VEGF was expressed by uterine luminal epithelium (Fig. 4a
, 4d
, 4e
, 4g
, 4h)
, endometrial glands (Fig. 4h
and 4j)
, vascular endothelium (Fig. 4b)
, stromal cells (Fig. 4e)
, and endometrial lymphocytes (Fig. 4c
and 4f)
. Fetal trophoblasts were also VEGF-reactive (Fig. 4d)
. Although quantitative assessment of differences in VEGF localization between healthy- and arresting-attachment sites was not possible, we observed greater intensity of VEGF expression in healthy compared with arresting sites (Fig. 4i
and 4j)
.
 View larger version (112K): [in a new window] |
Figure 4. Immunohistochemical localization of VEGF in the attachment sites associated with healthy and arresting conceptuses at gd20 and gd50. VEGF was expressed abundantly in the luminal epithelium (a, d, e, g, h, i), endometrial glands (h, j), vascular endothelium (b), and endometrial lymphocytes (c, f). VEGF was also expressed in the trophoblasts (d). Black, white, red, yellow, and green arrowheads represent uterine epithelium, uterine glands, endometrial lymphocytes, endometrial endothelium, and trophoblasts, respectively. *, Vessels. EN, Endothelium; SMC, smooth muscle cells; Tr, trophoblasts; LE, luminal epithelium; UG, uterine glands. Original magnifications: a, d, e, x100; b, j, x400; c, f, x1000; g–i, x200.
|
 View larger version (113K): [in a new window] |
Figure 5. Representative expression of PlGF in attachment sites. PlGF expression was prominent in luminal epithelium (a, d, g, h), endometrial glands (d, e), vascular endothelium (b), and endometrial lymphocytes (c, f). PlGF immunoreactivity was more intense in arresting sites. Original magnifications: a, e, g, h–j, x100; b, x400; d, x200; c, f, x1000.
|
 View larger version (114K): [in a new window] |
Figure 6. VEGFRI was found in the attachment sites at gd20 and gd50. VEGFRI was localized predominantly in luminal epithelium, endothelium, endometrial glands, and stroma. Original magnifications: a, d, g, i, x100; b, e, h, j, x200; c, x1000; f, x400.
|
 View larger version (110K): [in a new window] |
Figure 7. VEGFRII was found in the attachment sites at gd20 and gd50. The pattern of VEGFRII expression was mostly similar to VEGRI with the exception of endometrial glands, which appeared negative for VEGFRII. Original magnifications, a, b, d, e–h, x100; i, x200; c, x1000.
|
VEGFRI and -RII staining was more intense in specimens from healthy sites. In addition to reactivity with luminal epithelium, endometrial glands (Fig. 6b and 6f) , and endothelium (Fig. 6h) , VEGFRI had strong reactivity with endometrial lymphocytes (Fig. 6c) , and trophoblasts were weakly reactive (Fig. 6d) . VEGFRI staining was more intense in gd50- than in gd20-attachment sites. VEGFRII was expressed by luminal epithelium, endometrial lymphocytes, vascular endothelium (Fig. 7c) , and trophoblast. Endometrial glands were nonreactive for VEGFRII (Fig. 7i) . Expression of these proteins was not uniform in any attachment site.
Expression of HIF-1 in endometrial biopsies, endometrial lymphocytes, endothelium, and trophoblasts
HIF-1 transcription was strongly induced in endometrium of attachment sites with healthy conceptuses (Fig. 8
). Endometrium associated with arresting conceptuses had fewer transcripts than gd-matched endometrium from healthy sites [gd20 (P<0.001); gd50 (P<0.005)]. HIF-1 expression in endometrial lymphocytes from healthy sites exceeded lymphocytes from virgin uterus or gd20- and gd50-arresting sites (P<0.05).
 View larger version (46K): [in a new window] |
Figure 8. Gene expression analysis of HIF-1 in endometrium, lymphocytes, endothelial cells, and trophoblasts obtained from the attachment sites of three healthy or three arresting conceptuses from each gilt at gd20 and gd50 (n=4/gd time-point). HIF-1 expression was significantly higher (P<0.01; P<0.05) in endometrium, lymphocytes, and endothelium from healthy, conceptus-attachment sites than from arresting, conceptus-attachment sites at gd20 and gd50. Transcription of HIF-1 in trophoblasts from healthy sites was greater (P<0.01; P<0.001) than in trophoblasts from arresting sites between gd20 and gd50. Gene expression is relative to ß-actin.
|
transcription in endothelial cells from virgin uteri was low but elevated by pregnancy. More HIF-1 transcripts were present in gd20 versus gd50 endothelium from healthy sites. A significant drop in HIF-1 transcripts occurred in endothelium associated with gd20-arresting sites compared with healthy (P<0.05). A significant gain in HIF-1 transcripts occurred in endothelium from gd50-arresting compared with healthy sites (P<0.05). HIF-1 transcription in trophoblasts from healthy sites was greater at gd50 than gd20 and dropped significantly at arresting sites [gd20 (P<0.05); gd50 (P<0.001)]. Only at gd50 were HIF-1 transcripts were more abundant in trophoblasts than in maternal lymphocytes from healthy sites.
