| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Faculty of Life Sciences, University of Manchester, Manchester M13 9PT, United Kingdom
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
cytokines, immunology, implantation, pregnancy, uterus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
During early pregnancy, an influx of leukocytes into the uterus occurs that has been likened to an inflammatory response, and although this may be partially stimulated by factors in seminal fluid [12], the expression of ovulatory estrogen also is required [13]. After implantation, the paternal antigens expressed by the embryo become exposed to the maternal blood system and, potentially, could stimulate an inflammatory response [14]. However, many of the leukocytes in the uterus during pregnancy are thought to be modified compared to those in the periphery and may even be beneficial to the developing embryo [15]. Macrophages enter the uterus in response to colony stimulating factor 1 (CSF1) expression from the luminal epithelium on D1 of pregnancy [16]. Mice lacking CSF1 expression have fewer macrophages in the uterus compared to wild-type mice, and the stromal macrophages are less mature than normal [17]. Pregnancy is affected in these mice, because they have reduced litter sizes [18]. Other cytokines are also involved in macrophage recruitment to the uterus, such as granulocyte-macrophage colony-stimulating factor (CSF2), interleukin (IL)-6, [19], and possibly, LIF.
Natural killer (NK) cells become the major leukocyte population in the uterus after implantation in both mice [20] and humans [21, 22]. Specific subsets of NK cells migrate to the uterus from the periphery using molecules such as L-selectin and 
4-integrin and cell adhesion molecules on blood vessels that may include MADCAM-1 and VCAM-1 [23–25]. This also is true for humans, in which CD56BRIGHT cells form only a small subset of NK cells in the periphery but are the major subpopulation of NK cells in the uterus [21]. Recently, NK cells have been detected in the uterus using specific staining as early as D5 of pregnancy [26], although morphological studies have found granulated NK-like cells as early as D3.5 of pregnancy [27]. Additionally, these cells become localized to the mesometrial side of the uterus, as is typical of NK cells [20]. The NK cell-deficient mice are fertile but have disrupted decidualization that is odematous and hypocellular, and uterine blood vessel modification does not take place because of a lack of interferon-
, which is normally produced by NK cells [28–30]. Some evidence indicates that increased numbers of NK cells are implicated in spontaneous abortion [31], but this also has been seen in women with decreased proportions of specific types of NK cells [32].
The main role of eosinophils in the human uterus is thought to be the production of matrix metalloproteinases that influence menstruation [33, 34]. Large numbers of eosinophils are found in the murine uterus on D1 of pregnancy and are located throughout the stroma, but by D3 of pregnancy, they become restricted to the myometrium and the stromal-myometrial junction [35]. The Il5 knockout mice have decreased numbers of eosinophils, and mice overexpressing IL-5 have increased numbers of eosinophils in the uterus [36]. However, no adverse effects on pregnancy are apparent in either case.
LIF could affect leukocyte subpopulations in the uterus either directly, because macrophages and eosinophils both express the LIF receptor [37], or indirectly. If these leukocyte subpopulations are disrupted in the absence of LIF, then pregnancy may be adversely affected, and this may contribute to the infertility phenotype seen in the Lif knockout mice. Therefore, in the present study, we compared the uterine leukocyte subpopulations during early pregnancy in Lif knockout mice with those in wild-type mice to evaluate the role of LIF in influencing leukocyte distributions and dynamics in early pregnancy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Mice were maintained under conditions required by the British Home Office Licence as described previously [9]. The Lif knockout mice were obtained from Dr. Andrew Sharkey (Department of Pathology, University of Cambridge, Cambridge, U.K.) from a colony produced by Dani et al. [38]. The Lif knockout mice on an MF-1 background between 7 and 9 wk of age were placed with MF-1 males overnight and checked for vaginal plugs the following morning, and this was designated as D1 of pregnancy. Pregnancy and/or ovulation also were confirmed by either identification of embryos in the reproductive tract or identification of corpora lutea. Because Lif knockout females are infertile, the colony was sustained by breeding from heterozygote females and homozygote knockout males. Genotyping for identification of Lif knockout mice was carried out by polymerase chain reaction on DNA samples from mice after weaning using the following primer sequences: To identify the wild-type Lif sequence, forward-primer 5'-cgcaatgacacctccaggtctagaag and reverse-primer 3'-ggtgcctaggaatggctgacactc were used to produce a 300-base pair (bp) DNA fragment, and for the Lif knockout sequence, the 5' primer and knockout 3'-gctgtccatctgcacgagactagtga were used to produce a 490-bp DNA fragment.
