Research ArticleTranscriptional Regulation

HES1 Is a Master Regulator of Glucocorticoid Receptor–Dependent Gene Expression

See allHide authors and affiliations

Science Signaling  03 Dec 2013:
Vol. 6, Issue 304, pp. ra103
DOI: 10.1126/scisignal.2004389


Hairy and enhancer of split-1 (HES1) is a basic helix-loop-helix transcription factor that is a key regulator of development and organogenesis. However, little is known about the role of HES1 after birth. Glucocorticoids, primary stress hormones that are essential for life, regulate numerous homeostatic processes that permit vertebrates to cope with physiological challenges. The molecular actions of glucocorticoids are mediated by glucocorticoid receptor–dependent regulation of nearly 25% of the genome. Here, we established a genome-wide molecular link between HES1 and glucocorticoid receptors that controls the ability of cells and animals to respond to stress. Glucocorticoid signaling rapidly and robustly silenced HES1 expression. This glucocorticoid-dependent repression of HES1 was necessary for the glucocorticoid receptor to regulate many of its target genes. Mice with conditional knockout of HES1 in the liver exhibited an expanded glucocorticoid receptor signaling profile and aberrant metabolic phenotype. Our results indicate that HES1 acts as a master repressor, the silencing of which is required for proper glucocorticoid signaling.


Hairy and enhancer of split-1 (HES1) is a highly conserved basic helix-loop-helix transcriptional repressor that mediates its biological effects by binding to N-boxes (CACNAG) throughout the genome and recruiting chromatin-modifying factors to these sites (1, 2). HES1 is required for organogenesis and development of several species as a component of the Notch signaling pathway (36). The molecular function of HES1 in adult tissues, however, is less clear.

Glucocorticoids are primary stress hormones necessary for life, which are synthesized in the adrenal cortex and released into the bloodstream in response to environmental and physiological stress. Because of their hydrophobic nature, these hormones readily diffuse from the circulation into organs, tissues, and cells, where they orchestrate various physiological processes, including metabolism, energy production, immune system function, vascular tone, bone mineralization, and central nervous system function (7). Glucocorticoids are named for their vital role in glucose metabolism where they increase blood glucose concentrations by promoting liver gluconeogenesis and insulin insensitivity. Because of their potent anti-inflammatory and immunosuppressive actions, synthetic glucocorticoids are widely used in the clinic to treat inflammatory and autoimmune diseases, as well as hematological cancers.

Intracellularly, glucocorticoids interact with the ubiquitously distributed glucocorticoid receptor (GR) and promote its translocation from the cytoplasm into the nucleus. Hormone-bound GR binds to glucocorticoid response elements (GREs) in the DNA or interacts with various transcription factors to either increase or decrease the expression of nearly a quarter of the mammalian genome (8). In this manner, glucocorticoids elicit changes in the transcriptional profile of cells and alter the physiology of the organism (9).

We have observed in genome-wide microarray studies that glucocorticoids repress HES1 mRNA expression (10). The molecular mechanisms underlying the apparent mutual antagonism of HES1 and GR, however, remain unknown. Moreover, the genome-wide impact of the interplay of these two transcription factors on the physiological actions of glucocorticoids has not been explored.

Here, we show that glucocorticoids silence HES1 gene expression in multiple cell types and tissues. Glucocorticoids rapidly decreased HES1 mRNA abundance through a GR-dependent antagonism of nuclear factor κB (NFκB) at an NFκB regulatory element within the first exon of the HES1 gene. This repression resulted in the concomitant decline of the HES1 protein within a few hours and to its disappearance from the promoters of genes regulated by glucocorticoids. Overexpression of HES1 in human cells led to reduced glucocorticoid-mediated changes in gene expression, and knockdown of HES1 enhanced sensitivity to glucocorticoids without altering GR association with DNA. Impairment of glucocorticoid signaling by HES1 was abolished when mutations rendering HES1 incapable of DNA binding were introduced or when N-boxes are mutated. To evaluate the role of HES1 in vivo, we used Albumin-Cre mice and created mice bearing loxP sites in the HES1 gene to establish HES1 liver knockout (HESKOL) animals. The absence of HES1 in hepatocytes resulted in no gross physiological or morphological defects in the adult liver. However, genome-wide microarray analysis revealed that HESKOL mice display abnormal glucocorticoid-dependent signaling profiles that affected genes associated with various biological functions including energy production, lipid metabolism, and carbohydrate metabolism. As a result of this dysregulation, HESKOL mice exhibited impaired glucose tolerance. Removal of endogenous glucocorticoids by adrenalectomy corrected this phenotype, whereas injection of exogenous glucocorticoids restored it. These findings indicate that HES1 silencing is necessary to initiate GR-mediated changes in gene expression, and suggest that the dismissal of HES1 cooperates with the GR to regulate a large component of the transcriptional targets of glucocorticoids through a transcriptional derepression mechanism.


Glucocorticoid signaling rapidly silences HES1

Our laboratory identified HES1 as a potential repressed target of glucocorticoid signaling from genome-wide microarray studies performed on human U2OS cells expressing the GR and treated with dexamethasone (10). Reverse transcription polymerase chain reaction (RT-PCR) analysis of U2OS cells engineered to express GR (U2OS-GR) revealed that dexamethasone treatment reduced HES1 mRNA abundance in U2OS-GR cells but not in the parental line that lacks GR (Fig. 1A). HES1 was also silenced in dexamethasone-treated A549 cells, which have endogenous GR (Fig. 1B).

Fig. 1 GR signaling represses the expression of HES1.

(A) HES1 mRNA abundance in wild-type (WT) parental U2OS and U2OS-GR cells treated with dexamethasone (Dex). n = 4 independent biological replicates. (B) Dose response of the repression of HES1 by dexamethasone in U2OS-GR and A549 cells. n = 4 independent biological replicates. (C and D) Time course of the repression of HES1 by dexamethasone in U2OS-GR cells analyzed by RT-PCR (C) and Western blot (D). n = 4 independent biological replicates for (C) and n = 3 independent biological replicates for (D). (E) Abundance of HES1 nascent and mature RNA at early time points upon dexamethasone treatment. n = 4 independent biological replicates. (F) Pretreatment of U2OS-GR cells with cycloheximide (CHX) does not inhibit the repression of HES1 by dexamethasone. n = 4 independent biological replicates. *P < 0.05, one-way analysis of variance (ANOVA)/Tukey analysis.

