Biol Reprod
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Engelhardt, H.
Right arrow Articles by King, G. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Engelhardt, H.
Right arrow Articles by King, G. J.
Agricola
Right arrow Articles by Engelhardt, H.
Right arrow Articles by King, G. J.
Biology of Reproduction 66, 1185-1192 (2002)
© 2002 Society for the Study of Reproduction, Inc.

Regular Article

Evaluation of Natural Killer Cell Recruitment to Embryonic Attachment Sites During Early Porcine Pregnancy1

Heidi Engelhardt2,a, B. Anne Croyb, and Gordon J. Kinga

a Departments of Animal and Poultry Science b Biomedical Sciences, University of Guelph, Guelph, Ontario, Canada N1G 2W1


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Specialized natural killer (NK) lymphocytes are a feature of the pregnant uterus in humans and rodents. Conceptus-mediated recruitment of uterine (u)NK cells in the pig was proposed based on evidence that elevated uNK activity was temporally associated with increased leukocyte density in endometrium underlying conceptuses. The objective of this study was to determine whether uNK cells were more abundant at embryonic attachment sites during the early postattachment period. Mononuclear leukocytes were isolated from endometrium at attachment sites versus between attachment sites, and expression of CD16, a marker for NK cells, was assessed by flow cytometry. CD16 binding was normalized to leukocyte numbers in each sample. CD16+ small lymphocytes were more frequent in uterus than in blood (41% ± 2% versus 26% ± 4%). Differences between pregnant and luteal phase uterus (43% ± 2% versus 31% ± 7%, respectively) were not statistically significant. In pregnant animals, CD16+ lymphocytes were slightly but significantly more abundant in uterus at attachment sites versus between attachment sites at Days 15–17, 21–22, and 25–28. Before normalization, CD16+ large, granular cells were more abundant at attachment sites versus between attachment sites; however, these differences were removed when data were normalized according to leukocyte numbers. Further characterization showed that the proportion of large granular leukocytes expressing CD8, reactive with NK cells and T cell subsets, was 2-fold higher in pregnant uterus than in maternal blood. These results raise the possibility that uNK cells resembling those in blood may be transformed into larger, more granulated forms in the uterine microenvironment.

immunology, implantation, pregnancy, trophoblast, uterus


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The establishment of the mammalian placenta involves close apposition between maternal tissue and semiallogeneic fetally derived tissues. In the hemochorial placenta of humans, invasion of the uterine mucosa by trophoblast is associated with accumulation of specialized natural killer (NK) lymphocytes [1]. In the rodent, uterine (u)NK cells at each implantation site differentiate into large, highly granular forms and aggregate in the mesometrial region of the uterine musculature during the latter half of gestation [2]. Cells with the phenotype of uNK cells are rare or absent in the circulation of these species, implying that there is selective recruitment to, or differentiation within the uterus. Their location within the pregnant uterus would suggest that uNK cells are recruited or activated by embryo-derived factors. However, recruitment of uNK cells in these species accompanies decidualization of endometrial stroma, even in the absence of pregnancy [3–5].

The epitheliochorial placenta of the pig contrasts with that of humans and rodents in that there is no decidualization and no invasion. The maternal-fetal union consists of diffuse attachment between uterine epithelium and chorioallantoic membrane via interlocking microvilli. Functional studies have shown that NK activity, as measured by lysis of tumor target cells in vitro, is elevated in endometrial cell suspensions at Days 10 and 20 of pregnancy, relative to Days 10 and 20 of the cycle [6]. This elevation in NK activity requires the presence of a conceptus (estrogen-induced pseudopregnancy or intrauterine seminal plasma exposure does not induce increases in NK activity comparable to those seen in early pregnancy) [6]. The distribution of these cells in the porcine maternal-fetal interface has not been established. However, during the early postattachment period (Days 15–28), leukocytes are more numerous in endometrial stroma in direct association with conceptuses than in endometrium between attachment sites [7].

In contrast to rodents and humans, NK activity in conventional pigs is mediated by small- to medium-sized, agranular lymphocytes [8–10]. CD16, the low-affinity receptor for the Fc portion of immunoglobulin (Ig) G, is found on the majority of NK cells in several species and has been widely used to distinguish NK cells from other lymphocytes [11]. Myeloid cells, including macrophages and monocytes, also express CD16 [11], but are usually distinguished from CD16-positive lymphocytes by their larger size and more granular cytoplasm. In common with some T cell subsets, porcine NK cells express CD8 [12, 13]. The presence of this surface antigen was used to distinguish large, granular CD16+ NK cells from myeloid cells.

The objective of this study was to determine whether NK cells were recruited to endometrium that was directly associated with conceptuses during the early postattachment period. Our approach was to compare the frequencies of CD16-expressing lymphocytes in suspensions of endometrial cells from early pregnant versus luteal phase animals, and at embryonic attachment sites versus between embryonic attachment sites during the early postattachment period.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Animals

Yorkshire sows were housed under total confinement in a specific pathogen-free facility at the Arkell Swine Unit, University of Guelph. Animals were observed for estrus once daily. To obtain pregnant animals, sows were bred at the postweaning estrus, on the first day of standing estrus (Day 0), and 24 h later. Reproductive tracts were recovered at slaughter on Days 15–17 (n = 3), 21–22 (n = 11), and 25–28 (n = 4) of gestation and on Days 15–16 of the cycle in unmated animals (n = 4). Peripheral blood was collected from each animal at slaughter for paired study. Breeding and slaughter procedures were approved by the University of Guelph Animal Care Committee and complied with the guidelines of the Canadian Council on Animal Care.

