Mechanisms of FOXC2- and FOXD1 -mediated Regulation of the RI (cid:1) Subunit of cAMP-dependent Protein Kinase Include Release of Transcriptional Repression and Activation by Protein Kinase B (cid:1) and cAMP*

We have reported recently that mice overexpressing the forkhead/winged helix transcription factor FOXC2 are lean and show increased responsiveness to insulin due to sensitization of the (cid:2) -adrenergic cAMP-PKA (cid:3) pathway and increased levels of the RI (cid:1) subunit of cAMP-dependent protein kinase (PKA) (Cederberg, A., Grønning, L. M., Ahren, B., Taske´n, K., Carlsson, P., and Enerba¨ck, S. (2001) Cell 106, 563–573). In this present study, we reveal that FOXC2 and a related factor, FOXD1, specifically activate the 1b promoter of the RI (cid:1) gene in adipocytes and testicular Sertoli cells, respectively. By deletional mapping, we discovered two different mechanisms by which the Fox proteins activated expression from the RI (cid:1) 1b promoter. In 3T3-L1 adipocytes, an upstream region represses promoter activity under basal conditions. Bandshift

The forkhead/winged helix family of transcription factors is characterized by a highly conserved monomeric DNA-binding domain called the winged helix (reviewed in Ref. 2). A number of forkhead and forkhead-related genes have been isolated to date (3)(4)(5)(6)(7), and the Fox 1 nomenclature (Forkhead box) has now been adopted for all chordate forkhead genes (www.biology. pomona.edu/fox.html) (8). Among these are several forkhead related activators (FREACs) cloned from human (9 -13) that all share the minimum requirement for a 7-bp core binding motif (RTAAAYA). One of these factors, FOXD1 (FREAC4, FKHL8), has expression restricted to kidney, the central nervous system testis, and is regulated by Ets-1 and p53 in kidney-derived cell lines (11,14). FOXC2 (FREAC11, FKHL14, MFH-1) (15,16) is restricted to adipocytes in adults (1), whereas the prenatal form is important for development. Mice lacking Foxc2 die during embryogenesis or perinatally and exhibit aortic arch and skeletal defects (17,18). Instead, overexpression of FOXC2 in adipose tissue has been important in understanding its function in adult mice. FOXC2 transgenic mice develop a phenotype characterized by a high sensitivity to insulin and partial resistance to diet-induced obesity (1). This effect is partly due to up-regulation of ␤-adrenergic receptors and PKA type I␣ that lower the threshold for activation of PKA by cAMP, resulting in a hypersensitive ␤-adrenergic pathway.
Activation of the cAMP-dependent protein kinase (PKA) proceeds by a concerted reaction in which binding of the intracellular second messenger cAMP to the regulatory subunit dimer (R 2 ) in a positive, cooperative fashion results in dissociation and activation of two catalytic (C) subunits (reviewed in Ref. 19). Targeted disruption of the RII␤ regulatory subunit gene in mice leads to a lean phenotype with elevated levels of uncoupling protein 1 and increased metabolic rate due to a shift in the PKA composition from PKA II␤ (RII␤ 2 C 2 ) to type I␣ (RI␣ 2 C 2 ) holoenzyme (20 -22). The effect of this regulatory subunit shift was shown to lower the threshold for PKA activation by cAMP and to modulate lipolysis (23). The RII␤ knockout phenotype resembles that of the FOXC2 transgenic mice (1) and supports the notion that regulation of PKA isozyme composition, particularly RII␤ versus RI␣, is important for hormonal responsiveness and cAMP sensitivity.
The RI␣ gene is controlled by several promoters, giving rise to at least three mRNAs that differ in the first non-coding exon (24 -26). Two of these mRNAs (RI␣1a and -1b) are expressed in most tissues (25), and we have recently reported cAMP-mediated post-transcriptional regulation of RI␣1b in Sertoli cells (27). We have shown that both FOXC2 and RI␣ are induced by cAMP in mouse 3T3-L1 adipocytes (28) 2 and that FOXD1 and RI␣ are induced by cAMP in Sertoli cell primary cultures 3 (27). We have also reported recently (28) that basal levels of Foxc2 mRNA are down-regulated during differentiation of 3T3-L1 adipocytes, with a simultaneous increase in Foxc2 responsiveness to insulin and TNF␣.
