| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Department of Animal Sciences,4 University of Missouri, Columbia Missouri 65211
Department of Animal Sciences,5 University of Florida, Gainesville, Florida 32611
ABSTRACT
Interferon tau (IFNT) is secreted by the trophectoderm of ruminant conceptuses during the peri-implantation period and serves an anti-luteolytic function. The question as to whether IFNT is superior as an anti-luteolytic agent to closely related Type I IFNs, such as IFN alpha (IFNA), which have a different function, remains unanswered. Thus, the aim of this study was to determine whether equivalent antiviral (AV) units of ovIFNA and ovIFNT are equipotent in extending estrous cycle length. Four distinct ovIFNA mRNA (ovIFNA1–4) were cloned from ovine lymphocytes. Recombinant ovine IFNs (ovIFNT4 and ovIFNA1) were prepared in the yeast Pichia pastoris. The AV activity of the purified IFNs was determined on a bovine cell line (MDBK) and on transformed ovine luminal uterine epithelial cells. Indwelling uterine catheters were fitted into crossbred ewes on Day 3 postestrus (Day 0 = estrus). Between Days 10 and 18 postestrus, ewes received twice-daily infusions of 0.7 x 107 IU of either ovIFNA1 or T4, plus serum albumin. Control ewes received serum albumin only. Daily blood samples were collected for progesterone determination, and ewes were monitored twice daily for estrus. Both ovIFNA (P = 0.04) and ovIFNT (P = 0.01) caused estrous cycle extension in nonpregnant ewes compared to controls when administered at equivalent AV doses. In conclusion, the uniqueness of IFNT as an anti-luteolytic agent most likely resides in its unique expression pattern rather than its special biopotency.
conceptus, corpus luteum, endometrium, estrous cycle, pregnancy, progesterone, trophoblast
INTRODUCTION
The structural evolution of the placenta likely governs the way in which the early conceptus signals its presence to the mother, and an extensive range of structural and biochemical mechanisms have developed for fetal-maternal communication across species [1–3]. In ruminant artiodactyls (sheep, cattle, goats, deer, giraffe), where there is minimal trophoblast invasion of the maternal endometrium, this communication uniquely involves the production of a Type I interferon, termed interferon-tau (IFNT) [2, 4]. IFNT is produced by the mononucleate trophectoderm cells of the developing conceptus beginning at the blastocyst stage coincident with the generation of trophoblast cells as trophectoderm. Its level of production is greatly enhanced via transcriptional regulation as the conceptus elongates before uterine attachment and is reduced dramatically as definitive attachment of trophectoderm to the uterine epithelium is established (Day 21 or 25 of pregnancy in the sheep and cow, respectively) [2, 4].
By ensuring CL maintenance and the continued production of progesterone beyond the length of a normal estrous cycle, the conceptus is provided with a nurturing uterine environment in which to develop [5]. In the absence of IFNT, the CL regresses in response to the pulsatile release of luteolytic agent prostaglandin F2 alpha (PTGF2A), which is produced by the uterine endometrium [6, 7]. IFNT is considered to prevent luteolysis by blocking the upregulation in oxytocin receptors during diestrus [8, 9] through the activation and repression of genes responsive to Type I IFN [10–12].
Type I IFN genes are found in all vertebrate species but have evolved over the more than 300 million years into several subfamilies, many of which themselves contain multiple genes [13]. The main function of the Type I IFN is generally considered to be defense against pathogen infection, most notably viruses, although these cytokines are pleiotropic in their effects and also target the immune system. All Type I IFNs bind to a common receptor [14] through which they activate STAT transcription factors [15] and induce several common genes [10–12, 16–18]. The IFNT genes (denoted as IFNT), however, are unusual in many respects. They are not virally inducible and lack a well-organized viral response element in their upstream gene regulatory regions. Moreover, they evolved quite recently and are expressed only in emerging trophoblast of ruminant conceptuses and have clearly evolved a role as pregnancy hormones [2, 13]. Despite these unique properties, however, they have retained potent antiviral (AV), immunomodulatory, and antiproliferative activity, properties that raise two interesting and yet incompletely answered questions. Do these IFN exhibit unique biological activities relative to the other Type I IFN, including features that cannot be explained by differing affinities for the common Type I receptor [19]? Are the IFNT superior as anti-luteolytic agents to other Type I IFNs [8, 20], despite having sequence similarity and functional homology to the related IFN omega (IFNW) and IFNA [13, 21]?