Examination of the kinetics of HES1 silencing in U2OS-GR cells revealed that dexamethasone rapidly reduced HES1 mRNA within 3 hours of treatment (Fig. 1C). Maximal repression was reached at 6 hours and maintained thereafter until at least 24 hours (Fig. 1C). Under the same conditions, Western blot analysis showed that HES1 protein abundance decreased as early as 6 hours, a reduction that was sustained for at least 24 hours (Fig. 1D), indicating that HES1 is a short-lived protein. Comparing the newly transcribed “nascent” RNA to its “mature” form in the same samples at different time points showed that the abundance of nascent HES1 RNA decreased as early as 30 min after dexamethasone addition and remained low for the duration of the 3-hour experiment (Fig. 1E), suggesting that glucocorticoids repress HES1 gene expression at the level of transcription. In contrast, mature HES1 mRNA started to significantly decline only after 1 hour of dexamethasone treatment (Fig. 1E). These results establish that glucocorticoids can rapidly silence HES1 RNA expression in GR-containing cells.

The rapid repression of HES1 by glucocorticoids suggested that HES1 may be a primary rather than secondary target of the GR. To evaluate this prediction, we preexposed U2OS-GR cells to the protein synthesis inhibitor cycloheximide before dexamethasone treatment. Although HES1 was induced after exposure to cycloheximide alone, dexamethasone still silenced HES1 (Fig. 1F), establishing HES1 as a primary target of glucocorticoid signaling because new protein synthesis was not required for its repression by GR.

The GR can alter expression patterns of target genes by either directly binding to GREs in the DNA or interacting with transcription factors without itself binding to DNA (9). We did not find any potential functional GREs up to 2 kilo–base pairs (kbp) upstream of the HES1 gene transcriptional start site (TSS) when using gene promoter analysis computer programs, suggesting that the GR can repress the HES1 gene without needing to bind directly to the DNA. To address this question, we generated U2OS cells stably expressing a GR DNA binding mutant that interacts poorly with DNA (Dim4), but can still associate with transcription factors and inhibit their activity (11). After dexamethasone treatment, the Dim4 mutant GR decreased HES1 abundance to a similar extent as WT GR (fig. S1), demonstrating that DNA binding by the GR is indeed largely dispensable for silencing HES1 gene transcription.

Glucocorticoids repress HES1 by interfering with NFκB-mediated transcriptional activation of the HES1 gene

Having established that HES1 is a primary target of glucocorticoid signaling, and that DNA binding by GR was not required for repression of HES1 gene transcription, we reasoned that protein-protein interactions of the GR with transcription factors at the HES1 promoter were responsible for the observed repression. To define the site where GR might be acting on the HES1 promoter, we created luciferase reporter constructs carrying progressively smaller regions of the HES1 promoter and stably expressed them in U2OS-GR cells. In response to dexamethasone treatment, luciferase activity was reduced in all stable cell lines, except the one carrying the empty vector control (Fig. 2A), suggesting that GR acts on the HES1 gene near the TSS.

Fig. 2 Glucocorticoids require RelA (p65) to repress HES1.

(A) Activity of luciferase constructs bearing decreasing lengths of the HES1 promoter in U2OS-GR cells. n = 5 independent biological replicates. (B) ChIP results for anti-GR and anti-RNAP2 (RNA pol 2) of U2OS-GR cells treated with dexamethasone. n = 3 independent biological replicates. (C) Activity of luciferase constructs with WT or NFκB-mutated HES1 promoter upon dexamethasone treatment. n = 5 independent biological replicates. (D and E) Knockdown of RelA in U2OS-GR cells (D) and its effect on HES1 expression upon dexamethasone treatment (E). n = 3 independent biological replicates for (D) and n = 4 independent biological replicates for (E). (F) ChIP assay for p65 with RT-PCR probes directed at three areas on the HES1 promoter: −2 kb upstream of the TSS, −1 kb of TSS, and at the TSS. n = 3 independent biological replicates. *P < 0.05, one-way ANOVA/Tukey analysis (E and F) or the Mann-Whitney test (C).

Chromatin immunoprecipitation (ChIP) assays performed in U2OS-GR cells treated with dexamethasone also established that GR preferentially interacts with chromatin near the TSS of the HES1 gene (Fig. 2B). Gene promoter analysis computer programs identified a potential NFκB DNA binding site located at +26/+34 of the HES1 gene [GGAATCCCC] within the 5′ untranslated region of the first exon (fig. S2). This sequence exhibited high homology to the consensus NFκB binding site and is conserved across multiple mammalian species (fig. S2). Because GR and NFκB signaling antagonize one another and basal NFκB activity promotes gene expression in these cells (12), we speculated that the GR-mediated silencing of HES1 expression could be mediated by antagonism of basal NFκB activity on this site. Consistent with this idea, U2OS-GR cells stably expressing a luciferase construct in which this putative NFκB site was mutated exhibited lower basal expression and impaired glucocorticoid-dependent repression (Fig. 2C).

To further determine whether GR regulatory interactions with NFκB could affect HES1 gene expression, we knocked down RelA, which encodes p65, a main component of the NFκB complex, in U2OS-GR cells (Fig. 2D). HES1 expression was lower in untreated cells depleted of RelA than in cells transfected with nontargeting controls, and exposure to dexamethasone did not further decrease the expression of HES1 (Fig. 2E). p65 ChIP assays probed at different regions of the HES1 promoter revealed the presence of p65 near the TSS of the HES1 gene (Fig. 2F). These data indicate that basal NFκB activity at the conserved +26/+34 site of the HES1 gene promotes its expression, and that glucocorticoids can silence HES1 by inhibiting this activity.

Silencing of HES1 is necessary for glucocorticoid signaling

Having elucidated a mechanism by which glucocorticoids repress HES1, we next sought to understand the molecular implications of this silencing. We generated U2OS-GR cells that over- or underexpressed HES1 to avert or mimic the effects of glucocorticoids on this gene (Fig. 3A). Altering the expression of HES1 did not obviously affect GR abundance (Fig. 3A). However, genome-wide microarrays revealed that overexpression of HES1 resulted in inhibition of GR-mediated changes in the glucocorticoid-regulated transcriptome compared to nonoverexpressing controls (Fig. 3B). A comparison of the effects of HES1 overexpression in these cells in the absence of glucocorticoids revealed that many of the gene changes that were repressed by glucocorticoid signaling were also repressed upon HES1 overexpression, suggesting that the GR signaling inhibitory effects of HES1 on these genes may be due to the overexpression of HES1 alone (Fig. 3B and fig. S3). In contrast, genes that were induced by glucocorticoid signaling were largely unaffected by overexpression of HES1 in the absence of hormone, indicating that the removal of HES1 is necessary, but not sufficient, to elicit these GR-mediated changes in gene expression (Fig. 3B).