Isolation of Leukocytes

Uteri were trimmed free of mesentery and opened along the antimesometrial aspect to avoid disruption of embryonic attachment sites. By Day 17 of gestation, attachment sites could be easily identified. At earlier stages, attachment sites were detected by increases in vascularity, and confirmed by viewing with UV light to detect the conceptus-associated autofluorescence that is characteristic of early porcine pregnancy [14]. Embryos and their associated membranes were gently removed. Within each uterus, a sheet of endometrium weighing approximately 10 g was separated from the underlying myometrium and connective tissue, at attachment sites and at regions between attachment sites (designated "at" and "between," respectively). In nonpregnant animals, tissue was collected from random sites along the mesometrial aspect of the uterus.

Tissue was minced and incubated in PBS containing 0.1% collagenase (grade II; ICN Biomedicals Inc., Montreal, PQ, Canada), 0.02% hyaluronidase (Sigma; Oakville, ON, Canada) and 0.001% DNase (type I; Sigma) in a shaking water bath at 37°C for 1 h. Tissue digests were filtered through cheesecloth then through 100-µm nylon mesh (Nitex; Sefar Canada Inc., Scarborough, ON, Canada) and washed by centrifugation. Cells were then resuspended in approximately 3 ml of PBS, underlaid with 3 ml of Lympholyte-Mammal (density = 1.086; Cedarlane Laboratories Ltd., Hornby, ON, Canada), and centrifuged at 1200 x g for 15 min. The resulting band of cells, consisting of variable proportions of mononuclear leukocytes and nonleukocyte endometrial cells, was removed, washed, and assessed for viability by trypan blue-exclusion. Peripheral blood mononuclear cells (PBMCs) were isolated from blood collected at slaughter using Lympholyte Mammal according to the manufacturer's instructions. Both uterine-derived and blood-derived leukocytes were resuspended at approximately 2 x 107 viable cells per ml. Viability of PBMCs was consistently greater than 98%, whereas that in uterine cell suspensions ranged from 65% to 90%. Although the majority of uterine epithelial cells were nonviable, it is likely that some nonleukocyte endometrial cells contributed to the total count in uterine cell suspensions.

Antibodies

For the majority of the experiments, the primary antibodies were mouse monoclonals recognizing porcine CD16 (clone G7, IgG1; generously provided by Y.B. Kim, Chicago Medical School) and CD44 (clone BAT31A, IgG1; VMRD, Inc., Pullman, WA). Mouse monoclonal antibodies to CD4 (clone 74-12-4; IgG2b) and CD8 (clone 76-2-11, IgG2a) were used in later experiments. Both CD4 and CD8, produced by Pescovitz and coworkers [15], were purchased from VMRD. Goat F(ab')2 anti-mouse IgG (H+L) conjugated to fluorescein isothiocyanate (FITC; Cedarlane) was used as the secondary antibody. The suitability of CD44 as a pan-leukocyte marker for cells derived from porcine uterine tissue was confirmed by costaining with anti-porcine CD45 (clone 74-9-3, IgM; VMRD). CD44-positive and CD45-positive cells were detected using subisotype-specific secondary antibodies conjugated to different fluorochromes (goat anti-mouse IgG1-FITC and goat anti-mouse IgM-PE, respectively; Cedarlane).

Flow Cytometry

Incubations were carried out at 4°C in PBS plus 0.01% sodium azide. Cells (50 µl; 0.5–1.0 x 106) were added to 50 µl of diluted primary antibody in 96-well V-bottomed plates, and incubated for 20 min. Plates were then centrifuged (300 x g, for 5 min at 4°C), washed, and resuspended in secondary antibody for a further 20-min incubation. Cells were again centrifuged, washed, and then transferred to 12 x 75-mm polypropylene tubes for analysis using a Coulter Epics-XL/MCL analytical flow cytometer. A total of 10 000 events were collected per sample.

Binding of both uterine cells and PBMCs to CD16 and CD44 were assessed for each animal.

For each cell suspension, unstained cells (PBS only) and cells incubated with the IgG1 isotype control plus secondary antibody were included as controls for background autofluorescence and nonspecific binding, respectively. When greater than 1% of specific staining, these background values were subtracted. FCS Express, Version 1.0 (De Novo Software; Thornhill, ON, Canada) was used for post hoc analysis and plotting of flow cytometric data.

Statistical Analysis

Frequencies of cells expressing the pan-leukocyte marker, CD44, in samples from blood versus uterus, and from animals of differing reproductive status were compared by ANOVA, and computed using the general linear models procedure (PROC GLM) of the SAS system for Windows (version 8.01; SAS Institute Inc., Cary, NC).

For CD16 binding, data were normalized for numbers of leukocytes (CD44+ cells) in each sample and means were compared using PROC GLM. Although statistical analysis was performed on both normalized and nonnormalized data, interpretations and graphical presentations are based on normalized data. Any discrepancies in findings with nonnormalized data are indicated in the text. Within pregnant animals, effects of stage of gestation and proximity to attachment sites on CD16 expression in uterine leukocytes were assessed in a 3 x 2 factorial ANOVA. Because more than one data point was collected from each sow, "animal" was included in the model as a random effect and means were compared using the mixed linear models procedure of SAS (PROC MIXED). Data are presented as least squares means. Frequencies of cells reactive with CD8 and CD4 in blood, and uterus at and between attachment sites were also analyzed using this approach.