Protein products derived from different IFNT from the same species exhibit a range of AV and anti-luteolytic activities [21, 22]. In general, however, these two properties are found to be positively correlated [22]. In other words, those proteins with the highest AV activities are the most effective anti-luteolytic agents. Since all the IFNT that have been used in such comparisons are produced as recombinant products, usually from Escherichia coli or yeast, there remains the concern that activities vary not because the native proteins have intrinsically different properties, but because refolding and other aspects of their processing might complicate comparisons.
In this paper, we have addressed the question of comparative anti-luteolytic activity of different IFN subtypes more rigorously than before. First, rather than relying on intramuscular injection of the different IFN, which requires large doses, causes distressing side effects, and usually provides only modest estrous cycle extension [23–28], we have chosen to use the intrauterine route of administration. Since these experiments are much more easily carried out in ewes than in cows, we have also chosen to make our comparisons between different ovine IFN subtypes, a choice that necessitated the isolation of the ovine IFNA genes and the production of recombinant protein. We have also prepared the proteins in yeast to minimize folding artifacts, although full renaturation and stability still cannot be assured. Finally, and most importantly, we have made our comparisons not only on the basis of the mass of the proteins but also on their AV activities at the time they were administered to the ewes.
MATERIALS AND METHODS
Isolation and Characterization of Recombinant ovIFNA and T
Amplification and isolation of ovIFNA cDNA Blood (50–70 ml) was collected from mature Dorset rams (n = 4) by jugular venipuncture using heparinized vacutainers (Becton Dickinson). Blood was layered on top of an equal volume of Histopaque-1077 (Sigma Chemical Co.) and centrifuged (400 x g for 30 min at 25°C). The opaque interface containing mononucleated cells was isolated, diluted with 0.01 M phosphate-buffered saline (PBS; pH 7.2), and centrifuged (250 x g for 10 min at 25°C). The resulting cell pellets were washed twice with PBS and centrifugation (250 x g for 10 min at 25°C). Total cellular RNA was extracted by using the Trizol reagent according to the manufacturer's instructions (Invitrogen Corp.). RNA preparations were treated with RNase-free DNase (Promega Corp.) for 45 min at 37°C and then incubated at 75°C for 15 min to inactivate the enzyme. One microgram of total cellular (tc) RNA was incubated at 65°C for 5 min, then reverse-transcribed with SuperScript II reverse transcriptase (Invitrogen), oligo (dT) primer, and 250 µM dNTP at 42°C for 50 min. A total of 50 PCR cycles (94°C for 30 sec, 55°C for 30 sec, 74°C for 45 sec) were performed with ThermalAce DNA Polymerase (Invitrogen). This proofreading DNA polymerase was added at the beginning of PCR (2 U) and again after 25 cycles were completed (2 U) to minimize the loss of proofreading activity during PCR. Primers used in the reactions corresponded to degenerative sequences that border the coding sequences for boIFNA sequences (GenBank accession no. M10952) (5' primer = TGCCACCTGCCTCACWCCCAC; 3' primer = GTCCTTTCTCCTGAAWCTCTCC; W = A or T). Non-reverse-transcribed tcRNA was included as a control for genomic DNA contamination for each RNA preparation.