Pontamine Blue Injections for Identification of Implantation Sites
Tail-vein injection of 0.5 ml of 2% Pontamine blue (Sigma, Dorset, U.K.) was carried out on mice between D4 and D6 of pregnancy. Mice were killed by cervical dislocation after 5 min, and the uterine horns were dissected and inspected for sites of embryo implantation.
Cryosectioning and Immunocytochemistry
All mice were killed by cervical dislocation before dissection of the uteri. Cryosections of the uterus (thickness, 7 µm) were taken from the uteri of wild-type and Lif knockout mice from D2 to D6 of pregnancy for use in immunocytochemistry. On D5 and D6 of pregnancy, implantation sites were used, which were identified by pontamine blue injection and morphology. Sections were rinsed three times in PBS (Oxoid, Hampshire, U.K.) and then blocked in 20 µl of normal goat serum (NGS; 1:20 dilution; Sigma) for 45 min. Primary antibodies (all from BD Biosciences Pharmingen, Oxford, U.K., unless otherwise stated) were then added and incubated overnight at 4°C. Neutrophils were detected using a rat immunoglobulin (Ig) G antibody to LY6G (2.25 µg/ml), macrophages with a rat IgG antibody to F4/80 (1 µg/ml; Serotec, Oxford, U.K.), NK cells with a mouse IgM antibody to CD57 (1.3 µg/ml; clone NK-1; Neomarkers, CA), B cells with rat anti-CD19 (5 µg/ml), and T cells with rat anti-CD3 (5 µg/ml). Negative controls were carried out using rat IgG (Jackson Immunoresearch, distributed by Stratech Scientific, Bedfordshire, U.K.) and mouse IgM (NeoMarkers) at the same concentration as the primary antibodies. Primary antibodies were removed with six rinses of PBS, and either an anti-rat IgG secondary antibody conjugated to fluorescein isothiocyanate (FITC; Sigma) or IgM conjugated to FITC (Sigma) were used to detect the primary antibodies. Sections were incubated in the secondary antibodies for 90 min at room temperature in the dark before six rinses in PBS and mounting in Vectashield containing 1.5 µg/ml of diamidophenylindole (DAPI; Vector Laboratories, Inc., Burlingame, CA).
Diaminobenzidine Staining for Eosinophils
Uterine eosinophils were detected by a staining method that detects peroxidase within these cells [35]. One diaminobenzidine (DAB) tablet and one urea tablet (Sigma) were dissolved in 5 ml of PBS, and a 20-µl drop was placed on each section for 10 min at room temperature in a dark, humid chamber. Next, they were rinsed three times in PBS and left in PBS for 10 min, and this washing step was then repeated. The sections were counterstained with hematoxylin for approximately 30 sec before washing in water for 1 min and mounting in Vectashield containing DAPI.
