Ikaros Modulates Cholesterol Uptake: A Link between Tumor Suppression and Differentiation

The zinc finger transcription factor Ikaros was initially described as recognizing regulatory sequences of lymphoid-restricted genes (1–3). It influences hematopoietic stem cell lineage commitment, as evidenced by mouse models of Ik1 deficiency (4, 5). This founding member of a family of zinc finger DNA-binding proteins is implicated in chromatin remodeling (6). The Ikaros gene contains seven exons that, by alternative splicing, give rise to eight known isoforms (7). These isoforms differ in the number of NH₂-terminal zinc finger DNA-binding motifs, yielding members with variable DNA-binding properties. Isoforms 1 and 3 contain the requisite three domains that confer high-affinity binding to the Ikaros-specific DNA sequence GGAAA in promoters of target genes (3, 8). All Ikaros isoforms share a common COOH terminus with two zinc finger motifs that are required for heterodimerization or homodimerization and for interactions with other proteins. Homodimers and heterodimers among DNA-binding isoforms increase their affinity for DNA, whereas heterodimers between DNA-binding isoforms and non–DNA-binding isoforms are unable to bind DNA (4).

Ikaros is also expressed in the pituitary with highest expression during fetal and early postnatal development (9). In hormone-producing pituitary corticotroph cells, Ikaros binds the pro-opiomelanocortin promoter to activate the endogenous gene (10). Disruption of the Ikaros gene results in contraction of the pituitary corticomelanotroph population, reduced circulating adrenocorticotrophic hormone levels, and adrenal glucocorticoid hormone insufficiency (10).

Altered expression of Ikaros has been implicated in neoplasia. Ik^+/− mice exhibit hyperactive T-cell proliferation and develop leukemias and lymphomas (11, 12). Transgenic animals expressing the dominant-negative non–DNA-binding Ik6 develop T-cell lymphoproliferative disorders (11). In human neoplasia, Ik6 expression has been identified in a third of human T-cell leukemias (13–15). Ik6 is implicated in pituitary tumorigenesis by acting on fibroblast growth factor receptor 4 (FGFR4) transcription (9, 16). Consistent with its identification in pituitary tumors, the non–DNA-binding Ik6 isoform promotes pituitary cell survival in vitro and in vivo through enhanced antiapoptotic activity by inducing Bcl-XL transcription (17).

In this study, we examined targets of Ikaros action that could modulate cell differentiation, metabolism, and proliferation. Using cDNA microarray, we uncovered mediators of cholesterol uptake, including the low-density lipoprotein receptor (LDL-R) and sterol regulatory element-binding protein 2 (SREBP2), as targets of Ikaros action. Our data implicate Ikaros in modulating differentiation and maturation of neuroendocrine cells, at least partly through LDL-R–mediated cholesterol uptake.

Cell culture and pituitary specimens. Mouse corticomelanotroph AtT20 cells were propagated in DMEM (Life Technologies, Inc.) with high glucose supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich), 2 mmol/L glutamine, 100 IU/mL penicillin, and 100 μg/mL streptomycin. Human embryonic kidney (HEK) 293 cells were propagated in MEM (Life Technologies) supplemented with 10% horse serum (Sigma-Aldrich), 1.0 mmol/L sodium pyruvate, 100 IU/mL penicillin, and 100 μg/mL streptomycin. Cells were transfected with Ik1, Ik6, or their empty vector pcDNA, and stable clones were generated (10).

For LDL-R and SREBP2 protein expression by Western blotting as well as LDL uptake by fluorometry, cells were incubated either in DMEM as above with 10% FBS containing cholesterol (CC) or in 10% lipoprotein-deficient, cholesterol-free (CF) FBS (Intracel).

Normal human pituitaries with no morphologic abnormalities were obtained from patients with no endocrine abnormalities and autopsied within 12 h of death. Human pituitary tumors were obtained at surgery following informed consent and approval by the Ethics Committee of the University Health Network and classified according to current Armed Forces Institute of Pathology and WHO criteria (18, 19). The adenoma samples examined included six somatotroph adenomas, four lactotroph adenomas, four oncocytic null cell adenomas, three gonadotroph adenomas, and three corticotroph adenomas.