PCR products were ligated into the pCR4 Blunt-TOPO vector (Invitrogen) and used to transform TOP10 Competent E. coli (Invitrogen). Cells were plated on Luria Broth plates containing kanamycin (50 µg/ml). Bacterial colonies were picked and propagated in 4 ml Luria Broth containing kanamycin at 37°C overnight. Plasmid was isolated by using the Quantum Prep Plasmid Miniprep Kit (Bio-Rad Laboratories, Inc.). Presence of cDNA inserts was verified by restriction digestion with EcoR1. A total of 40 clones (10 clones/ram) were sequenced in both directions with vector primers (Davis Sequencing).
Amplification and isolation of ovIFNT4 cDNA The ovIFNT used, ovIFNT4 (GenBank accession no. X56341), was originally cloned from an ovine conceptus cDNA library into a pGEX-2T vector in E. coli [21]. The ovIFNT4 coding region, minus its signal peptide, was then subcloned into pGEM-T before template production.
Production of Recombinant ovIFNs
One of the ovIFNA clones identified (ovIFNA1; GenBank accession no. AY802984) and the previously characterized ovIFNT isoform (ovIFNT4; GenBank accession no. X56341) were used as templates for recombinant protein production. The coding sequences for ovIFNA1 and ovIFNT4, minus their signal peptide regions, were amplified with oligonucleotides matching a pPICZ alpha A vector (Invitrogen) by using Pfu (Stratagene). Amplified products were purified on spin columns (WizKit purification, Promega). The vector and PCR products were digested with Xho1 and ethanol precipitated (facilitated by the addition of Genelute LPA; Sigma) and ligated in Rapid Ligation Buffer (Promega). The ligase was subsequently heat inactivated, and the resulting constructs were digested with EcoR1 and Xba1. Restriction sites were converted to blunt ends by addition of a mixture of dNTP and Klenow fragment. The blunt-end products were purified on agarose gels. Purified fragments were circularized and used to transform DH5 alpha Competent E. coli (Invitrogen). Cells were plated on Luria Broth plates containing kanamycin (50 µg/ml). Bacterial colonies were picked and propagated in 4 ml Luria Broth containing kanamycin at 37°C overnight. Plasmids were isolated by using a Wizard Plus SV minipreps DNA Purification System (Promega), and the presence of cDNA inserts was verified by DNA sequencing.
Plasmids were digested with Sac1, and the product was purified on Wizard Plus SV minipreps columns as described previously. Yeast (Pichia pastoris) strain SMD1168H (Invitrogen) and culture media (yeast extract peptone dextrose medium [YPD], YPD with sorbitol [YPDS], and Buffered Media with Glycerol for Yeast [BMGY]) were prepared according to the manufacturer's protocol. One microgram of plasmid was combined with the prepared yeast in a final volume of 100 µl and the yeast transformed by electroporation at 8 kV/cm. Cells were allowed 2 h to recover in 1 M sorbitol before being diluted with YPD broth (Teknova Inc.) and incubated at 30°C overnight. Transformations (0.5 ml) were spread on YPDS plates containing zeocin (100 µg/ml) and incubated at 30°C for 3 days. Colonies were picked, placed into 24-well plates, and propagated in BMGY containing zeocin (100 µg/ml) and chloramphenicol (20 µg/ml) overnight. Induced protein production was accomplished over 72 h by feeding methanol as outlined for Pichia pastoris (Invitrogen). The resulting culture medium was tested for AV activity (see the following discussion) and the most productive clone used for larger scale fermentation.