Dolichos Biflorus Agglutinin Lectin Staining for NK Cells
This method of staining for NK cells is adapted from a protocol described by Paffaro et al. [26]. Uterine sections were incubated with 1% H2O2 for 30 min to inactivate endogenous peroxidase and then rinsed six times with PBS before blocking in 20 µl of NGS (1:20 dilution; Sigma) for 30 min. The NGS was removed, and 20 µl of a 6.6-µg/ml solution of Dolichos Biflorus Agglutinin (DBA) lectin conjugated to biotin (Sigma) were added and incubated at 4°C overnight. Sections were then rinsed six times in PBS and incubated in 20 µl of a 40-µg/ml solution of ExtraAvidin peroxidase conjugate (Sigma) for 30 min at room temperature in the dark. This was washed off in six changes of PBS before adding 20 µl of 1x DAB solution for a further 30 min and then rinsing in six changes of PBS. The sections were counterstained with hematoxylin for approximately 30 sec before washing with tap water for 1 min and mounting in Vectashield containing DAPI. Controls were carried out by substituting PBS for H2O2 or DBA lectin.
Toluidine Blue Staining for Mast Cells
A 20-µl drop of acidified toluidine blue (0.5 g of toluidine blue [Sigma] and 5 ml of concentrated HCl in 95 ml of distilled H2O) was added to uterine sections for 20 min and then rinsed off in tap water for 30 sec. Sections were then mounted in Vectashield.
Cell Counting
After immunocytochemistry or histological staining to detect different leukocyte subpopulations, cells in specific areas of uterine sections 100 x 200 µm2 in size were counted as shown in Figure 1. Percentages of different leukocyte subpopulations were calculated for each area by counting the FITC-labeled cells and then dividing this by the total cell count obtained from counting the DAPI-stained nuclei of all the cells in that area. Leukocyte subpopulations as a percentage of total cells counted over all the areas examined also were calculated for each section. Cells that were only partially within the grid were included only if they fell on the upper or the left side, not if they fell on the lower or the right side. For each day of pregnancy, three animals were used, with three sections per animal for both wild-type and Lif knockout mice.
|
To determine whether the anti-CD57 antibody and DBA lectin used for NK cell detection recognized the same population of cells, cell counts were carried out on serial sections stained alternately with either anti-CD57 or DBA lectin and analyzed statistically. Sections also were compared visually for individual cells stained by both methods of NK cell detection.
Statistics
A repeated-measures ANOVA was used to test all data for significant differences between the wild-type mice and the Lif knockout mice and to see if any differences occurred between areas on different days of pregnancy. Differences were accepted as being significant at P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Several leukocyte populations were analyzed by immunofluorescence in uterine sections. Neutrophils were detected from D1 of pregnancy. However, a sharp drop between D2 and D3 of pregnancy was observed (Fig. 2, A and B) and confirmed by flow cytometry, although the decrease was not as steep in the Lif knockout mice as in the wild-type mice (data not shown). From D3 onward, neutrophils were detected rarely.
|
The T and B cells were not detected by staining of uterine sections during early pregnancy in wild-type or Lif knockout mice (data not shown). Mast cells were detected on all the days of pregnancy analyzed (D2 to D6) by staining of uterine sections but were found only in very low numbers in the myometrium, and no obvious differences were seen between the wild-type and Lif knockout mice (data not shown).
Macrophages, NK cells, and eosinophils also were detected by immunofluorescence in uterine sections, and the dynamics of these cells as well as the differences between wild-type and Lif knockout mice were analyzed in more detail.