All data were subjected to arcsine transformation for analysis, but are presented as untransformed means ± SEM.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Establishing Parameters for Flow Cytometric Analysisof Porcine Uterine Cells

Density gradient centrifugation substantially enriched uterine cell suspensions with respect to leukocytes. Shown in Figure 1 are plots of forward scatter, side scatter, and fluorescence for unpurified uterine cells (subjected to hypotonic lysis to remove erythrocytes; A), uterine cells from the same suspension after density gradient centrifugation (B), and PBMCs from the same animal (C).



View larger version (50K):
[in this window]
[in a new window]
  		FIG. 1. Flow cytometry of suspensions of uterine cells, unpurified (A) or enriched for mononuclear leukocytes (B), and peripheral blood mononuclear cells (C) from the same animal (Day 15 of gestation). Plots shown are forward scatter (proportional to cell size) versus side scatter (proportional to granularity; first column), binding to CD44, a pan-leukocyte marker (center column) and filled histograms depicting the percentage of CD44-positive cells in the leukocyte region (R1). Unfilled histograms indicate isotype-matched negative controls

Leukocyte gates were established in the forward-versus-side scatter plots (first column) by back-gating from CD44+ cells, which were "FL1-bright" due to their coupling to the FITC-conjugated secondary antibody. The forward and side scatter characteristics of leukocytes in the three samples (indicating cell size and granularity, respectively) were determined by first creating a region (region 2) to include CD44+ cells in the plots in the center column. Back-gating from cells in R2 to the scatter plots in the first column created a leukocyte region (R1). Gradient centrifugation of uterine cells resulted in a lower proportion of small events (low forward scatter; small cells plus cell fragments), and a higher proportion of CD44+ cells (indicated by histograms in the third column). As expected, a very high percentage of PBMCs were CD44+ (94%; Fig. 1C). In uterine cell suspensions subjected to hemolysis only, only 19% of the cells in the leukocyte region were reactive with CD44 (Fig. 1A). This was increased more than 4-fold by gradient centrifugation (79%, Fig. 1B). All samples were subsequently purified by gradient centrifugation. Cell viability (trypan blue-exclusion) was similar in unpurified and gradient-purified uterine preparations (82% and 84%, respectively).

Subpopulations of leukocytes in peripheral blood can be readily distinguished by their scatter characteristics. Figure 2 shows plots of leukocyte samples derived from uterus (Fig. 2A) and peripheral blood (Fig. 2B). Lymphocytes appear as a distinct population of small, relatively agranular cells and leukocytes of the monocyte/macrophage lineage form a cluster of large, granular cells. Granulocytes are removed by the gradient processing step. CD44+ cells typical of blood lymphocytes and macrophages were present in uterine samples. However, uterus also contained large numbers of very large, highly granular cells that did not express CD44. In contrast to blood, most uterine samples also contained numerous cells that fluoresced in the absence of antibodies or fluorochromes. Some of this autofluorescence was similar in intensity to that of positively stained cells. Autofluorescent cells were excluded by gating on the lymphocyte region in the forward versus side scatter plot, but these cells had similar scatter characteristics to the large, granular population. Based on their morphology in Wrights stained smears, many of the large, granular cells in the uterine samples were identified as epithelial cells (not shown).



View larger version (38K):
[in this window]
[in a new window]
  		FIG. 2. Plots showing subpopulations among uterine-derived and blood-derived leukocytes (Day 25 of gestation). Histograms in fourth and fifth columns are gated on lymphocyte (LY) and large, granular leukocyte (LGL) regions, respectively, shown in the plots on the left. Note the autofluorescent cells (auto) in the uterine cell suspension that are not present in the blood sample. Unfilled histograms indicate isotype-matched negative controls

Effect of Pregnancy and Proximity to Embryoson Proportion of Leukocytes in Uterine Cell Suspensions

Because differences in endometrial leukocyte distribution relative to pregnancy status and embryonic attachment sites have been observed [7], CD44-expressing cells in uterine samples with respect to gestational stage and intrauterine location were examined. Among presumptive lymphocytes (LY gate), the proportion expressing CD44 was not influenced by pregnancy or proximity to attached conceptuses (cycling, 84% ± 10%; pregnant, at, and between attachment sites, respectively, 85% ± 2%; and 81% ± 3%; P > 0.10). In contrast, among large, granular uterine cells, CD44+ cells were significantly higher at attachment sites (58% ± 3%; P < 0.05), relative to either between attachment sites (40% ± 6%) or cycling uterus (40% ± 7%). When uterine samples from pregnant animals were subdivided according to stage of gestation, there was a significant interaction between cell size (small versus large, granular) and stage (Fig. 3). In samples obtained at attachment sites, CD44 binding among cells with lymphocyte morphology was not affected by gestational stage. However, between attachment sites, the proportion of small cells reactive with CD44 was significantly reduced by Days 25–28 (P < 0.05). Among large, granular cells, there was a marked decline in the proportion of CD44-expressing cells in samples obtained between attachment sites as gestation progressed.



View larger version (34K):
[in this window]
[in a new window]
  		FIG. 3. Effect of reproductive status and proximity to the conceptus on uterine leukocyte recovery among small, lymphocyte-sized cells (A) and large, granular cells (B). Luteal phase, n = 4; Day 15–17, n = 3; Day 21–22, n = 11; Day 25–28, n = 4. Asterisks indicate differences significant at P < 0.05

In order to detect differences in CD16+ uterine lymphocytes relative to total leukocytes, data for CD16 binding were normalized to the number of CD44+ cells in each sample for statistical analysis and graphical presentation. Any situations in which findings differed with uncorrected data are indicated in the text.