Large-scale fermentation to produce the recombinant proteins was undertaken as detailed by the distributor's protocol (Invitrogen) in medium adapted from Brady et al. [29]. Briefly, increasing volumes of BMGY media containing zeocin (100 µg/ml) and chloramphenicol (20 µg/ml) were inoculated with the respective clones and cultured at 30°C for 2 days. This inoculated medium was added in to a 10-L fermentor (New Brunswick Scientific) containing 4 L of 0.25x sterile fermentation basal salts (Invitrogen) modified with 4% (w:v) glycerol, 2% (w:v) tryptone. When this mixture cooled, sterile filtered 0.01 M sodium hexametaphosphate, PTM1 (4 ml/L; Invitrogen), 500x biotin (2 ml/L; Invitrogen), Antifoam 289 (0.25 ml/L; Sigma), histidine 0.1% (w:v), and chloramphenicol (20 µg/L) were added, and the pH was adjusted to 6. Every 24 h, the culture was supplemented with histidine (1 g 100 g–1 L–1 fresh weight). The culture was subsequently fed at regular intervals from 14 to 26 h with 3:1 (v:v) methanol:glycerol (11 ml) with PTM1 (4 ml). The feed rate was increased to 7:1 and added every hour between 26 and 45 h postinoculation, after which the feed was altered to methanol with PMSF (15 mM) every hour. In addition, PTM1 (3 ml) was added every 5 h until the end of culture. At 50 h, feeding frequency was further increased to every 45 min and again at 68 h to 30-min intervals, although the PMSF content of the supplement after 68 h was reduced to 7.5 mM. Once the cell density reached 330 g/L (fresh weight) at approximately 80 h, the culture was terminated, and the yeast cells were removed by centrifugation (7600 x g for 30 min).
Purification of Recombinant ovIFNs
The resulting supernatant (4.5 L) was adjusted to 20 mM Tris, pH 8, 0.1 mM PMSF, 20 mM EDTA, and 0.02% sodium azide and passed through a glass-fiber filter (Pall Corp.). The filtrate was then dialyzed against 28 volumes of 20 mM Tris pH 8, 1 mM EDTA, and 0.02% sodium azide overnight and a further 17 volumes for 7 h. The dialyzed solution was run through a DEAE 52 cellulose column (Whatman) in a batch process and the column washed with buffer until a stable baseline absorption at 280 nM was achieved. Proteins were eluted from the column with a gradient (4 L) of 0–200 mM NaCl in 20 mM Tris-HCl buffer pH 8, containing 2 mM EDTA and 0.02% sodium azide. Fractions containing protein were collected, and the presence of IFN was verified by SDS-PAGE under reducing conditions. Fractions with the highest concentrations of IFN were dialyzed against 10 mM Tris pH 8, 2 mM EDTA, and 0.02% sodium azide to remove NaCl before a second (FPLC) anion exchange chromatography step on a DEAE Sepharose CL-6B (Bio-Rad Laboratories). Fractions with the highest concentrations of IFN were concentrated by dialysis against polyethylene glycol (MW 14–18 kDa; Sigma). The concentrate was then dialyzed against 0.1 M PBS, passed through a DetoxiGel column (10 ml, Pierce Biotechnology Inc.) to remove endotoxins and filter-sterilized through a 0.22-µM filter (Millex-GV; Millipore Corp.). Samples were subjected to MALDI-TOF MS to verify the exact molecular weights of the proteins (http://www.biotech.missouri.edu/dnacore/).
AV and Protein Content of the IFN
AV activity of purified filtrate was determined primarily via cytopathic reduction assays involving Madin-Darby Bovine Kidney (MDBK) cells challenged with a vesicular stomatitis virus [30]. The standard used was recombinant boIFNT1A of known AV activity (7.5 x 107 IU/mg) [31] that had been standardized against human IFNA (PBL Biomedical Laboratories). Results were confirmed by the use of ovine uterine epithelial cells (kindly donated by Drs. G.A. Johnson and T.E. Spencer, Texas A&M University) by following essentially the same protocol. Total protein concentration of purified filtrate was measured [32] with bovine serum albumin as a standard.
