Published online before print January 18, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1,1
,,2
* Departments of Pediatrics and
Obstetrics, Gynecology and Reproductive Sciences, Center for Reproductive Sciences, University of California, San Francisco, California, USA; and
Department of Gynecology and Obstetrics, Emory University School of Medicine, Atlanta, Georgia, USA
2 Correspondence: Department of Gynecology and Obstetrics, Emory University School of Medicine, 1639 Pierce Drive, Suite 4217-WMB, Atlanta, GA 30322, USA. E-mail: robert.n.taylor{at}emory.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B inhibition. Western blotting of U937 and primary monocyte lysates demonstrated that rGd activated caspase-8, -2, and -3 to execute programmed cell death in these cells. We postulate that infiltrating monocytes and potentially other innate immune cells of the decidua might be manipulated by this glycoprotein to enhance embryonic implantation rates or conversely, to develop novel contraceptive strategies.
Key Words: apoptosis • embryonic implantation • monocytes • pregnancy • decidua
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
GdA, previously referred to as placental protein 14 and progesterone-associated endometrial protein, has been the most extensively studied isoform [1 ]. It is the predominant protein secreted by human endometrial epithelial cells and decidua in the secretory and early pregnancy phases, respectively, of women and other primates [3 , 4 ]. Recent cDNA microarray experiments verify that the endometrial gene encoding GdA is markedly up-regulated following ovulation [5 6 7 ]. The postulated regulators of cyclical GdA secretion in the endometrium include progesterone, relaxin, and human chorionic gonadotropin (hCG) [8 , 9 ]. Ligand-activated progesterone receptors A and B can activate GdA gene transcription [10 ], but it appears that consensus progesterone response elements (PREs) in the proximal gene promoter are synergized by other transcription factors [4 , 11 , 12 ].
The GdA glycoform bears a pattern of glycosylation that is specific to the endometrium. The primary glycosylation sites are Asn 28 and Asn 63 of the 162-aa backbone of the Gd protein. GdA, purified from secretory endometrium and midtrimester amniotic fluid, contains complex oligosaccharides with sialylated lactosamine (lacNAc and lacdiNAc) antennae at their terminal ends. As a result, Wisteria floribunda and Sambuccus nigra lectins can bind to GdA. By contrast, GdS, derived from seminal plasma, is devoid of sialylated oligosaccharides and also lacks immunomodulatory bioactivity [1 ]. The endometrial GdA glycoform potently and dose-dependently suppresses NK cell, macrophage, and T cell function [13 14 15 ] and also inhibits human sperm-egg binding at the zona pellucida. The mechanism of macrophage inhibition by GdA remains unknown and is the subject of the present series of experiments.
The immunosuppressive properties of GdA are highly relevant to the unique physiology of embryonic implantation and early pregnancy in Old World primates. In these species, the hemochorial placental interface consists of a complex admixture of allogeneic fetal and maternal cells, including a host of innate immune cells that invades the endometrium during the secretory phase and peri-implantation period. The major classes of immune cells present at the implantation site are NK cells, macrophages, and T cells [16
]. GdA has been shown to inhibit the activities of all three of these lineages [13
14
15
, 17
], as well as B cell proliferation [18
]. Putative Gd receptors with a density of 
20,000/cell, Kd of 50 nM, and an apparent molecular weight of 250 kDa have been described on human monocytic cells [14
, 17
].
It is proposed that GdA confers immunotolerance at the fetomaternal interface [1 , 13 ]. Reduced GdA expression has been found in cases with recurrent pregnancy loss and repetitive in vitro fertilization failures and in endometriosis [19 ], a syndrome increasingly associated with progesterone resistance and impaired secretory differentiation [20 ]. Each of these clinical conditions is associated with reduced embryonic implantation and impaired maintenance of pregnancy. The precise mechanism of immune cell-induced pregnancy loss is unknown in humans, but data from murine models indicate that NO-free radicals released from activated macrophages contribute to abortion [21 ]. It was the purpose of these studies to first clone, express, and characterize recombinant Gd (rGd) in Pichia pastoris and second, to analyze the effects of natural GdA and rGd on monocytic U937 cells and primary peripheral human monocytes as in vitro models for immune cell regulation in the decidua. Our findings indicate that GdA and rGd affect cell cycle kinetics and programmed cell death pathways in human monocytes. We describe several of the molecular mediators and pathways involved and propose how GdA might play an important immunoregulatory role in the maintenance of human pregnancy.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A (Invitrogen, Carlsbad, CA, USA) and verified by DNA sequencing of both DNA strands. The recombinant expression plasmid (pPICZA-Gd), encoding a secreted fusion protein with yeast -factor, human Gd, c-myc, and polyhistidine tags, was transformed into competent KM71H cells by electroporation and grown in media containing 1% yeast extract, 2% peptone, 2% dextrose, 100 mM potassium phosphate, pH 6, 1.34% yeast nitrogen base with ammonium sulfate without amino acids, 0.00004% biotin, and 1% glycerol [buffered minimal glycerol complex medium (BMGM)] at 30°C until the culture reached an OD at 600 nm = 2–6. Recombinant gene expression was induced by adding 0.5% methanol to the BMGY media for 5 days, with the addition of methanol (0.5% final concentration) each day. The cells were centrifuged, and the supernatant, containing secreted rGd, was used for purification or stored at –80°C.
