Portal de Periódicos da CAPES

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

1999; Elsevier BV; Volume: 274; Issue: 50 Linguagem: Inglês

10.1074/jbc.274.50.35607

1083-351X

Frank J. Delfino, William H. Walker,

Reproductive System and Pregnancy

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 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. follicle-stimulating hormone cAMP-response element-binding protein cAMP-response element nuclear factor κB inhibitor of κB tumor necrosis factor α base pair(s) cytomegalovirus electrophoretic mobility shift assay polyacrylamide gel electrophoresis 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 (1Sharpe R.M. Knobil E. Neil J.D. The Physiology of Reproduction. Raven Press, New York1994: 1363-1434Google Scholar, 2de Kretser D.M. Risbridger G.P. Kerr J. DeGroot L. Endocrinology, 3. 3rd Ed. W. B. Saunders, Philadelphia1995: 2307-2335Google Scholar). 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) (3Gonzalez G. Montminy M.R. Cell. 1989; 59: 675-680Abstract Full Text PDF PubMed Scopus (2041) Google Scholar).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, (4Waeber G. Meyer T.E. LeSieur M. Hermann H. Gérard N. Habener J.F. Mol. Endocrinol. 1991; 5: 1418-1430Crossref PubMed Scopus (139) Google Scholar, 5West A.P. Sharpe R.M. Saunders P.T.K. Biol. Reprod. 1994; 50: 869-881Crossref PubMed Scopus (30) Google Scholar, 6Walker W.H. Daniel P.B. Habener J.F. Mol. Cell. Endocrinol. 1988; 143: 167-178Crossref Scopus (45) Google Scholar). FSH-induced changes in cAMP levels have been implicated in the cyclical control of Sertoli cell CREB expression through CREs within the CREB promoter (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar, 8Walker W.H. Habener J.F. Trends Endocrinol. Metab. 1996; 7: 133-138Abstract Full Text PDF PubMed Scopus (68) Google Scholar). Additional signaling pathways may also control stage-specific expression of CREB. Specifically, NF-κB transcription factors that were recently identified as activating gene expression in Sertoli cells (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar) are potential regulators of CREB expression as the CREB promoter contains consensus NF-κB binding motifs.The family of NF-κB 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 (10Siebenlist U. Franzoso G. Brown K. Annu. Rev. Cell Biol. 1994; 10: 405-455Crossref PubMed Scopus (2010) Google Scholar, 11Baeuerle P.A. Henkel T. Annu. Rev. Immunol. 1994; 12: 141-179Crossref PubMed Scopus (4582) Google Scholar, 12Baldwin A.S. Annu. Rev. Immunol. 1996; 14: 649-681Crossref PubMed Scopus (5544) Google Scholar, 13Verma I.M. Stevenson J.K. Scwartz E.M. Van Antwerp D. Miyamoto S. Genes Dev. 1995; 9: 2723-2735Crossref PubMed Scopus (1654) Google Scholar). In most cells NF-κB dimers remain sequestered in the cytoplasm by inhibitor proteins (IκB-α, IκB-β, IκB-γ, IκB-ε, and IκB-δ). Upon stimulation by diverse stimuli such as TNF-α, phorbol myristic acid, viral proteins, and interleukins, IκB is phosphorylated and ubiquinated leading to proteosome-mediated degradation. The NF-κB nuclear localization signal is then unmasked, and NF-κB is free to translocate to the nucleus and regulate gene expression via interactions with κB enhancer elements (reviewed in Ref. 14Ghosh S. May M.J. Kopp E.B. Annu. Rev. Immunol. 1998; 16: 225-260Crossref PubMed Scopus (4572) Google Scholar). In addition to being regulated via stimulus-induced release from IκB, the activity of free NF-κB can also be modulated through direct phosphorylation of the RelA subunit by protein kinase A (15Zhong H. Voll R.E. Ghosh S. Mol. Cell. 1998; 1: 661-671Abstract Full Text Full Text PDF PubMed Scopus (1022) Google Scholar, 16Naumann M. Scheidereit C. EMBO J. 1994; 13: 4597-4607Crossref PubMed Scopus (325) Google Scholar). The cytokine TNF-α is a regulator of NF-κB activity (reviewed in Ref. 11Baeuerle P.A. Henkel T. Annu. Rev. Immunol. 1994; 12: 141-179Crossref PubMed Scopus (4582) Google Scholar) and therefore is a candidate regulator of CREB gene expression. In the testis TNF-α is secreted primarily by round spermatids within the seminiferous tubules, and the 55-kDa TNF-α receptor has been detected in Sertoli cells (17De S.K. Chen H.-L. Pace J.L. Hunt J.S. Terranova P.F. Enders G.C. Endocrinology. 1993; 133: 389-396Crossref PubMed Scopus (150) Google Scholar,18Mauduit C. Besset V. Caussanel V. Benahmed M. Biochem. Biophys. Res. Commun. 