Fig. 3 HES1 alters GR signaling.

(A) U2OS-GR cells with different abundances of HES1 but similar amounts of GR. n = 3 independent biological replicates. (B) HES1 overexpression promotes the silencing of many genes and inhibits GR-induced changes in gene expression, as seen by microarray analysis. n = 3 independent biological replicates. Con, control. (C to E) HES1 silencing was necessary, but not sufficient, to elicit GR-mediated changes in gene expression in NRO-B1 (C), FLRT3 (D), and RGS2 (E). n = 4 independent biological replicates for (C) to (E). (F) Percent of propidium iodide (PI)–positive cells as determined by flow cytometry of WT, green fluorescent protein (GFP)–overexpressing, and HES1-overexpressing cells untreated or treated with dexamethasone. n = 5 independent biological replicates for (F). *P < 0.05, one-way ANOVA/Tukey analysis.

From this genome-wide microarray analysis, we chose to focus on the following three genes that are targets of GR signaling (1315) and were strongly induced by glucocorticoids in the control cell line, but not in the HES1 overexpressors, and analyzed their expression by RT-PCR: nuclear receptor subfamily 0, group B, member 1 (NRO-B1); fibronectin leucine-rich transmembrane protein 3 (FLRT3); and regulator of G protein (heterotrimeric guanine nucleotide–binding protein) signaling 2 (RGS2). Overexpressing or knocking down HES1 did not affect the basal expression of these genes in cells not treated with glucocorticoids (Fig. 3, C to E). However, the expression of these genes did not increase in a time-dependent manner upon exposure to dexamethasone in HES1-overexpressing cells compared to controls (Fig. 3, C to E). In contrast, in HES1 knockdown cells, these genes exhibited a greater induction 24 hours after dexamethasone treatment (Fig. 3, C to E), a greater dexamethasone sensitivity (fig. S4), and faster induction kinetics (fig. S5). These results establish that repression of HES1 is necessary for proper glucocorticoid signaling.

Because glucocorticoid-mediated changes in gene expression are largely responsible for promoting glucocorticoid-induced apoptosis (16), we also examined if HES1 overexpressors would show an impaired response to glucocorticoid-induced apoptosis. Upon treatment with vehicle or dexamethasone, U2OS-GR cells overexpressing HES1 were more resistant to glucocorticoid-induced apoptosis (Fig. 3F). These data demonstrate that the silencing of HES1 is also essential for the functional actions of glucocorticoids in these cells, such as glucocorticoid-induced apoptosis.

HES1 plays a crucial role during development as one of the main targets of the NOTCH signaling pathway (17, 18); therefore, we evaluated whether NOTCH could impair the actions of glucocorticoids through HES1. Overexpression of the NOTCH intracellular domain increased HES1 expression and decreased the dexamethasone-mediated gene induction of FLRT3, RGS2, and NRO-B1 (fig. S6). These inhibitory effects were abolished when HES1 was knocked down (fig. S6), suggesting that the NOTCH impairment of glucocorticoid signaling is dependent on HES1.

HES1 DNA binding is required to inhibit glucocorticoid actions

To elucidate the molecular mechanism by which HES1 represses glucocorticoid signaling, we initially evaluated the ability of HES1 to physically associate with GR in a complex. Results from coimmunoprecipitation assays indicated that HES1 and GR do not interact directly with one another (fig. S7), and that HES1 does not impair GR from associating with other binding partners, such as the steroid receptor coactivator 1 (SRC1) (fig. S7). Because HES1 can directly bind to N-boxes (CACNAG) in the DNA, we next evaluated whether this binding activity could be responsible for the inhibitory effects of HES1 on glucocorticoid signaling. We transiently transfected luciferase reporter constructs containing WT or mutant N-boxes, GREs, and a TATA box in their promoters into U2OS-GR cells (Fig. 4A). Glucocorticoids induced expression of the luciferase construct carrying mutant N-boxes in a dose-dependent manner, and cotransfection of HES1 minimally repressed luciferase activity under the same conditions (Fig. 4B). However, when a construct carrying WT N-boxes was used, luciferase activities at all dexamethasone concentrations tested were lower, and cotransfection of additional HES1 further repressed these activities (Fig. 4B). These data demonstrate that HES1 mediates its inhibitory actions on glucocorticoid signaling by binding to N-boxes in the DNA.

Fig. 4 HES1 requires DNA binding to inhibit GR signaling.

(A and B) Constructs bearing 2× N-boxes or 2× mutated N-boxes, and 2× GRE sites controlling the expression of a luciferase reporter (A) and their effects on luciferase expression by dexamethasone and HES1 overexpression after dexamethasone treatment (B). n = 5 independent biological replicates for (B). (C) Fold enrichment of luciferase constructs by ChIP of GR or immunoglobulin G (IgG) after dexamethasone treatment. n = 3 independent biological replicates. (D to F) U2OS cells stably expressing GR and Flag-tagged WT HES1 or a mutant HES1 unable to bind to DNA and the effect of dexamethasone on NRO-B1 (D), FLRT3 (E), and RGS2 (F) expression in these cells. n = 4 independent biological replicates for (D) to (F). (G) Fold enrichment of promoters of these genes by ChIP of HES1 or IgG after dexamethasone treatment. n = 3 independent biological replicates. *P < 0.05, one-way ANOVA/Tukey analysis.

To investigate if the molecular basis for these inhibitory effects on GR signaling could be explained by HES1 promoting a conformational change in the DNA that renders it inaccessible to the GR, we performed ChIP assays on the luciferase constructs containing either WT or mutant N-boxes in dexamethasone-treated cells. We found that GR immunoprecipitates were enriched on both of these constructs compared to IgG controls (Fig. 4C). Moreover, neither the overexpression of HES1 nor the presence of N-boxes nearby affected the ability of the GR to interact with these constructs, suggesting the GR can still bind GREs under these conditions (Fig. 4C).