Purification of rGd
Conditioned media were loaded onto a 1-ml Talon affinity resin (Clontech, Palo Alto, CA, USA), previously equilibrated in buffer A: 50 mM sodium phosphate, 300 mM NaCl, 10% glycerol. After loading, the column was washed with 10 bed vol equilibration buffer containing 10 mM imidazole, 2.5 mM β-ME, and 0.1 mM PMSF. The retained proteins were eluted by adding to the latter 100 mM KCl plus 100 mM imidazole.
Pooled fractions containing the proteins of interest were dialyzed against PBS, pH 7.4, at 4°C overnight. Immediately after dialysis, protein concentration was determined using the bicinchoninic acid protein assay kit from Pierce (Rockford, IL, USA) with BSA used as a standard. In some experiments, conditioned media were passed over a Con A affinity sepharose column (Sigma Chemical Co., St. Louis, MO, USA), previously equilibrated with 10 bed vol buffer B: 10 mM Tris-HCl, 150 mM NaCl, 1 mM CaCl2, 1 mM MnCl2. After loading, the column was washed with 10 vol buffer B, and then retained proteins were eluted by adding -methyl-mannose to buffer B at a final concentration of 0.5 M. To assure removal of trace β-ME, pooled fractions containing the proteins of interest were dialyzed against 500 vol PBS at 4°C overnight before protein concentrations were determined.
Gel electrophoresis and N-terminal sequence analysis
Purified proteins were analyzed by SDS-PAGE. In some experiments, gels were stained using Coomassie brilliant blue R-250. In other experiments, proteins were transferred to polyvinylidene difluoride (PVDF; 0.2 µm, Bio-Rad Laboratories, Hercules, CA, USA) membranes for protein sequencing in the presence of 10 mM 3-(cyclohexylamine)-1-propane sulfonic acid buffer, pH 11, and 10% methanol.
Blotted proteins were stained with 0.1% Coomassie brilliant blue R-250 in 50% methanol and 1% acetic acid and destained in 50% methanol. Peptide bands were excised from the PVDF membrane and directly microsequenced on a vapor-phase Beckman-Porton PI2090 sequencer (Beckman Instruments, Inc., San Ramon, CA, USA), using the Edman degradation procedure. Amino acid sequences obtained were searched in the Protein Information Resource/Swiss protein databank.
Deglycosylation of rGd
Purified rGd was incubated with 500 U N-glycosidase F (PNGase F) at 37°C for 18 h in 50 mM sodium phosphate, pH 7.5, containing 1% Nonidet P-40. After treatment, SDS sample-loading buffer was added, and the samples were analyzed by SDS-PAGE.
Western blotting
The conditioned medium from cells expressing rGd was subjected to 12% SDS-PAGE and blotted onto a PVDF membrane, which was blocked in 5% milk, 10 mM Tris-HCl, pH 7, 150 mM NaCl, and 0.1% Tween 20 and probed with anti-myc (1:2000) or anti-Gd (1:2000) antibodies.
For characterization of caspase-8, -2, and -3, lysates obtained from rGd-treated and untreated U937 cells and primary monocytes were separated by 10% SDS-PAGE, blotted onto PVDF membranes, and probed with specific rabbit antisera (1:1000) that recognize only the activated, cleaved products of caspase-8 (#9496) or -3 (#9661), respectively, obtained from Cell Signaling Technology (Danvers, MA, USA). Mouse anti-caspase-2 (sc-5292) mAb (1:200) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA).
Protein bands were visualized using an enhanced chemiluminescence detection system (Amersham, Piscataway, NJ, USA). To calculate the yields of purified proteins, the signal of each band was quantified using Scion Image software (Houston, TX, USA).