1996; 224: 631-637Crossref PubMed Scopus (47) Google Scholar). Recently, we demonstrated that TNF-α increases the activity of NF-κB in rat Sertoli cells and that NF-κB levels in the nuclei of Sertoli cells are highest during the stages in which round spermatids are present (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar).In this study we test the hypothesis that NF-κB and TNF-α are regulators of CREB expression. We demonstrate that Sertoli cell NF-κB proteins interact with NF-κB binding sites in the CREB promoter. We also show that overexpression of NF-κB subunits in Sertoli cells and NIH 3T3 cells increases CREB promoter activity. Stimulation of primary Sertoli cells with the cytokine TNF-α mediates a reduction in IκB-α and IκB-β levels, a concomitant increase in RelA nuclear translocation and the induction of NF-κB binding to a CREB promoter NF-κB enhancer motif. Transient transfection analyses demonstrate that TNF-α also stimulates CREB gene promoter activity. These data suggest that NF-κB may be an important regulator of genes required for spermatogenesis and a general regulator of CREB gene expression in non-testis cells.DISCUSSIONIn this study, we have demonstrated that NF-κB proteins present in Sertoli cells specifically bind κB enhancer motifs within the CREB promoter. Four potential κB enhancer elements were identified in the CREB promoter by computer-assisted sequence analysis. The two gene proximal κB motifs bound NF-κB proteins more effectively than the distal consensus sequences. The binding of NF-κB was functionally significant as overexpression of NF-κB proteins in Sertoli cells stimulated transcription from the CREB promoter. Although NF-κB is constitutively expressed in the nucleus of Sertoli cells, TNF-α was shown to induce the degradation of IκB and further increase the levels of nuclear NF-κB in Sertoli cells. Furthermore, TNF-α 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 −1264CREB promoter was 2- and 4-fold higher than the activities of the −537CREB and −278CREB promoter fragments, respectively. The decrease in −537CREB promoter activity in Sertoli cells but not NIH 3T3 cells may reflect the loss of two distal κB elements and the higher relative nuclear levels of NF-κB in Sertoli cells compared with NIH 3T3 cells. The more dramatic decrease in activity for the −278CREB promoter may be because of the elimination of Sp1 motifs as described earlier (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar) as well as the proximal κB enhancer motifs. Although the relative basal activities of the −1264CREB, −537CREB and −278CREB promoter constructs in Sertoli cells were similar to that of an earlier report (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar), 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-κB proved to be a potent inducer of the CREB promoter as overexpression of NF-κB p50 and RelA or RelA alone stimulated transcription from the full-length −1264CREB promoter six-fold. Although the distal κB3 and κB4 motifs may contribute to the induction of the CREB promoter, deletion analysis of the CREB promoter showed that the region containing the proximal κB1 and κB2 motifs is required to maintain high basal expression and full NF-κB induction of the CREB promoter. The induction through κB1 and κB2 is not unexpected as these sequences most effectively bind NF-κB present in Sertoli extracts. Deletion analysis of the −680 to −1160 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-κB in Sertoli cells suggested that NF-κB may regulate this gene in other cell types. In this regard, we found that the CREB gene is up-regulated by NF-κB overexpression in NIH 3T3 cells (Figs. 2 B and3 B), and TNF-α induces the CREB promoter in HEK 293 cells (Fig. 6 B). The regulation of CREB by NF-κB may be a new method of cross-talk between NF-κB and CREB signaling pathways to compliment the competition for the CREB-binding protein/p300 coactivator displayed by these factors (28Chrivia J.C. Kwok R.P. Lamb N. Hagiwara M. Montminy M.R. Goodman R.H. Nature. 1993; 365: 855-859Crossref PubMed Scopus (1758) Google Scholar, 29Gerritsen M. Williams A. Neish A. Moore S. Shi Y. Collins T. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 2927-2932Crossref PubMed Scopus (710) Google Scholar). Alternatively, it is possible that CREB and NF-κB 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-κB interactions on the CREB promoter exists as the CREB promoter contains binding sites for CREB in close proximity to κB motifs (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar). 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 (30Stein B. Cogswell P.C. Baldwin A.S. Mol. Cell. Biol. 1993; 13: 3964-3974Crossref PubMed Google Scholar, 31Stein B. Baldwin A.S. Ballard D.W. Greene W.C. Angel P. Herrlich P. EMBO J. 1993; 12: 3879-3891Crossref PubMed Scopus (567) Google Scholar). Furthermore, in glutathione S-transferase-pulldown and co-immunoprecipitation experiments RelA has been shown to also directly interact with CREB. 2W. Walker, unpublished results. Studies are underway to investigate potential cooperativity of NF-κB and CREB in stimulating transcription from the CREB promoter.In this initial characterization of IκB proteins from pure cultures of rat Sertoli cells, both IκBα and IκBβ were determined to be present in the cytoplasm, and the levels of both were dramatically reduced after the addition of TNF-α. A previous study failed to detect IκBα mRNA in extracts from whole mouse testis (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar). The inability to detect Sertoli cell-derived IκBα 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 (19Bellvé A.R. Millette C.F. Bhatnagar Y.M. O'Brien D.A. J. Histochem. Cytochem. 