To further elucidate how HES1 inhibits GR signaling, we generated U2OS-GR cells overexpressing WT HES1 and a DNA binding–deficient mutant HES1 (fig. S8), and examined the effects on endogenous gene expression upon exposure to dexamethasone. RT-PCR analysis revealed that the DNA binding–deficient HES1, unlike WT HES1, did not inhibit the GR-mediated induction of NRO-B1, FLRT3, and RGS2 (Fig. 4, D to F). Because these glucocorticoid-responsive genes have N-boxes in their proximal promoters, we performed ChIPs of endogenous HES1 to determine whether HES1 associated with these promoters. Compared to IgG controls, HES1 was greatly enriched at these gene promoters in untreated U2OS-GR cells, and dexamethasone exposure decreased these enrichments (Fig. 4G). PPIB is a gene with robust expression that is unaffected by the expression of HES1 or glucocorticoid signaling, and did not show significant HES1 enrichment in ChIP assays (Fig. 4G), indicating that HES1 associates specifically with the promoters of the glucocorticoid-responsive genes NRO-B1, FLRT3, and RGS2. In summary, these data demonstrate that binding of HES1 to N-boxes in the DNA is required to inhibit the actions of glucocorticoids.

Glucocorticoids silence HES1 in vivo

To examine how glucocorticoids affect the expression of HES1 in vivo, we performed time course experiments in which we injected dexamethasone intraperitoneally to mice that had been adrenalectomized to remove endogenous glucocorticoids. Dexamethasone treatment rapidly reduced the abundance of liver HES1 mRNA and protein by 3 hours (Fig. 5, A and B). By 24 hours, HES1 mRNA and protein had returned to basal amounts. Glucocorticoid exposure also rapidly and robustly repressed HES1 mRNA expression in heart, lung, and kidney (fig. S9). The transient nature of the dexamethasone effects on HES1 expression is likely due to the metabolic breakdown of the steroid. We analyzed the effect of dexamethasone treatment on rat primary hepatocytes to determine whether glucocorticoids acted directly on liver cells. Compared to controls, dexamethasone treatment decreased HES1 mRNA abundance within 3 hours, a repression that was sustained for at least 24 hours (Fig. 5C). These data indicate that the glucocorticoid-dependent repression of HES1 gene expression occurs in vivo in multiple tissues. In addition, consistent with the conservation across species of the NFκB binding site in the HES1 5′ untranslated region (fig. S2), silencing of HES1 by glucocorticoids occurs in human, mouse, and rat cell types.

Fig. 5 Glucocorticoids repress HES1 expression in hepatocytes.

(A and B) Time course of adrenalectomized male mice injected with dexamethasone or PBS, and HES1 mRNA abundance analyzed by RT-PCR (A) or HES1 protein abundance analyzed by Western blot (B). n = 4 independent biological replicates for (A) and n = 3 independent biological replicates for (B). (C) Effect of dexamethasone treatment on the expression of HES1 mRNA in rat primary hepatocytes (RPH). n = 6 independent biological replicates. (D) Scheme showing the approximate location of the loxP sites in the HES1 gene and PCR of liver DNA from control or HESKOL animals. KO, knockout. (E) RT-PCR analysis of HES1 mRNA in liver, kidney, brown adipose tissue (BAT), and muscle in WT, control, and HESKOL animals. n = 3 independent biological replicates. (F) Western blot analysis of control and HESKOL liver extracts. n = 3 independent biological replicates. *P < 0.05, one-way ANOVA/Tukey analysis.

HESKOL mice exhibit abnormal glucocorticoid signaling

To evaluate the role of HES1 during glucocorticoid signaling in vivo, we generated mice bearing loxP sites in the HES1 gene and crossed them with Albumin-Cre mice (19) to specifically knock out HES1 in hepatocytes (Fig. 5, D to F). This conditional approach avoids the lethal developmental defects observed in mice with global knockout of HES1 (3). The HESKOL mice developed normally, and their adult livers exhibited no gross morphological or physiological alterations, compared to loxP/loxP control mice. We performed genome-wide microarrays on liver mRNA extracted from control and HESKOL males that had been adrenalectomized to remove endogenous glucocorticoids. In mice devoid of glucocorticoids, knockout of HES1 in hepatocytes resulted in only 319 statistically different probes when compared to control mice (Fig. 6A). However, dexamethasone treatment resulted in 5411 probe differences in the livers of control mice and 7288 probe differences in the livers from the HESKOL (Fig. 6A). The proportion of induced compared to repressed probes differed between the dexamethasone-treated control and HESKOL mice (Fig. 6B). In the control mice, only 393 probes were induced (7%), whereas 5018 were repressed (93%) by glucocorticoid treatment. In marked contrast, 3644 probes were induced (50%) and 3644 were repressed (50%) in the HESKOL mice treated with glucocorticoids. The large increase in GR-induced genes that accompanies the loss of HES1 is consistent with the notion that HES1 acts primarily as a transcriptional repressor (20). A comparison between dexamethasone-treated control and dexamethasone-treated HESKOL livers revealed 5162 probes that were differentially regulated (Fig. 6A). These data demonstrate that HES1 modulates the mode, extent, and diversity of glucocorticoid signaling in the liver, but the absence of HES1 by itself does not significantly alter basal gene expression patterns in hepatocytes.

Fig. 6 Ablation of liver HES1 alters GR signaling.

(A) Male control (loxP/loxP) or HESKOL mice were adrenalectomized and either injected with dexamethasone or left untreated. Number of probes statistically different (P < 0.01, ANOVA) between the mRNA from livers of control and HESKOL, untreated or treated. n = 3 mice per condition. (B) Number of probes organized as either induced or repressed in control and HESKOL samples. (C) IPA of carbohydrate metabolism pathway by using the gene lists of dexamethasone-treated control and HESKOL mice. Red denotes gene induction, and green denotes gene repression.

To better understand the functional importance of HES1 during glucocorticoid signaling in the liver, we performed Ingenuity Pathway Analysis (IPA) on the set of genes regulated by dexamethasone in control or HESKOL mice. Marked differences were observed in the molecular and cellular functions associated with the regulated genes, because only 3 of the top 10 functions were common to both gene lists (Table 1 and Table 2). Among the biological functions most affected by the absence of HES1 were energy production, lipid metabolism, and carbohydrate metabolism. Because glucocorticoids play a critical role in glucose homeostasis, we chose to further study carbohydrate metabolism. A large number of genes associated with carbohydrate metabolism were regulated by glucocorticoid signaling only in HESKOL livers, suggesting that disrupting HES1 in hepatocytes had expanded the number of carbohydrate metabolism genes that are targets of glucocorticoid signaling (Fig. 6C).

Table 1 Biological functions regulated by glucocorticoids in control mice.