Cell culture and treatments of the U937 cell line
The U937 monocytic cell line was obtained from American Type Culture Collection (Manassas, VA, USA) and grown in RPMI-1640 medium with 25 mM HEPES, 2 g/L NaHCO3, and 10% FCS. The Ishikawa cell line (generously provided by Dr. Frederick Schatz, Yale University, New Haven, CT, USA) was grown in DMEM/F12 (1:1) medium with 25 mM HEPES, 10 nM estradiol, supplemented with 10% FCS, as described previously [22
]. The cells (0.25x106 in 0.5 ml) were treated in the above RPMI medium with 0.1% FCS in a 24-well plate up to 8 h with 3.2 µg/ml GdA purified from first trimester human decidual homogenates [12
], 0–20 µg/ml rGd, or 5 µg/ml actinomycin D (Sigma Chemical Co.), a concentration that we and others previously demonstrated to induce apoptosis [22
].
Cell viability assay
The U937 and Ishikawa cell lines were incubated with 3.2 µg/ml purified decidual GdA [14
] or various doses of rGd for up to 6 h. A 50-µl aliquot of each cell suspension was taken immediately after treatment and diluted with 50 µl 0.4% trypan blue solution (Invitrogen). The numbers of stained and unstained (viable) cells were counted on a hemacytometer using an optical microscope. All treatments and cell counts were done in triplicate.
Cell-death ELISA
U937 cells were treated with 5 µg/ml rGd, 3.2 µg/ml GdA, or 5 µg/ml actinomycin D for 6 h. The cells were collected, lysed, and centrifuged, and 20 µl of the resulting supernatant containing the cytoplasmic fraction was used to assess apoptosis by TUNEL of DNA with a cell-death detection ELISA (Roche, Nonnenwald, Germany), performed according to the manufacturer’s instructions. The absorbance measurement at 405 nm (A405) represents nucleosomal enrichment in the cytoplasm, induced after apoptotic cell death. For in situ evaluation of apoptosis in U937 cells, cytocentrifugation onto glass slides using a Shandon Cytospin 3 instrument was performed, and TUNEL was detected directly using fluorescein deoxy-UTP (dUTP) to label the DNA strand breaks.
DNA laddering assay
U937 cells were treated with 5 µg/ml rGd for up to 6 h. The cells were centrifuged and lysed for 20 min in 100 µl DNAzol (Invitrogen), and the lysate was precipitated with ethanol, washed, resuspended in 10 µl 8 mM NaOH, and neutralized with 1 µl 0.1 M HEPES. The isolated genomic DNA samples were analyzed on 1.2% agarose-Tris-boric acid-EDTA buffer (TBE) gels and visualized after ethidium bromide staining.
Isolation, culture, and flow cytometry of primary human monocytes
Heparinized blood (10 U/ml) was collected from the antecubital veins of volunteers providing written, informed consent under protocols approved by the University of California, San Francisco (UCSF; CA, USA) and Emory University (Atlanta, GA, USA) Institutional Review Boards. The mononuclear cells were separated by centrifugation over Histopaque solution and plated at 107 cells/ml onto polystyrene in six-well culture dishes in RPMI-1640 medium containing 1% nonessential amino acids, 1% sodium pyruvate, and 0.1 mg/ml penicillin/streptomycin/L-glutamine (Sigma Chemical Co.). After 1 h at 37°C, monocytes are adherent, whereas lymphocytes remain in suspension and are removed. The serum-free medium was subsequently replaced by RPMI 1640 containing 20% human serum and the above-mentioned supplements for 4 days. The cells were then exposed to control medium (RPMI-1640 medium with 25 mM HEPES, 2 g/L NaHCO3, and 0.5% FCS) or medium containing 3.2 µg/ml GdA purified from first trimester human decidual homogenates [14
], 5 µg/ml rGd, or 5 µg/ml actinomycin D (Sigma Chemical Co.), a concentration that we and others previously demonstrated to induce apoptosis [22
].
In Situ Cell Death Detection® immunocytochemistry in primary human monocytes
In some experiments, primary monocytes prepared as described above were plated directly onto Lab-Tek chamber slides, and TUNEL was performed in situ using the double immunocytochemical staining protocol according to the manufacturer’s specifications (Roche).