1977; 25: 480-494Crossref PubMed Scopus (295) Google Scholar). In addition to the immunodetection of IκBβ in Sertoli cells in the present report, IκBβ mRNA was previously shown to be enriched in mouse testis (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar). Together, these data suggest that whereas IκBα and IκBβ are both present in Sertoli cells, the developing germ cells, which account for greater than 90% of testis cells, likely contain IκBβ but not IκBα. Differences in the levels of the two IκB isoforms may be important in cell-specific gene regulation as stimulation of IκBα degradation is rapid and transient, but IκBβ degradation can be delayed and persistent (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar).One factor that may maintain high nuclear levels of NF-κB in Sertoli cells and account for additional NF-κB translocation to the nucleus is the cytokine TNF-α. In Sertoli cells, TNF-α activates NF-κB via the elimination of IκB from the cytoplasm and the subsequent translocation of additional NF-κB to the nucleus. TNF-α-mediated stimulation of NF-κB in Sertoli cells may be physiologically important because germ cells adjacent to Sertoli cells secrete TNF-α in a stage-specific manner. Although it is possible that TNF-α may also act through mitogen-activated protein kinase or other pathways, it is significant that NF-κB is activated by TNF-α in Sertoli cells. In the absence of nuclear NF-κB, apoptosis pathways may be initiated by TNF-α (32Wang C.-Y. Mayo M.W. Baldwin A.S. Science. 1996; 274: 784-789Crossref PubMed Scopus (2500) Google Scholar). In contrast to germ cell development in which some apoptosis is required to constrain the expansion of germ cells and maintain spermatogenesis (33Rodriguez I. Ody C. Araki K. Garcia I. Vassalli P. EMBO J. 1997; 16: 2262-2270Crossref PubMed Scopus (489) Google Scholar, 34Bartke A. Endocrinology. 1995; 136: 3-4Crossref PubMed Google Scholar), nuclear NF-κB may protect Sertoli cells from apoptosis inducers. NF-κB-mediated protection from apoptosis agents would explain why few or no apoptotic Sertoli cells are detected in the testis (32Wang C.-Y. Mayo M.W. Baldwin A.S. Science. 1996; 274: 784-789Crossref PubMed Scopus (2500) Google Scholar, 35Beg A. Baltimore D. Science. 1996; 274: 782-784Crossref PubMed Scopus (2926) Google Scholar, 36Antwerp D.J. Martin S.J. Kafri T. Green D. Verma I.M. Science. 1996; 274: 787-789Crossref PubMed Scopus (2441) Google Scholar).In a previous study TNF-α alone was able to stimulate transcription from a minimal promoter containing two consensus κB enhancers (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar). In contrast, TNF-α 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-α 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-α.Although other germ cell types secrete some TNF-α, most of the TNF-α is secreted by round spermatids (17De S.K. Chen H.-L. Pace J.L. Hunt J.S. Terranova P.F. Enders G.C. Endocrinology. 1993; 133: 389-396Crossref PubMed Scopus (150) Google Scholar). 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-α secretion may represent stage-specific communication between spermatids and Sertoli cells. In this regard, our previous studies (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar) have shown that there are significant increases in the nuclear expression of NF-κB during stages I-VII of spermatogenesis, which would correspond to the time when TNF-α producing spermatids are present. Therefore, spermatid-Sertoli communication via TNF-α may signal Sertoli cells to activate NF-κB causing CREB to induce the production of specific factors that are required by spermatids or other germ cells. The stage-specific expression of TNF-α may be relevant to the previously reported dramatic cyclical changes in Sertoli cell CREB mRNA expression (4Waeber G. Meyer T.E. LeSieur M. Hermann H. Gérard N. Habener J.F. Mol. Endocrinol. 1991; 5: 1418-1430Crossref PubMed Scopus (139) Google Scholar). 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 (4Waeber G. Meyer T.E. LeSieur M. Hermann H. Gérard N. Habener J.F. Mol. Endocrinol. 