In this study, we examine the effect of FOXD1 and FOXC2 on the RI␣ promoters, showing that the expression of RI␣1b but not RI␣1a is induced. Deletional mapping reveals that FOXC2 and FOXD1 regulate expression from the RI␣1b promoter through two different mechanisms, and EMSA and co-transfection experiments showed that in 3T3-L1 adipocytes FOXC2 induces a release of transcriptional repression. In Sertoli cells, PKB␣ strongly increases the effect of FOXD1, and a truncated FOXD1 inhibits cAMP-mediated induction of the RI␣1b promoter, indicating that FOXD1 may function as a mediator of signaling by both cAMP and PKB in a cell-specific manner.
Cell Cultures-Primary cultures of rat Sertoli cells were prepared from testes of 19 days old Sprague-Dawley rats (B&K Universal AS, Nittedal, Norway) according to the method of Dorrington et al. (32) with some modifications (33). The cells were plated in 6-well plates (35-mm/ well) for transfections or in 10-cm culture dishes for preparation of nuclear extracts. Cells were grown in Eagle's minimal essential medium (Invitrogen) with addition of streptomycin (100 g/liter), penicillin (70 mg/liter), fungizone (0,25 mg/liter), L-glutamine (2 mM), and 10% fetal bovine serum at 32°C in a humidified atmosphere with 5% CO 2 . After 3 days, the cells were incubated further in serum-free modified Eagle's minimal essential medium. After 2 days of culture in serum-free medium, LipofectAMINE-mediated transfections were performed as described elsewhere (34) using 2 g of DNA (1.5 g of CAT or luciferase reporter and 0.5 g of internal luciferase or CAT control) with 5 l of LipofectAMINE (Invitrogen) per 35-mm well for 3 h, after which media were changed. Mouse 3T3 L1 cells (American Type Culture Collection) were plated in 6-well plates for transfections or in 10-cm culture dishes for preparation of nuclear or whole cell extracts. Cells were grown in Dulbecco's modified Eagle's medium (Invitrogen) with 4.5 g/liter glucose with addition of streptomycin (100 mg/liter), penicillin (70 mg/liter), fungizone (0.25 mg/liter), anti-PPLO agent, and 10% fetal bovine serum at 37°C, and transfected at ϳ80% confluency as described above.
Luciferase and CAT Assays-All cells were harvested in reporter lysis buffer 48 h after transfection and assayed for luciferase activity (Promega, Madison, WI). CAT activities were measured using an organic phase extraction method (35) and normalized for expression of luciferase.
Preparation of Nuclear Extracts-Sertoli cells or preconfluent 3T3-L1 cells (10-cm 2 culture dish) were scraped in Hanks' balanced salts solution containing 0.1% fatty acid-free bovine serum albumin, harvested by centrifugation at 320 ϫ g for 5 min (4°C), and washed once in cold phosphate-buffered saline. Cell pellets were resuspended in 450 l of hypotonic buffer (10 mM Tris, pH 7.6, 10 mM NaCl, 3 mM MgCl 2 ) and lysed by addition of 50 l of 5% Nonidet P-40 in hypotonic buffer. The nuclei were pelleted by centrifugation (130 ϫ g, 5 min, 4°C); pellets were carefully washed in 1 ml of hypotonic buffer, centrifuged (130 ϫ g), and then resuspended in 100 l of a buffer containing 5 mM Hepes, pH 7.9, 26% glycerol, 1.5 mM dithiothreitol, and the protease inhibitors phenylmethylsulfonyl fluoride (0.5 mM), Complete protease inhibitor mix (1 tablet/10 ml), and calpain inhibitor I (50 M) (Roche Molecular Biochemicals). High salt extraction was accomplished by addition of NaCl to a final concentration of 400 mM while mixing for 30 min at 4°C. Extracts were centrifuged (30,000 ϫ g, 20 min, 4°C), and the supernatants were stored at Ϫ70°C until analysis.