Microarray and real-time quantitative RT-PCR (qRT-PCR)
U937 cells were treated with 5 µg/ml rGd for up to 6 h and harvested for DNA with DNAzol and RNA with Trizol (Invitrogen) and RNEasy columns (Qiagen, Valencia, CA, USA) using the manufacturer’s instructions. Internucleosomal DNA fragmentation of rGd-treated cells was confirmed by ethidium bromide staining of isolated genomic DNA in 1.2% agarose–TBE gels as described above. The apoptosis gene array (SuperArray, Frederick, MD, USA) was performed according to the manufacturer’s specifications. Briefly, 1 µg total RNA was reverse-transcribed at 37°C for 30 min using random primers and RT. The cDNAs were amplified and labeled with biotin-16–dUTP by DNA polymerase using PCR with an initial denaturation at 85°C, 5 min; 30 cycles of 85°C, 1 min; 55°C, 1 min; 72°C, 1 min; followed by 72°C for 5 min. The array blots were prehybridized for 1 h at 60°C by blocking with 100 µg/ml denatured salmon sperm DNA in 2 ml GEAprehyb. The labeled cDNAs were denatured at 95°C, 5 min, and hybridized to the microarray blots in 0.75 ml GEAprehyb containing denatured salmon sperm DNA for 16 h at 60°C. Each probe was hybridized with two different blots. Washing was performed at 60°C, 20 min, with 2x SSC, 1% SDS, and twice with 0.1x SSC, 0.5% SDS. A chemiluminescent detection system was used, and the blots were subjected to autoradiography. The X-ray films were scanned on a flat-bed scanner and analyzed using the GEArray Expression Suite software. Scatter-plots were used to compare the relative expression pattern of genes in rGd-treated versus untreated groups. Genes that gave fold-change values beyond the preassigned ± 1.3-fold boundary, based on the intergroup data range (P<0.05), were considered as candidate genes for further analysis by real-time PCR.
For real-time qRT-PCR, first-strand cDNA was synthesized using random primers and SuperScript III polymerase (Invitrogen), according to the manufacturer’s instructions. A 1:10 dilution of the cDNA was used in real-time PCR with the iQ SYBR Green Supermix on a Bio-Rad iCycler thermal cycler. The following upstream and downstream (respectively) PCR primer pairs were used: Bcl-2A1, 5' CAA AAC GTC CAG AGT GCT ACA AAA and 5' CCC AGT TAA TGA TGC CGT CTT C; a proliferation-inducing ligand [APRIL; or TNF superfamily (TNFSF)-13a], 5' GGC TCT GCT GAC CCA ACA AAC A and 5' GGG AAC CAG GTG CAG GAC AGA GT; TNF-R1, 5' AGC GCC CAC AAG CCA CAG AGC and 5' CAG GCA GCC CAG CAG GTC CAT; Bad, 5' AGC CAA CCA GCA GCA GCC ATC AT and 5' CTC CCC CAT CCC TTC GTC GTC; and Bax, 5' ACT GGG GCC GGG TTG TCG and 5' GGC AGG GGC GGT GGT GAG. β-Glucuronidase, 5' CTC ATT TGG AAT TTT GCC GAT T and 5' CCG AGT GAA GAT CCC CTT TTT A, was used as an internal control. The temperature profile was 95°C, 5 min, followed by 40 cycles of 94°C, 10 s; 62°C, 10 s; and 72°C, 20 s. The data were collected and analyzed using the comparative threshold cycle method.
Reporter gene construct and transfection
The NF-B response element reporter vector was a kind gift from Dr. Dale C. Leitman (UCSF). This 125-bp construct contains the HIV-1 long-terminal repeat with two NF-B-binding sites cloned in the phase upstream of the firefly luciferase cDNA. The NF-B response element reporter vector was transfected in triplicate into U937 cells by Effectene (Qiagen) and treated overnight in the absence or presence of 5 µg/ml rGd, 10 nM phorbol 12,13 dibutyrate (PDB), or 0.6 nM TNF-. These concentrations previously were determined to approximate the EC50 for each agent.