CD16-Expressing Leukocyte Subpopulations in Uterus and Blood

CD16 expression in uterine and peripheral blood cells from an individual animal (Day 15 of gestation) is shown in Figure 4. In peripheral blood, CD16+ cells were present among both lymphocyte and large granular populations (presumptive NK cells and monocytes, respectively; Fig. 4B). Both of these CD16+ populations were present in uterine samples (Fig. 4A). In contrast to findings in miniature pigs using this antibody [16], CD16 expression in monocytes was typically more intense (i.e., a brighter signal) than in lymphocytes (Fig. 4B).



View larger version (35K):
[in this window]
[in a new window]
  		FIG. 4. CD16 expression among leukocyte subpopulations in uterus (A) and blood (B) from a pig at Day 15 of gestation. Previously established CD16+ populations in blood are indicated. NK, Natural killer lymphocytes; M{phi}, monocytes/macrophages. Unfilled histograms indicate isotype-matched negative controls

The frequency of CD16+ small lymphocytes in peripheral blood (24% ± 3%) was highly variable (range 7% to 53%), but was not affected by reproductive stage (P > 0.10). Blood data were therefore considered as a single mean for all animals. CD16 was expressed by a higher proportion of lymphocytes in uterus (overall mean) than in blood (41% ± 2% versus 26% ± 4%, respectively; P < 0.0001). There was no correlation between NK cell frequencies in blood versus uterus in individual animals (P > 0.05). CD16+ small lymphocytes tended to be more abundant in pregnant uterus compared with luteal phase uterus, but this difference was not statistically significant (43% ± 2% versus 31% ± 7%, respectively; P > 0.05). Data from pregnant uteri were analyzed as a factorial with respect to stage of gestation, and position relative to conceptuses. Although there was no effect of stage (P > 0.05), CD16+ lymphocytes were slightly but significantly more abundant at versus between attachment sites at each stage of gestation (Fig. 5A).



View larger version (36K):
[in this window]
[in a new window]
  		FIG. 5. Effect of reproductive status and proximity to the conceptus on the proportion of small lymphocytes (A) and large, granular leukocytes (B) expressing CD16. Because frequencies of CD16+ cells in peripheral blood were not affected by reproductive stage (P > 0.10), blood data are shown as a single mean for all animals (n = 22). Animal numbers for uterine samples are as described in the legend for Figure 3

Before normalization to number of leukocytes, CD16+ large, granular cells (high forward scatter, high side scatter) were less frequent in uterus (27% ± 3%) than in maternal blood (51% ± 5%), and more abundant at versus between attachment sites (30% ± 3% and 21% ± 3%, respectively; P < 0.05). However, these differences reflected variations in leukocyte content and were removed when data were normalized to CD44 binding (Fig. 5B).

Expression of CD8 and CD4 by Uterine Leukocytes

Uterine leukocytes were further characterized for expression of the lymphocyte markers CD8 and CD4. In the pig, CD8 is expressed by NK cells and some T cell subsets; CD4 is expressed by T helper cells [12, 13, 17]. Uterine leukocytes and peripheral blood were obtained from eight pigs slaughtered during Weeks 2–4 of gestation.

Although there were no significant differences at versus between attachment sites, there were clear differences in CD4 and CD8 expression in uterine versus maternal blood leukocytes. Among small lymphocytes (Fig. 6A), CD4 binding was similar in blood and uterus (16% ± 1% and 15% ± 3%, respectively), whereas CD8 binding was significantly higher in uterine cells. Virtually all uterine lymphocytes were CD8+ (95% ± 1%), whereas only 67% ± 11% of blood lymphocytes expressed this antigen.



View larger version (25K):
[in this window]
[in a new window]
  		FIG. 6. Expression of CD8 and CD4 among small lymphocytes (A) and large, granular leukocytes (B) obtained from uterus and blood of pigs during Weeks 2–4 of pregnancy (n = 8; 2 on Day 15, 1 on Day 19, 4 on Days 21–22, and 1 on Day 26). Data are normalized to CD44+ cells in each size category. Asterisks indicate significant differences within each grouping (P < 0.05)

Among large, granular cells, which were expected to be largely CD4-, CD8-monocytes/macrophages, surprisingly high proportions of leukocytes expressed these lymphocyte markers, particularly in uterus (Fig. 6B). There were approximately twice as many CD4+ cells in uterus versus blood (38% ± 4% and 18% ± 4%, respectively). CD8 was expressed by even higher proportions of large, granular uterine leukocytes (uterus, 51% ± 5%; blood, 20% ± 5%).

CD44 as a Pan-Leukocyte Marker

CD44, a transmembrane glycoprotein involved in a broad range of leukocyte activities [18], is present on approximately 95% of porcine peripheral blood leukocytes [19]. Coexpression of CD44 and CD45 by uterine cells is shown in Figure 7. Similar to blood (not shown), 94%–97% of uterine cells reactive with either of these antibodies were dually stained. As expected, most (96%) small uterine cells were CD45+ (Fig. 7A). Only a small percentage (4%) of CD45+ uterine lymphocytes did not express CD44. Consistent with data in Figure 3, the large, granular uterine cell fraction contained a relatively high proportion of nonleukocytes, indicated by 28% double-negative cells in Figure 7B. Approximately 60% of the large, granular cells were CD44+CD45+ (plots on the right of Fig. 7 are isotype-matched negative controls). A small proportion of large, granular cells (~6%) expressed only one of the two markers. Some types of nonhematopoietic cells have been shown to express certain CD44 isoforms [20], including uterine epithelial cells [21]. However, the antibody used in the present study reacted with leukocytes but not epithelial cells in sections of porcine uterine tissue [7]. These findings confirmed that CD44 was an appropriate pan-leukocyte marker in uterine preparations.