FOXC2 Induces RI␣ Protein in 3T3-L1
Cells-In order to establish whether FOXC2 expression could induce RI␣ protein levels in 3T3-L1 cells as in the FOXC2 transgene, we harvested cells transfected with human FOXC2 expression vector or the corresponding empty vector and prepared whole cell extracts after 6, 12, and 24 h. RI␣, the PKA catalytic subunit, C␣, and FOXC2 protein levels were then examined by immunoblotting ( Fig. 1). Basal levels of endogenous Foxc2 were observed in mouse 3T3-L1 cells, and a stronger band (2-fold) was observed in extracts of 3T3-L1 cells transfected for 6 h with the human FOXC2 construct. The levels of FOXC2 protein were strongly increased after 12 and 24 h of expression (6-fold), and the mobility (ϳ62 kDa with the appearance of a faster migrating band) was similar to that observed previously (16). RI␣ levels were ϳ2-fold induced in the presence of FOXC2 at 6 -24 h of expression as determined by densitometric scanning. In contrast, FOXC2 had no effect on the levels of C␣.
FOXD1 and FOXC2 Induce Reporter Activity Driven by the RI␣1b Promoter in 3T3-L1 Cells-To map the effect of FOXC2 on RI␣ promoter activity in adipocytes, we cotransfected 3T3-L1 cells at 80% confluency with reporter constructs containing the RI␣1aϩ1b promoter, the 1a or 1b promoters alone, and a construct containing 4500 bp of the RII␤ promoter together with pCB6ϩ/FOXC2 or the empty expression vector ( Fig. 2A). We simultaneously tested whether FOXD1, which is not endogenously expressed in adipocytes, had a similar effect on the RI␣ promoter (Fig. 2B). FOXC2 expression induced a 4 -5-fold increase in reporter activity directed from the RI␣1aϩ1b and the RI␣1b promoter constructs, and in contrast, the RI␣1a or RII␤ promoters were slightly down-regulated by the presence of FOXC2. The same pattern of regulation was observed with FOXD1. We next analyzed expression from five 3Ј and/or 5Ј deletion constructs of RI␣1b in the pCAT basic reporter vector (Fig. 2C, left panel) (27). We observed a 7-fold induction of basal reporter expression when a 100-bp region in exon 1A was absent (RI␣1b Aϩ). This deletion raised basal expression to the same levels as in the presence of FOXC2 and thereby abolished the induction by FOXC2 indicating that FOXC2 regulation may involve release of repression residing in this region. In construct RI␣1b B, however, basal levels were down to the level of the longest construct (RI␣1b), but the inductive effect of FOXC2 was not reconstituted. Interestingly, the elevated basal levels of the shortest constructs (RI␣1b C and RI␣1b Cϩ) were again higher, which indicates that repressive effects may reside in RI␣1b B as well, although not regulated by FOXC2.