Intrauterine Infusion of Recombinant ovIFNs
Introduction of IFN into the uterine lumen of ewes has been described previously [21, 33, 34] and was completed in accordance with University of Missouri Animal Care and Use Committee Protocol 2745. On Day 10 postestrus, ewes (n = 16) were assigned by breed and age into groups to receive twice-daily infusions (1.5 ml, 12 h apart). Rectal temperatures were taken at 0, 1, 3, 6, and 12 h relative to first infusion in all ewes to identify any initial hyperthermic response to the infusions. Protein solutions for group 1 (ovIFNT4; n = 6) and group 2 (ovIFNA1; n = 6) were prepared in 0.01 M PBS (pH 7.2) and contained identical amounts of AV activity (0.7 x 107 IU, determined on MDBK cells) and protein (100 µg adjusted with ovine albumin). Ewes in groups 1 and 2 received twice-daily infusions from Day 10 until Day 18 postestrus, while group 3 ewes (control; n = 4) were given 0.01 M PBS (pH 7.2) containing only ovine albumin (100 µg) twice daily from Day 10 until returning to estrus. Catheters were flushed after each infusion with 1.0 ml 0.01 M PBS containing 0.1% (w:v) ovine albumin, 50 IU/ml penicillin, and 25 µg/ml streptomycin pH 7.2. All ewes were monitored twice daily for estrus behavior with vasectomized rams from Day 13 until they returned to estrus. Daily blood samples (10 ml) were collected via jugular venipuncture from initial estrus (Day 0) until a return to estrus and stored at –20°C until used to determine plasma progesterone concentrations. Once ewes had returned to estrus, uterine cannulae were removed and the animals returned to the University flock.
Progesterone Assay
Plasma progesterone concentrations were determined via the use of a Progesterone Coat-A-Count kit (DPC) following the manufacturer's instructions. Assay sensitivity (n = 3) was 0.01 ng/ml and the specific binding 51.3%. The intra- and interassay coefficients of variation were 3.3% and 6.8%, respectively.
Statistical Analysis
Estrous cycle length was defined in days from synchronized estrus to the return to estrus, and data were transformed by nonparametric ranking. Progesterone concentrations were compared by using repeated measures, and all data, including the number of CL per ewe and Day 10 temperatures, were analyzed by the PROC MIXED procedure of SAS (SAS Institute).
RESULTS
Isolation and Characterization of Recombinant ovIFNA1 and T4
Identification of ovIFNA cDNA Each of the RT-PCR reactions contained a single band of ~500 bp (data not shown; expected size = 498 bp). No amplified products were detected from reaction mixtures that had not been reverse transcribed, indicating that the amplified products were derived from reverse-transcribed mRNA and not from genomic DNA contaminants. Compiled sequences were compared with each other and with existing bovine and ovine sequences in GenBank. Eleven different variants were identified through this screening procedure. However, seven of the 11 novel sequences were identified only within a single RT-PCR reaction. Since it was unclear whether nucleotide mutations in these sequences were real or had resulted from errors that occurred during either reverse transcription or PCR amplification [21], these clones were discarded. Four of the 11 sequences were identified in two or more different RT-PCR reactions and were considered to represent either different ovIFNA genes or possibly polymorphic forms of the same genes (ovIFNA isoforms A1–A4, GenBank accession nos. AY802984–AY802987).
Each of the ovIFNA cDNA sequences is novel, and the open reading frames encode distinct proteins (Fig. 1). Like most other IFNA, they are 166 amino acids (aa) in length, unlike the IFNT, which are invariably 172 aa long. Nucleotide sequences differ by 0.4%–2.0% (2–10 nucleotide substitutions) and, when translated, yield protein isoforms that are 95.7%–98.7% identical in amino acid sequence (two to seven distinct amino acid residue changes). As expected, the ovIFNA sequences are more similar to bovine (bo) IFNA than to porcine (po) IFNA (Fig. 2). None of the ovIFNA isoforms, unlike poIFNA, possesses a potential site for N-glycosylation at position 78. The four ovIFNA sequences share approximately 89% identity with bovine IFNA, which has been used in some previous estrous cycle extension studies [21], but less than 50% identity with ovIFNT4, in the regions of sequence overlap (Fig. 2). As expected on phylogenetic grounds, the ovIFNA sequences are more similar to bovine (bo) IFNA than to porcine (po) IFNA (Fig. 2). Curiously, ovIFNT4 resembles poIFNA more closely than ovIFNA. Together, these data confirm the considerable phylogenetic distance of the IFNA from the IFNT.