View larger version (53K):
[in this window]
[in a new window]
  		FIG. 7. Expression of CD44 and CD45 by small, lymphocyte-sized cells (A) and large granular cells (B) in uterus from a pig at Day 15 of gestation. Plots on the right are isotype-matched negative controls


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
This study was the first to investigate local effects of the conceptus on leukocyte distribution in the pig. Our hypothesis was that NK cells would be more abundant in endometrium in contact with trophoblast than in tissue that was distant from attachment sites. CD16+ lymphocytes similar to NK cells in the circulation were more frequent in uterus than in blood, and slightly but significantly more abundant at versus between attachment sites from Days 15 to 28 of gestation. However, these differences were small given the reported differences in NK lytic activity [6]. Before normalization with respect to leukocytes, large, granular cells expressing CD16 were more abundant at versus between attachment sites throughout the early postattachment period. However, these differences reflected variations in leukocyte content and were removed when data were normalized to CD44 binding. Although leukocytes with these characteristics (high forward scatter, high side scatter) would be expected to be primarily macrophages, a substantial proportion of the large, granular uterine cells expressed the CD8 surface marker, indicating either T or NK cell lineage [12, 13]. This "spreading" of porcine lymphocytes into the "monocyte region" was also noted in the proceedings of the latest swine CD workshop [22]. These findings suggest the intriguing possibility that within the porcine uterus, small NK cells may be transformed into large, granulated forms not present in the peripheral blood.

Quantification of NK cells, particularly in livestock species, presents a challenge due to the scarcity of appropriate surface markers. Although NK cell frequency in the porcine uterus has not been addressed before, estimates in blood have been reported. With the antibody used in the present study, an NK cell frequency of 13% was reported in Minnesota miniature swine [16]. We found that the frequency of CD16+ small lymphocytes in peripheral blood of mature Yorkshire sows (24% ± 3%) was not affected by reproductive status, but was highly variable among animals (7%–53%, n = 23). To our knowledge, no previous studies have used CD16 to identify NK cells in conventional pigs. Using other surface markers, mean NK cell frequencies have been reported to be somewhat lower than that in our study (CD18a, 15% [23]; 5C6, 10% [24]). However, high variabilities have also been noted in studies using larger numbers of animals (PNK-E, 6%–33% [25]; human CD11b, 7%–27% [22]) and in the most recent Workshop on Swine Leukocyte Differentiation Antigens (CD16, 3%–62%; MIL-4, 5%–36% [26]).

Leukocytes in the porcine uterus had not previously been analyzed using flow cytometry. This technique has allowed us to use an antibody, CD16, which in our hands, has exhibited no binding to sections of porcine uterus using conditions under which other leukocyte surface markers have been successful. It has also provided objective information about the relative size and granularity of leukocytes expressing CD16 and CD8. However, several potential pitfalls must be considered when interpreting our flow cytometric findings. The presence of CD44-negative cells in uterine digests necessitated careful attention to negative controls. These cells, primarily epithelial cells, had scatter characteristics similar to large, granular leukocytes, and tended to be autofluorescent. In the human, autofluorescent cells have been observed in suspensions of intestinal epithelial cells [27] and decidual cells [28]. In the pig, green autofluorescence has been observed on the luminal surface of the endometrium in pregnant but not in nonpregnant animals [14]. Another factor to consider is that because mild enzymatic digestion selectively disaggregates cells least tightly bound [29], our uterine cell suspensions would have had higher proportions of leukocytes than the original tissue. Although enrichment for mononuclear leukocytes and normalization to leukocyte content enabled comparisons of relative changes in CD16+ cells relative to total leukocytes, true differences in abundance in the original tissue may have been masked.

Our findings do not support the notion of conceptus-mediated recruitment of lytically active NK cells. Evidence based on peripheral blood had indicated that the cells mediating NK lytic activity in conventional pigs were small, agranular lymphocytes [8–10]. The antibody we used to identify NK cells was reactive with lytic NK cells, because it was originally selected for its ability to alter NK-lytic activity [16] before being identified as the porcine homologue of CD16 [30]. Whereas CD16+ small lymphocytes were found to be more abundant in tissue obtained at embryonic attachment sites than in nonpregnant animals, differences in frequencies of these presumptive NK cells at versus between attachment sites were very small. Furthermore, the proportion of these cells remained constant over the period studied (Days 15–28), whereas lytic activity was reported to decline sharply between Day 20 and Day 30 [6]. It is possible that conceptus-mediated recruitment of uNK cells occurs diffusely throughout the porcine uterus. This interpretation is consistent with evidence that lytic activity is already elevated by Day 10 [6] when blastocysts would be freely moving within and between uterine horns. As a result of dramatic elongation on Day 12 just before attachment [31], porcine conceptuses are in contact with, and thus potentially influence large areas of luminal surface even during the early postattachment period. However, even differences between pregnant and luteal phase animals were small in light of the 25-fold differences in NK-lytic activity that have been reported [6].

Although lysis of tumor target cells in vitro provides a means of detecting NK cells when surface markers are not available, this function, which requires cell-cell contact, is unlikely to be important in vivo given the anatomical separation between uNK cells and potential targets. An alternative viewpoint is that the conceptus mediates other changes in uNK cells not related to lytic activity. We propose that local differentiation of small NK cells into large, granular forms, as has been described in the mouse [32, 33], may occur in the pig. Before normalization for leukocyte content, the proportion of large granular uterine cells expressing CD16 was 1.5-fold and 3-fold higher at attachment sites at Days 21–22 and 25–28, respectively, compared with between attachment sites at these stages. These findings are consistent with differences in leukocyte abundance in uterine tissue at these stages [7]. Although further characterization is required to identify these cells, the finding that the proportion of large granular leukocytes expressing CD8 was 2-fold higher in pregnant uterus than in maternal blood supports the notion that some of these large granular cells are NK cells rather than macrophages.

