Tumor necrosis factor alpha inhibits transcriptional activity of the porcine P-45011A insulin-like growth factor response element.

We investigated the effects of tumor necrosis factor α (TNFα) on the transcriptional activity of the porcine P-45011A (P450scc) insulin-like growth factor response element (IGFRE). TNFα inhibited insulin-like growth factor-I (IGF-I)-stimulated P450scc mRNA concentrations in cultures of porcine granulosa cells. Transient transfection experiments in granulosa cells with deletion P450scc/luciferase constructs showed that TNFα inhibited the transcriptional activity of the IGFRE. IGF-I binding and IGF-I receptor mRNA concentrations in porcine granulosa cells were not inhibited by TNFα. Electrophoretic mobility shift assay with nuclear extract protein from porcine granulosa cells treated with IGF-I and TNFα showed that Sp1 and a second transcription factor, P2, bound to the IGFRE. While IGF-I treatment increased the binding activity of both factors, TNFα specifically inhibited the IGF-I-stimulated binding activity of P2. Transient transfection studies done in mouse fibroblasts overexpressing the IGF-I receptor (NWTb3) with the porcine IGFRE (three repeats) in an SV40/luciferase construct also showed TNFα inhibited IGF-I-stimulated reporter gene expression. We conclude that TNFα inhibits the transcriptional activity of the porcine P450scc IGFRE by preventing IGF-I-stimulated binding of P2.

We investigated the effects of tumor necrosis factor ␣ (TNF␣) on the transcriptional activity of the porcine P-45011A (P450scc) insulin-like growth factor response element (IGFRE). TNF␣ inhibited insulin-like growth factor-I (IGF-I)-stimulated P450scc mRNA concentrations in cultures of porcine granulosa cells. Transient transfection experiments in granulosa cells with deletion P450scc/luciferase constructs showed that TNF␣ inhibited the transcriptional activity of the IGFRE. IGF-I binding and IGF-I receptor mRNA concentrations in porcine granulosa cells were not inhibited by TNF␣. Electrophoretic mobility shift assay with nuclear extract protein from porcine granulosa cells treated with IGF-I and TNF␣ showed that Sp1 and a second transcription factor, P2, bound to the IGFRE. While IGF-I treatment increased the binding activity of both factors, TNF␣ specifically inhibited the IGF-I-stimulated binding activity of P2. Transient transfection studies done in mouse fibroblasts overexpressing the IGF-I receptor (NWTb3) with the porcine IGFRE (three repeats) in an SV40/luciferase construct also showed TNF␣ inhibited IGF-I-stimulated reporter gene expression. We conclude that TNF␣ inhibits the transcriptional activity of the porcine P450scc IGFRE by preventing IGF-I-stimulated binding of P2.
Insulin-like growth factor I (IGF-I) is a growth factor that also regulates steroidogenesis in the ovary. In porcine granulosa cells, IGF-I increases P450scc mRNA concentrations (4). Isolation of the porcine P450scc gene and transient transfection studies in porcine granulosa cells identified a 30-base pair IGF-responsive region (IGFRE) in the gene (5). The porcine P450scc IGFRE is a GC-rich domain that binds Sp1 and an-other transcription factor, P2 (6).
This study determined that TNF␣ inhibits the function of the porcine P450scc IGFRE in porcine granulosa by preventing IGF-I-stimulated binding of P2. This finding presents a mechanism whereby TNF␣ can induce luteolysis by inhibiting IGF-I-supported steroidogenesis in the CL. Moreover, it establishes P2 as the transcription factor mediating the IGF-I response. Reporter Gene Constructs-The reporter gene constructs Ϫ100, Ϫ130, and Ϫ2320 P450scc/luc have been described (5). Briefly, these constructs contain 5Ј deletions of the upstream region of porcine P450scc, the core porcine P450scc promoter, and the entire coding region of the firefly luciferase gene with a polyadenylation tract (10). rWT pSVPLUC contains three repeats of the porcine IGFRE cloned into pSVPLUC, a modified pGEM3 plasmid containing the luciferase gene described above, and the enhancerless SV40 early region promoter (10). The plasmids were obtained from Dr. Allan Brasier, University of Texas Medical Branch, Galveston, TX.

Materials
Porcine Granulosa Cell Culture, RNA Isolation, and P450scc cDNA Hybridization-Granulosa cells were isolated from 1-5-mm follicles of ovaries from immature swine (60 -70 kg). The ovaries were collected from a local slaughterhouse. The granulosa cells were plated in Eagle's minimum essential medium and 3% fetal calf serum for 12-16 h to facilitate granulosa cell attachment to the tissue culture plates as described previously (11). After granulosa cell attachment, all culture conditions were done in serum-free medium for experiments that measured mRNA concentrations.