1991; 5: 1418-1430Crossref PubMed Scopus (139) Google Scholar, 7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar), it is possible that spermatid-derived, TNF-α-mediated NF-κB activation may contribute to the high levels of CREB mRNA that accumulate in Sertoli cells at specific stages of spermatogenesis.In summary, NF-κB and TNF-α are able to stimulate expression of the CREB gene in Sertoli cells. In the testis, NF-κB 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-κB and TNF-α to induce the CREB gene promoter in NIH 3T3 and HEK 293 cells suggests that NF-κB may be a general regulator of CREB expression 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 (1Sharpe R.M. Knobil E. Neil J.D. The Physiology of Reproduction. Raven Press, New York1994: 1363-1434Google Scholar, 2de Kretser D.M. Risbridger G.P. Kerr J. DeGroot L. Endocrinology, 3. 3rd Ed. W. B. Saunders, Philadelphia1995: 2307-2335Google Scholar). 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) (3Gonzalez G. Montminy M.R. Cell. 1989; 59: 675-680Abstract Full Text PDF PubMed Scopus (2041) Google Scholar). 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, (4Waeber G. Meyer T.E. LeSieur M. Hermann H. Gérard N. Habener J.F. Mol. Endocrinol. 1991; 5: 1418-1430Crossref PubMed Scopus (139) Google Scholar, 5West A.P. Sharpe R.M. Saunders P.T.K. Biol. Reprod. 1994; 50: 869-881Crossref PubMed Scopus (30) Google Scholar, 6Walker W.H. Daniel P.B. Habener J.F. Mol. Cell. Endocrinol. 1988; 143: 167-178Crossref Scopus (45) Google Scholar). FSH-induced changes in cAMP levels have been implicated in the cyclical control of Sertoli cell CREB expression through CREs within the CREB promoter (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar, 8Walker W.H. Habener J.F. Trends Endocrinol. Metab. 1996; 7: 133-138Abstract Full Text PDF PubMed Scopus (68) Google Scholar). Additional signaling pathways may also control stage-specific expression of CREB. Specifically, NF-κB transcription factors that were recently identified as activating gene expression in Sertoli cells (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar) are potential regulators of CREB expression as the CREB promoter contains consensus NF-κB binding motifs. The family of NF-κB 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 (10Siebenlist U. Franzoso G. Brown K. Annu. Rev. Cell Biol. 1994; 10: 405-455Crossref PubMed Scopus (2010) Google Scholar, 11Baeuerle P.A. Henkel T. Annu. Rev. Immunol. 1994; 12: 141-179Crossref PubMed Scopus (4582) Google Scholar, 12Baldwin A.S. Annu. Rev. Immunol. 1996; 14: 649-681Crossref PubMed Scopus (5544) Google Scholar, 13Verma I.M. Stevenson J.K. Scwartz E.M. Van Antwerp D. Miyamoto S. Genes Dev. 1995; 9: 2723-2735Crossref PubMed Scopus (1654) Google Scholar). In most cells NF-κB dimers remain sequestered in the cytoplasm by inhibitor proteins (IκB-α, IκB-β, IκB-γ, IκB-ε, and IκB-δ). Upon stimulation by diverse stimuli such as TNF-α, phorbol myristic acid, viral proteins, and interleukins, IκB is phosphorylated and ubiquinated leading to proteosome-mediated degradation. The NF-κB nuclear localization signal is then unmasked, and NF-κB is free to translocate to the nucleus and regulate gene expression via interactions with κB enhancer elements (reviewed in Ref. 14Ghosh S. May M.J. Kopp E.B. Annu. Rev. Immunol. 1998; 16: 225-260Crossref PubMed Scopus (4572) Google Scholar). In addition to being regulated via stimulus-induced release from IκB, the activity of free NF-κB can also be modulated through direct phosphorylation of the RelA subunit by protein kinase A (15Zhong H. Voll R.E. Ghosh S. Mol. Cell. 1998; 1: 661-671Abstract Full Text Full Text PDF PubMed Scopus (1022) Google Scholar, 16Naumann M. Scheidereit C. EMBO J. 1994; 13: 4597-4607Crossref PubMed Scopus (325) Google Scholar). The cytokine TNF-α is a regulator of NF-κB activity (reviewed in Ref. 11Baeuerle P.A. Henkel T. Annu. Rev. Immunol. 1994; 12: 141-179Crossref PubMed Scopus (4582) Google Scholar) and therefore is a candidate regulator of CREB gene expression. In the testis TNF-α is secreted primarily by round spermatids within the seminiferous tubules, and the 55-kDa TNF-α receptor has been detected in Sertoli cells (17De S.K. Chen H.-L. Pace J.L. Hunt J.S. Terranova P.F. Enders G.C. Endocrinology. 1993; 133: 389-396Crossref PubMed Scopus (150) Google Scholar,18Mauduit C. Besset V. Caussanel V. Benahmed M. Biochem. Biophys. Res. Commun. 1996; 224: 631-637Crossref PubMed Scopus (47) Google Scholar). Recently, we demonstrated that TNF-α increases the activity of NF-κB in rat Sertoli cells and that NF-κB levels in the nuclei of Sertoli cells are highest during the stages in which round spermatids are present (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar). In this study we test the hypothesis that NF-κB and TNF-α are regulators of CREB expression. We demonstrate that Sertoli cell NF-κB proteins interact with NF-κB binding sites in the CREB promoter. We also show that overexpression of NF-κB subunits in Sertoli cells and NIH 3T3 cells increases CREB promoter activity. Stimulation of primary Sertoli cells with the cytokine TNF-α mediates a reduction in IκB-α and IκB-β levels, a concomitant increase in RelA nuclear