The presence of large, granular lymphocytes distinct from their counterparts in the circulation is a feature of the pregnant uterus in humans [28], mice [32, 33], rats [34], and sheep [35]. Increases in size and granularity are morphological correlates of lymphocyte activation or "blastogenesis." Accordingly, fully differentiated uNK cells in the mouse, which reach diameters of up to 80 µm, are morphologically and phenotypically similar to the "anomalistically large" cells in cultures of lymphokine-activated NK cells derived from spleen [33, 36, 37]. The parallels between the pig and the mouse are striking given the contrasts in placentation between these two species. In the cycling and early pregnant uterus in both pigs (present study) and mice [38], small NK cells similar to those in blood are present in the endometrial stroma. In both species, uterine NK lytic activity is transiently elevated through the implantation/attachment period (pigs, [6, 39]; mice, [40]), and the presence of embryos is required for this activity (pigs, [6]; mice, [41]). In the mouse, small, relatively agranular NK cells are dispersed throughout the nonpregnant endometrium, but differentiation into nonlytic cells that are larger and more granular than NK cells in circulation occurs in discrete regions within each implantation site [32, 33, 40]. A striking difference is that decidualization of endometrial stroma without pregnancy is sufficient for apparently normal uNK cell differentiation in mice [4, 5], whereas decidualization does not occur in the pig.

The question of what may mediate differentiation of these large, granular uNK cells is open to speculation. In the mouse, blastogenesis of NK cells in spleen and nonlymphoid sites (liver and peritoneal cavity) can be induced by in vivo treatment with interferons (IFNs; type I or type II) or IFN inducers [42, 43]. IFN-{gamma} and IFN-{delta}, a type I IFN unique to the porcine conceptus, are synthesized by porcine trophoblast during the establishment of maternal-fetal contact (Days 11–17) [44, 45]. It is tempting to speculate that trophoblast-derived IFNs may, via interactions with maternal epithelium, mediate activation of uNK cells in the porcine uterus. NK cells themselves are major producers of IFN-{gamma}. Activation of stromal uNK cells may require IFN or other conceptus-derived cytokines as an initial signal, which is then amplified by autocrine actions of IFN-{gamma} produced by uNK cells within the endometrium. In the mouse uterus, mature uNK cells are the major source of IFN-{gamma} in the metrial triangle [46, 47]. In mice that are genetically deficient in uNK cells, IFN-{gamma}, or IFN-{gamma} signaling pathways, abnormalities in the major decidual arteries become evident at Days 8–10 (term = 19 days) [47–49], the period over which small, lytic precursors differentiate into very large, nonlytic uNK cells [33]. Unlike humans and rodents, porcine conceptuses do not erode maternal tissues to gain better access to maternal blood. Success of this noninvasive placental strategy requires well-developed vascular beds on both maternal and fetal surfaces [50]. Although a link between porcine uNK cells and vascular remodeling in pregnancy has not been established, variations in placental vascularity in late gestation have been associated with differences in embryo survival [51, 52].

Clearly, there are many unanswered questions about the NK cells in the porcine uterus and their role in pregnancy. The results of the present study, viewed in light of findings in the mouse, suggest that the small, relatively agranular uterine NK cells resembling those in blood may be transformed into larger, more granulated forms under the influence of the uterine microenvironment. Early embryonic mortality is 30%–40% in the pig, most of which occurs before Day 30 [53]. Recruitment of uNK cells early in gestation and their subsequent differentiation may be key determinants of the ability of individual fetuses to adjust blood flow to meet future metabolic demands. Further defining these cells and determining the factors that regulate their development will be addressed in future investigations.


   ACKNOWLEDGMENTS
 
We thank Dr. Y.B. Kim, Finch University of Health Sciences, Chicago Medical School, for generously providing the CD16 antibody. Thanks also to Teresa Storer for technical assistance, the staff of the Arkell Swine Research Unit for care of the animals, the staff of the Meats Wing, Department of Animal and Poultry Science for slaughter of animals, and Doug Wey for the coordination of all animal-related aspects of this study.


   FOOTNOTES
 
First decision: 19 September 2001.

1 This study was funded by the Natural Sciences and Engineering Research Council of Canada, Ontario Pork, and the Ontario Ministry of Agriculture, Food and Rural Affairs. H.E. was supported in part by an Ontario Veterinary College Postdoctoral Fellowship. Back

2 Correspondence: FAX: 519 767 0573; engel{at}uoguelph.ca Back

Accepted: November 19, 2001.