Genes significantly regulated by glucocorticoids in the livers of control mice were analyzed by IPA software. Shown are the top 10 molecular and cellular functions that were most significantly associated with these genes.

View this table:
Table 2 Biological functions regulated by glucocorticoids in HESKOL mice.

Genes significantly regulated by glucocorticoids in the livers of HESKOL mice were analyzed by IPA software. Shown are the top 10 molecular and cellular functions that were most significantly associated with these genes.

View this table:

To delineate how the absence of HES1 in the HESKOL mice could affect the GR gene expression profile, we focused our studies on insulin-like growth factor binding protein 1 (IGFBP1), a gene that encodes a key factor involved in carbohydrate metabolism (21, 22) and that is targeted by glucocorticoid signaling (23, 24). Compared to control animals, livers of HESKOL mice show higher basal expression of IGFBP1 as well as a greater response to dexamethasone treatment (Fig. 7A). Analysis of the IGFBP1 promoter revealed a conserved N-box at position ~−100 bp. In IGFBP1 promoter–driven luciferase assays performed in U2OS-GR cells, greater basal and dexamethasone-induced activities were observed when this N-box was mutated (Fig. 7B), suggesting that HES1 can influence the expression of IGFBP1 through this N-box. RT-PCR analysis of endogenous IGFBP1 expression further showed that dexamethasone elicited a faster and stronger response in U2OS-GR cells when HES1 had been knocked down (Fig. 7, C and D), suggesting that HES1 regulates both the timing and robustness of the glucocorticoid-mediated IGFBP1 gene induction. In addition, glucose 6-phosphatase (G6P) and phosphoenolpyruvate carboxykinase (PEPCK), genes that also encode factors that play a critical role in liver carbohydrate metabolism, exhibited a more rapid and greater induction in response to dexamethasone treatment in HESKOL animals (fig. S10). These results indicate that the absence of HES1 in the liver potentiates glucocorticoid signaling of carbohydrate metabolism–related genes.

Fig. 7 HES1 regulates glucocorticoid-mediated IGFBP1 gene induction.

(A) Male control (loxP/loxP) and HESKOL mice were adrenalectomized and injected with dexamethasone for the indicated periods or left untreated. Liver RNA was extracted, and IGFBP1 abundance was measured by RT-PCR assays. n = 4 independent biological replicates. (B) Activity assays of luciferase in U2OS-GR cells under the control of the rat IGFBP1 promoter with a WT or mutated N-box in the presence of absence of dexamethasone. n = 5 independent biological replicates. RLU, relative luciferase units. (C and D) Abundance of mature (C) and nascent (D) RNA of endogenous IGFBP1 from U2OS-GR cells treated with dexamethasone for the indicated periods. n = 4 independent biological replicates for (C) and (D). *P < 0.05, one-way ANOVA/Tukey analysis (A, C, and D) or Mann-Whitney test (B).

HESKOL mice exhibit glucocorticoid-dependent impaired glucose tolerance

Glucocorticoids are essential for carbohydrate metabolism because they increase blood glucose concentration during fasting, and thus, we examined whether HESKOL mice had abnormal glucose homeostasis. Both male and female HESKOL animals exhibited normal fed and fasted blood glucose concentrations, consistent with the fact that many factors can regulate glucose concentrations. However, intraperitoneal glucose tolerance tests revealed significant impairment of glucose clearance in HESKOL mice compared to controls (Fig. 8, A and B). Removal of endogenous glucocorticoids by adrenalectomy completely abolished these phenotypes in both male and female HESKOL mice (Fig. 8, C and D), demonstrating that this phenotype was glucocorticoid-dependent. Indeed, treatment of the adrenalectomized mice with dexamethasone restored glucose tolerance in both male and female HESKOL mice (Fig. 8, E and F). Moreover, plasma insulin concentrations did not differ between HESKOL and control mice before or after glucose injection and throughout these experiments (fig. S11), suggesting that insulin sensitivity, rather than insulin release, is likely responsible for these phenotypes.

Fig. 8 Glucose tolerance tests.

(A to F) Mice were fasted overnight and injected with glucose, and their blood glucose concentrations were measured. Each group contained between 8 and 11 animals. Control and HESKOL intact males (A) and females (B). Adrenalectomized control and HESKOL males (C) and females (D). Dexamethasone was injected 14 hours before the test to adrenalectomized control and HESKOL males (E) and females (F). *P < 0.05, **P < 0.01, two-tailed Student’s t test.


Here, we provide new evidence demonstrating that HES1 operates as a master regulator of genome-wide glucocorticoid signaling. First, glucocorticoids can directly and rapidly decrease HES1 mRNA and protein abundance. A mechanism for this silencing involves GR-mediated repression of HES1 transcription through an NFκB binding site in the 5′ untranslated region of the HES1 gene that is conserved among mammalian species. Second, overexpression of HES1 in human cells leads to resistance to glucocorticoid-mediated changes in gene expression, whereas knockdown of HES1 increases sensitivity to glucocorticoids. These inhibitory effects of HES1 on glucocorticoid signaling are mediated by its ability to bind to N-boxes in the DNA. Third, HESKOL mice exhibited abnormal changes in liver gene expression when exposed to glucocorticoids, even though the absence of HES1 by itself leads to few transcriptional changes. In particular, glucocorticoids affected more genes associated with carbohydrate metabolism in HESKOL mice than in controls, suggesting an expansion of glucocorticoid signaling. Finally, HESKOL mice are glucose-intolerant when subjected to intraperitoneal glucose tolerance tests. This phenotype disappears upon removal of the adrenal glands, the site of endogenous glucocorticoid synthesis, and is restored when exogenous glucocorticoids are injected. Together, these results reveal that glucocorticoid-mediated silencing of HES1 derepresses large sections of the glucocorticoid-responsive component of the genome.

The widespread actions of HES1 as an inhibitor of glucocorticoid signaling could be explained by the abundance of N-boxes throughout the genome, occurring about once every 1000 bp of DNA. Our data imply that HES1 inhibits glucocorticoid signaling by acting as a cellular brake that must be first removed if the GR is to elicit changes in gene expression. In this scenario, hormone-bound GR interacts with the chromatin in a gene promoter and attempts to activate gene transcription, but if HES1 were also present in this promoter, it would counteract the effects of the GR by promoting a transcriptionally silenced chromatin structure. Appropriately, HES1 is silenced by the GR, and only after it has disappeared from the chromatin can the GR induce expression of that gene. We also recognize that by acting in a similar manner, HES1 could also influence the magnitude, kinetics, and duration of glucocorticoid signaling. Additionally, the reexpression of HES1, after having been silenced by glucocorticoids, could act to reset the system in preparation for the next signaling event.

Consistent with glucocorticoids targeting NFκB for many of their repressive actions (25), our data identify a highly conserved NFκB element located near the TSS of HES1. Moreover, we show that this site is necessary both for the basal expression of HES1 and for its silencing by glucocorticoids. The mutual antagonism of NFκB and GR likely silences HES1 by GR directly blocking the positive transcriptional actions of NFκB on the HES1 gene (9). Therefore, the transcriptional inhibition of the HES1 gene by glucocorticoids could be sufficient to concomitantly reduce HES1 mRNA and protein because both HES1 mRNA and protein have half-lives of only 30 min (26). However, we cannot rule out that other mechanisms could also be involved. For instance, in the one other study on glucocorticoids and HES1, Lemke et al. (27) have suggested that glucocorticoids silence HES1 in the liver through a GR binding site (TGTTCC) located at position −422 bp in the mouse HES1 promoter. However, this potential GRE half site is not conserved in either the rat or human HES1 promoters, organisms in which we also showed that glucocorticoids silence HES1. Moreover, deletion of this region in our human HES1 promoter reporter constructs did not affect the ability of glucocorticoids to inhibit luciferase gene expression (Fig. 2A).

We have begun to explore the effects of NOTCH signaling on glucocorticoid actions mediated through HES1. Our data suggest that NOTCH signaling, by promoting the expression of HES1, can inhibit glucocorticoid signaling. Although the functions of HES1 in tissues and homeostasis outside NOTCH signaling and development are unclear, the mice we have generated bearing loxP sites in the HES1 gene for tissue- and time-specific knockout experiments should help shed new light on this matter. The deletion of HES1 in the adult liver using Albumin-Cre mice revealed that, in the absence of glucocorticoids, disruption of HES1 in the adult liver does not generate widespread changes in gene expression. However, HES1 alters the regulation of several thousand glucocorticoid-regulated genes, and among those, many are associated with carbohydrate metabolism, energy production, and lipid metabolism. Physiologically, we discovered that disrupting HES1 in hepatocytes impairs glucose tolerance, suggesting that HES1 plays an important role during glucose homeostasis in the adult animal. IPA revealed that many other glucocorticoid-regulated processes are also likely to be altered in the HESKOL mice, suggesting an even broader role for HES1 in the modulation of GR responses. A previous study exploring the function of HES1 in the adult organism used overexpression of HES1 by viral vectors in the liver of db/db mice, and suggested that HES1 prevents glucocorticoid-induced liver dyslipidemia (27).

In summary, we have discovered that HES1 is a master regulator of global GR signaling in both cells and in the whole animal. Glucocorticoid-dependent repression of HES1 and its consequent dismissal from glucocorticoid-responsive genes is necessary for the full transcriptional complement of GR signaling. These findings reveal a role for HES1 in controlling the mode, timing, and magnitude of glucocorticoid-dependent stress responses and suggest that aberrant expression of HES1 may contribute to acquired forms of glucocorticoid resistance frequently encountered in the clinic.



Dexamethasone was purchased from Steraloids. Cycloheximide was purchased from Sigma and used as indicated. Rabbit anti-HES1 (AB15740) was purchased from Chemicon (now Millipore). Rabbit anti-GR57 was produced as previously described (28). Rabbit anti-p65 (sc-109) was purchased from Santa Cruz Biotechnology. Rabbit anti–β-actin was purchased from Millipore. Human HES1 complementary DNA (cDNA) was a gift from T. Sudo, and it was cloned into the pcDNA3.1-zeocin, pcDNA3.1-hygro, or pcDNA3.1-puro (a gift from A. Thorburn, University of Colorado Health Sciences Center) for expression in mammalian cells and subsequent selection. Flag-HES1-WT and Flag-HES1-mut cDNAs were gifts from the Kadesch laboratory, and they were cloned into pcDNA3.1-puro. The mutant version contains three point mutations (E43A, K44A, and R47A) that render HES1 incapable of DNA binding (29). For gene silencing, we used the Mission shRNA (short hairpin RNA) system from Sigma. After testing numerous constructs, we used the following: shRNA (SHC202): shHES1, CCGGTGGCCAGTTTGCTTTCCTCATCTCGAGATGAGGAAAGCAAACTGGCCATTTTT (TRCN0000018993); shRelA, CCGGCACCATCAACTATGATGAGTTCTCGAGAACTCATCATAGTTGATGGTGTTTTT (TRCN0000014686).

Cell culture

A549 and U2OS cells (American Type Culture Collection) were maintained in Dulbecco’s modified Eagle’s medium/F-12 supplemented with 10% fetal calf serum, penicillin (100 IU/ml), and streptomycin (100 mg/ml). U2OS-GR cells (10) were maintained in the same medium, except that geneticin (500 μg/ml; Invitrogen) and hygromycin (200 μg/ml; Invitrogen) were also included in the medium. U2OS-GR cells bearing HES1-overexpressing constructs, shRNA, or shHES1 were generated and maintained by including puromycin (2 μg/ml) from Sigma in the medium. Rat primary liver hepatocytes were isolated from adult male Sprague-Dawley rats, digested with the collagenase perfusion method, and treated with dexamethasone. All cells were transfected as recommended by the manufacturer with the Transit-LT1 transfection reagent (Mirus). All dexamethasone treatments were performed in medium containing 10% charcoal-dextran–stripped fetal calf serum.

Luciferase assays

For luciferase assays on the HES1 promoter, the indicated sections of human HES1 promoter were cloned into the reporter plasmid pGL4.20[luc2/Puro] (Promega). For mutagenesis of the NFκB site, the sequence GGAATCCCCC on the HES1 gene was altered to GGAATCCTTT. Constructs were transfected into U2OS-GR cells, and 48 hours after transfection, cells were selected with puromycin (2 μg/ml) for 2 weeks. Stable cells were then plated onto 24-well plates at ~30,000 cells per well, allowed to rest for 24 hours in 10% charcoal-dextran–stripped fetal calf serum, and then treated with 100 nM dexamethasone or vehicle for 16 hours. For luciferase assays on the promoter containing GREs and N-boxes, the pGRE2-luciferase construct (10) was modified by introducing sequences at the Hind III site upstream of the two GRE elements in the following manner: for 2× WT-N-Boxes, AAGCTTacaacaCTTGTGacaGAATTCacaCTTGTGacaacaAAGCTT, and for 2× mut-N-Boxes, AAGCTTacaacatatctcacaGAATTCtatctcacaacaAAGCTT, where AAGCTT is a Hind III site, CTTGTG is an N-box sequence, and GAATTC is an Eco RI used for selection. For luciferase assays on the promoter containing the rat IGFBP1 promoter, the Addgene plasmid 12146 (30) was left unchanged (WT) or modified (mutant) by altering the N-box sequence CACAAG to TGTAAG by site-directed mutagenesis. U2OS-GR cells plated onto 24-well plates at ~30,000 cells per well were transiently transfected with these constructs for 1 day and allowed to rest for another day in medium containing 10% charcoal-dextran–stripped fetal calf serum before treatment with dexamethasone at the indicated concentrations. Luciferase activities were measured as previously described (10). Each experiment was repeated at least three times, and firefly luciferase activity was normalized to the protein concentration.

Primers and real-time RT-PCR analysis

RNA was isolated and treated with deoxyribonuclease with a Qiagen RNeasy Mini kit as suggested by the manufacturer. Real-time PCR was measured on a 7900HT sequence detection system with custom-made or predesigned primer-probe sets from Applied Biosystems. The measurement from each primer-probe set was normalized to that of PPIB, an unregulated housekeeping gene. Values obtained from tissues isolated from mice injected for 3, 6, 12, or 24 hours with phosphate-buffered saline (PBS) or dexamethasone were normalized to untreated mice. The following primer-probe sets from Applied Biosystems were used: LYPD1 (Hs00375992_m1), GLI1 (Hs01110766_m1), IL11 (Hs00174148_m1), NOX4 (Hs00276431_m1), NROB1 (Hs00230864_m1), PPIB (Hs00168719_m1), HES1 (Hs00172878_m1), FLRT3 (Hs00183798_m1), RGS2 (Hs00180054_m1), mHES1 (Mm01342805_m1 and Mm00468601_m1), rHES1 (Rn00577566_m1), mPPIB (Mm00478295_m1), rPPIB (Rn03302274_m1), IGFBP1 (Hs00236877_m1), mIGFBP1 (Mm00515154_m1), mPEPCK (Mm01247058), and mG6P (Mm00839363). To detect nascent HES1 RNA, we used the following custom primers-probe: probe, /56-FAM/CCTGTATCTCTTTGCAGCCCCTCA/3IABkQ/; primers, CAGAAAGGTAAGGGCGGTAC and AAGAGTTCTGTGTTCCCATGG. To detect nascent IGFBP1 RNA, we used the following custom primers-probe: probe, /56-FAM/AATGCCTCTTTCTCTACTCCAGCCC/3IABkFQ/; primers, GCAAGCAGTCCAGATGAGG and TGTTTGTAGCGGGAAGTGG. To detect nascent FLRT3 RNA, we used the following custom primers-probe: probe, /56-FAM/AGGGTTCTG/ZEN/AAGTAACGGAAGCTACCT/3IABkfFQ/; primers, TTCAGTATGCTGGCCTTATTGT and GTCAGCAGTGTTGAGGTCTTTA. To detect nascent PPIB RNA (used to normalize nascent FLRT3 gene expression), we used the following custom primers-probe: probe, /56-FAM/TTTGTGGCC/ZEN/TTAGCTACAGGAGAGGT/3IABKFQ/; primers, TGAACTCTGCAGGTCAGTTTGCTG and ATGAAGATGTAGGCCGGGTGATCT. To detect pGRE2-luciferase construct bearing WT or mutant N-boxes by real-time RT-PCR from ChIP assays, we used the following custom primers-probe: probe, /56-FAM/TGGACAAACCACAACTAGAATGCAGTGA/3IABkFQ/; primers, AGCATCACAAATTTCACAAATAAAGC and GGATCCAGACATGATAAGATACATTG. To detect the promoter of NROB1 by real-time RT-PCR from ChIP assays, we used the following custom primers-probe: probe, /56-FAM/CCCGTAGCCCAGTTCTGCCC/3IABkFQ/; primers, ATGTTGTAGAGGATGCTGCC and CGCGCTAGGTATAAATAGGTCC. To detect the promoter of RGS2 by real-time RT-PCR from ChIP assays, we used the following custom primers-probe: probe, /56-FAM/TGCTGTAGGACTCATTCGACACCC/3IABkFQ/; primers, ATTGCCTCAGTTCACAGACC and CGCGCCTCATTTCTTGTTTG. To detect the promoter of FLRT3 by real-time RT-PCR from ChIP assays, we used the following custom primers-probe: probe, /56-FAM/CATGTTGGTCAGGCTGGTCTCGAA/3IABkFQ/; primers, TCTAATTCCGGCACTTTGGG and CAAGTGATTCTCCTGCCTCAG. To detect the promoter of PPIB by real-time RT-PCR from ChIP assays, we used the following custom primers-probe: probe, /56-FAM/TCTGATACCAATCCCAACGCTGCCTT/3IABkFQ/; primers, AGTCTGAAAGTTGGATGGGCAGGT and TCTGATTGGGTATGTCAAGGCGGT. To detect the promoter of HES1 by PCR in agarose gels from ChIPs assays, we used the following custom primers: ~2 kb upstream of TSS, TCTGGCGAAATCAATGACAACGT and CGTCTTGTTTGATGTGGCCTCC; ~1 kb upstream of TSS, GCAATAAAACATCCTGGCACGTG and TTTAAGAGCTACACCAGCCGAGC; and at TSS, GTTCGCGTGCAGTCCCAGATATAT and GTTCCAGGACCAAGGAGAGAGG. To detect the promoter of HES1 by real-time RT-PCR from ChIP assays, we used the following custom primers-probe: ~2 kb upstream of TSS: probe, /56-FAM/AATTCCCCACTC/3IABkFQ/; primers, GCCAAGGTCAGCTCTTCC and AGTGAAAACCCCAAGCCC; ~1 kb upstream of TSS: probe, /56-FAM/CCACCCCGTCTT/3IABkFQ/; primers, GTAGTTCTTGAATCCCACCCC and TCTAAGGCCCCAAATCCAAAC; and at TSS: probe, /56-FAM/CCAGAGGGAGAG/3IABkFQ/; primers, 5-CGGAGGCTACAACGTCAATC and GACAAGATCAAGACCAAAGCG.