translocation and the induction of NF-κB binding to a CREB promoter NF-κB enhancer motif. Transient transfection analyses demonstrate that TNF-α also stimulates CREB gene promoter activity. These data suggest that NF-κB may be an important regulator of genes required for spermatogenesis and a general regulator of CREB gene expression in non-testis cells. DISCUSSIONIn this study, we have demonstrated that NF-κB proteins present in Sertoli cells specifically bind κB enhancer motifs within the CREB promoter. Four potential κB enhancer elements were identified in the CREB promoter by computer-assisted sequence analysis. The two gene proximal κB motifs bound NF-κB proteins more effectively than the distal consensus sequences. The binding of NF-κB was functionally significant as overexpression of NF-κB proteins in Sertoli cells stimulated transcription from the CREB promoter. Although NF-κB is constitutively expressed in the nucleus of Sertoli cells, TNF-α was shown to induce the degradation of IκB and further increase the levels of nuclear NF-κB in Sertoli cells. Furthermore, TNF-α 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 −1264CREB promoter was 2- and 4-fold higher than the activities of the −537CREB and −278CREB promoter fragments, respectively. The decrease in −537CREB promoter activity in Sertoli cells but not NIH 3T3 cells may reflect the loss of two distal κB elements and the higher relative nuclear levels of NF-κB in Sertoli cells compared with NIH 3T3 cells. The more dramatic decrease in activity for the −278CREB promoter may be because of the elimination of Sp1 motifs as described earlier (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar) as well as the proximal κB enhancer motifs. Although the relative basal activities of the −1264CREB, −537CREB and −278CREB promoter constructs in Sertoli cells were similar to that of an earlier report (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar), 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-κB proved to be a potent inducer of the CREB promoter as overexpression of NF-κB p50 and RelA or RelA alone stimulated transcription from the full-length −1264CREB promoter six-fold. Although the distal κB3 and κB4 motifs may contribute to the induction of the CREB promoter, deletion analysis of the CREB promoter showed that the region containing the proximal κB1 and κB2 motifs is required to maintain high basal expression and full NF-κB induction of the CREB promoter. The induction through κB1 and κB2 is not unexpected as these sequences most effectively bind NF-κB present in Sertoli extracts. Deletion analysis of the −680 to −1160 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-κB in Sertoli cells suggested that NF-κB may regulate this gene in other cell types. In this regard, we found that the CREB gene is up-regulated by NF-κB overexpression in NIH 3T3 cells (Figs. 2 B and3 B), and TNF-α induces the CREB promoter in HEK 293 cells (Fig. 6 B). The regulation of CREB by NF-κB may be a new method of cross-talk between NF-κB and CREB signaling pathways to compliment the competition for the CREB-binding protein/p300 coactivator displayed by these factors (28Chrivia J.C. Kwok R.P. Lamb N. Hagiwara M. Montminy M.R. Goodman R.H. Nature. 1993; 365: 855-859Crossref PubMed Scopus (1758) Google Scholar, 29Gerritsen M. Williams A. Neish A. Moore S. Shi Y. Collins T. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 2927-2932Crossref PubMed Scopus (710) Google Scholar). Alternatively, it is possible that CREB and NF-κB 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-κB interactions on the CREB promoter exists as the CREB promoter contains binding sites for CREB in close proximity to κB motifs (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar). 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 (30Stein B. Cogswell P.C. Baldwin A.S. Mol. Cell. Biol. 1993; 13: 3964-3974Crossref PubMed Google Scholar, 31Stein B. Baldwin A.S. Ballard D.W. Greene W.C. Angel P. Herrlich P. EMBO J. 1993; 12: 3879-3891Crossref PubMed Scopus (567) Google Scholar). Furthermore, in glutathione S-transferase-pulldown and co-immunoprecipitation experiments RelA has been shown to also directly interact with CREB. 2W. Walker, unpublished results. Studies are underway to investigate potential cooperativity of NF-κB and CREB in stimulating transcription from the CREB promoter.In this initial characterization of IκB proteins from pure cultures of rat Sertoli cells, both IκBα and IκBβ were determined to be present in the cytoplasm, and the levels of both were dramatically reduced after the addition of TNF-α. A previous study failed to detect IκBα mRNA in extracts from whole mouse testis (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar). The inability to detect Sertoli cell-derived IκBα 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 (19Bellvé A.R. Millette C.F. Bhatnagar Y.M. O'Brien D.A. J. Histochem. Cytochem. 1977; 25: 480-494Crossref PubMed Scopus (295) Google Scholar). In addition to the immunodetection of IκBβ in Sertoli cells in the present report, IκBβ mRNA was previously shown to be enriched in mouse testis (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar). Together, these data suggest that whereas IκBα and IκBβ are both present in Sertoli cells, the developing germ cells, which account for greater than 90% of testis cells, likely contain IκBβ but not IκBα. Differences in the levels of the two IκB isoforms may be important in cell-specific gene regulation as stimulation of IκBα degradation is rapid and transient, but IκBβ degradation can be delayed and persistent (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar).One factor that may maintain high nuclear levels of NF-κB in Sertoli cells and account for additional NF-κB translocation to the nucleus is the cytokine TNF-α. In Sertoli cells, TNF-α activates NF-κB via the elimination of IκB from the cytoplasm and the subsequent translocation of additional NF-κB to the nucleus. TNF-α-mediated stimulation of NF-κB in Sertoli cells may be physiologically important because germ cells adjacent to Sertoli cells secrete TNF-α in a stage-specific manner. Although it is possible that TNF-α may also act through mitogen-activated protein kinase or other pathways, it is significant that NF-κB is activated by TNF-α in Sertoli cells. In the absence of nuclear NF-κB, apoptosis pathways may be initiated by TNF-α (32Wang C.-Y. Mayo M.W. Baldwin A.S. Science. 1996; 274: 784-789Crossref PubMed Scopus (2500) Google Scholar). In contrast to germ cell development in which some apoptosis is required to constrain the expansion of germ cells and maintain spermatogenesis (33Rodriguez I. Ody C. Araki K. Garcia I. Vassalli P. EMBO J. 1997; 16: 2262-2270Crossref PubMed Scopus (489) Google Scholar, 34Bartke A. Endocrinology. 1995; 136: 3-4Crossref PubMed Google Scholar), nuclear NF-κB may protect Sertoli cells from apoptosis inducers. NF-κB-mediated protection from apoptosis agents would explain why few or no apoptotic Sertoli cells are detected in the testis (32Wang C.-Y. Mayo M.W. Baldwin A.S. Science. 1996; 274: 784-789Crossref PubMed Scopus (2500) Google Scholar, 35Beg A. Baltimore D. Science. 1996; 274: 782-784Crossref PubMed Scopus (2926) Google Scholar, 36Antwerp D.J. Martin S.J. Kafri T. Green D. Verma I.M. Science. 1996; 274: 787-789Crossref PubMed Scopus (2441) Google Scholar).In a previous study TNF-α alone was able to stimulate transcription from a minimal promoter containing two consensus κB enhancers (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar). In contrast, TNF-α 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-α 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-α.Although other germ cell types secrete some TNF-α, most of the TNF-α is secreted by round spermatids (17De S.K. Chen H.-L. Pace J.L. Hunt J.S. Terranova P.F. Enders G.C. Endocrinology. 1993; 133: 389-396Crossref PubMed Scopus (150) Google Scholar). 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-α secretion may represent stage-specific communication between spermatids and Sertoli cells. In this regard, our previous studies (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar) have shown that there are significant increases in the nuclear expression of NF-κB during stages I-VII of spermatogenesis, which would correspond to the time when TNF-α producing spermatids are present. Therefore, spermatid-Sertoli communication via TNF-α may signal Sertoli cells to activate NF-κB causing CREB to induce the production of specific factors that are required by spermatids or other germ cells. The stage-specific expression of TNF-α may be relevant to the previously reported dramatic cyclical changes in Sertoli cell CREB mRNA expression (4Waeber G. Meyer T.E. LeSieur M. Hermann H. Gérard N. Habener J.F. Mol. Endocrinol. 1991; 5: 1418-1430Crossref PubMed Scopus (139) Google Scholar). 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 (4Waeber G. Meyer T.E. LeSieur M. Hermann H. Gérard N. Habener J.F. Mol. Endocrinol. 