Distributions and Dynamics of Macrophages
Macrophages were detected in uterine sections by immunocytochemistry on all days of pregnancy analyzed (D1 to D6) in both wild-type and Lif knockout mice (Fig. 2, C and D). Cell counts showed that on D3 of pregnancy, 17% of the cells were macrophages in the wild-type uteri, and this was more than double the percentage in the Lif knockout mice (P = 0.0003). The areas around the lumen (areas A, B, and C; areas are as shown in Fig. 1) had the highest proportion of macrophages in the wild-type mice, but the percentages of macrophages in these areas in the Lif knockout mice were only one-third as high in comparison (P = 0.013, 0.014, and 0.031, respectively) (Figs. 2, C and D, and 3A). The areas in the midstroma (areas D and E) on D3 also had significantly fewer macrophages in the Lif knockout mice, with only half the percentage seen in the wild-type mice (P = 0.034 and 0.016, respectively). A significant decrease of one-third in the total percentage of macrophages occurred in the wild-type mice between D3 and D4 (P < 0.0001), especially in the areas around the lumen. The stromal areas had significantly fewer macrophages than the myometrial areas on D4 of pregnancy in the wild-type mice (P = 0.019) (Fig. 3B). A significantly lower percentage of macrophages was found overall in the Lif knockout mice on D4 compared to the wild-type mice (P = 0.016), but differences in none of the individual areas reached significance. On D5, the areas around the lumen had the lowest proportion of macrophages in the wild-type mice, and no significant differences were found between wild-type and Lif knockout mice (Fig. 3C). On D6, the areas in the myometrium (areas H, I, and J) had significantly higher percentages of macrophages than the areas in the stroma (P = 0.043) (Fig. 3D) in the wild-type mice. The total percentage of macrophages was significantly lower in the Lif knockout mice compared to the wild-type mice (P = 0.008), as were those in areas C and I (P = 0.038 and 0.014, respectively).
|
Distributions and Dynamics of NK Cells
NK cells were not detected in uterine sections on D1 or D2 in wild-type or Lif knockout mice. However, from D3 onward, they were found throughout the stroma but not the myometrium, and by D7, they showed the typical mesometrial distribution associated with these cells. This was confirmed by statistical analysis of cell counts after alternate section staining with anti-CD57 and DBA lectin to show that both methods of NK cell detection recognized the same cell population (Fig. 2, E–H).
Cell counts showed that NK cells comprised 8% of the total cell population in wild-type mice on D3 (Fig. 3E) and were located mainly on the mesometrial side of the uterus (area A), which had a significantly higher percentage of NK cells than the other areas investigated (P = 0.009). However, Lif knockout mice had a significantly higher percentage of NK cells on the antimesometrial side of the uterus (areas C, E, and G) on D3 of pregnancy compared to the wild-type mice (P = 0.034, 0.021, and 0.04, respectively) (Fig. 3E). The total percentage of NK cells in wild-type animals more than doubled, to more than 16%, on D4 of pregnancy in wild-type mice (P = 0.04) (Fig. 3F), especially in areas A and B next to the lumen, which more than tripled in percentage of NK cells, and these areas also were significantly different from D3 onward (P = 0.037 and 0.007, respectively). The Lif knockout mice showed a similar increase in percentage of NK cells on D4, but area A had a significantly higher percentage than that in the wild-type mice (P = 0.027) (Fig. 2, G and H, and 3F). On D5, the total percentage of NK cells in the wild-type mice decreased slightly, but they were still located mainly in the mesometrial region. Indeed, area A had at least double the percentage of NK cells compared to any other area (P = 0.004). The Lif knockout mice showed a similar decrease in the proportion of NK cells on D5, and no significant differences were found between them and the wild-type mice (Fig. 3G). On D6 in area A, the proportion of NK cells decreased to two-thirds of what it was on D4 in the wild-type mice (P = 0.025), but NK cells were still located mainly in the areas around the lumen (areas A, B, and C) (Fig. 3H). The midstromal region (area E) contained only 4% NK cells, which was significantly lower than the average (P = 0.019). No significant differences in NK cells were found between the wild-type mice and the Lif knockouts on D6 of pregnancy.
Distributions and Dynamics of Eosinophils
Eosinophil staining showed that these cells were present in the uterus throughout the time period investigated (D2 to D6) (Fig. 2, I and J). Cell counting revealed that on D2 of pregnancy, 5% of the uterine cells were eosinophils in wild-type mice. The highest concentration of eosinophils was in area H in the mesometrial triangle, which contained 10% eosinophils, a significantly higher percentage than found in the other areas (P = 0.026) (Fig. 3I). No significant differences were found in the percentage of eosinophils between the Lif knockout and wild-type mice on D2 of pregnancy. On D3 of pregnancy, the total number of eosinophils dropped to 3% in wild-type mice, and few were detected in the stroma. Most areas contained less than 1% eosinophils. From D3 onward, eosinophils were mainly found in the myometrium, especially in the mesometrial triangle (area H) (Fig. 3J). However, on D3 in the Lif knockout mice, the areas in the outer stroma (areas F and G) had at least double the percentage of eosinophils seen in the wild-type mice (P = 0.018 and 0.048, respectively) (Figs. 2, I and J, and 3J). The proportion and distribution of eosinophils in wild-type mice did not alter significantly from D4 onward; the highest percentage of eosinophils was always located in area H in the mesometrial triangle. On D4, area H had more than 8% eosinophils, which was significantly higher than the percentage in the other areas (P < 0.0001) (Fig. 3K). In the Lif knockout mice on D4 of pregnancy, areas C and E in the antimesometrial stroma had significantly higher percentages of eosinophils (P = 0.027 and 0.034, respectively). The overall proportion of eosinophils was significantly higher than that for the wild-type mice (P = 0.034) (Fig. 3K). On D5 (Fig. 3L) and D6 (Fig. 3M), no significant differences were found between the wild-type mice and the Lif knockout mice.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The NK cells were detected in the uterus as early as D3 by both DBA lectin staining and anti-CD57 staining. Recently, DBA lectin has been shown to detect only the NK cell subset of leukocytes in the uterus and not those elsewhere in the body. Immature uterine NK cells, which previously have been difficult to identify, are also recognized by DBA lectin, although it recognizes endothelial cells as well [26]. The present study has shown that anti-CD57 also can be used to identify these cells in mice. The first appearance of NK cells on D3 of pregnancy is earlier than previously reported [26], but this may result from the more sensitive detection methods used for the present study. The detection of NK cells on D3 correlates with the rise in progesterone expression from the corpora lutea, and this hormone may influence NK cell migration. The Lif knockout mice had higher percentages of NK cells in several areas of the uterus on D3 of pregnancy compared with wild-type females, suggesting that the recruitment of these cells is normally restricted by LIF. One possible role of LIF may be to reduce the expression of cell adhesion molecules on blood vessels that are used by NK cells for migration into the uterus, such as MADCAM-1 and VCAM-1 [23– 25]. However, this remains to be investigated. Although the proportion of NK cells was altered in the Lif knockout mice, the overall distribution of NK cells was similar to that in the wild-type mice. The NK cells became located preferentially at the mesometrial side of the uterus, especially around the lumen, as has been shown previously [41]. Therefore, the present study shows that LIF has a role in limiting recruitment of NK cells to the uterus, but does not affect their distribution within the uterus.
As previously shown, eosinophils become located preferentially to the myometrium and the stromal-myometrial border during early pregnancy [36, 42]. The Lif knockout mice had higher percentages of eosinophils in the outer stroma on D3 and in the midstroma on D4, and the total percentages also were higher than those in the wild-type mice, suggesting that LIF restricts the migration of eosinophils into the uterus. As for NK cells, LIF may restrict the expression of molecules on blood vessels that facilitate recruitment of eosinophils. However, it also is possible that LIF induces inhibitory cytokines in the uterus or represses chemoattractant molecules.
Neutrophils also were affected by the lack of LIF in the uterus, because the decrease that was seen between D2 and D3 in the Lif knockout mice was not as large as that seen in the wild-type mice. This suggests that LIF may normally promote the apoptosis of neutrophils in the uterus. Although to our knowledge no previous evidence supports this action of LIF on neutrophils, it can promote apoptosis in other cells types, such as mammary epithelial cells [43].