|
|
Purification and MALDI-TOF MS Analysis
After final passage of the recombinant IFN through the DetoxiGel column, the preparations were largely free of contaminating proteins, and each had an apparent molecular weight (Mr) of ~17000 when analyzed on a 12% polyacrylamide gel (Fig. 3).
|
When subjected to MALDI-TOF MS, ovIFNA1 had an actual molecular mass of 19452 Da, which is greater than the predicted value of 18890 Da, a difference of 562 Da. Although the ovIFNA1 lacks a site for N-glycosylation, this extra mass is assumed to be attributable to some covalent modification, possibly O-linked glycosylation acquired during biosynthesis. The molecular mass of the ovIFNT4, which possesses six more amino acids than ovIFNA1, was determined at 19957 Da, a value slightly greater than the predicted molecular mass of 19940 Da. This discrepancy is within the margin of error of the technology.
AV Activities
The specific AV activity of ovIFNT4 as measured on MDBK cells was approximately 6.8 x 108 IU–1 mg protein–1, which is comparable to that noted for ovIFNT4 produced as a recombinant product from E. coli [21, 33, 34]. The specific activity of ovIFNA1 (4.0 x 107 IU–1 mg protein–1) was less than that of ovIFNT4 but was comparable to that of boIFNA1 used as a standard. Essentially identical AV activities were obtained on ovine epithelial cells as on MDBK cells (see Table 1). Since the ovIFNT4 was 16.6-fold more active than the ovIFNA1, much less had to be used to provide the required identical amounts of AV activity in the preparation introduced into the uteri of the ewes.
|
Intrauterine Infusion of Recombinant ovIFNs
Two ewes were excluded from the ovIFNA1 group because they exhibited short estrous cycles (9–12 days), determined on the basis of a premature fall in progesterone (below 0.5 ng/ml). All four control ewes returned to estrus at the expected time around Day 16 after previous estrus (Fig. 4). Cycle length was significantly longer (21.7 ± 4.0; P = 0.04) in the four ewes treated with ovIFNA1, although the extension was modest, and one ewe came into estrus at the expected time. OvIFNT4 significantly extended cycle length in a group of six ewes (33.3 ± 13.9; P = 0.01) when compared to the controls (Fig. 4), although again one of these ewes had a cycle of normal length. The difference in cycle length between A1- and T4-treated ewes was not significant, however (P > 0.1). Nevertheless, Figure 4 demonstrates that the magnitude of responses differed considerably within treatments from no response (Day 16) to partial response (around Day 21) or a long extension (>Day 40), especially in ewes infused with T4. This variability is consistent with the findings of numerous studies from this [21, 33, 34] and other laboratories [20, 24, 35, 36]. The number of CL per ewe did not differ between treatments. Progesterone concentrations, including the early rise observed between Day 0 and Day 6, and the pattern observed throughout the estrous cycle were not significantly (P > 0.1) different among the three groups of ewes (Fig. 5). Furthermore, no significant (P > 0.1) increases in ewe body temperatures were observed during the treatments.