Expression of IFN- and TNF- in endometrial biopsies, lymphocytes, endothelium, and trophoblasts
Equivalent, low expression of IFN- was found in endometrial biopsies from virgin and gd20- and gd50-healthy, conceptus-attachment sites (Fig. 9A
). This level was also measured in endometrium from gd50-failing, conceptus-attachment sites. In contrast, significantly elevated IFN- transcription was detected in endometrial biopsies from gd20-arresting sites (P<0.05). None of this elevation could be attributed to endothelium, which showed no IFN- transcripts under any condition studied. Pregnancy elevated IFN- transcription in endometrial lymphocytes above that seen in lymphocytes from virgin uteri. Similar levels were found at gd20 and gd50 in lymphocytes from healthy sites. Lymphocytes from gd20-arresting sites had significantly elevated IFN- transcription. This elevation was not seen in lymphocytes from arresting gd50 sites. IFN- transcription in healthy trophoblasts was much higher at gd20 than gd50 (P<0.001). In arresting sites, gd20 trophoblast transcribed more IFN- than in healthy sites (P<0.01). No significant difference was found in trophoblast IFN- transcription between healthy and arresting at gd50.
 View larger version (20K): [in a new window] |
Figure 9. (A) Transcription of IFN- in endometrium, lymphocytes, and trophoblasts collected from the attachment sites of three healthy and three arresting conceptuses from each gilt at gd20 and gd50 (n=4/gd time-point). Lymphocytes and trophoblasts transcribed more IFN- compared with endometrial biopsies. The expression of IFN- in endometrium and trophoblasts from arresting sites was significantly higher (P<0.05) at gd20. (B) Expression of TNF- in endometrial microdomains. The TNF- transcription decreased in endometrium from healthy sites through gd20 and gd50, but its expression was elevated (P<0.05) in endometrium from arresting sites at gd50, compared with endometrium from healthy sites. In endometrial lymphocytes, endothelium, and trophoblasts from arresting, conceptus-attachment sites at gd50, elevated expression of TNF- was found.
|
transcripts were detected in all samples tested. Pregnant endometrium expressed more TNF- than virgin endometrium. In biospies from healthy, conceptus-attachment sites, TNF- transcription was greater at gd50 than gd20 (P<0.001; Fig. 9B
). TNF- transcription in endometrium from arresting sites was not different from healthy sites at gd20 but was higher at gd50 (P<0.05). Endometrial lymphocytes from virgin uteri transcribed TNF- at levels similar to those in lymphocytes from gd20- or gd50-healthy sites and gd20-arresting sites. Elevated lymphocyte TNF- transcription was only found in arresting gd50 sites (P<0.001). TNF- transcription in endothelium was low in virgin uteri and in healthy sites. It was elevated significantly in endothelium from arresting, conceptus-attachment sites (P<0.01 at gd20 and gd50). Thus, in maternal and fetal tissue, elevated TNF- transcription predominated during midgestation fetal arrest, and elevated IFN- transcription dominated during gd20 arrest. |
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Porcine endometrial lymphocytes were shown to produce VEGF and PlGF, functions shared with uNK cells found in the decidualized endometria of humans [14 15 16 ] and mice [13 , 27 ]. Although a threefold enrichment of potentially angiogenic uNK cells occurs in peri-implantation porcine endometrial lymphocytes [6 ], there are no data characterizing porcine endometrial lymphocyte subsets at gd50. A possible continuing role for uNK cell-promoted angiogenesis at gd50 can be postulated from the presence of such cells in women [11 ] and in mice up to midgestation [8 ]. As we captured total endometrial lymphocytes, VEGF and PlGF production cannot be attributed exclusively to uNK cells. At attachment and midgestation, lymphocytes transcribed more VEGF than endometrial endothelium or trophoblast, and they appeared to be a dominating angiogenic cell type within endometrium. In contrast, PlGF transcription was more abundant in endothelial cells. In humans, uNK cells produce VEGF-C, a molecule promoting proliferation of lymphatic endothelium [14 ]. The porcine VEGF gene family has not been characterized sufficiently to provide reagents for extending this aspect of comparison. Pregnancy failure was associated with large declines in maternal transcription of VEGF, and trophoblast levels remained relatively constant. In contrast, maternal PlGF levels remained relatively high, particularly within lymphocytes at the failed implantation sites. This is the first report of PlGF transcription in porcine gestation.