1991; 5: 1418-1430Crossref PubMed Scopus (139) Google Scholar, 7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar), it is possible that spermatid-derived, TNF-α-mediated NF-κB activation may contribute to the high levels of CREB mRNA that accumulate in Sertoli cells at specific stages of spermatogenesis.In summary, NF-κB and TNF-α are able to stimulate expression of the CREB gene in Sertoli cells. In the testis, NF-κB 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-κB and TNF-α to induce the CREB gene promoter in NIH 3T3 and HEK 293 cells suggests that NF-κB may be a general regulator of CREB expression in non-testis tissues. In this study, we have demonstrated that NF-κB proteins present in Sertoli cells specifically bind κB enhancer motifs within the CREB promoter. Four potential κB enhancer elements were identified in the CREB promoter by computer-assisted sequence analysis. The two gene proximal κB motifs bound NF-κB proteins more effectively than the distal consensus sequences. The binding of NF-κB was functionally significant as overexpression of NF-κB proteins in Sertoli cells stimulated transcription from the CREB promoter. Although NF-κB is constitutively expressed in the nucleus of Sertoli cells, TNF-α was shown to induce the degradation of IκB and further increase the levels of nuclear NF-κB in Sertoli cells. Furthermore, TNF-α 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 −1264CREB promoter was 2- and 4-fold higher than the activities of the −537CREB and −278CREB promoter fragments, respectively. The decrease in −537CREB promoter activity in Sertoli cells but not NIH 3T3 cells may reflect the loss of two distal κB elements and the higher relative nuclear levels of NF-κB in Sertoli cells compared with NIH 3T3 cells. The more dramatic decrease in activity for the −278CREB promoter may be because of the elimination of Sp1 motifs as described earlier (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar) as well as the proximal κB enhancer motifs. Although the relative basal activities of the −1264CREB, −537CREB and −278CREB promoter constructs in Sertoli cells were similar to that of an earlier report (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar), 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-κB proved to be a potent inducer of the CREB promoter as overexpression of NF-κB p50 and RelA or RelA alone stimulated transcription from the full-length −1264CREB promoter six-fold. Although the distal κB3 and κB4 motifs may contribute to the induction of the CREB promoter, deletion analysis of the CREB promoter showed that the region containing the proximal κB1 and κB2 motifs is required to maintain high basal expression and full NF-κB induction of the CREB promoter. The induction through κB1 and κB2 is not unexpected as these sequences most effectively bind NF-κB present in Sertoli extracts. Deletion analysis of the −680 to −1160 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-κB in Sertoli cells suggested that NF-κB may regulate this gene in other cell types. In this regard, we found that the CREB gene is up-regulated by NF-κB overexpression in NIH 3T3 cells (Figs. 2 B and3 B), and TNF-α induces the CREB promoter in HEK 293 cells (Fig. 6 B). The regulation of CREB by NF-κB may be a new method of cross-talk between NF-κB and CREB signaling pathways to compliment the competition for the CREB-binding protein/p300 coactivator displayed by these factors (28Chrivia J.C. Kwok R.P. Lamb N. Hagiwara M. Montminy M.R. Goodman R.H. Nature. 1993; 365: 855-859Crossref PubMed Scopus (1758) Google Scholar, 29Gerritsen M. Williams A. Neish A. Moore S. Shi Y. Collins T. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 2927-2932Crossref PubMed Scopus (710) Google Scholar). Alternatively, it is possible that CREB and NF-κB 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-κB interactions on the CREB promoter exists as the CREB promoter contains binding sites for CREB in close proximity to κB motifs (7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar). 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 (30Stein B. Cogswell P.C. Baldwin A.S. Mol. Cell. Biol. 1993; 13: 3964-3974Crossref PubMed Google Scholar, 31Stein B. Baldwin A.S. Ballard D.W. Greene W.C. Angel P. Herrlich P. EMBO J. 1993; 12: 3879-3891Crossref PubMed Scopus (567) Google Scholar). Furthermore, in glutathione S-transferase-pulldown and co-immunoprecipitation experiments RelA has been shown to also directly interact with CREB. 2W. Walker, unpublished results. Studies are underway to investigate potential cooperativity of NF-κB and CREB in stimulating transcription from the CREB promoter. In this initial characterization of IκB proteins from pure cultures of rat Sertoli cells, both IκBα and IκBβ were determined to be present in the cytoplasm, and the levels of both were dramatically reduced after the addition of TNF-α. A previous study failed to detect IκBα mRNA in extracts from whole mouse testis (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar). The inability to detect Sertoli cell-derived IκBα 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 (19Bellvé A.R. Millette C.F. Bhatnagar Y.M. O'Brien D.A. J. Histochem. Cytochem. 