|
|
DISCUSSION
The identification of four novel ovIFNA mRNA species in ovine lymphocytes was not unexpected given that multiple IFNA exist in every mammalian species examined to date. Several duplication events followed by rapid gene diversification are a feature of Type I IFN. In the human, for example, there are 13 expressed IFNA whose expression varies according to the inducing stimulus and to the cell type examined [37]. Several human IFNA have AV activities on MDBK cells that are similar to, or greater than, that of the ovIFNT4 examined here. The diversification of the Type I IFN and the nucleotide divergence of their gene promoters are most likely a consequence of attempting to adapt to the ever-evolving threat of viral pathogenesis. What was somewhat surprising about the present study was the relatively low AV activity of the ovIFNA1 on both MDBK and uterine epithelial cells when compared to ovIFNT4. While the latter are generally considered to have a role as a reproductive hormone and not to have a major role in viral pathogenesis, the IFNA, like most other Type I IFN, are considered to function mainly as AV agents [37, 38]. The differences in AV activity between ovIFNA1 and T4 were observed on ovine as well as bovine cells and were unlikely, therefore, to be due to differences in cross-species reactivity. On the other hand, there are components to an AV response other than the dilution at which an IFN can evoke resistance to viral attack. Only a single strain of a single virus was examined in this study, and it is conceivable that outcomes would differ if other viruses were tested. That being said, another explanation for the low activity of the ovIFNA1 is that the recombinant form folded less than optimally. The fact that the ovIFNA1 had an unexpectedly high molecular weight indicated that it had been posttranslationally modified by the yeast, whereas ovIFNT4 had not. Despite this concern, the AV activity of the ovIFNA1 was not much different than those of several other IFNT forms that had been tested for their abilities to extend estrous cycle length in ewes [21, 22, 34]. In addition, comparable AV activities of both ovIFNA1 and ovIFNT4 prepared in the same manner to the present study have enhanced MHC class I expression in ovine endometrial cells in vitro [39].
Since it had relatively low AV activity, more ovIFNA1 protein was used in the infusion studies. Nevertheless, on an AV basis, this protein was as capable of causing an extension of estrous cycle length as ovIFNT4. Three out of four ewes had extended cycles with ovIFNA1, while five out of six showed extensions with ovIFNT4. Since the AV activity of an IFN is related to its ability to bind to the Type I IFN receptor and activate downstream signal transduction events, this relationship between anti-luteolytic and AV activity is perhaps not surprising. In an unpublished experiment, recombinant ovIFNW (ovine IFN-omega) and ovIFNT4, both prepared in E. coli and used at comparable AV activities to those employed here, were able to extend estrous cycle lengths when injected into the uterine lumen of ewes from Day 11 to Day 17 postestrus (A.D. Ealy, J.A. Green, and R.M. Roberts, unpublished observations). Unfortunately, the laboratory was unable to generate sufficient quantities of recombinant ovIFNW in either bacterial or yeast hosts to include in the current study. A study in cattle, however, has shown that boIFNW can successfully extend estrous cycle length [28]. Together, these data imply that any Type I IFN could accomplish the task of initially extending the life span of the ovine CL provided that the protein used has AV activity on ovine cells and is administered to the ewes above a minimum AV threshold. They support the contention that IFNT are unique not because of their unusual ability to cause estrous cycle extension but because they are produced in the right place at the right time and in sufficient quantities to cause the mother to respond. Since no other Type I IFN is produced simultaneously by trophectoderm for those few critical early days of pregnancy, the uniqueness of IFNT must lie in the temporal expression of their genes.
The data illustrated in Figure 4 show that considerable variability exists in the response of individual ewes to both ovIFNA1 and T4 infusion. This variability in response to both intramuscular and intrauterine injection of IFN has been commented on several times previously [20, 21, 34, 35] and has not been adequately explained. Examination of data from these and the current study indicates that ewe responsiveness can be classified into three main types: nonresponders (with luteolysis and no extension in cycle length), modest responders (those that prevent luteolysis but fail to extend past approximately Day 21), and good responders (those that prevent luteolysis and maintain pseudopregnancy way beyond Day 21). The ewes that failed to respond at all to IFN (e.g., one in each treated group in the present study) would possibly not have extended their cycles even if successfully bred during that particular cycle. The remaining treated ewes exhibited either modest or good responses. For example, three IFNA ewes and two IFNT-treated ewes showed modest responses, while no IFNA ewes but three IFNT-treated ewes showed good extensions (Fig. 4).