Data presentation and analysis
Unless stated otherwise, experiments were repeated three times, and the results are presented as mean ± SE. Statistical analyses were performed using ANOVA and t-tests, and significance was accepted for two-tailed tests, where P < 0.05.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To obtain larger quantities of purified Gd for the present study, we constructed a P. pastoris expression plasmid pPICZA-Gd containing an N-terminal yeast -factor signal sequence to direct the secretion of rGd into the culture medium (Fig. 1A
). The secreted rGd product was designed to contain c-myc and polyhistidine tags in the C terminus, permitting ready detection and purification of the recombinant protein. Western blotting of rGd showed that it dose-dependently reacted with anti-myc (Fig. 1B
, Lanes 1–3) and anti-GdA (Fig. 1B
, Lane 4) antibodies, respectively. These banding patterns were similar to those obtained by Coomassie blue staining of the culture medium of pPICZA-Gd-expressing cells, where the predominant product migrated at 28 kDa (data not shown). N-terminal sequencing of the predominant 28-kDa band and the weaker 23-kDa product showed identical amino acid sequences: EFDIPQTKQDLE, which match exactly the amino terminal sequence of human Gd. We used the Talon metal affinity chromatography resin to purify histidine-tagged rGd secreted into the culture medium. Although 28 kDa and 23 kDa proteins were detected in the culture medium of pPICZA-Gd-expressing cells (Fig. 1C
, Lane 1), the 28-kDa product was bound selectively to and eluted from the Talon resin (Lanes 2 and 3, respectively). This observation suggested that the 23-kDa band represents a rGd product that was cleaved near the C terminus and lacked a portion of the 6x His tag or alternatively, is a partially glycosylated species that binds less well to the metal column. Our affinity purification method yielded an estimated 20- to 30-fold enrichment of the 28-kDa protein (Fig. 1C
, Lane 1 versus 3).
 View larger version (34K): [in a new window] |
Figure 1. (A) The expression construct pPICZA-Gd contained the cDNA sequence encoding aa 2–162 of the mature GdA cloned into a Pichia expression vector that has an N-terminal -factor signal sequence and C-terminal c-myc and polyhistidine tags (6xHis). The full-length fusion protein and the secreted rGd product have estimated molecular weights of 32 kDa and 22.7 kDa, respectively. The downward-pointing arrowhead represents the signal sequence cleavage site. (B) Western blot of conditioned medium from cells expressing rGdA, probed with anti-myc (Lanes 1–3) and anti-Gd (Lane 4) antibodies. The conditioned medium was diluted 1:4 (Lane 1), 1:16 (Lane 2), 1:64 (Lane 3), and 1:5 (Lane 4). The first 12 aa in the N terminus of each protein band are shown on the figure. (C) Coomassie blue staining of a 12% SDS-PAGE analyzed by electrophoresis containing undiluted conditioned medium from cells expressing rGd (Lane 1), flow-through material after passing conditioned medium through the Talon affinity chromatography column (Lane 2), and eluted fraction (Lane 3). M, molecular weight markers. (D) Conditioned medium from cells expressing rGd was subjected to Con A affinity chromatography. The eluted fractions in Lanes 1–3 were analyzed by 12% SDS-PAGE, followed by Coomassie blue staining. (E) Purified rGd secreted by P. pastoris was incubated without and with PNGase F (Lanes 1 and 2, respectively), the samples loaded onto a 12% SDS-PAGE gel, separated by electrophoresis, and the gel was stained with Coomassie blue. These data were replicated in two independent experiments.
|
GdA and rGd induce cell death of U937 monocytes but not of endometrial Ishikawa cells
Previously, using GdA purified from first trimester human decidua, we showed that the 28-kDa GdA protein inhibited U937 monocyte chemotaxis in Boyden chamber migration assays [14
]. To determine whether the inhibition in chemotaxis might be a result of reduced cell viability by GdA, we treated U937 cells with 3.2 µg/ml GdA purified from first trimester decidua. As determined by the trypan blue dye exclusion assay, the viability of U937 monocytes was reduced 32% when treated with purified GdA (Fig. 2
, P<0.05).
 View larger version (26K): [in a new window] |
Figure 2. U937 cells were treated for 6 h with 3.2 µg/ml GdA purified from first trimester decidual homogenates. Cells were stained with trypan blue, and the number of stained and unstained cells was counted on a hemacytometer glass plate using an optical microscope. Each treatment and cell counts were done in triplicate. These data were replicated in three independent experiments. *, Significant decrease relative to control (P<0.05).
|
2.5 µg/ml and with a maximum effect after 6 h of treatment (Fig. 3B
, P<0.05). Using Limulus amebocyte lysate tests, we verified that our purified, decidual GdA and yeast rGd preparations contained less than 1 European unit/ml endotoxin or less than 100 pg/ml, according to the U.S. Pharmacopeia (Rockville, MD, USA), indicating that the effects observed above were not induced by residual endotoxin contamination of our preparations. Furthermore, we verified that rGd had no effect on Ishikawa cells treated under identical conditions (Fig. 3C)
, consistent with our previous failure to detect specific GdA binding to this cell line [14
].