1977; 25: 480-494Crossref PubMed Scopus (295) Google Scholar). In addition to the immunodetection of IκBβ in Sertoli cells in the present report, IκBβ mRNA was previously shown to be enriched in mouse testis (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar). Together, these data suggest that whereas IκBα and IκBβ are both present in Sertoli cells, the developing germ cells, which account for greater than 90% of testis cells, likely contain IκBβ but not IκBα. Differences in the levels of the two IκB isoforms may be important in cell-specific gene regulation as stimulation of IκBα degradation is rapid and transient, but IκBβ degradation can be delayed and persistent (27Thompson J.E. Phillips R.J. Erdjument-Bromage H. Tempst P. Ghosh S. Cell. 1995; 80: 573-582Abstract Full Text PDF PubMed Scopus (692) Google Scholar). One factor that may maintain high nuclear levels of NF-κB in Sertoli cells and account for additional NF-κB translocation to the nucleus is the cytokine TNF-α. In Sertoli cells, TNF-α activates NF-κB via the elimination of IκB from the cytoplasm and the subsequent translocation of additional NF-κB to the nucleus. TNF-α-mediated stimulation of NF-κB in Sertoli cells may be physiologically important because germ cells adjacent to Sertoli cells secrete TNF-α in a stage-specific manner. Although it is possible that TNF-α may also act through mitogen-activated protein kinase or other pathways, it is significant that NF-κB is activated by TNF-α in Sertoli cells. In the absence of nuclear NF-κB, apoptosis pathways may be initiated by TNF-α (32Wang C.-Y. Mayo M.W. Baldwin A.S. Science. 1996; 274: 784-789Crossref PubMed Scopus (2500) Google Scholar). In contrast to germ cell development in which some apoptosis is required to constrain the expansion of germ cells and maintain spermatogenesis (33Rodriguez I. Ody C. Araki K. Garcia I. Vassalli P. EMBO J. 1997; 16: 2262-2270Crossref PubMed Scopus (489) Google Scholar, 34Bartke A. Endocrinology. 1995; 136: 3-4Crossref PubMed Google Scholar), nuclear NF-κB may protect Sertoli cells from apoptosis inducers. NF-κB-mediated protection from apoptosis agents would explain why few or no apoptotic Sertoli cells are detected in the testis (32Wang C.-Y. Mayo M.W. Baldwin A.S. Science. 1996; 274: 784-789Crossref PubMed Scopus (2500) Google Scholar, 35Beg A. Baltimore D. Science. 1996; 274: 782-784Crossref PubMed Scopus (2926) Google Scholar, 36Antwerp D.J. Martin S.J. Kafri T. Green D. Verma I.M. Science. 1996; 274: 787-789Crossref PubMed Scopus (2441) Google Scholar). In a previous study TNF-α alone was able to stimulate transcription from a minimal promoter containing two consensus κB enhancers (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar). In contrast, TNF-α 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-α 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-α. Although other germ cell types secrete some TNF-α, most of the TNF-α is secreted by round spermatids (17De S.K. Chen H.-L. Pace J.L. Hunt J.S. Terranova P.F. Enders G.C. Endocrinology. 1993; 133: 389-396Crossref PubMed Scopus (150) Google Scholar). 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-α secretion may represent stage-specific communication between spermatids and Sertoli cells. In this regard, our previous studies (9Delfino F.J. Walker W.H. Mol. Endocrinol. 1998; 12: 1696-1707Crossref PubMed Scopus (89) Google Scholar) have shown that there are significant increases in the nuclear expression of NF-κB during stages I-VII of spermatogenesis, which would correspond to the time when TNF-α producing spermatids are present. Therefore, spermatid-Sertoli communication via TNF-α may signal Sertoli cells to activate NF-κB causing CREB to induce the production of specific factors that are required by spermatids or other germ cells. The stage-specific expression of TNF-α may be relevant to the previously reported dramatic cyclical changes in Sertoli cell CREB mRNA expression (4Waeber G. Meyer T.E. LeSieur M. Hermann H. Gérard N. Habener J.F. Mol. Endocrinol. 1991; 5: 1418-1430Crossref PubMed Scopus (139) Google Scholar). 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 (4Waeber G. Meyer T.E. LeSieur M. Hermann H. Gérard N. Habener J.F. Mol. Endocrinol. 1991; 5: 1418-1430Crossref PubMed Scopus (139) Google Scholar, 7Walker W.H. Fucci L. Habener J.F. Endocrinology. 1995; 136: 3534-3545Crossref PubMed Google Scholar), it is possible that spermatid-derived, TNF-α-mediated NF-κB activation may contribute to the high levels of CREB mRNA that accumulate in Sertoli cells at specific stages of spermatogenesis. In summary, NF-κB and TNF-α are able to stimulate expression of the CREB gene in Sertoli cells. In the testis, NF-κB 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-κB and TNF-α to induce the CREB gene promoter in NIH 3T3 and HEK 293 cells suggests that NF-κB may be a general regulator of CREB expression in non-testis tissues. We thank Drs. Dean Ballard and Stefan Dorre for supplying NF-κB and IκB antisera and Dr. Simon Watkins of the University of Pittsburgh Structural Biology Imaging Center for assistance with immunofluorescence imaging. We are indebted to Dr. Tony Zeleznik for critical reading of the manuscript.