Received: August 27, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. King A, Burrows T, Loke YW. Human uterine natural killer cells. Nat Immun 1996–1997; 15:41-50[Medline]
  2. Head JR. Uterine natural killer cells during pregnancy in rodents. Nat Immun 1996–1997; 15:7-21[Medline]
  3. King A, Wellings V, Gardner L, Loke YW. Immunocytochemical characterization of the unusual large granular lymphocytes in human endometrium throughout the menstrual cycle. Hum Immunol 1989; 24::195-205[CrossRef][Medline]
  4. Stewart I. Granulated metrial gland cells in the lungs of mice in pregnancy and pseudopregnancy. J Anat 1985; 140:551-563
  5. Zheng LM, Joag SV, Parr MB, Parr EL, Young JD. Perforin-expressing granulated metrial gland cells in murine deciduoma. J Exp Med 1991; 174:1221-1226[Abstract/Free Full Text]
  6. Yu Z, Croy BA, Chapeau C, King GJ. Elevated endometrial natural killer cell activity during early porcine pregnancy is conceptus-mediated. J Reprod Immunol 1993; 24:153-164[CrossRef][Medline]
  7. Swatman J, King GJ, Engelhardt H. Leukocyte distribution in the pig uterus relative to sites of conceptus attachment during early pregnancy. Biol Reprod 1996; 54:(suppl 1):475
  8. Martin S, Wardley RC. Natural cytotoxicity detected in swine using Aujeszky's disease virus infected targets. Res Vet Sci 1984; 37:211-218[Medline]
  9. Yang WC, Schultz RD, Spano JS. Isolation and characterization of porcine natural killer (NK) cells. Vet Immunol Immunopathol 1987; 14:345-356[CrossRef][Medline]
  10. Pinto A, Ferguson F. Characteristics of Yorkshire swine natural killer cells. Vet Immunol Immunopathol 1988; 20:15-29[CrossRef][Medline]
  11. Perussia B. Fc receptors on natural killer cells. Curr Top Microbiol Immunol 1998; 230:63-68[Medline]
  12. Pescovitz MD, Lowman MA, Sachs DH. Expression of T-cell associated antigens by porcine natural killer cells. Immunology 1988; 65::267-271[Medline]
  13. Yang H, Parkhouse ME. Phenotypic classification of porcine lymphocyte subpopulations in blood and lymphoid tissues. Immunology 1996; 89:76-83[CrossRef][Medline]
  14. Keys JL, King GJ, LaForest JP. Autofluorescence of the porcine endometrium during early pregnancy. Biol Reprod 1989; 40:220-222[Abstract]
  15. Pescovitz MD, Lunney JK, Sachs DH. Preparation and characterization of monoclonal antibodies reactive with porcine PBL. J Immunol 1984; 133:368-375[Abstract]
  16. Dato ME, Wierda WG, Kim YB. A triggering structure recognized by G7 monoclonal antibody on porcine lymphocytes and granulocytes. Cell Immunol 1992; 140:468-477[CrossRef][Medline]
  17. Saalmuller A, Reddehase MJ, Buhring HJ, Jonjic S, Koszinowski UH. Simultaneous expression of CD4 and CD8 antigens by a substantial proportion of resting porcine T lymphocytes. Eur J Immunol 1987; 17:1297-1301[Medline]
  18. Goodison S, Urquidi V, Tarin D. CD44 cell adhesion molecules. Mol Pathol 1999; 52:189-196[Abstract]
  19. Zuckermann FA, Binns RM, Husmann R, Yang H, Carr MM, Kim YB, Davis WC, Misfeldt M, Lunney JK. Analyses of monoclonal antibodies reactive with porcine CD44 and CD45. Vet Immunol Immunopathol 1994; 43:293-306[CrossRef][Medline]
  20. Borland G, Ross JA, Guy K. Forms and functions of CD44. Immunology 1998; 93:139-148[CrossRef][Medline]
  21. Behzad F, Seif MW, Campbell S, Aplin JD. Expression of two isoforms of CD44 in human endometrium. Biol Reprod 1994; 51:739-747[Abstract]
  22. Haverson K, Simon A, LeFlufy L, Banfield G, Chen Z, Hollemweguer E, Ledbetter G. Monoclonal antibodies raised to human cells—specificity for pig leukocytes. Vet Immunol Immnopathol 2001; 80:175-186
  23. Dato ME, Kim YB. Characterization and utilization of a monoclonal antibody inhibiting porcine natural killer cell activity for isolation of natural killer and killer cells. J Immunol 1990; 144:4452-4462[Abstract]
  24. Camenisch TD, Jaso-Friedmann L, Evans DL, Harris DT. Expression of a novel function-associated molecule on cells mediating spontaneous cytolysis in swine. Dev Comp Immunol 1993; 17:277-282[CrossRef][Medline]
  25. Johnson BD, Wierda WG, Kim YB. Further characterization of PNK-E: a monoclonal antibody enhancing porcine natural killer cell activity. Cell Immunol 1991; 134:378-389[CrossRef][Medline]
  26. Haverson K. Appendix: flow cytometric data included in first round cluster analysis. Vet Immunol Immunopathol 2001; 80:187-208[CrossRef]
  27. Ebert EC, Roberts AI. Pitfalls in the characterization of small intestinal lymphocytes. J Immunol Methods 1995; 178:219-227[CrossRef][Medline]
  28. King A, Balendran N, Wooding P, Carter NP, Loke YW. CD3-leukocytes present in the human uterus during early placentation: phenotypic and morphologic characterization of the CD56++ population. Dev Immunol 1991; 1:169-190[Medline]
  29. Ritson A, Bulmer JN. Extraction of leukocytes from human decidua. A comparison of dispersal techniques. J Immunol Methods 1987; 104::231-236[CrossRef][Medline]
  30. Halloran PJ, Sweeney SE, Strohmeier CM, Kim YB. Molecular cloning and identification of the porcine cytolytic trigger molecule G7 as a Fc{gamma}RIII{alpha} (CD16) homologue. J Immunol 1994; 153:2631-2641[Abstract]