Specific Binding of a Protein to an Upstream Region of the RI␣1b Promoter Is Abolished in the Presence of FOXC2
Expression-Specific binding of nuclear proteins from 3T3-L1 cells transfected with FOXC2 expression vector or the corresponding empty vector (basal) to a 32 P-labeled forkhead oligonucleotide or an oligonucleotide containing 100 bp of RI␣ exon 1A were examined (Fig. 3). Because FOXC1/FREAC3 and FOXC2 have identical DNA binding domains, we used the forkhead-binding site characterized for FOXC1 to detect FOXC2 by EMSA (9). Binding to the labeled forkhead consensus site was induced in  3A). However, this induction was not as profound as the increase in FOXC2 proteins levels observed by immunoblotting (Fig. 1), which can be explained by the fact that the labeled FOXC1 consensus probe is not specific for FOXC2 (9). An optimal DNA binding sequence for FOXC2 has not been determined. Furthermore, in extracts from FOXC2-transfected cells, we observed a small shift in complex I to a faster mobility compared with basal extracts. By supershift experiments using a human FOXC2 antibody, we identified FOXC2 as the protein forming complex I (lanes 7 and 13). Complex I binding to the labeled DNA fragment could only be displaced by the homolo-  These observations indicate that FOXC2 does not bind directly to the RI␣1b promoter. We next labeled oligos from the two responsive promoter regions located in exon 1a (100 bp) and exon 1b (72 bp). Although no changes in DNA-protein complex formation was observed with the 32 P-labeled RI␣ exon 1b oligo (not shown), the exon 1a probe that corresponds to the FOXC2responsive region formed a specific complex in 3T3-L1 nuclear extracts from untransfected cells (Fig. 3B, lane 2), which was strongly reduced by addition of the homologous unlabeled probe (lane 3). This specific DNA-protein complex was nearly abolished in extracts from 3T3-L1 cells transfected with FOXC2 for 24 and 48 h (lanes 4 and 5), which may indicate that FOXC2 is implicated in regulating release of a transcriptional repressor from the RI␣ promoter.
FOXD1 and FOXC2 Induce Reporter Activity Driven by the RI␣1b Promoter in Sertoli Cells-We next wanted to investigate if FOXD1, which is expressed in Sertoli cells of the testis, 3 regulated RI␣ levels through the same mechanism as in adipocytes. The level of transfection is much lower in Sertoli cell primary cultures (2-5% as opposed to 20 -30% in 3T3-L1 cells), and we detected ectopically expressed HA-tagged FOXC2 and FOXD1 with FITC-conjugated anti-HA antibody (Fig. 4A), showing that expression and localization is restricted to the nucleus in Sertoli cells. To study the effect of FOXD1 on the RI␣ promoter region, we co-transfected reporter constructs containing the RI␣1aϩ1b promoter, the 1a or 1b promoters alone, or a construct containing 4500 bp of the RII␤ promoter into rat Sertoli cell primary cultures together with expression vector for FOXD1 or the corresponding empty vector (Fig 4B). We also FIG. 3. A protein binding to the RI␣1b promoter is released following FOXC2 expression. EMSA was performed with a doublestranded forkhead oligonucleotide (5Ј-GATCCCTTAAGTAAACAG-CATGAGATC-3Ј) as the 32 P-labeled probe (A). Complex (Comp.) formation was analyzed using 2.5 g of nuclear extracts from 3T3-L1 cells transfected with FOXC2 expression vector (lanes 8 -13) or empty vector (Basal, lanes 2-7). Lane 1 shows probe in the absence of nuclear extract. 250-Fold excess of homologous unlabeled probe (ϩ, lanes 3 and 9), a forkhead oligo with altered flanking regions around the core sequence (5Ј-GATCCAGGCCGTAAACAGCATGAGATC-3Ј (mut, lanes 4 and 10), RI␣ exon 1a (lanes 5 and 11), and RI␣ exon 1b (lanes 6 and 12) were added as competitors. Supershifts with FOXC2 antibody (ϩ, lanes 7 and 13) were obtained by incubation of antibody for 1 h at room temperature. B, EMSA experiment with a 32 P-labeled 100-bp oligonucleotide from RI␣ exon 1a incubated with 2.5 g of nuclear extracts from 3T3-L1 cells transfected with FOXC2 expression vector (lanes 4 and 5) or the empty vector (lanes 2 and 3). Lane 1 is probe in the absence of nuclear extract. 250-Fold excess of homologous unlabeled probe was added as competitor in lane 3. Results shown are representative of 3 observations using extracts made from separate cultures.

FOXD1 and FOXC2 Regulate Expression of PKA/RI␣
tested if expression of FOXC2 had the same effect on the RI␣ promoter (Fig. 4C). In Sertoli cell cultures, reporter activity from the RI␣1b promoter construct was increased 8-fold in the presence of FOXD1 and 5-fold with FOXC2. The RI␣1aϩ1b construct was not similarly induced. The RI␣ and RII␤ constructs were also induced to a smaller extent by the presence of FOXD1 or FOXC2 (about 2-fold). Deletion of the exon 1a region (RI␣1b Aϩ, Fig. 4D) elevated basal activity 2-fold, whereas a further 3Ј deletion of the region downstream of transcription start (RI␣1b A) again reduced basal activity. Some activating effect of FOXD1 (3-fold) was restored in constructs RI␣1b A and Cϩ indicating that a downstream region of the promoter mediated activation of RI␣1b by FOXD1. The presence of both upstream-and downstream-responsive regions appeared to have the maximal effect on activity of the 1b promoter. In Sertoli cells, the activation mediated by the downstream region appeared to be more profound than the cis-acting repressor activity residing in the upstream promoter region.
A Truncated FOXD1 Mutant Reduces cAMP-dependent Induction of the RI␣ Gene in Sertoli Cells-Observing that the pattern of RI␣ regulation by FOXD1 in Sertoli cells mapped to the same downstream region that was identified as responsible for regulation by the cAMP pathway in our previous studies (27), we wished to examine if FOXD1 was implicated in cAMPdependent regulation of expression from RI␣1b. We expressed full-length FOXD1 as well as a FOXD1 mutant containing only the DNA-binding region and no transactivating domains, which by overexpression would displace endogenous FOXD1 by occupying Fox-binding sites (Fig. 5). The RI␣1b reporter construct (1 g/well) was transfected into primary rat Sertoli cells together with empty expression vector, the full-length FOXD1 expression vector, or the truncated FOXD1 vector (FOXD1-DBD) (0.5 g/well) and left untreated (open bars) or stimulated by 8-(4-chlorophenyl)thio-cAMP (100 M) for 28 h (filled bars). The induction of RI␣ by cAMP was ϳ6-fold in cells transfected with empty expression vector and 8-fold in the presence of full-length FOXD1. In the presence of FOXD1-DBD, however, cAMP-induced levels were reduced, which may indicate that FOXD1 is implicated in cAMP-mediated regulation of the RI␣1b promoter.

Protein Kinase B (PKB␣) Induces the Effect of FOXD1 in
Sertoli Cells-The insulin-activated serine/threonine kinase PKB/Akt (reviewed in Ref. 37) has been reported to phosphorylate and negatively regulate forkhead/winged helix transcription factors of the FOXO-family (AFX, FKHR, and FKHRL1) by inducing transport out of the nucleus (38 -42). Three highly homologous isoforms of PKB have been characterized and termed PKB␣/Akt1, PKB␤/Akt2, and PKB␥/Akt3 (43)(44)(45). To test if PKB affected FOXD1 and FOXC2 in our system, we transfected primary cultures of Sertoli cells with pCMV expression vectors for wild type PKB␣/Akt1 (0.5 g/well) together with the FOXD1 expression vector (0.5 g/well) and pCAT reporter constructs containing the RI␣1a or RI␣1b promoters (1.0 g/well). We found that PKB␣ induced expression from the RI␣1b promoter 15-fold in the presence of FOXD1 but had no significant effect when FOXD1 was absent (Fig 6A). No effect was observed on the RI␣1a promoter. In contrast, overexpression of PKB␣ had no effect in the presence of FOXC2 or FOXD1 in preconfluent or fully differentiated 3T3-L1 adipocytes (not shown). Furthermore, expression of PKB␣ did not affect the level of FOXD1 as shown by immunoblotting. When we expressed a constitutively active and membrane-bound myristoylated PKB␣ mutant (Fig. 6B, dark gray bars), we observed a reduced activation of RI␣1b compared with cells expressing wild type PKB␣ (black bars). This might indicate that the activating mechanism of PKB partly depends on detachment of PKB from the membrane. Immunofluorescence data (Fig. 6C) show that HA-tagged wild type PKB␣ is located throughout the cell, whereas myristoylated PKB␣ mutant is restricted to the plasma membrane region as expected. DISCUSSION In the present work, we have performed detailed studies on regulation of the RI␣ subunit of PKA by FOXC2 and FOXD1 in 3T3-L1 adipocytes and Sertoli cell primary cultures, and we found that both transcription factors induce expression from the RI␣1b promoter in both Sertoli cells and adipocytes, whereas the RI␣1a and RII␤ promoters were not similarly affected. The reasons for differentially regulated promoters and alternative splicing of leader exons in the RI␣ gene are not known, but one possible explanation is that the relative levels of RI␣ transcripts containing different 5Ј-UTRs create a finetuned control mechanism for rapid translation of RI␣ protein as a response to various signals. Sequences in the 5Ј-UTR may also affect subcellular localization.
Deletional mapping of the RI␣1b promoter in 3T3-L1 cells indicate that an upstream region of the 1b promoter (Ϫ400 to Ϫ300, part of exon 1a) represses basal transcription, and EMSA experiments show that FOXC2 mediates release of a factor from this region with subsequent activation of RI␣ expression. The responsive part of promoter 1b contains binding sites for the zinc finger transcriptional repressor Ikaros (IK1 and IK2) at positions Ϫ401, Ϫ392, and Ϫ315 (46,47). We also found a binding site for GATA1 at position Ϫ382, another zinc finger protein known to interact with CBP/p300. Recently, an atypical GATA protein, TRPS1, was identified as a transcriptional repressor (48). The IK2 and GATA1 sites in the RI␣ promoter are thereby potential repressor binding elements.
In Sertoli cells, a region downstream of the transcription start site is essential for the inductive effect of FOXD1/FOXC2. We have reported recently (27) that the RI␣1b promoter is induced 8-fold by cAMP at the post-transcriptional level, and the cAMP responsiveness is mapped to the same downstream region. Here we suggest that FOXD1 may partly and indirectly be involved in cAMP-mediated regulation through this region, based on the observation that a truncated FOXD1 lacking the transactivating region and functioning as an inhibitor of FOXD1 activity lowers cAMP-induced expression from the RI␣1b promoter. FOXD1 and FOXC2 both contain putative PKA phosphorylation sites in the DNA binding domain, identically positioned 5 amino acids downstream of the 15-amino acid N-terminal forkhead signature. This indicates the possibility of a direct regulation of FOXD1 by PKA in addition to induction of the FOXD1 transcript. 3 Another winged helix transcription factor is also reported to be regulated by the FSH signal in Sertoli cells (49).
In our study, we found that overexpression of the insulinactivated kinase PKB␣/Akt1 induced the FOXD1-mediated activation of RI␣1b in Sertoli cells but did not affect regulation by FOXD1 or FOXC2 in preconfluent or fully differentiated 3T3-L1 cells. There is a possibility that different PKB isozymes mediate this effect in these cells. PKB␣/Akt1 seems to regulate spermatogenesis with no effect on glucose tolerance and insulin sensitivity (50,51), whereas PKB␤/Akt2 was identified as the PKB isoform required for insulin to maintain normal glucose homeostasis (52,53).
The FOXO family and several other transcription factors are regulated by direct phosphorylation by PKB/Akt. However, no consensus phosphorylation sites for PKB were found in the FOXC2 or FOXD1 sequences. Thus, we anticipate that a yet unknown factor mediates the effect of PKB␣ on FOXD1 or that FOXD1 and PKB␣ act in concert on downstream effectors that regulate RI␣. Studies on the dynamics of PKB following activation in B-cells show that shortly after activation, PKB is distributed throughout the cytosol and nucleus (54). We found that ectopically expressed FOXD1 and FOXC2 are strictly localized to the nucleus, and a direct interaction between PKB and FOXD1 should thereby depend on the ability of PKB to dissociate from the membrane and travel to the nucleus. This is supported by the observation that a myristoylated, constitutively membrane-bound PKB mutant has a reduced effect on FOXD1-mediated induction of RI␣1b.