RNA extraction. RNA was extracted using Trizol (Life Technologies). The integrity of RNA was monitored on agarose gels. Before reverse transcription, RNA samples were purified using an RNeasy kit (Qiagen, Inc.) to eliminate genomic DNA contamination. Phosphoglycerate kinase mRNA transcripts were coamplified to verify integrity of RNA (20).

Oligonucleotide microarray analysis. Total RNA was extracted from two independent stably transfected clones expressing Ik1, Ik6, or their empty control vector (pcDNA; ref. 10).

Hybridization to the Affymetrix Mouse Genome 430A 2.0 Array was conducted at The Centre of Applied Genomics (Hospital for Sick Children, Toronto, Ontario, Canada). Sample RNA in vitro transcription, labeling, and hybridization used standard Affymetrix protocols. Hybridized chips were washed and scanned on an Affymetrix GeneChip 3000 confocal scanner. Raw microarray data were analyzed by ArrayAssist software (Iobion) using the PLIER algorithm. Genes were considered to be differentially expressed if the signal changed at least 2-fold (signal log2 ratio, 1) and statistical analysis was performed with Spotfire DecisionSite software package.

Real-time quantitative reverse transcription-PCR. Expression of LDL-R was analyzed by real-time quantitative reverse transcription-PCR (qRT-PCR) using the ABI 7500 Real-Time PCR System (Applied Biosystems). cDNA was generated using the Taqman Reverse Transcription Reagent kit (Applied Biosystems). The cDNA was amplified in 25 μL of PCR buffer with IQ SYBR Green Master Mix (Bio-Rad Laboratories) and intron-spanning primers for LDL-R (+)5′-ATGAGTCCCCAGAGACATGC-3′ and (−)5′-CGAGATTTGCCTAGATGGCTC-3′ (product, 215 bp). Thermocycling conditions were optimized as initial denaturation (95°C for 3 min), amplification (40 cycles: 95°C for 15 s, 55°C for 20 s, 72°C for 30 s), and melting curve analysis (95°C for 3 min, 60–95°C, increasing 0.5°C every 10 s). Standard curves were run as 10× serial dilutions from amplified target cDNA [quantified using PicoGreen dsDNA (Molecular Probes)] over the inclusive range 1 × 10⁷ to 10 copies. Estimates were computed from C_T (cycle threshold) and normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in the same sample. The quality of the real-time PCR product was monitored by agarose gel electrophoresis and specificity by the melting temperature. PCR efficiency exceeded 85% as calculated from the slope of the standard curves (r > 0.95). The median precision estimated from absolute copy number was 7.9% (intraassay) and 11.5% (interassay). All results were obtained from at least two independent experiments from two different clones.

Western blotting. Cultured cells were lysed in radioimmunoprecipitation assay lysis buffer with proteinase inhibitors. Protein concentration of samples and bovine serum albumin (BSA) standard were determined using the Bio-Rad protein assay kit by measurement of absorbance at 595 nm. Equal amounts of protein (40 μg) solubilized in sample buffer were separated on 10% SDS-polyacrylamide gels and transferred electrophoretically to polyvinylidene difluoride membranes. Membranes were blocked in TBS containing 0.5% Tween 20 (TBS-T) plus 5% nonfat dried milk for 1 h at room temperature and probed with primary antibodies at 4°C overnight.

Primary antibodies were used at the specified dilutions: LDL-R (chicken polyclonal, 1:1,000; Abcam), SREBP2 (rabbit polyclonal, 1:1,000; Abcam), 3-hydroxy-3-methylglutaryl CoA reductase (HmGCoAR; rabbit polyclonal, 1:1,000; Upstate Biotechnology, Inc.), adrenocorticorticotropic hormone (ACTH; chicken polyclonal, 1:5,000; Abcam), Ikaros (mouse monoclonal 4E9, 1:1,000; kindly provided by Dr. K. Georgopoulos, Harvard Medical School, Boston, MA), and actin (mouse monoclonal, 1:1,000; Sigma-Aldrich). Membranes were washed thrice for 10 min each in TBS-T and incubated with horseradish peroxidase–conjugated goat anti-rabbit (1:2,000; Santa Cruz Biotechnology), anti-chicken (1:1,000; Abcam), or anti-mouse (1:2,000; Santa Cruz Biotechnology) secondary antibody for 1 h at room temperature and visualized using enhanced chemiluminescence detection (Amersham).