As PlGF is a competitive antagonist for VEGF and has direct effects on cell proliferation and migration [28 ], it is unclear whether the action of elevated PlGF in lymphocytes from failing sites would be indirect through elevation of bioavailable VEGF or is autocrine. In mice, PlGF has a role in the terminal maturation of uNK cells [27 ]. Thus, elevated PlGF may be a mechanism to enhance porcine lymphocyte maturation and production of VEGF. Up-regulation of PlGF has been associated with induction of inflammation and edema in pathologic angiogenesis [29 ]. Thus, at sites of arrest, elevated, lymphocyte-derived PlGF may represent initiation of a strong, inflammatory response. PlGF also activates mononuclear phagocytes, promoting production of proinflammatory cytochemokines [30 ]. Maintenance of PlGF in the face of declining VEGF may represent a lymphocyte-regulated switch mechanism to move from controlled angiogenesis to inflammation at sites of conceptus arrest and to contribute to recruitment of debris-removal cells such as neutrophils or macrophages. Endometrial lymphocytes not only produced angiogenic factors but also showed dynamic, gd-dependent changes in expression of VEGFRI and VEGFRII. This suggests that PlGF and VEGF will have autocrine actions on receptor-expressing endometrial lymphocytes. As VEGFRI was transcribed more abundantly by lymphocytes than VEGFRII, and trophoblast showed the reverse, VEGF may have greater effects on trophoblast than PlGF. Reduction in receptor gene expression in lymphocytes from gd20 and gd50 may indicate that adequate change to the endometrial vasculature has occurred by mid-pregnancy and that angiogenesis is in decline.
Endothelial cells are major components of uterine stroma. Endothelial cells respond to VEGF by proliferation and sprouting during growth of blood vessels [31 ]. We have proposed that maternal endothelial sprouts are targets of pathological immune attack in processes, which compromise fetal outcome [25 ]. To advance this hypothesis, transcription profiles from endometrial endothelial cells were analyzed. PlGF was transcribed far more highly by endometrial endothelial cells than VEGF in virgin and pregnant uteri. The much-lower transcription of VEGFRI than VEGFRII by endothelial cells could indicate that endothelial PlGF has no or limited autocrine activity. The overall pattern of endothelial cell transcription during fetal arrest did not suggest widespread destruction of this cell type but rather, loss of its ability to respond to VEGF-related signaling. The histopathology of failing implantation sites indicated that the significant changes occur in vessel walls rather than in endothelial cells. In mice, VEGF promotes synthesis of peri-vascular stromal cell-derived factor-1, a chemokine, which retains circulating, proangiogenic, myeloid cells and endothelial cell progenitors at sites of neoangiogenesis, as these cells egress from circulation [32 ].
Porcine trophoblast transcribed VEGF and PlGF. These molecules could promote angiogenesis in the endometrium as well as within the placenta, as trophoblast expressed VEGFRI and VEGFRII. It is surprising that major differences were not found in transcription profiles for angiogenic genes in trophoblast from healthy and arresting conceptuses. This suggests the angiogenic compromise of porcine pregnancy is initiated by endometrium. Hypoxia provides a potent stimulus for VEGF production and is essential for development of maternal and placental vasculature in early human pregnancy [33
, 34
]. HIF-1, a prime regulator of oxygen homeostasis, binds to the hypoxia response element in the VEGF promoter [34
, 35
] and regulates peri-implantation angiogenesis in many species [36
37
38
]. The reduction of HIF-1 transcripts in endometrial biopsies, lymphocytes, and endothelium at arresting-attachment sites is directly proportional to the reduced VEGF at the failing sites [20
, 25
].
Another major finding of this study is the dominance of different cytokines during gd20 versus gd50 fetal arrest. IFN- and TNF- were transcribed in endometria from virgin and from healthy conceptus sites and in trophoblast. With failure, endothelium greatly up-regulated transcription of TNF- but never transcribed IFN-, which was dominant during early loss (gd20) in all other tissues, and TNF- dominated during midgestational loss (gd50). This differential cytokine expression implies that the immune-mediated pathology participating in fetal loss differs at different stages of gestation. This might be a result of the normal, dynamic changes in composition of the endometrial leukocyte population, which occur at different stages of pregnancy [6
7
8
9
10
11
], or to the recruitment and activation of newly immigrated immune effector cells.
Our study has not answered the questions about whether porcine embryo failure is maternally or fetally initiated or when heterogeneity for conceptus survival becomes established. Coincidence of endometrial lymphocyte-mediated compromises in angiogenesis and elevations in proinflammatory cytokine elevation with early stages of fetal arrest have been shown for the first time. These changes are highly localized to endometrial microdomains of the arresting fetuses and appear to be unique to lymphocytes quantitatively and dynamically. It will be important to establish whether this information is relevant to pregnancy loss in all mammals, but particularly in women.
Received May 25, 2007; revised June 22, 2007; accepted June 26, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
contributes to initiation of uterine vascular modification, decidual integrity, and uterine natural killer cell maturation during normal murine pregnancy J. Exp. Med. 192,259-270 by interleukin-1ß involves the extracellular signal-regulated kinase 1/2 pathway in normal human cytotrophoblast cells Biol. Reprod. 70,1822-1827This article has been cited by other articles:
 |
|
 |
|
  S. P. Murphy, C. Tayade, A. A. Ashkar, K. Hatta, J. Zhang, and B. A. Croy Interferon Gamma in Successful Pregnancies Biol Reprod, May 1, 2009; 80(5): 848 - 859. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Chang, D. Mukherjea, R. M. Gobble, K. A. Groesch, R. J. Torry, and D. S. Torry Glial Cell Missing 1 Regulates Placental Growth Factor (PGF) Gene Transcription in Human Trophoblast Biol Reprod, May 1, 2008; 78(5): 841 - 851. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