 View larger version (17K): [in a new window] |
Figure 3. U937 cells were treated for up to 6 h with increasing doses of rGd. Cell viability was determined using the trypan blue exclusion assay, as described in Materials and Methods. Each treatment and cell counts were done in triplicate. (A) U937 cells were treated for 6 h with 0–20 µg/ml rGd. (B) U937 cells were treated for up to 6 h with 5 µg/ml rGd. (C) Endometrial Ishikawa cells were treated for up to 6 h with 5 µg/ml rGd. No effect on viability was noted in the endometrial cell line. These data were replicated in four independent experiments. *, Significant decrease relative to 0 (P<0.05).
|
 View larger version (40K): [in a new window] |
Figure 4. To confirm an apoptotic effect of Gd on cultured U937 cells, the In Situ Cell Death Detection® kit, an immunocytochemical method to identify free 3'-OH ends localized in apoptotic bodies of cytocentrifuged U937 monocytes, was used. U937 cells were incubated in the absence (control) or presence of 5 µg/ml rGd and then cytocentrifuged onto glass slides using a Shandon Cytospin 3 instrument. Direct fluorescence cytochemistry for TUNEL staining was performed using fluorescein dUTP. The left upper image represents control U937 cells, whereas the right upper image shows the immunofluorescence features of cellular apoptosis following incubation with rGd. The lower images are the phase-contrast views of the same microscopic fields. These data were replicated in three independent experiments.
|
 View larger version (26K): [in a new window] |
Figure 5. (A) U937 cells were treated with 5 µg/ml rGd, 3.2 µg/ml purified decidual GdA, or 5 µg/ml actinomycin D (Act-D) for 6 h. The release of cytoplasmic mono- and oligonucleosomes in the samples was detected by the cell death detection ELISA (Roche). Each treatment and ELISA assay was done in triplicate. (B) U937 cells were treated with 5 µg/ml rGd for up to for 6 h in duplicate. Isolated genomic DNA was loaded onto a 1.2% agarose–TBE gel and visualized by ethidium bromide staining. These data were replicated in three independent experiments. *, Significant increase relative to control (P<0.05).
|
rGd and GdA induce apoptosis in primary human monocytes
Having demonstrated proapoptotic effects of GdA and rGd in U937 cells by a variety of techniques, we wanted to verify that this glycoprotein also had a direct biological effect on primary human monocytes. Using the immunocytochemical method of the In Situ Cell Death Detection® kit (Fig. 6
), we validated that GdA and rGd induced perinuclear TUNEL staining in primary monocytes cultured on Lab-Tek chamber slides. Exposure to a low concentration of actinomycin D (5 µg/ml) was used as a positive control for apoptosis, as we demonstrated previously [22
].
 View larger version (61K): [in a new window] |
Figure 6. Primary peripheral human monocytes were isolated and cultured on Lab-Tek chamber slides for 4–6 days and subsequently treated in the absence (control) or presence of purified GdA (3.2 µg/ml), rGd (5 µg/ml), or actinomycin D (5 µg/ml). The Roche In Situ Cell Death Detection® kit was used according to the double immunocytochemical method with 3'-diaminobenzidine tetrahydrochloride (DAB) as the chromogen. The perinuclear brown DAB signal is present in the treated cells, particularly in those exposed to actinomycin D. The cell nuclei were counterstained with hematoxylin. These data were replicated in three independent experiments.
|
 View larger version (18K): [in a new window] |
Figure 7. U937 cells were treated with rGd for 1 h and 6 h. Microarrays were performed using the SuperArray apoptosis gene array containing 96 genes involved in apoptotic pathways. Biotin-labeled cDNA probe was synthesized from each sample and hybridized with two independent blots. Scatter-plots were used to compare the relative expression pattern of genes in the rGd-treated versus untreated cells. Genes, which gave fold-change values beyond the preassigned ± 1.3-fold boundary value, were considered as candidate genes. (Upper panel) As determined from the microarray software analyses of multiple gene spots, the relative fold-change in expression patterns after 1 h and 6 h of rGd treatment is shown. (Lower panel) The expression level of target genes was further analyzed using real-time qRT-PCR. These data were replicated in two independent experiments.
|
U937 cell apoptosis appears to be mediated by an inhibition of NF-B
The activation of immune cell apoptosis via the extrinsic pathway, which is mediated by ligation of cell death receptors, often involves translocation of the transcription factor NF-B. To determine whether NF-B activation was affected by rGd, U937 cells were transfected with the NF-B response element–luciferase vector. As shown in Figure 8
, rGd inhibited PDB, and TNF- induced activation of NF-B by 48% and 31%, respectively (P<0.05).