Transfection and luciferase assay. Expression vectors encoding Ik1 (CDM8-1) or Ik6 (CDM8-6) were provided by Dr. K. Georgopoulos.

The human LDL-R promoter (−1,563 to −58) was a gift of Dr. Sato (University of Tokyo, Tokyo, Japan; ref. 21). Nucleotide +1 is assigned to A of the translation initiation codon (ATG). This convention is used because multiple transcription start sites located between positions −93 and −79 have been identified; thus, +1 cannot be used to refer to a single site of transcription initiation (22). The 1,507-bp promoter (−1,563 to −58; pLDLR-Luc 1507) and a 329-bp construct (−387 to −58; pLDLR-Luc 329) generated by restriction enzyme digestion with Hpa1 were placed into the basic pGL3 vector (Promega). The shorter construct contains two potential Ikaros-binding sites (−71/−76; −241/−246); the 1,507-bp promoter encompasses one additional potential Ikaros-binding site (−1,004/−1,009). The proximal region of the LDL-R promoter contains a SRE-1 for SREBP and is responsible for its up-regulation by SREBP, whereas the distal promoter may be important for sterol-independent regulation of LDL-R expression (23).

Plasmid reporters were prepared by column chromatography (Qiagen) for sequencing and transfection. Cells were plated into six-well dishes (7 × 10⁵ per well) and transfected with 5 μL/well Lipofectamine (Invitrogen) and 1 μg of either the short or long LDL-R vector. Transfection efficiency was monitored by simultaneous cotransfection with a β-galactosidase control expression plasmid (cytomegalovirus)-β-galactosidase (20 ng/well). Forty-eight hours after transfection, cells were lysed in buffer containing 25 mmol/L glycylglycine, 15 mmol/L MgSO₄, 4 mmol/L EDTA, 1% Triton X-100, and 1 mmol/L DTT. Luciferase activity was measured for 20 s in a luminometer. Promoter activity was expressed as relative firefly luciferase/β-galactosidase activity compared with pcDNA. Each experiment was performed in triplicate on three separate occasions.

Chromatin immunoprecipitation assay. Histone was cross-linked to DNA by addition of 37% formaldehyde to cells in culture (20, 24) that were washed with cold PBS containing protease inhibitors (Roche), pelleted, and resuspended in SDS lysis buffer containing inhibitors. Lysates were sonicated on ice to release 200 to 1,000 bp DNA fragments. After centrifugation, cell suspensions were diluted and 20 μL of each sample were retained for PCR of DNA input. The remaining lysate was cleared with salmon sperm DNA/protein G-agarose beads. Immunoprecipitation was performed using anti-Ikaros (4E9) or anti-methyl-histone 3 Lys⁹ [MeH3 (K9)], anti-acetyl-histone 3 Lys⁹ [AcH3 (K9)], anti-acetyl-histone 3 (AcH3), or acetyl-histone 4 (AcH4) antibodies (Upstate Biotechnology) incubated overnight at 4°C with rotation. Negative controls included either omission of antibody or an anti-IgG antibody. For PCR analysis, the histone-DNA cross-links of eluates and of input samples were reversed at 65°C, immunocomplexes were digested with proteinase K for 1 h at 50°C, and DNA was purified by phenol extraction. Purified DNA was resuspended in 20 μL of H₂O for PCR using LDL-R promoter-specific primers (+)5′-CCCTAGTACTGGGAATGACTCTGG-3′ and (−)5′-GGTGCTCATCCTTAGCTTCTGC-3′, which encompass two potential Ikaros-binding sites. Experiments were performed on three independent occasions.

Fluorometry. LDL and transferrin uptake were assayed according to a previously described method (25) using two different clones of each group. Cells were plated into six-well dishes (4 × 10⁵ per well) and incubated for 48 h in CC or CF medium. As a positive control and to show the range of cholesterol uptake, we also serum starved a subset of cells for 6 h and then added FGF1 (10 μg/mL final concentration) + heparin (1 μL/mL) for 18 h. Cells were washed and incubated for 5 h (established by a time course analysis) at 37°C with DiI-LDL in CF medium (10 μg/mL final concentration; Molecular Probes). After the incubation period, cells were washed in 3% BSA and in PBS and then lysed in 500 μL of 1% SDS and 1 mol/L NaOH. Fluorescence intensity of each sample was quantified by SpectraMax M5 (excitation λ at 530 nm, emission λ at 580 nm; Molecular Devices) and normalized to protein concentration (Bio-Rad detergent-compatible protein assay) and to the autofluorescence of the lysing medium. The results are expressed as percent change compared with controls. As a negative control for cholesterol uptake, cells were incubated with Alexa Fluor 488-transferrin (50 μg/mL final concentration; Molecular Probes) and processed as above. Fluorescence intensity was measured with excitation λ at 495 nm and emission λ at 519 nm by SpectraMax M5.

Immunofluorescence. Cells were cultured on chamber slides in CC or CF medium for 48 h and then incubated with DiI-LDL (10 μg/mL) for 15 min at 37°C. Cells were washed thrice in 3% BSA and thrice in PBS and then fixed in 4% paraformaldehyde for 20 min at room temperature. Cells were washed with PBS for 5 min, permeabilized with 0.5% Triton X-100 in PBS for 5 min, blocked with 3% BSA in PBS for 15 min, and washed again in PBS. Fixed cells were then incubated with primary antibody [chicken anti-LDL-R (1:500) or mouse anti-Ikaros (1:1,000)] or nonimmune control IgG in 0.1% BSA in PBS for 1 h at 37°C. Coverslips were washed thrice in PBS, incubated with secondary antibody (either anti-chicken Alexa Fluor 488 or anti-mouse Alexa Fluor 488, 1:1,000; Molecular Probes) for 30 min at 37°C, washed thrice in PBS, and mounted in 4′,6-diamidino-2-phenylindole (DAPI)-Vectashield (1:1; Vector Laboratories). Microscopy was performed on a Zeiss LSM510 using a 63×/1.2 numerical aperture water immersion objective lens.

Electron microscopy. Lightly centrifuged cell pellets were fixed in 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 mol/L phosphate buffer (pH 7.4) for 2 to 4 h followed by thorough washing with phosphate buffer and storage at 4°C in PBS containing 10 mmol/L azide. Cell pellets were embedded in 15% gelatin followed by infusion with 2.3 mol/L sucrose. Samples were then mounted on aluminum pins and frozen in liquid nitrogen, and ultrathin cryosections were prepared and transferred to copper grids and then stained with saturated uranyl acetate for 15 min and lead citrate for 5 min. The grids were examined in a JEOL JEM 1230 transmission electron microscope (JEOL USA, Inc.), and images were acquired with an AMT charge-coupled device digital camera (AMT Corp.).

Pituitary corticotroph function of LDL-R knockout (LDL-R^−/−) mice. LDL-R^−/− mice backcrossed at least 10 generations into the C57BL/6 strain were kindly provided by Dr. M. Cybulsky (University Healthcare Network, Toronto, Ontario, Canada; ref 26). At postnatal day 6, LDL-R^−/− and their age- and sex-matched control C57BL/6 mice were sacrificed by decapitation for serum collection. ACTH was measured as described previously (10).

Statistical analysis. All data are presented as mean ± SE. Significance was assessed by ANOVA followed by appropriate t test using SAS/STAT software. P < 0.05 was considered significant.

Ikaros regulates the expression of factors crucial for cellular cholesterol homeostasis in AtT20 pituitary corticotroph cells. To identify factors that are targeted by Ikaros in pituitary cells, we performed cDNA microarray analysis comparing corticotroph AtT20 cells transfected with Ik1, Ik6 (non–DNA-binding isoform), and their empty vector (pcDNA) using two independent clones of each. We identified a significant 9- to 22-fold up-regulation of LDL-R, a significant 2- to 5-fold up-regulation of SREBP2, and significant 2-fold up-regulation of HmGCoAR in AtT20 cells overexpressing Ik1 compared with controls and those overexpressing Ik6 (Supplementary Table S1). These factors are crucial for cellular cholesterol homeostasis, as SREBP2 regulates LDL-R expression in response to intracellular cholesterol levels to either increase or decrease uptake of extracellular cholesterol, whereas HmGCoAR is the rate-limiting enzyme in intracellular cholesterol synthesis and is also regulated by SREBPs in response to intracellular cholesterol levels (27, 28).

We focused on LDL-R and validated the oligonucleotide microarray findings by qRT-PCR. This confirmed a significant 4-fold increase of LDL-R mRNA levels in cells overexpressing Ik1 compared with controls and cells overexpressing Ik6 (Fig. 1). We confirmed mRNA expression of LDL-R in human pituitary samples with stronger expression in tumors (Supplementary Fig. S1). LDL-R protein levels, determined by Western blotting, also revealed up-regulation by Ik1, but not Ik6, in cells grown in CC medium (Fig. 2A).

To determine if pituitary corticotroph cells are able to respond predictably to an increased demand for cholesterol following lipid depletion, we examined the LDL-R response under CF conditions. AtT20 cells transfected with Ik1 or empty vector (pcDNA) responded to lipid depletion by increasing LDL-R protein levels (Fig. 2A). In contrast, cells transfected with Ik6 failed to mount a LDL-R response to cholesterol depletion (Fig. 2A). Cells overexpressing Ik1 showed the most robust up-regulation of LDL-R. Conversely, inhibitory conditions using medium containing high cholesterol levels resulted in down-regulation of the LDL-R (data shown for Ik1-overexpressing cells).

Expression of SREBP2 protein revealed parallel changes in response to lipid depletion and down-regulation by incubation with high sterol-containing medium (data not shown). Ik1 cells showed the highest expression of the 120-kDa SREBP2 protein; Ik6 cells had the lowest levels of expression (Fig. 2B). HmGCoAR expression exhibited the same pattern and response to lipid depletion (Fig. 2C).

To determine whether Ikaros mediates cholesterol uptake, we examined the expression of Ikaros itself following lipid depletion (Fig. 2D). We observed that levels of Ik1 but not Ik6 increased in empty vector (pcDNA)-transfected cells in response to lipid depletion. Confocal microscopy to localize Ikaros also identified increased Ikaros in a cytoplasmic location consistent with active synthesis during cholesterol depletion (CF), consistent with up-regulation of endogenous Ikaros expression in control cells (Fig. 2D , bottom). These data suggest a role for Ikaros as a mediator of LDL-R and cellular cholesterol homeostasis.

Ikaros regulates LDL-R gene expression by binding and activating the LDL-R promoter and through histone modifications. To establish that Ikaros is able to induce LDL-R gene transcription, we performed luciferase reporter assays using the LDL-R promoter (−1,563 to −58) fused to a pGL3 basic vector. Because most cis-acting elements, including SRE-1 and Sp1, are present between −234 and −58 bp, we transfected AtT20 cells overexpressing Ik1 or Ik6 with two different LDL-R promoter-luciferase constructs, pLDLR-Luc 329 and pLDLR-Luc 1507. Ik1 induced both pLDLR-Luc 329 and pLDLR-Luc 1507 activity compared with controls (Fig. 3A). Similarly, transient cotransfection of Ik1 in HEK 293 cells, which do not endogenously express Ikaros, resulted in activation of pLDLR-Luc 329 and pLDLR-Luc 1507. These effects were not mimicked by Ik6 (Fig. 3B).

Binding of Ik1 to two potential GGGAA binding sites in the proximal LDL-R promoter was confirmed by chromatin immunoprecipitation (ChIP) assays (Fig. 3C).

Ikaros is an integral component of a functionally diverse chromatin remodeling network and is known to interact with nucleosome remodeling and deacetylation (NURD) complexes in a gene-specific manner (29, 30). We therefore examined whether histone modifications are important for Ikaros-mediated control of the LDL-R using ChIP assays. Forced expression of Ik1 in AtT20 cells resulted in decreased methylation of histone H3 (Lys⁹) and increased acetylation of histone H3 (Lys⁹), whereas no significant changes were observed for overall acetylation of histones H3 and H4 (Fig. 4A–D). These findings are consistent with the positive effect of Ikaros on LDL-R gene expression.

LDL endocytosis is regulated by Ikaros. The ability of hepatocytes to endocytose fluorescent or radioactive LDL is well established. However, the role of LDL uptake in nonsteroidogenic endocrine cells has only been examined in one report of rat mammosomatotroph GH3 cells (31). To determine whether AtT20 cells can take up cholesterol via receptor-mediated endocytosis and whether Ikaros influences endocytosis of LDL, we examined the binding and uptake of LDL particles labeled with the fluorophore DiI. AtT20 cells stably transfected with Ik1, Ik6, or empty vector were grown in CF or CC medium. A third group of cells was serum starved for 6 h and then incubated with FGF1 before incubation with DiI-LDL. This ligand was selected based on its recognized ability to enhance LDL uptake in hepatic and arterial smooth muscle cells through induction of LDL-R transcription (32) and the fact that pituitary cells express FGFRs (9, 20, 33). LDL uptake by AtT20 cells was significantly increased by ∼2-fold following cholesterol depletion and by ∼3-fold by FGF1, showing that AtT20 cells can respond to known stimuli of LDL uptake (data not shown). We then compared LDL uptake in cells overexpressing Ik1 or Ik6 and controls (Fig. 5, top). LDL uptake was significantly increased in cells overexpressing Ik1 in CC and CF medium (1.5- and 1.7-fold, respectively). LDL uptake in the presence of cholesterol was not significantly different in cells overexpressing Ik6 compared with controls and was slightly reduced by Ik6 in CF medium. When treated with FGF1, cells overexpressing Ik1 showed a marked increase in LDL uptake compared with controls (2-fold increase), whereas cells forced to express Ik6 showed an abrogated response to FGF stimulation.

To confirm that Ik1 specifically up-regulates LDL uptake and does not affect the general endocytosis machinery, we assayed binding and uptake of fluorescent transferrin. There was no change of transferrin uptake in cells overexpressing Ik1 or Ik6 compared with controls (data not shown).

We also examined LDL endocytosis by confocal microscopy. AtT20 cells were incubated with DiI-LDL in CF medium. Cells overexpressing Ik1 showed increased uptake of cholesterol, as evidenced by the detection of the fluorophore throughout the cytoplasm (Fig. 5, middle). Moreover, LDL-R in cells overexpressing Ik1 showed strong cytoplasmic as well as enhanced membranous staining (Fig. 5, bottom).

Ik1 induces ultrastructural changes reflective of enhanced biosynthetic activity in AtT20 cells. We analyzed AtT20 cells by electron microscopy for morphologic evidence of changes in response to altered cholesterol uptake. Cells overexpressing Ik1 showed more numerous secretory granules, more abundant rough endoplasmic reticulum, and larger Golgi complexes (Fig. 6) compared with control cells (data not shown) and with cells overexpressing Ik6, which displayed more abundant mitochondria (Fig. 6). These findings and our previous observations in Ikaros-null mice (10) are consistent with a differentiating role for Ikaros in pituitary neuroendocrine cells that require abundant lipid-rich membrane structures for hormone synthesis and packaging. To determine if these changes that reflect increased cholesterol integration into subcellular membranes correlate with cholesterol uptake, we examined the effect of cholesterol depletion on the same ultrastructural variables. Control cells incubated in CF medium had an increased number of mitochondria but less endoplasmic reticulum, smaller Golgi complexes, and fewer secretory granules than cells incubated in CC medium (data not shown). The same response was seen in cells expressing Ik1 but not in cells expressing Ik6 that already exhibited these changes.

Loss of LDL-R in LDL-R knockout mice (LDL-R^−/−) impairs pituitary corticotroph ACTH secretion. To determine the physiologic relevance of our findings on the role of the LDL-R in pituitary corticotroph function, we examined LDL-R^−/− mice. Quantification of ACTH in the serum of LDL-R^−/− mice (n = 10) showed a reduction in circulating ACTH levels by 25% compared with age- and sex-matched wild-type C57BL/6 mice (n = 12), a finding that was statistically significant with P < 0.005. These findings are the first to support a role for the LDL-R in differentiated pituitary cell function.

We present here the first evidence that Ikaros, the hematopoietic zinc finger stem cell factor with chromatin remodeling properties, also plays an important role in the regulation of a metabolic function by altering cellular cholesterol homeostasis. We show that this action is mediated through the LDL-R and related cholesterol uptake genes in pituitary corticotroph cells.

Although the LDL-R is known to regulate hepatic uptake of CC LDL particles from plasma by endocytosis, its role in extrahepatic cells is less well appreciated. Cholesterol homeostasis is vital for cellular structure and function because cholesterol is a major component of cell wall membranes and cellular organelles (34). In malignant cells, it has been noted that increased cell proliferation parallels increased LDL-R expression (35–37). Conversely, it has been shown in human hepatoma cells that Raf-1/mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase kinase/p42/44^MAPK uncouples regulation of LDL-R induction from cell proliferation, suggesting that the two processes can be dissociated in neoplasia (38). The growth-independent regulation of LDL-R expression is consistent with the induction of LDL-R promoter activity by cytokines through the p42/44^MAPK pathway without affecting cellular proliferation or DNA synthesis (39, 40). Furthermore, different members of the LDL-R family have important functions in modulating synaptic plasticity, are necessary for hippocampus-specific learning and memory, and are implicated in several human neurodegenerative and neuropsychiatric disorders (41). Indeed, LDL-R–null mice display impaired spatial memory associated with less differentiated CA1 hippocampus neurons and a decreased synaptic density in the hippocampus (42). As shown here, LDL-R–deficient mice also exhibit impaired pituitary ACTH secretion.

In the pituitary, estrogen can regulate LDL-R expression and LDL uptake in GH3 mammosomatotrophs, providing a potential mechanism for hypertrophic and proliferative responses of pituitary cells to this hormone (31). It was, therefore, intriguing that microarray analysis of AtT20 cells overexpressing Ik1 identified increased mRNA expression of not only the LDL-R but also other related genes involved in cellular sterol metabolism, including SREBP2 and HmGCoAR. We confirmed that LDL-R mRNA expression was increased by qRT-PCR, and we verified that LDL-R, SREBP2, and HmGCoAR proteins were increased in AtT20 cells overexpressing Ik1, whereas they were decreased in cells overexpressing the dominant-negative Ik6 isoform. LDL-R, SREBP2, and HmGCoAR protein levels increased in response to cholesterol depletion, indicating a physiologic response of pituitary cells to cholesterol deficiency. Interestingly, we also found an increase of Ikaros itself in response to cholesterol depletion, raising the possibility of cellular nutrient-induced regulation of Ikaros expression.

We show that Ik1, a DNA-binding Ikaros isoform, induced LDL-R promoter activation by binding the LDL-R promoter, whereas Ik6, a non–DNA-binding Ikaros isoform, did not. We transfected two LDL-R promoter constructs, a 329-bp construct that contains the important cis-acting elements, including SRE-1, which are sufficient for expression as well as negative regulation by steroids, and the 1,507-bp construct, which is considered the full-length promoter and also contains repressive cis-acting elements (43, 44). Two potential Ikaros-binding sites are located in the shorter construct, whereas the long construct contains one additional potential Ikaros-binding site. Both constructs were activated by Ik1, but the relative activation was more pronounced in the 1,507-bp construct, suggesting that Ik1 interacts with binding sites in the 329-bp region as well as more proximally.

As Ikaros has been shown to interact with NURD complexes targeting them to DNA by its DNA-binding capacity, we analyzed the effect of Ik1 and Ik6 on histone acetylation and methylation at the LDL-R promoter. It is important to note that the Ikaros-NURD complexes are active in both chromatin remodeling and histone deacetylation in differentiating thymocytes and mature T cells and that Ikaros, via its ability to recruit NURD complexes, can repress the expression of genes that promote myeloid differentiation (29, 30). However, gene expression studies on Ikaros-null lymphocytes have revealed that a significant number of genes is also up-regulated; subsequently, it was shown that Ikaros can also function as a potentiator of gene expression through recruitment of the Swi/Snf chromatin remodeling complexes (45, 46). Our findings that Ik1 enhances acetylation and diminishes methylation at H3 (Lys⁹) are consistent with the observed induction of LDL-R gene transcription and with Ikaros acting as a potentiator of this gene. In pituitary GH4 mammosomatotrophs, Ik1 induces transcription of the Prolactin gene by acetylating histone H3 and facilitating binding of the pituitary transcription activator-1 to the proximal promoter (47).

To show the functional consequences of increased LDL-R expression induced by Ik1, we studied LDL binding and uptake by fluorometry. We established that these nonsteroidogenic AtT20 pituitary corticotroph cells are able to increase their LDL uptake to two well-known stimuli: incubation with CF medium leading to depletion of intracellular lipids and incubation with FGF1 (32, 48). Cells overexpressing Ik1 showed a 53% increase of LDL uptake in CC medium and a 67% increase in CF medium compared with controls, indicating that Ik1 is able to increase LDL uptake under conditions of intracellular cholesterol availability as well as under conditions of cellular cholesterol depletion. As Ik1 is also able to induce SREBP2, the major contributor to sterol-dependent regulation of LDL-R, and as Ikaros expression itself increases in response to cholesterol depletion, it can be hypothesized that Ikaros mediates transcriptional regulation of the LDL-R gene in response to intracellular cholesterol availability. Conversely, the presence of the dominant-negative Ik6 isoform abrogated the response to cholesterol depletion and FGF1 stimulation, further supporting the pivotal role for Ikaros in modulating lipid uptake.

Cells overexpressing Ik1, in contrast to the tumor-derived, dominant-negative Ik6, do not show increased proliferation and tumor progression (17). This indicates that Ik1-mediated regulation of cholesterol is not linked to DNA synthesis. AtT20 cells overexpressing Ik1 do, however, produce more ACTH (10). We hypothesized, therefore, that increased LDL uptake in Ik1 cells allows increased differentiated function rather than proliferation. The ultrastructural findings characterize a more active endocrine phenotype with increased rough endoplasmic reticulum, larger Golgi complexes, and more secretory granules in cells overexpressing Ik1. In contrast, cells expressing Ik6 display reduced biosynthetic organelles and more numerous mitochondria (17). Similarly, Ikaros has been recognized to exert more profound effects on differentiation than proliferation of lymphocyte precursors (5, 11).

In summary, our results indicate that Ikaros plays an important role in LDL-R regulation and function in pituitary corticotrophs. Ik1 up-regulates LDL-R through increased promoter activity mediated by diminished methylation and enhanced acetylation of histone H3 (Lys⁹) and increased acetylation of histone H3 (Lys⁹). By fluorometry and immunofluorescence, we show that Ik1-mediated LDL-R protein levels promote LDL uptake. These actions are not associated with increased cell proliferation but are more in keeping with a hormonally active phenotype. Furthermore, two other genes important in cellular cholesterol homeostasis, SREBP2 and HmGCoAR, are up-regulated by Ik1, implicating Ikaros in the regulation of multiple diverse cellular metabolic functions. In accordance with a role for the LDL-R and LDL uptake in pituitary differentiation, analysis of LDL-R^−/− mice revealed diminished ACTH secretion.

Our findings raise the possibility that Ikaros exerts its effects on cell differentiation, maturation, and function at least partly through regulation of cellular cholesterol homeostasis. Given the therapeutic potential of cholesterol-lowering medications in Ikaros-dependent disorders, including leukemias (49) and pituitary tumors (50), the current findings unmask a novel mechanism for the action of such therapies.

No potential conflicts of interest were disclosed.

Grant support: Canadian Institutes of Health Research grant MOP 79340 (S.L. Asa and S. Ezzat). S. Loeper holds personnel support from the Deutsche Forschungsgemeinschaft (LO 1178/1-1).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank K. So for technical assistance.