At the time of cell harvesting, medium for measurement of progesterone concentrations and cells for DNA content were collected for each condition as described previously (11). Total cellular RNA was prepared by the method of Chirgwin et al. (12), and 15 g was used to make Northern blots for hybridization to a P450scc cDNA clone. Membranes were hybridized with 50 ng of P450scc cDNA clone radioactively labeled by random priming with [␣-32 P]dCTP to a specific activity of 1 ϫ 10 8 cpm/g (11). After washing, filters were exposed to film as described previously (4).
Progesterone and DNA Assay-Progesterone concentrations in media were measured by radioimmunoassay after celite microcolumn chromatography as described previously (11). Progesterone antiserum used in the assay was rabbit-produced using progesterone-11-succinate/bovine serum albumin as described (13). All samples from each experiment were assayed in a single assay. Total cellular DNA was measured by fluorometric assay using Hoechst 33258 dye (14). Calf thymus DNA was used as standard. The assay has a sensitivity of 20 ng/tube and was linear to 400 ng/tube.
Densitometry-The 18 S ribosomal RNA band from photographs of ethidium-stained total RNA formaldehyde gels and hybridization bands from autoradiograms of corresponding membranes were measured for integrated optical density with a BVI 4,000 digital analysis system (Applied Imaging, Santa Clara, CA). The 18 S ribosomal band was used to correct for differences in RNA loading for Northern blots as described previously (11).

IGF-I Binding Assay and Receptor mRNA Concentrations-
The IGF-I binding assay was done as described previously (11). Briefly, cells were cultured at a concentrations of 20 -30 million cells/dish and received 0.3 ng/ml 125 I-IGF to give 200,00 cpm/ dish. Competition with unlabeled IGF-I was done at the following concentrations: 0.3, 1.0, 3.0, 10.0, and 30.0 ng/ml. Granulosa cells were maintained in serum-free medium (control) or treated with TNF␣ (30 ng/ml) for 48 h prior to binding assays. Type I IGF receptor mRNA concentrations were determined by Northern blot hybridization using a porcine riboprobe as described previously (11).
Transient Transfection in Porcine Granulosa Cells-Porcine granulosa cells were cultured in 60-mm culture dishes at a concentration of 3 ϫ 10 7 cells/dish (5). At the time of transfection, cells received 30 g of P450scc/luc construct (divided among three 60-mm culture plates) by the calcium phosphate precipitation technique (5,15). After 24 h, the precipitate was removed and fresh medium was added with specific hormonal treatments. Cultures were maintained for designated treatment times, harvested, and measured for light production (5). Reporter gene activity for porcine transfection experiments was normalized by the measurement of protein concentrations from the supernatant of the samples using Bio-Rad Bradford protein assay kit.
Transient Transfection in Mouse Fibroblast Cells-The two variants (NWTb3, KR1) of mouse fibroblast (NIH-3T3) cells were cultured in Dulbecco's modified Eagle's medium ϩ 10% fetal bovine serum and 500 g/ml Geneticin ® (Life Technologies, Inc.). Transient transfection was done by lipofection (Tfx-50 Reagent, Promega, Madison, WI). Transfection experiments were done on 60-mm plates as per the Promega protocol for Tfx-50 Reagent. A 3 to 1 ratio (1 g of DNA/2.5 l of Tfx-50) was used for each transfection. The control plasmid pSV2Apap (containing the SV40 early promoter and the human placental alkaline phosphatase gene; Ref. 16) was cotransfected with the chimeric construct of interest. After transfection, cells were maintained in 2.5% fetal bovine serum without Geneticin ® . Cells were harvested and measured for luminescence as described previously (5). The remaining lysate was measured for alkaline phosphatase activity using p-nitrophenyl phosphate (Sigma) and measuring absorbance at 405 nm.
Statistical Methods-One-way analysis of variance (ANOVA) with Tukey multiple comparison test was used to determine differences in experiments. Probability values of Յ0.05 were considered statistically significant. Data are presented as mean Ϯ S.E.

Effects of TNF␣ on IGF-I-stimulated P450scc mRNA Concentrations and Progesterone
Production-TNF␣ inhibits insulinstimulated P450scc mRNA concentrations and progesterone production in porcine granulosa cells (2), but the effects of TNF␣ on IGF-I-stimulated P450scc mRNA concentrations and progesterone production had not been tested. Porcine granulosa cells were treated for 48 h with IGF-I (20 nM), TNF␣ (30 ng/ml), and IGF-I and TNF␣. TNF␣ inhibited IGF-I-stimulated P450scc mRNA concentrations and progesterone production (Fig. 1).
TNF␣ Effects on the Porcine P450scc IGFRE in Porcine Granulosa Cells-Transient transfections of deletion constructs of the upstream region of the porcine P450scc gene were done in porcine granulosa cells during treatment with TNF␣ to determine whether TNF␣ inhibited the activity of the porcine IGFRE. Experiments were done with three deletion constructs of porcine P450scc as follows: 1) the sequenced upstream region including the IGFRE (Ϫ2320), 2) the IGFRE and porcine P450scc core promoter (Ϫ130), and 3) the core promoter only (Ϫ100). These constructs have been described (5). Transfected cells were treated with IGF-I, TNF␣, or both for 48 h. TNF␣ inhibited the transcriptional activity of the porcine IGFRE, but had no effect on the core porcine P450scc promoter (Fig. 2).
Effects of TNF␣ on IGF-I Binding and Receptor mRNA Concentrations-TNF␣ could inhibit IGF-I-stimulated porcine P450scc mRNA concentrations by decreasing IGF-I binding or reducing IGF type I receptors. IGF-I binding assays performed in porcine granulosa cells after treatment with TNF␣ for 48 h found that TNF␣ had no effect on IGF-I binding (Fig. 3). Moreover, Northern blot hybridization with a riboprobe of the porcine P450scc IGF type I receptor (11) showed no decrease in receptor mRNA concentrations in porcine granulosa cells treated with TNF␣ for 48 h (Fig. 3).
Effects of TNF␣ on the Binding Activity of IGFRE Transcription Factors-EMSA was done with nuclear extract proteins from porcine granulosa cells treated with IGF-I or TNF␣ for 48 h to determine whether the binding affinity of the IGFRE proteins were changed with treatment. Supershift assay with an Sp1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was also done to confirm that Sp1 bound to the IGFRE. As shown in Fig. 4, IGF-I increased the binding activity of both transcription factors to the IGFRE, but TNF␣ specifically prevented IGF-I-induced binding of P2. The binding activity of Sp1 was increased with IGF-I and TNF␣ treatment over IGF-I treatment alone (Fig. 4). The Sp1 antibody did not completely supershift the upper band, as seen previously (6).

Effects of TNF␣ on the Function of the Porcine P450scc IG-FRE in Mouse Fibroblasts-Mouse fibroblast (NIH-3T3) cells
were also used to assess TNF␣ effects on the porcine IGFRE. This cell model was used to show reproducibility of the granulosa cells results and also allowed the use of a control plasmid to normalize transfection experiments. Two variations of the NIH-3T3 cells were used in these studies. NWTb3 cells overexpress the IGF-I receptor and have been described previously (8). KR1 cells express an inactive IGF-I receptor from a mutation to the lysine residue (9). TNF␣ produces cytotoxic effects on NIH-3T3 cells, but this effect is minimal at a concentration of 1 ng/ml (17). Therefore, we first transfected the rWTpSV-PLUC (three repeats of the IGFRE and an SV40 promoter) construct in KR1 cells and treated the cells with IGF-I (20 nM) and three doses of TNF␣ (1, 10, and 30 ng/ml). As shown in Fig.  5, IGF-I treatment did not induce an increase in reporter gene activity and TNF␣ at a concentration of 1 ng/ml did not significantly inhibit reporter gene expression from control cells. Therefore, this experiment showed that a functional IGF-I receptor was necessary for activation of the porcine IGFRE and established a concentration of TNF␣ that was not toxic to the mouse fibroblasts. The rWTpSVPLUC construct was next transfected into NWTb3 cells and the cells treated with IGF-I (20 nM), TNF␣ (1 ng/ml), and both together. IGF-I treatment induced a significant stimulation of the porcine IGFRE, and this effect was inhibited by TNF␣ (Fig. 6). The pSVPLUC plasmid alone was not stimulated with IGF-I treatment (data not shown). The significant decrease in reporter gene expression from control by TNF␣ treatment alone is caused by the stimulation of the porcine IGFRE by the IGF-I present in the 3% fetal bovine serum necessary for cell culture. DISCUSSION This study found that TNF␣ inhibited IGF-I-stimulated P450scc mRNA concentrations and progesterone production in porcine granulosa cells by suppressing the transcriptional activity of the porcine P450scc IGFRE. This inhibition was caused by a reduction in the IGF-I-stimulated binding of the transcription factor designated P2. We showed in a previous study (6) that the binding of both Sp1 and P2 in MCF-7 cells was necessary for transcriptional activity of the porcine IG-FRE. These results extend this finding by showing that the binding of P2 mediates the cell-specific effects of TNF␣ on the porcine IGFRE.
From the results of this study and the previous study in MCF-7 cells (6), we can develop the mechanism of how the porcine IGFRE is able to stimulate gene expression. The IG-FRE requires the binding of Sp1 for basal function of the element. However, while necessary, this binding is not specific. This is evident from the EMSA results during treatment with TNF␣ and IGF-I. Despite a reduction in P450scc gene expression (decreased mRNA concentrations), the binding of Sp1 to the IGFRE was increased over the binding that occurred with IGF-I treatment alone. It is the binding of P2 to the IGFRE that correlates with gene expression of the IGFRE. Therefore, P2 is mediating the cell-specific actions of IGF-I on gene expression.
The increase in the binding activity of Sp1 and P2 to the porcine P450scc IGFRE during IGF-I treatment is consistent with results found in MCF-7 cells (6). However, these findings in porcine granulosa cells were not found using EMSA in a previous study by our group (5). In the previous study, the EMSA was done using a 7% polyacrylamide gel and samon sperm as carrier DNA (5). The current method (4% polyacrylamide gel and 0.5 g of poly(dI-dC)⅐poly(dC-dI)) results in DNA-protein complexes that can be supershifted with an Sp1 antibody (this could not be done with the 7% gel). Therefore, the refinement in methods is the most likely explanation for the differences in our findings.
We were also unable to completely supershift the upper band with the Sp1 antibody. This is similar to our findings in MCF-7 nuclear extract protein (6), but the remaining band is more evident in porcine granulosa cells. This may be because a human antibody is being used in porcine cells. A mutant oligonucleotide to the IGFRE that binds only P2 showed no binding of the upper band (6).
The transfection experiments done in mouse fibroblasts are significant because they show that the findings in porcine granulosa cells can be reproduced in a different cell. This is made even more significant in that with the mouse fibroblast cell lines, a control plasmid can be co-transfected and used to normalize results. Porcine granulosa cells transfection results (due to a lack of expression of most viral promoters) were normalized with protein concentrations.
Another significant finding in this study was the TNF␣ signaling pathway interacting with the IGF-I pathway in porcine granulosa cells. There is evidence in other cells of interactions between the pathways. TNF␣ inhibits IGF-I-stimulated proteoglycan synthesis in cartilage from hypophysectomize rats (18). In human obesity, TNF␣ expression from adipose tissue can be correlated with the level of hyperinsulinemia (19). Studies indicate that TNF␣ inhibits insulin receptor signaling by causing a modified form of IRS-1 that suppresses rather than enhances the signaling pathway (20). However, other studies have indicated that TNF␣ modulates insulin receptor signaling by protein-tyrosine phosphatase activation (21). Mutations to the IGF-I receptor show that the initial mechanisms of activation of the insulin and IGF-I receptor are almost identical (22). Therefore, understanding the mechanisms of TNF␣ inhibition of IGF-I-stimulated transcriptional activity of the porcine IG-FRE may further our understanding of obesity and insulin resistance.
These findings in porcine granulosa cells must also be viewed from their physiologic significance regarding corpus luteum function. Corpus luteum that are regressing have an increase in resident ovarian macrophages and TNF␣ concentrations (23). The site of synthesis of TNF␣ in the CL has not been absolutely determined, but, in addition to resident macrophages, the endothelial cells of the CL have been implicated as a source of TNF␣ (24). In the bovine, an increase in intraluteal TNF␣ is related to luteolysis and the peak activity of luteal TNF␣ occurs before the decline in progesterone production (25). IGF-I is also important in the regulation of CL function. The porcine CL expresses mRNA concentrations for IGF-I and IGFbinding proteins (26). In women, elevated concentrations of insulin-like growth factor-binding protein-1 (IGFBP-1) occur in luteinizing follicles, further suggesting the importance of the IGF-I autocrine/paracrine system in regulation of CL function (27). This study presents a plausible mechanism for the interaction of TNF␣ and IGF-I on CL function. These experiments were done in granulosa cells, so these results must be extrapolated to the small luteal cell. Increasing concentrations of TNF␣ from either machrophages or epithelial cells in CL would inhibit IGF-I-supported progesterone production. This inhibition of steroidogenesis would be the initial step in the regression of the CL.
In summary, this study shows that TNF␣ inhibits the transcriptional activity of the porcine P450scc IGFRE in porcine granulosa and mouse fibroblasts cells. This finding supports a plausible mechanism for TNF␣ induction of CL luteolysis and further defines P2 as the transcription factor that mediates IGF-I-stimulated gene expression. The control plasmid pSV2Apap was co-transfected for normalization. Cells were treated for 48 h after transfection with IGF-I (20 nM) and TNF␣ (1 ng/ml). Arbitrary units, luminescence of the lysate after treatment divided by absorbance (alkaline phosphatase assay). The data represent the mean Ϯ S.E. from five replicates. The single asterisk indicates a statistical significance increase, and the double asterisk indicates a significant decrease from control as determined by ANOVA.