Alterations in the proportions of leukocytes have been shown to adversely affect pregnancy. For example, decreased numbers of NK cells during early pregnancy result in acellular and odematous decidual zones [20], and increased numbers of NK cells have been associated with embryo resorption [44]. Decreased numbers of macrophages in the uterus lead to decreased litter sizes and birth weights [17]. Although mice with altered proportions of single leukocyte subpopulations remain capable of producing viable offspring, it is unknown whether the combination of decreased macrophage and increased NK cell proportions seen in the Lif knockout mice is enough to severely perturb pregnancy, but it could at least be a contributing factor to the infertility seen in these mice. It is interesting to note that these differences precede D4 expression of LIF and, therefore, must be a result of D1 LIF expression, which has not been attributed previously with any major functions in the uterus during pregnancy. Previous work has indicated that ovulation in Lif knockout mice is normal [3] (unpublished data) and that progesterone levels in plasma are equivalent to those in wild-type mice [7]. Some of the differences seen on D6 of pregnancy could be a result of decidualization in wild-type mice, because this does not occur in the Lif knockout mice. Other work in the laboratory has shown differences in uterine cell density and stromal cell proliferation between wild-type mice and Lif knockout mice during early pregnancy (unpublished data), but these differences do not correlate with the differences in the leukocyte subpopulations reported in the present study.
Future work should test whether D1 injection of LIF into Lif knockout mice can reverse these differences from the wild-type mice in the proportions of leukocyte subpopulations. Dose-response experiments injecting LIF into the knockout uterus also would establish if LIF is a chemokine for macrophages as well as if it can restrict NK cell and eosinophil migration into the uterus. Additionally, immunocytochemistry for endothelial cell adhesion molecules involved in NK cell migration could be carried out to see if they are increased in the uterus of Lif knockout mice.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: Susan J. Kimber, Faculty of Life Sciences, University of Manchester, 3.239 Stopford Building, Oxford Road, Manchester M13 9PT, U.K. FAX: 0161 275 3915; sue.kimber{at}man.ac.uk
3 Current address: Institute of Human Genetics, International Centre for Life, Central Parkway, Newcastle upon Tyne, U.K
Received: 9 August 2004.
First decision: 13 September 2004.
Accepted: 10 November 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. J Reprod Immunol 2001 49:33-47[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  M. S. M. van Mourik, N. S. Macklon, and C. J. Heijnen Embryonic implantation: cytokines, adhesion molecules, and immune cells in establishing an implantation environment J. Leukoc. Biol., January 1, 2009; 85(1): 4 - 19. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. W. Agnew, L. B. Corbeil, L. Munson, B. A. Byrne, and R. H. BonDurant A Pregnant Mouse Model for Bovine Tritrichomonas foetus Infection Vet. Pathol., November 1, 2008; 45(6): 849 - 864. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.A Fouladi-Nashta, L Mohamet, J.K Heath, and S.J Kimber Interleukin 1 Signaling Is Regulated by Leukemia Inhibitory Factor (LIF) and Is Aberrant in Lif-/- Mouse Uterus Biol Reprod, July 1, 2008; 79(1): 142 - 153. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. V. Ingman and R. L. Jones Cytokine knockouts in reproduction: the use of gene ablation to dissect roles of cytokines in reproductive biology Hum. Reprod. Update, March 1, 2008; 14(2): 179 - 192. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Tonello and G. Poli Tubal ectopic pregnancy: macrophages under the microscope Hum. Reprod., October 1, 2007; 22(10): 2577 - 2584. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Barbaric, G. Miller, and T. N. Dear Appearances can be deceiving: phenotypes of knockout mice Brief Funct Genomic Proteomic, June 20, 2007; (2007) elm008v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Song and H. Lim Evidence for heterodimeric association of leukemia inhibitory factor (LIF) receptor and gp130 in the mouse uterus for LIF signaling during blastocyst implantation Reproduction, February 1, 2006; 131(2): 341 - 349. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Dimitriadis, C.A. White, R.L. Jones, and L.A. Salamonsen Cytokines, chemokines and growth factors in endometrium related to implantation Hum. Reprod. Update, November 1, 2005; 11(6): 613 - 630. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J Kimber Leukaemia inhibitory factor in implantation and uterine biology Reproduction, August 1, 2005; 130(2): 131 - 145. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