 View larger version (9K): [in a new window] |
Figure 8. U937 cells were transfected with a 125-bp NF-B response element–luciferase vector and treated overnight in the absence or presence of 5 µg/ml rGd, 10 nM PDB, or 0.6 nM TNF-. The experiment was performed in triplicate. These data were replicated in two independent experiments. RLU, Relative light units; *, significant decrease (P<0.05).
|
 View larger version (43K): [in a new window] |
Figure 9. (A) Lysates were prepared from U937 cell cultures incubated without (–) or with (+) 5 µg/ml rGd for 6 h. Western blotting showed activated caspase-8 fragments of 43 kDa and 18 kDa, which were faint in control cultures. Treatment with 5 µg/ml actinomycin D, our positive control, strongly induced the activated fragments. Total protein (50 µg) was loaded in each lane, and actin staining confirmed equal loading of the gels. (B) U937 cells were incubated in the absence (–) or presence (+) of 5 µg/ml rGd for 6 h. The top band (46 kDa) reflects procaspase-2, whereas the middle and bottom bands (38 kDa and 37 kDa, respecitvely) are the cleaved intermediates of the caspase-2 enzyme. (C) U937 cells were incubated in the absence (–) or presence (+) of 5 µg/ml rGd for 6 h. Rabbit anti-human, cleaved caspase-3 antibodies showed a dominant 19-kDa band and a minor higher molecular weight fragment not observed in control cells. Total protein (50 µg) was loaded in each lane, and equal actin signals were noted. (D) Primary peripheral human monocytes also were exposed to 5 µg/ml rGd for 6 h, and their lysates were probed with rabbit anti-human, cleaved caspase-3 antibodies. An activated, 19-kDa fragment was detected in the rGd-treated (+) but not the control (–) cells. Total protein (50 µg) was loaded in each lane, and actin staining confirmed equal loading of the gels. These data were replicated in three independent experiments.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To better understand the immunomodulatory function of this glycoprotein, Gd cDNA was cloned and expressed in P. pastoris, and the secreted protein was purified to homogeneity and evaluated in human monocyte cultures. The recombinant, yeast-derived protein harbored the same antiproliferative and proapoptotic activities as those of GdA purified from human decidual tissue. We also demonstrated—by two different methods, binding to a lectin and digestion with a specific glycosidase—that the recombinant protein is glycosylated. Despite the fact that proteins expressed in P. pastoris are not sialylated [24 ], the yeast-derived rGd was able to induce apoptosis in U937 cells and primary monocytes. It had been suggested that the difference in immune cell apoptogenic activity between GdA and GdS was a result of sialylation [25 , 26 ]. However, our results [14 ] are consistent with those of Jayachandaran et al. [27 ], using rGd expressed in Sf21 cells devoid of any glycosylation, and indicate that in our models, neither sialylation nor glycosylation appears to be absolutely required for apoptogenic activity. Nevertheless, the different behaviors of GdA and GdS may be explained in part by differences in their sugar moieties. It is well known that glycans play an important role in maintaining the stability, conformation, and biological activities of glycoproteins [28 , 29 ], and high-sensitivity scanning calorimetry, which tests the thermodynamic parameters of reversible unfolding of GdA and GdS at a physiological pH, did show subtle differences [30 ].
Programmed cell death can occur in a variety of cell types in the endometrium and plays a key role in the extensive remodeling observed in this dynamic tissue [31 ]. Among these, the innate immune cell lineages that populate the endometrium have become a focus of increasing attention. Uterine epithelial and stromal cells are rich sources of chemokines, and the chemoattraction of leukocytes into premenstrual, endometriotic, and carcinomatous endometrium appears to play an important role in the angiogeneisis and repair of these tissues [32 33 34 ]. Our findings indicate that GdA and rGd can induce apoptosis in the monocytic cell line U937 as well as in primary human monocytes. Given the effects of trophoblast products on GdA synthesis, the physiological function of this glycoprotein within the endometrium may be localized to areas in close proximity to the implanting embryo.
Apoptosis is initiated by a pathway following ligation and multimerization of cell surface death receptors [35 ] or alternatively, via the mitochondria-dependent pathway [36 ]. Recent evidence supports caspase-8 as an apical or initiator caspase in apoptosis. Once activated, it transduces its proteolytic signal to distal effector caspases, such as caspase-3, and the latter enzyme subsequently degrades nucleocytoplasmic proteins including actin, lamin, and poly-ADP-ribose polymerase characteristic of apoptosis [37 ]. In U937 cells, caspase-8 activation appears to be an early and critical event, as overexpression of dominant-negative caspase-8 protein dramatically reduced the apoptotic effect of the kinase inhibitor sorafenib in this cell line [38 ].