  31. King GJ. Comparative placentation in ungulates. J Exp Zool 1993; 266:588-602[CrossRef][Medline]
  32. Parr E, Parr M, Zheng L, Young J. Mouse granulated metrial gland cells originate by local activation of uterine natural killer lymphocytes. Biol Reprod 1991; 44:834-841[Abstract]
  33. Linnemeyer PA, Pollack SB. Stage-specific expression of activation antigens on NK cells at uterine implantation sites in mice. J Immunol 1994; 153:1478-1486[Abstract]
  34. Head JR, Kresge CK, Young JD, Hiserodt JC. NKR-P1+ cells in the rat uterus: granulated metrial gland cells are of the natural killer cell lineage. Biol Reprod 1994; 51:509-523[Abstract]
  35. Meeusen E, Fox A, Brandon M, Lee CS. Activation of uterine intraepithelial gamma/delta T cell receptor-positive lymphocytes during pregnancy. Eur J Immunol 1993; 23:1112-1117[Medline]
  36. Linnemeyer PA, Pollack SB. Monoclonal antibody 4H12 recognizes subsets of adherent-lymphokine activated killer cells and splenic natural killer cells from pregnant and neonatal mice. J Immunol 1991; 146:3729-3735[Abstract]
  37. Ginsburg H, Coleman R, Davidson S, Khoury C, Mor R. Lymphokine-activated killer (LAK) cells are identical to the uterine granulated metrial gland (GMG) cells. Transplant Proc 1989; 21:186-189[Medline]
  38. Pollack SB, Linnemeyer PA. Natural killer cells in the nonpregnant murine uterus. Nat Immunol 1996; 15:34-40
  39. Croy BA, Waterfield A, Wood W, King GJ. Normal murine and porcine embryos recruit NK cells to the uterus. Cell Immunol 1988; 115::471-480[CrossRef][Medline]
  40. Gambel P, Croy BA, Moore WD, Hunziker RD, Wegmann TG, Rossant J. Characterization of immune effector cells present in early murine decidua. Cell Immunol 1985; 93:303-314[CrossRef][Medline]
  41. Croy BA, Gambel P, Rossant J, Wegmann TG. Characterization of murine decidual natural killer (NK) cells and their relevance to the success of pregnancy. Cell Immunol 1985; 93:315-326[CrossRef][Medline]
  42. Biron CA, Sonnenfeld G, Welsh RM. Interferon induces natural killer cell blastogenesis in vivo. J Leukoc Biol 1984; 35:31-37[Abstract]
  43. McIntyre KW, Natuk RJ, Biron CA, Kase K, Greenberger J, Welsh RM. Blastogenesis of large granular lymphocytes in nonlymphoid organs. J Leukoc Biol 1988; 43:492-501[Abstract]
  44. Lefèvre F, Martinat-Botté F, Guillomot M, Zouari K, Charley B, La Bonnardière C. Interferon-gamma gene and protein are spontaneously expressed by the porcine trophectoderm early in gestation. Eur J Immunol 1990; 20:2485-2490[Medline]
  45. Lefèvre F, Boulay V. A novel and atypical type one interferon gene expressed by trophoblast during early pregnancy. J Biol Chem 1993; 268:19760-19768[Abstract/Free Full Text]
  46. Platt JS, Hunt JS. Interferon-gamma gene expression in cycling and pregnant mouse uterus: temporal aspects and cellular localization. J Leukoc Biol 1998; 64:393-400[Abstract]
  47. Ashkar A, Croy BA. Interferon-{gamma} contributes to the normalcy of murine pregnancy. Biol Reprod 1999; 61:493-502[Abstract/Free Full Text]
  48. Guimond MJ, Luross JA, Wang B, Terhorst C, Danial S, Croy BA. Absence of natural killer cells during murine pregnancy is associated with reproductive compromise in TgE26 mice. Biol Reprod 1997; 56::169-179[Abstract]
  49. Ashkar AA, Di Santo JP, Croy BA. Interferon gamma contributes to initiation of uterine vascular modification, decidual integrity, and uterine natural killer cell maturation during normal murine pregnancy. J Exp Med 2000; 192:259-270[Abstract/Free Full Text]
  50. Leiser R, Dantzer V. Structural and functional aspects of porcine placental microvasculature. Anat Embryol (Berl) 1988; 177:409-419[CrossRef][Medline]
  51. Biensen NJ, Wilson ME, Ford SP. The impact of either a Meishan or Yorkshire uterus on Meishan or Yorkshire fetal and placental development to days 70, 90, and 110 of gestation. J Anim Sci 1998; 76::2169-2176[Abstract/Free Full Text]
  52. Wilson ME, Biensen NJ, Ford SP. Novel insight into the control of litter size in pigs, using placental efficiency as a selection tool. J Anim Sci 1999; 77:1654-1658[Abstract/Free Full Text]
  53. Pope WF. Embryonic mortality in swine. In: Zavy MT, Geisert RD (eds.), Embryonic Mortality in Domestic Species. Boca Raton, FL: CRC Press, Inc.; 1994: 53–77



This article has been cited by other articles:


Home page
 Vet PatholHome page
S. Arrighi, F. Cremonesi, G. Bosi, D. Groppetti, and A. Pecile
Characterization of a Population of Unique Granular Lymphocytes in a Bitch Deciduoma, Using a Panel of Histo- and Immunohistochemical Markers
Vet. Pathol., July 1, 2007; 44(4): 521 - 524.
[Abstract] [Full Text] [PDF]

Home page
 Biol. Reprod.Home page
H. Engelhardt, B. A. Croy, and G. J. King
Conceptus Influences the Distribution of Uterine Leukocytes During Early Porcine Pregnancy
Biol Reprod, June 1, 2002; 66(6): 1875 - 1880.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Engelhardt, H.
Right arrow Articles by King, G. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Engelhardt, H.
Right arrow Articles by King, G. J.
Agricola
Right arrow Articles by Engelhardt, H.
Right arrow Articles by King, G. J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS