NF-κB Induces cAMP-response Element-binding Protein Gene Transcription in Sertoli Cells*

Spermatogenesis is dependent upon Sertoli cells, which relay hormonal signals and provide factors required for the differentiation and proliferation of germ cells. NF-κB transcription factors are constitutively expressed in the nuclei of Sertoli cells in rodent testis. Electrophoretic mobility shift assays demonstrated that Sertoli NF-κB proteins specifically bind to κB enhancer motifs within the promoter of the cAMP-response element-binding protein (CREB) gene, an important mediator of hormonal signals that control spermatogenesis. Overexpression of NF-κB proteins in primary Sertoli and NIH 3T3 fibroblast cells induced the CREB promoter in transient transfection assays. Stimulation of Sertoli cells with tumor necrosis factor-α, an NF-κB-activating cytokine produced by round spermatids located adjacent to Sertoli cells, stimulated the elimination of IκB, the translocation of additional NF-κB to the nucleus, and increased NF-κB binding to CREB promoter κB enhancer elements. Tumor necrosis factor-α also stimulated transcription from the CREB promoter. These data demonstrate that NF-κB contributes to the up-regulation of CREB expression in Sertoli cells and raises the possibility that NF-κB may induce other Sertoli genes required for spermatogenesis. Furthermore, the CREB promoter is also inducible by NF-κB in NIH 3T3 cells suggesting that NF-κB may be a general regulator of CREB in non-testis tissues.

Spermatogenesis is a multistep process by which spermatogonial germ cells differentiate into mature spermatozoa within the seminiferous tubules of the mammalian testis. In addition to germ cells at various developmental stages, the seminiferous tubules contain peritubular myoid cells, which line the outer wall of the tubule, and somatic Sertoli cells, which provide elements essential for germ cell maturation in response to endocrine and paracrine factors. The primary hormonal inputs regulating spermatogenesis are follicle stimulating hormone (FSH) 1 and luteinizing hormone (1,2). Luteinizing hormone stimulates testicular Leydig cells to secrete testosterone, which diffuses into and acts upon Sertoli cells. FSH binding to Sertoli cells results in the elevation of cAMP levels and activation of protein kinase A, which can phosphorylate a number of proteins including the cAMP-response element-binding protein (CREB) transcription factor. Phosphorylation of CREB allows the induction of genes containing a cAMP-responsive element (CRE) (3).
The modulation of CREB expression levels represents a potential mechanism to alter Sertoli cell responsiveness to FSH. CREB expression in Sertoli cells has been demonstrated to vary in a stage-specific manner during the spermatogenesis cycle, (4 -6). FSH-induced changes in cAMP levels have been implicated in the cyclical control of Sertoli cell CREB expression through CREs within the CREB promoter (7,8). Additional signaling pathways may also control stage-specific expression of CREB. Specifically, NF-B transcription factors that were recently identified as activating gene expression in Sertoli cells (9) are potential regulators of CREB expression as the CREB promoter contains consensus NF-B binding motifs.
The family of NF-B or Rel transcription factors consists of five known mammalian subunits (RelA, RelB, c-Rel, p50, and p52). Multiple combinations of homo-and heterodimers are possible, thus providing the potential to generate both transactivators and transrepressors of transcription (10 -13). In most cells NF-B dimers remain sequestered in the cytoplasm by inhibitor proteins (IB-␣, IB-␤, IB-␥, IB-⑀, and IB-␦). Upon stimulation by diverse stimuli such as TNF-␣, phorbol myristic acid, viral proteins, and interleukins, IB is phosphorylated and ubiquinated leading to proteosome-mediated degradation. The NF-B nuclear localization signal is then unmasked, and NF-B is free to translocate to the nucleus and regulate gene expression via interactions with B enhancer elements (reviewed in Ref. 14). In addition to being regulated via stimulus-induced release from IB, the activity of free NF-B can also be modulated through direct phosphorylation of the RelA subunit by protein kinase A (15,16). The cytokine TNF-␣ is a regulator of NF-B activity (reviewed in Ref. 11) and therefore is a candidate regulator of CREB gene expression. In the testis TNF-␣ is secreted primarily by round spermatids within the seminiferous tubules, and the 55-kDa TNF-␣ receptor has been detected in Sertoli cells (17,18). Recently, we demonstrated that TNF-␣ increases the activity of NF-B in rat Sertoli cells and that NF-B levels in the nuclei of Sertoli cells are highest during the stages in which round spermatids are present (9).
In this study we test the hypothesis that NF-B and TNF-␣ are regulators of CREB expression. We demonstrate that Sertoli cell NF-B proteins interact with NF-B binding sites in the CREB promoter. We also show that overexpression of NF-B subunits in Sertoli cells and NIH 3T3 cells increases CREB promoter activity. Stimulation of primary Sertoli cells with the cytokine TNF-␣ mediates a reduction in IB-␣ and IB-␤ levels, a concomitant increase in RelA nuclear translocation and the induction of NF-B binding to a CREB promoter NF-B enhancer motif. Transient transfection analyses dem-onstrate that TNF-␣ also stimulates CREB gene promoter activity. These data suggest that NF-B may be an important regulator of genes required for spermatogenesis and a general regulator of CREB gene expression in non-testis cells.

EXPERIMENTAL PROCEDURES
Isolation of Primary Sertoli Cells and Cell Culture-Sertoli cells were isolated from 16 day Harlan Sprague Dawley rats as described previously (7). Decapsulated testes were digested with collagenase (0.5 mg/ ml, 37°C, 12 min) in enriched Krebs-Ringer bicarbonate medium (19), followed by three washes in enriched Krebs-Ringer bicarbonate medium (1 ϫ g, 3 min) to isolate seminiferous tubules. Tubules were digested with trypsin (0.5 mg/ml, 37°C, 12 min), and cell aggregates were passed repeatedly through a Pasteur pipette. An equal volume of Dulbecco's modified Eagle's medium containing 10% fetal bovine serum was added to the Sertoli cells, which were then pelleted (500 ϫ g, 5 min) and resuspended in serum-free media containing 50% Dulbecco's modified Eagle's medium, 50% Ham's F-12, 5 mg/ml insulin, 5 mg/ml transferrin, 10 Ϫ6 M retinoic acid, 10 ng/ml epidermal growth factor, 3 mg/ml cytosine ␤-D-arabinofuranosidase, 1 mM sodium pyruvate, 100 units/ml penicillin, and 100 mg/ml streptomycin. Sertoli cells were cultured on matrigel (Collaborative Research, Bedford, MA) coated dishes (33°C, 5% CO 2 ). Sertoli cells were routinely Ͼ95% pure as determined by phase microscopy and alkaline phosphatase staining (20). In some cases cells were also cultured in the presence of TNF-␣ (20 ng/ml). NIH 3T3 fibroblast cells and human embryonal kidney HEK 293 cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum, 100 units/ml penicillin, and 100 mg/ml streptomycin. Animals used in these studies were maintained and sacrificed according to the principles and procedures described in the NIH Guide for the Care and Use of Laboratory Animals.
Protein Extract Preparation-Nuclear and cytoplasmic extracts were prepared by detergent lysis (21). Cells were incubated in Buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and a protease mixture consisting of 0.5 ng/ml pepstatin A, and 5 ng/ml each of leupeptin, antipain, soybean trypsin inhibitor, and aprotinin) for 15 min on ice followed by the addition of 0.06% Nonidet P40. Cells were vortexed for 10 s, and nuclei were collected by centrifugation (12,000 ϫ g, 30 s). The supernatant containing cytoplasmic proteins was removed and frozen in 20% glycerol. To prepare nuclear extracts, pelleted nuclei were washed once with Buffer A and resuspended in Buffer C (20 mM HEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 20% glycerol, and the protease mixture used for Buffer A). Cells were mixed vigorously on a shaking platform (4°C, 15 min). The cell debris was pelleted (5 min 12,000 ϫ g), and the supernatant containing nuclear proteins was frozen. Protein concentrations were determined by the Bradford method (Bio-Rad protein assay, Bio-Rad).
DNA-Protein Binding Studies-32 P-radiolabeled DNA probes were generated by annealing nucleotide templates containing CREB B enhancer elements plus promoter flanking sequences to complementary 10-base primers. The overhangs were filled in with Klenow in the presence of [␣-32 P]dATP and 5 mM dCTP, dGTP, and dTTP. The template sequences are as follows (B sequences are underlined): CREBB1 (5Ј-C-GACACCCCTCGGGAATTCCCCCACTGGGCC-3Ј), CREBB2 (5Ј-AGG-CCTGAGCGGGGGTTTCCACCAAGTCGCC-3Ј), CREBB3 (5Ј-CGTCC-CCACGGGGGTCCCACGACGCCCC-3Ј), CREBB4 (5Ј-CAGCCACGGA-AGTCCCCCTGCTGGGTT-3Ј). Electrophoretic mobility shift assays (EMSA) were performed as described (22). Briefly, 32 P-labeled B probes were incubated with 1-10 g of nuclear or cytoplasmic extracts from untreated 16 day rat primary Sertoli cell cultures or from those treated with TNF-␣ (20 ng/ml). Binding reactions were incubated 15 min at room temperature in the presence of 1 g of poly(dI-dC), 5 g bovine serum albumin, 5 mM dithiothreitol, 50 -100 mM NaCl or KCl, 20 mM HEPES, and 1 mM EDTA. For competition EMSAs, a 20-fold excess of double-stranded unlabeled competitor oligonucleotides containing a consensus B site (5Ј-CGGCAGGGGAATTCCCCTCTC-3Ј), an Sp1 binding site (5Ј-GCTGCCTGTGGCCCGGGCGGCTGGGAGAAGCGG-3Ј), or a CRE motif (5Ј-GATCCGGCTGACGTCATCAAGCTAGATC-3Ј) were co-incubated with labeled B probe and nuclear extracts. Protein-DNA complexes were resolved via 5% polyacrylamide gel electrophoresis (PAGE) under nondenaturing conditions in a Tris borate/EDTA buffer. In binding reactions involving detergent treatment to release NF-B proteins from IB, proteins were preincubated with 0.5% deoxycholate for 10 min followed by the addition of 1% Nonidet P40 (Nonidet P-40) immediately prior to the addition of the radiolabeled probe. For protein binding comparisons to various B motifs, labeled probes with similar specific activities (Ϯ20%) were used in simultaneous reactions. DNA-protein complex formation was quantified using NIH Image 1.6 software analysis of digitized images.
Western Immunoblotting and Immunocytochemistry-Sertoli cell nuclear and cytoplasmic extracts (50 g) were resolved by denaturing discontinuous SDS-PAGE followed by electroblotting to a Polyscreen polyvinylidene difluoride transfer membrane (NEN Life Science Products). Membranes were incubated in blocking buffer (10% nonfat dry milk and 2% Tween in phosphate-buffered saline) for 30 min at room temperature followed by an overnight incubation at room temperature with primary antibodies in 50% radioimmune precipitation buffer and 50% wash buffer (1% nonfat dry milk and 2% Tween in PBS). Membranes were then probed with horseradish peroxidase-conjugated goat anti-rabbit secondary antibodies 4 h at room temperature in 50% radioimmune precipitation buffer and 50% wash buffer, and detection was performed using the Renaissance Western blot chemiluminescence reagent detection system (NEN Life Science Products). The RelA-and IB-␣-specific primary antibodies are polyclonal antiserum directed against the amino-terminal 21 amino acids of RelA or the aminoterminal 29 amino acids of IB-␣ (provided by Dean Ballard, Vanderbilt University, Nashville, TN). The IB-␤ antibody is a rabbit polyclonal directed against amino acids 339 -358 (Santa Cruz Biotechnology, provided by Stefan Dorre, Boston University, Boston, MA). NIH Image 1.6 software was used to quantitate the levels of immunodetected IB proteins from digitized images. Immunocytochemistry was performed on primary rat Sertoli cells cultured on glass coverslips and fixed with 4% paraformaldehyde. Cells were stained with preimmune sera or antisera directed against the amino-terminal 21 amino acids of RelA and Cy3 fluorescent secondary anti-rabbit antisera.
Plasmid Constructs, Transfections, and Luciferase Assays-For transient reporter transfections, the Ϫ1264 CREBLUC vector was constructed by inserting a Sau3A CREB promoter fragment from Ϫ1264 CREBCAT (7) into the BglII site directly upstream of the luciferase gene in the pGL2-Basic vector (Promega, Madison, WI) by blunt end ligation. This CREB promoter fragment included the entire region from Ϫ1264 to Ϫ51 bp upstream of the CREB translation start site. To generate the Ϫ537 CREBLUC plasmid an Asp718-StuI fragment containing sequences upstream of Ϫ537 bp was removed from the Ϫ1264 CREBLUC construct. The Ϫ278 CREBLUC was constructed by eliminating a XhoI fragment of Ϫ1264 CREBLUC containing CREB promoter sequences upstream of Ϫ278 bp. The following CREB promoter deletion constructs were also generated from the Ϫ1264 CREBLUC vector: ⌬B III-IV (a 480-bp PstI-RsrII deletion from Ϫ1160 to Ϫ680 bp), ⌬B I-II (a 285-bp RsrII-Esp1 deletion from Ϫ680 to Ϫ385 bp), and ⌬BI-IV (a 775-bp PstI-EspI deletion from Ϫ1160 to Ϫ385 bp). The Ϫ1264 CREBLUC I-IImt was created by mutating the CREB promoter B1 and B2 binding motifs according to the Chameleon doublestranded, site-directed mutagenesis kit instructions (Stratagene). The following oligonucleotides were utilized to introduce the indicated underlined point mutations: 5Ј-ACACCCCTCATGAGGGCCCCCACTGG-GCC-3Ј for the B1 site and 5Ј-GCCTGAGCGGCCGGGCCCACCAAG-TCGCC-3Ј for the B2 site. The NF-B expression vectors pCMV4p50 (23) and pCMV5p65 (24) contain cDNAs for NF-B p50 and p65, respectively, positioned downstream of the CMV promoter/enhancer in the pCMV4 and pCMV5 expression vectors (25,26).
Primary Sertoli cells were transfected by calcium phosphate coprecipitation as described (7) using 1 g of luciferase reporter plasmid and 1 g of empty pCMV expression vector or 1 g of pCMV expression vectors encoding p50 and p65. DNA precipitates were added to Sertoli cells in the presence of 2% fetal bovine serum, and serum was removed after the addition of DNA to the cells for 4 h. For studies employing TNF-␣, DNA was added to the cells in the presence of FuGENE reagent (Roche Molecular Biochemicals) and 2% serum for 24 h. After washing off the FuGENE reagent, the cells were maintained in 2% serum until harvesting 6 h later. NIH 3T3 cells and HEK 293 cells were transfected by the calcium phosphate coprecipitation method with the cells being maintained in 10% serum throughout the transfection experiment. Total cellular proteins were extracted in reporter lysis buffer (Promega), and luciferase assays were performed using a luminometer and the Promega luciferase assay reagent. Luciferase activity of the extracts was normalized for total protein as determined by Bradford assay.

RESULTS
Sertoli Cell Proteins Bind CREB NF-B Enhancer Elements-Four potential B enhancer elements within the CREB promoter (CREBB1-4) (Fig. 1A) were tested for interactions with Sertoli cell proteins using EMSA. A series of radiolabeled oligonucleotide probes having similar specific activities (Ϯ20%), each containing one of the CREB promoter B enhancer elements plus 22 bp of promoter flanking sequences, were used in EMSA studies. Following incubation of the probes with nuclear protein extracts prepared from rat primary Sertoli cell cultures, DNA-protein complexes exhibiting varying degrees of binding affinity were generated (Fig. 1B). In the CREB promoter, CREBB1 and CREBB2 containing probes were most effective at forming complexes with Sertoli cell nuclear extracts; however, lower levels of probe-protein interactions were also observed with the CREBB4 probe. Specifically, if CREBB1 binding activity was arbitrarily set to 100%, then the binding activity of CREBB2 and CREBB4 was 95 Ϯ 10.0% and 15.9 Ϯ 2.9%, respectively. No detectable binding was observed following incubation with the CREBB3 probe. All complexes observed for each probe represented specific protein interactions with the B enhancer elements as complex formation was effectively inhibited by a 20-fold excess of an unlabeled probe containing a consensus B binding site (B) but not by an excess of competitor oligonucleotides containing an Sp1 binding site or a CRE (Fig. 1C). These data suggest that the complexes formed are because of the specific binding of NF-B proteins to the probes.

NF-B Regulates CREB Promoter Activity in Sertoli Cells-
The hypothesis that CREB transcription is regulated by NF-B proteins in Sertoli cells was tested in transient transfection assays. Primary cultures of rat Sertoli cells were initially transfected with luciferase reporter plasmids driven by either the full-length CREB promoter extending from 51-1264 bp upstream of the translation start site (Ϫ1264 CREBLUC), a 5Ј-deletion mutant extending only 537 bp upstream of the translation start site (Ϫ537 CREBLUC), or a CREB promoter construct that included only the proximal 278 base pairs of the CREB promoter (Ϫ278 CREBLUC) (Fig. 2A). The Ϫ1264 CRE-BLUC construct contains all four B enhancer elements, whereas Ϫ537 CREBLUC contains only the downstream two B motifs, and Ϫ278 CREBLUC lacks all NF-B binding ele- ments. In agreement with earlier studies (7) basal activity was reduced nearly 50 and 75% for the Ϫ537 CREBLUC and Ϫ278 CREBLUC deletion mutants, respectively. Cotransfection of Sertoli cells with expression vectors encoding either p50 and RelA or RelA alone resulted in a 6-fold induction of Ϫ1264 CREBLUC and a 3-fold induction of Ϫ537 CREBLUC. The Ϫ278 CREBLUC construct was not induced by NF-B. The NF-B-mediated induction of the CREB promoter does not appear to be specific to Sertoli cells as similar results were obtained in transient transfections utilizing the NIH 3T3 cell line (Fig. 2B). Overexpression of RelA alone in NIH 3T3 cells induced Ϫ1264 CREBLUC and Ϫ537 CREBLUC, 14-and 8-fold, respectively. However, the combination of p50 and RelA was a less effective activator of the CREB promoter in NIH 3T3 cells as Ϫ1264 CREBLUC and Ϫ537 CREBLUC were induced only 3-and 2-fold, respectively.
To identify NF-B inducible regions of the CREB promoter, a series of CREB promoter deletion constructs were generated. Primary Sertoli cells were transfected with luciferase reporter vectors driven by various CREB promoter mutants in the presence or absence of RelA and p50 expression vectors (Fig. 3). Deletion of a region containing the two distal B enhancer elements (⌬BIII-IV) resulted in a 2-fold increase in basal promoter activity suggesting that the distal B enhancer motifs were not required for basal CREB promoter activity and/or negative regulatory elements may be present in the deleted region. Overexpression of RelA and p50 induced ⌬BIII-IV CREB promoter activity 3-fold over the elevated basal levels and slightly higher than the NF-B-induced wild type promoter thus providing further evidence that the distal B enhancer sequences were not essential for CREB promoter induction. In contrast, deletion of a fragment including the two proximal B enhancer elements (⌬BI-II) reduced basal activity by 50% and abolished NF-B-mediated induction of the CREB promoter. Introduction of 5 bp changes into each of the two proximal B enhancer motifs (BI-IImt) also reduced basal activity by 50% and reduced NF-B induction of the CREB promoter to 3-fold. Basal activity was dramatically reduced and induction by NF-B transcription factors was completely eliminated following removal of a fragment containing all four B enhancers (⌬BI-IV). The response of the CREB promoter mutants to overexpression of RelA or p50 and RelA in NIH 3T3 cells was very similar to that observed in primary Sertoli cells. However, one apparent difference was the further diminished response of BI-IImt to the overexpression of NF-B in NIH 3T3 cells. Together, the data in Fig. 3 suggest that basal and NF-B inducible promoter activity can be regulated through B motifs I-IV but that the region containing BI and BII is more responsive to NF-B.

TNF-␣ Induces Degradation of IB-␣ and IB-␤, Nuclear Translocation of RelA, and NF-B Binding to CREB Promoter B Motifs in Sertoli Cells-TNF-␣, a known activator of NF-B
in Sertoli cells (9), is secreted by the adjacent round spermatids (17). Stimulation of cells with TNF-␣ has been demonstrated to result in phosphorylation-dependent degradation of IB and subsequent translocation of NF-B from the cytoplasm to the nucleus. Therefore, the TNF-␣-mediated effects on IB family members in Sertoli cells was investigated. Because IB expression has not been previously characterized in Sertoli cells, studies focused on IB-␣ and IB-␤, as IB-␣ degradation is a prerequisite for NF-B translocation in most cells and high levels of IB-␤ mRNA have been detected in whole testis (27). Cytoplasmic extracts from untreated primary rat Sertoli cells or from cells treated with TNF-␣ for 30 min were subjected to Western blot analyses using antiserum directed against IB-␣ (Fig. 4A) or IB-␤ (Fig. 4B). TNF-␣ induced a reduction in the cytoplasmic levels of both IB-␣ (4.33 Ϯ 1.1-fold, n ϭ 3) and IB-␤ (3.4 Ϯ 0.1-fold, n ϭ 3).
The ability of TNF-␣ to induce nuclear translocation of NF-B in Sertoli cells was studied using indirect immunofluorescence and Western blot analyses. The RelA subunit was chosen for investigation because it is required to maximize NF-B transactivation activity (see Figs. 2 and 3). In western immunoblotting assays, a 30-min TNF-␣ (20 ng/ml) treatment of rat Sertoli cell cultures caused a significant increase in nuclear RelA levels that was accompanied by a reduction in cytoplasmic RelA (Fig. 5A). The TNF-␣-induced translocation of RelA was confirmed in immunofluorescence localization studies. These studies demonstrated that RelA was present in the nucleus of untreated Sertoli cells; however, a rapid and dramatic increase in nuclear RelA levels occurred following TNF-␣ treatment (Fig. 5B). To confirm that TNF-␣ treatment of Sertoli cells could increase the levels of NF-B available to bind to the CREB promoter, EMSA analyses were performed with a probe containing the CREBB1 binding site, which exhibited the most efficient binding of Sertoli cell NF-B proteins. Stimulation of primary rat Sertoli cells with TNF-␣ (20 ng/ml) for 30 min caused a dramatic induction of NF-B binding activity in nuclear extracts, whereas a loss of NF-B binding activity was observed in cytoplasmic extracts (Fig. 5C). Together, the data in Figs. 4 and 5 suggest that TNF-␣ stimulates rapid IB-␣ and IB-␤ degradation in Sertoli cells leading to the translocation of NF-B to the nucleus and subsequent binding to B enhancer elements.
TNF-␣ Stimulates the CREB Promoter-Because TNF-␣ efficiently stimulated NF-B activation in Sertoli cells, TNF-␣ induction of the NF-B-responsive CREB promoter was investigated. Transient transfections of primary Sertoli cells (Fig.  6A) and HEK 293 cells (Fig. 6B) were performed with the Ϫ1264 CREBLUC and ⌬B1-4 CREBLUC constructs. The Ϫ1264 CREBLUC construct was induced 2.2-fold in both cell lines following stimulation with TNF-␣ for 6 h. In contrast, the ⌬B1-4 CREBLUC construct was not inducible by TNF-␣. Interestingly, TNF-␣ induction of Ϫ1264 CREBLUC in Sertoli cells was only possible after employing a new, more efficient transfection protocol. Transfections in which Sertoli cells were incubated with FuGENE reagent (Roche Molecular Biochemicals) and plasmid DNA for 24 h in the presence of 2% fetal bovine serum followed by continued incubation with TNF-␣ in 2% serum resulted in significant induction of Ϫ1264 CRE-BLUC. Serum alone did not activate the CREB promoter and in the absence of serum, the CREB promoter was not stimulated by TNF-␣ (data not shown).

DISCUSSION
In this study, we have demonstrated that NF-B proteins present in Sertoli cells specifically bind B enhancer motifs within the CREB promoter. Four potential B enhancer elements were identified in the CREB promoter by computerassisted sequence analysis. The two gene proximal B motifs bound NF-B proteins more effectively than the distal consensus sequences. The binding of NF-B was functionally significant as overexpression of NF-B proteins in Sertoli cells stimulated transcription from the CREB promoter. Although NF-B is constitutively expressed in the nucleus of Sertoli cells, TNF-␣ was shown to induce the degradation of IB and further increase the levels of nuclear NF-B in Sertoli cells. Furthermore, TNF-␣ was found to stimulate CREB gene expression in primary Sertoli cells and the HEK293 cell line.
In transient transfection studies of primary Sertoli cells, the basal activity of the Ϫ1264CREB promoter was 2-and 4-fold higher than the activities of the Ϫ537CREB and Ϫ278CREB promoter fragments, respectively. The decrease in Ϫ537CREB promoter activity in Sertoli cells but not NIH 3T3 cells may reflect the loss of two distal B elements and the higher relative nuclear levels of NF-B in Sertoli cells compared with NIH 3T3 cells. The more dramatic decrease in activity for the Ϫ278CREB promoter may be because of the elimination of Sp1 motifs as described earlier (7) as well as the proximal B enhancer motifs. Although the relative basal activities of the Ϫ1264CREB, Ϫ537CREB and Ϫ278CREB promoter constructs in Sertoli cells were similar to that of an earlier report (7), it is difficult to make precise comparisons of the basal activity between various constructs as transfection efficiencies were not standardized using unregulated, control reporter constructs.
Nevertheless, NF-B proved to be a potent inducer of the CREB promoter as overexpression of NF-B p50 and RelA or RelA alone stimulated transcription from the full-length Ϫ1264CREB promoter six-fold. Although the distal B3 and B4 motifs may contribute to the induction of the CREB promoter, deletion analysis of the CREB promoter showed that the region containing the proximal B1 and B2 motifs is required to maintain high basal expression and full NF-B induction of the CREB promoter. The induction through B1 and B2 is not unexpected as these sequences most effectively bind NF-B present in Sertoli extracts. Deletion analysis of the Ϫ680 to Ϫ1160 upstream region of the CREB promoter also suggests that this region may contain negative elements as removal of this region results in a 2-fold increase in basal activity.
The demonstration of CREB promoter stimulation by NF-B in Sertoli cells suggested that NF-B may regulate this gene in other cell types. In this regard, we found that the CREB gene is up-regulated by NF-B overexpression in NIH 3T3 cells (Figs. 2B and 3B), and TNF-␣ induces the CREB promoter in HEK 293 cells (Fig. 6B). The regulation of CREB by NF-B may be a new method of cross-talk between NF-B and CREB signaling pathways to compliment the competition for the CREB-binding protein/p300 coactivator displayed by these factors (28,29). Alternatively, it is possible that CREB and NF-B may cooperate for recruiting CREB-binding protein/p300 to the promoter as CREB and RelA interact with different regions of the coactivator. The opportunity for CREB-NF-B interactions on the CREB promoter exists as the CREB promoter contains binding sites for CREB in close proximity to B motifs (7). RelA has been shown to directly interact with other bZIP family members related to CREB (ATF-2, c-Jun, and c-Fos) through a mini leucine zipper located in the Rel homology domain of RelA (30,31). Furthermore, in glutathione S-transferase-pulldown and co-immunoprecipitation experiments RelA has been shown to also directly interact with CREB. 2 Studies are underway to investigate potential cooperativity of NF-B and CREB in stimulating transcription from the CREB promoter.
In this initial characterization of IB proteins from pure cultures of rat Sertoli cells, both IB␣ and IB␤ were determined to be present in the cytoplasm, and the levels of both were dramatically reduced after the addition of TNF-␣. A previous study failed to detect IB␣ mRNA in extracts from whole mouse testis (27). The inability to detect Sertoli cell-derived IB␣ in whole testis is likely because of the small proportion of Sertoli cells in the mammalian testis. Fewer than 5% of adult mouse testis cells are Sertoli cells (19). In addition to the immunodetection of IB␤ in Sertoli cells in the present report, IB␤ mRNA was previously shown to be enriched in mouse testis (27). Together, these data suggest that whereas IB␣ and IB␤ are both present in Sertoli cells, the developing germ cells, which account for greater than 90% of testis cells, likely contain IB␤ but not IB␣. Differences in the levels of the two IB isoforms may be important in cell-specific gene regulation as stimulation of IB␣ degradation is rapid and transient, but IB␤ degradation can be delayed and persistent (27).
One factor that may maintain high nuclear levels of NF-B in Sertoli cells and account for additional NF-B translocation to the nucleus is the cytokine TNF-␣. In Sertoli cells, TNF-␣ activates NF-B via the elimination of IB from the cytoplasm and the subsequent translocation of additional NF-B to the nucleus. TNF-␣-mediated stimulation of NF-B in Sertoli cells may be physiologically important because germ cells adjacent to Sertoli cells secrete TNF-␣ in a stage-specific manner. Although it is possible that TNF-␣ may also act through mitogen-2 W. Walker, unpublished results.
FIG. 6. TNF-␣ stimulates the CREB promoter. Primary Sertoli cells were transfected with the Ϫ1264 CREBLUC or ⌬B1-4 CRE-BLUC reporter plasmids in the presence or absence of TNF-␣ (20 ng/ml). Luciferase activity was measured using a luminometer, and relative light units were adjusted for total protein levels and expressed as fold induction over the untreated activity for each plasmid. Data represent the mean Ϯ S.E. of three independent experiments in duplicate. activated protein kinase or other pathways, it is significant that NF-B is activated by TNF-␣ in Sertoli cells. In the absence of nuclear NF-B, apoptosis pathways may be initiated by TNF-␣ (32). In contrast to germ cell development in which some apoptosis is required to constrain the expansion of germ cells and maintain spermatogenesis (33,34), nuclear NF-B may protect Sertoli cells from apoptosis inducers. NF-B-mediated protection from apoptosis agents would explain why few or no apoptotic Sertoli cells are detected in the testis (32,35,36).
In a previous study TNF-␣ alone was able to stimulate transcription from a minimal promoter containing two consensus B enhancers (9). In contrast, TNF-␣ induction of the CREB promoter in primary Sertoli cells required the addition of serum. These latest findings suggest that serum factors in addition to TNF-␣ are required to allow the activation of the more complex CREB promoter or that serum-dependent signaling pathways must be activated to allow some Sertoli cell genes to be regulated by TNF-␣.
Although other germ cell types secrete some TNF-␣, most of the TNF-␣ is secreted by round spermatids (17). Because round spermatids are present during only the first eight stages (stages I-VIII) of the 14 stages of rat spermatogenesis, it is possible that TNF-␣ secretion may represent stage-specific communication between spermatids and Sertoli cells. In this regard, our previous studies (9) have shown that there are significant increases in the nuclear expression of NF-B during stages I-VII of spermatogenesis, which would correspond to the time when TNF-␣ producing spermatids are present. Therefore, spermatid-Sertoli communication via TNF-␣ may signal Sertoli cells to activate NF-B causing CREB to induce the production of specific factors that are required by spermatids or other germ cells. The stage-specific expression of TNF-␣ may be relevant to the previously reported dramatic cyclical changes in Sertoli cell CREB mRNA expression (4). The timing for the induction of CREB mRNA expression correlates with the presence of round spermatids. Although CREB mRNA expression has been shown to be induced by FSH-mediated elevation of cAMP levels in stages II-V (4, 7), it is possible that spermatidderived, TNF-␣-mediated NF-B activation may contribute to the high levels of CREB mRNA that accumulate in Sertoli cells at specific stages of spermatogenesis.
In summary, NF-B and TNF-␣ are able to stimulate expression of the CREB gene in Sertoli cells. In the testis, NF-B regulation of the CREB gene may prove to be important in providing the appropriate timing for the expression of downstream genes required for spermatogenesis. Furthermore, the ability of NF-B and TNF-␣ to induce the CREB gene promoter in NIH 3T3 and HEK 293 cells suggests that NF-B may be a general regulator of CREB expression in non-testis tissues.