Our data suggest that Gd-induced apoptosis occurs via the TNFR1 signaling pathway, as caspase-8, -2, and -3 and NF-B all signal downstream of TNF-R1. The recruitment of Fas-associated death domain to TNF-R-associated death domain (TRADD) permits the activation of caspase-8, which can then activate the executioner caspase-3, and the association of receptor-interacting protein (RIP) and TRADD is reported to induce apoptosis via caspase-2, which is recruited to the RIP-associated Ich-1-homologous death domain [39
, 40
]. Members of the Bcl family, such as the antiapoptotic Bcl-2, Bcl-xL, and Bcl-2A1 and the proapoptotic Bad and Bax, also play important roles in the regulation of mitochondrial stability [41
], and caspase-2 has been implicated in the release of cytochrome c from mitochondria. Transgenic expression of rGd in human endometrial carcinoma cells inhibited expression of the antiapoptotic Bcl-xL gene and induced cellular differentiation [42
]. Our data in U937 cells show that rGd can regulate the expression of Bcl protein family members Bad, Bax, and Bcl-2A1 as well as TNFSF genes, TNF-R1 and APRIL. Moreover, it appears that the induction of these genes is mediated via an inhibition of NF-B.
APRIL is expressed on dendritic cells, monocytes, and macrophages and plays a role in T cell survival and the proliferation of certain tumor cell lines [43
, 44
]. APRIL-mediated signaling involves the TNFR family members (e.g., B cell-activating factor of the TNF family or TNFSF 13b; B cell maturation antigen or TNFSF 17), leading to an activation of NF-B [44
]. The addition of soluble APRIL proteins has been shown to protect lymphocytes from spontaneous and drug-induced apoptosis via NF-B activation [45
]. Further, the activation of NF-B increased the expression of prosurvival Bcl-2 and Bcl-xL and inhibited Bax [46
]. Bcl-2A1 gene expression also is regulated by NF-B [47
], and Pagliari et al. [48
] demonstrated that NF-B activation was essential for monocyte survival, as the inhibition of NF-B led to a decrease in Bcl-2A1 expression and initiated apoptosis. Bcl-2A1 also was shown to protect cells from TNF-induced apoptosis [47
, 49
]. We postulate that rGd induces apoptosis in U937 cells by decreasing APRIL and Bcl-2A1 expression, as shown in Figure 7
and reducing NF-B activation (Fig. 8)
. Procaspase-8 is activated, and ultimately, procaspase-3 is cleaved to its functional 19 kDa isoform (Fig. 9)
, whereby nucleosomal digestion and programmed cell death are effected. Future experiments are designed, using human endometrial-monocyte cocultures, to test if endogenous GdA production can induce monocyte apoptosis via paracrine or juxtacrine effects.
Our model of focal immunosuppression at the human embryonic-maternal interface incorporates the up-regulation of GdA by corpus luteum-derived progesterone and relaxin, as well as local hCG emanating from trophoblasts of the implanting conceptus. We propose that these three pregnancy hormones induce GdA regionally, which subsequently initiates a proapoptotic cascade (manifested by reduced NF-B, APRIL, and Bcl-2A1 and increased TNF-R1, Bad, and Bax), and are executed by caspase-8, -2, and -3 in resident and infiltrating endometrial monocytes. Should such a pathway be verified in clinical studies, it would suggest a variety of targets that might be manipulated therapeutically to enhance embryonic implantation or conversely, lead to new modes of fertility regulation.
Received April 25, 2006; revised December 20, 2007; accepted December 21, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
expression in human endometrial adenocarcinoma cells J. Soc. Gynecol. Investig. 12,370-375[CrossRef][Medline] 2,6-sialyltransferase gene in Pichia pastoris J. Biotechnol. 81,55-61[CrossRef][Medline]-induced cell death Expert Rev. Mol. Med. 4,1-15[Medline]B that blocks TNF-induced apoptosis Genes Dev. 13,382-387B activation to maintain A1 expression and mitochondrial homeostasis Mol. Cell. Biol. 20,8855-8865This article has been cited by other articles:
 |
|
 |
|
  G. Weiss, L. T. Goldsmith, R. N. Taylor, D. Bellet, and H. S. Taylor Inflammation in Reproductive Disorders Reproductive Sciences, February 1, 2009; 16(2): 216 - 229. [Abstract] [PDF]
|
|
|
|
 |
|
 M Seppala, H Koistinen, R Koistinen, L Hautala, P C Chiu, and W S Yeung Glycodelin in reproductive endocrinology and hormone-related cancer Eur. J. Endocrinol., February 1, 2009; 160(2): 121 - 133. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