One explanation for the variability among ewes to treatment is that IFN administration causes a two-phase response. While the primary role of trophoblast IFN in pregnancy is the initial rescue of the CL, which is accomplished with relatively high efficiency, there is a second phase that is initiated at around Day 21 that ensures a longer cycle extension. The variability observed in this phase, which is ewe based, may reflect the fact that infusions of IFN are generally continued only to about Day 18 or 20 and then halted. Conceivably, not all ewes have responded fully by then. IFNT production by the conceptus does not cease this early and may continue past Day 21 [40, 41] and possibly longer [36]. Such an explanation for modest responders is supported by an experiment in which IFNT was administered for two periods (Days 11–15 and Days 21–25). This treatment regimen resulted in, on average, a greater cycle extension in animals that had received a second IFNT infusion (32 versus 55 days) [36]. Of course, IFNT is not the only hormone produced by the trophoblast and important in fetal-maternal communication. In addition, migration and fusion of the binucleate cells with the luminal epithelial around this time may also have a role in supporting the pregnancy to term or at least to the point where placental-derived progesterone is produced in sufficient quantities to maintain the pregnancy (approximately Days 45–60 in sheep) [42].
An alternative explanation for varied extensions may rely not on the timing of IFN administration but on the susceptibility to and occurrence of negative influences that override the effects elicited by the decreasing levels of IFNT around this period. For example, estradiol secreted by follicular waves, emerging every 4–5 days [43], therefore peaking around Day 21, coincides with the return to estrus of modest responders. Previous studies have demonstrated that ovine endometrial estrogen and oxytocin receptor genes are regulated following IFNT administration [8] and that the concentration of the latter increases [44] because of estradiol of follicular origin during the luteal phase. Interestingly, follicular dynamics have not been monitored throughout any IFN administration study.
Although it remains an attractive possibility that IFNT has some unique bioactivity as a pregnancy hormone that distinguishes it from other Type I IFN, the present experiment was unable to resolve this dilemma. What does seem certain is that the initial ability of any Type I IFN to extend the estrous cycle for a few days will be correlated with the AV activity of the IFN used. The more potent the AV activity, the less protein will be needed to achieve a positive outcome. Whether there is a unique second phase of IFN action during pregnancy for which IFNT is especially equipped remains to be proven.
ACKNOWLEDGMENTS
The authors thank Ms. C.A. Gibson and Ms. O.M. Ocón (Pennsylvania State University) for their technical assistance with cloning ovine IFNA isoforms and Drs. G.A. Johnson and T.E. Spencer (Texas A&M University) for kindly donating the ovine epithelia cells. The authors particularly thank Dr. C.N. Murphy and Dr. J.A. Green for conducting the majority of the surgeries and Mr. J. Bader, Mr. N. Carroz, Mrs. T. Lamprecht, and Miss G. Vawter, as well as personnel in Dr. R.M. Roberts's laboratory, for providing assistance with surgical procedures and the animal husbandry.
FOOTNOTES
1 Supported by a grant from NIH grant HD21896 (to R.M.R.) and the Molecular Biology Program, Life Sciences, University of Missouri (salary support for M.P.G.). 
2 Correspondence: R.M. Roberts, 105F Life Sciences Center, 1201 East Rollins Rd., Columbia, MO 65211-7310. FAX: 573 884 9395; robertsrm{at}missouri.edu
3 Current address: The Liggins Institute, Faculty of Medicine and Health Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand.
Received: 9 May 2005.
First decision: 13 June 2005.
Accepted: 20 July 2005.
REFERENCES
and interferon-
interact with the same receptors in bovine endometrium: use of a readily iodinatable form of recombinant interferon- for binding-studies. J Biol Chem 1994 269:13544-13550 with mutations near the carboxyl terminus on luteal life span in sheep. Biol Reprod 1997 56:214-220[Abstract]-interferon (trophoblastin) in sheep. J Endocrinol 1990 127:R5-R8This article has been cited by other articles:
|
|
 |
|
  Y. Chen, J. A. Green, E. Antoniou, A. D. Ealy, N. Mathialagan, A. M. Walker, M. P. Avalle, C. S. Rosenfeld, L. B. Hearne, and R. M. Roberts Effect of Interferon-{tau} Administration on Endometrium of Nonpregnant Ewes: A Comparison with Pregnant Ewes Endocrinology, May 1, 2006; 147(5): 2127 - 2137. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