The effects of insulin and PKB on translation is well accounted for (55) and facilitates translational initiation of mRNAs containing strong cap-proximal secondary structures (56). We have reported recently (27) that the 5Ј-UTR of RI␣1b mRNA, in contrast to RI␣1a mRNA, contains a strong stemloop and is regulated at the post-transcriptional level by cAMP in Sertoli cells, an effect that appeared to be cell-specific. The observation that both PKA and PKB appear to signal through FIG. 7. Tuning of cAMP sensitivity, metabolism, and reproductive function by integration of signals in adipocytes and Sertoli cells. Stimulation of the ␤-adrenergic receptor (␤-AR) by catecholamines (CAT) in adipocytes or the follicle-stimulating hormone receptor (FSHR) by FSH in Sertoli cells leads to increased levels of cAMP and activation of PKA. In both cells, PKA phosphorylates and activates hormone-sensitive lipase leading to enhanced lipolysis. In adipocytes, PKA signaling up-regulates levels of uncoupling protein-1 (UCP1) that contributes to increase energy expenditure. In Sertoli cells, PKA activity is essential for expression of several factors vital for spermatogenesis (e.g. androgen-binding protein, ABP). Furthermore, cAMP induces levels of the RI␣ regulatory subunit of PKA in both cell types and of the winged helix transcription factors Foxc2 in adipocytes and FoxD1 in Sertoli cells. Conversely, Fox proteins increase RI␣ levels by regulating the RI␣1b promoter. Insulin (Ins) and tumor necrosis factor ␣ (TNF␣) in adipocytes and various growth factors (GrF) in Sertoli cells signal through corresponding receptors (insulin receptor (IR), TNF receptor (TNFR), growth factor receptor (GrFR)) and phosphatidylinositol 3Ј-kinase (PI3K) to activate PKB and also to induce the levels of Foxc2 mRNA in adipocytes. In Sertoli cells, PKB␣ acts in concert with FoxD1 to further activate RI␣ levels, and possibly a similar function of PKB␤ may be found in adipocytes. In addition to this, stimulation of PKB activity by FSH has also been reported in Sertoli cells. Taken together, similar mechanisms regulate RI␣ expression and thereby increase the sensitivity of PKA to cAMP in both cell types (marked by bold arrow with *). FOXD1 and the RI␣1b 5Ј-UTR in regulation of RI␣ in Sertoli cells could indicate that a post-transcriptional regulatory mechanism may be involved.
A PKA-independent, FSH/cAMP-mediated activation of PKB, mimicking the insulin-like growth factor I response, has been reported in granulosa cells (57). This activation was found to involve cAMP-regulated guanine exchange factors that may activate phosphatidylinositol 3Ј-kinase. Other reports have suggested that PKA may activate PKB in a phosphatidylinositol 3Ј-kinase-independent manner (58). Our observations that activation of the RI␣1b promoter by FOXD2 is significantly elevated in the presence of both PKB and cAMP may implicate that cAMP activates PKB by PKA-dependent or PKA-independent mechanisms. Both PKA-dependent and PKA-independent cAMP-signaling pathways may stimulate proliferation in several cell types (59).
Sertoli cells and adipocytes are both highly responsive to hormonal stimuli and share many common features, such as glucose uptake, lipid storage, as well as hormone-sensitive lipase activity. Whereas adipocyte metabolism is regulated by catecholamines and sympatic tonus through the ␤-adrenergic receptors, mature Sertoli cells support and control spermatogenesis in response to the pituitary follicle-stimulating hormone (FSH) testosterone, insulin, and insulin-like growth factors I/II (60,61). Apparently, downstream pathways are parallel in the two cell types and involve cAMP, PKA, as well as Fox family members. This in turn tunes cAMP sensitivity by up-regulation of RI␣ levels (Fig. 7).
In conclusion, FOXD1 and FOXC2 both induce the RI␣1b promoter to elevate RI␣/PKAI levels. In adipocytes the effect is mainly mediated through an upstream promoter region and the possible release of a repressor, whereas in Sertoli cells, FOXD1 primarily activates through a downstream region and mediates signals from the cAMP/PKA and PKB signaling pathways.