ChIP assays

For ChIP assays, we used the EZ ChIP (catalog no. 17-408) or the Magna ChIP (catalog no. 17-610) kit from Millipore and followed the manufacturer’s protocol. Briefly, 0.5 × 106 U2OS-GR cells were plated on 10-cm dishes and allowed to rest for 2 days or were transiently transfected with the indicated plasmids the following day. Cells were then exposed to 10% charcoal-stripped fetal calf serum for 1 day and subsequently treated with vehicle or dexamethasone (100 nM) as indicated. Cells were washed with ice-cold PBS and fixed for 15 min at room temperature by adding formaldehyde to the medium to a final concentration of 1%. Cross-linking was then stopped by addition of glycine and incubating at room temperature for 10 min. Cells were harvested, and their DNA was sheared by sonication on ice (Branson Sonifier 150, five sets of 10-s pulses at setting 7). The antibodies for each immunoprecipitation reaction were 6 μg of anti-HES1, 4 μg of anti-GR, and 4 μg of anti-RNAP2 (clone 8WG16, Millipore).

Western blot analysis

After the indicated treatments, cells were washed once with ice-cold PBS and lysed for 60 min in lysis buffer [50 mM tris (pH 8.0), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100] supplemented with inhibitor cocktail set II (Calbiochem). The whole-cell extract was later cleared of cellular debris by centrifugation, and protein concentrations were quantitated by the Bradford assay (Bio-Rad). Tissue extracts were isolated and grinded with a tissue homogenizer in 1× Laemmli sample buffer (Bio-Rad), and debris was cleared by centrifugation. Cellular and tissue samples containing 20 to 30 μg per lane were resolved on 4 to 12% SDS–polyacrylamide gel electrophoresis gels (Bio-Rad) and transferred onto a polyvinylidene difluoride membrane (Invitrogen). The dilutions for antibodies were as follows: anti-HES1, 1:1000; anti-GR57, 1:1000; anti–β-actin, 1:10,000; and anti-p65, 1:1000.

Animal studies

All experimental protocols were approved by the animal review committee of the National Institute of Environmental Health Sciences (NIEHS) and were performed in accordance with the guidelines set forth in the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. C57BL/6 mice at 3 to 6 months of age, which were fed ad libitum, were sacrificed by carbon dioxide asphyxiation. Organs were immediately collected, homogenized in 1× Laemmli’s SDS sample buffer, and boiled for 5 min. Samples were then centrifuged to remove debris, and protein concentrations were measured by the Bradford assay (Bio-Rad).

Production and characterization of HESKOL mice

Mice bearing two loxP sites for tissue-specific deletion of the HES1 gene were designed and generated by Xenogen on a C57BL/6 background. The loxP sites are located ~1800 bp upstream (A site) and ~1200 bp downstream (B site) of the HES1 TSS. Cre-mediated recombination of these sites deletes a significant portion of the HES1 promoter and exons 1 to 3. These mice were crossed with Albumin-Cre mice from The Jackson Laboratory [strain name: B6.Cg-Tg(Alb-cre)21Mgn/J; stock number: 003574] to generate HES1 liver knockout animals (Cre/+; HES1 loxP/HES1 loxP). For genotyping, we used a nested PCR to genotype the HESKOL mice with the following primers: A-site external primers, CTTGCCCAAGGTCATGCAGTCAAGGAAGCA and GTCACGAGATGCCTGACCGCACTTAGGAAG; A-site internal primers, TCCGGGACCAGAGCTGGAGAAATCTTTCAC and AGGCTGCCGGCGGACGGCTGGGAAGAGA; B-site external primers, TCAGCTACATTTTACTGCCTTGGCTCACTC and ATAGCCACAGCTCCCAAGTTGTTACTGCTC; B-site internal primers, GTTGGGAGGGTTGGGTAGGCTAAGAACAG and CTCCATCTAAACCGATCTCAGCTCCAGATC; Cre external, GGAAGGTGTCCAATTTACTGACCGTACACC and GGATTAACATTCTCCCACCGTCAGTACGTG; and Cre internal, GCATTACCGGTCGATGCAACGAGTGATGAG and GAGTGAACGAACCTGGTCGAAATCAGTGCG.

Intraperitoneal glucose tolerance tests

After the mice were injected with PBS or dexamethasone (1 mg/kg) and fasted for 14 hours, dextrose (2 g/kg) was injected intraperitoneally, and blood glucose concentrations were measured at 0, 15, 30, 60, and 120 min after injection. Plasma was also collected at 0 and 15 min after glucose injection and used to measure insulin concentrations by enzyme-linked immunosorbent assay (80-INSMS-E01, Alpco).

Microarray analysis

For U2OS-GR cell lines, gene expression analysis was conducted with Agilent Whole Human arrays (Agilent Technologies). Two separate biological replicates of cytoplasmic RNA samples were purified from the U2OS-GR parental cells and the two independent hHES1-overexpressing cell lines with RNeasy Midi kits (Invitrogen) after treatments with 100 nM dexamethasone or vehicle for 6 hours. Total RNA was amplified with the Agilent Low RNA Input Fluorescent Linear Amplification Kit protocol. Starting with 500 ng of total RNA, two complementary RNAs (cRNAs), one labeled with Cy3 and the other labeled with Cy5, were produced according to the manufacturer’s protocol. For each two-color comparison, 750 ng each of Cy3- and Cy5-labeled cRNAs was mixed and fragmented with the Agilent In Situ Hybridization Kit protocol. Hybridizations were performed for 17 hours in a rotating hybridization oven at 65°C at 4 rpm. Slides were washed with 6× saline sodium phosphate EDTA (SSPE) + 0.005% N-lauroylsarcosine for 1 min and then with 0.06× SSPE + 0.005% N-lauroylsarcosine for 1 min at 37°C. The slides were dried by slowly removing from second wash solution, then scanned with an Agilent G2565 Scanner (10 μm and with extended dynamic range), and processed with Agilent Feature Extraction v9.1. Twenty chips in total were used for these studies: 5 comparisons (control compared to dexamethasone for each of the three cell lines and parental clone compared to each of the two hHES1-overexpressing clones) × 2 biological replicates × 2 chips per replicate = 20. The resulting files were imported into the Rosetta Resolver system (version 6.0, Rosetta Biosoftware). Intensity ratios (treated/controls) were analyzed at the Entrez gene level in Resolver (version 6.0, Rosetta Biosoftware) and were considered differentially expressed if the P value was less than 0.0001. Clustering analysis was performed with the Rosetta Resolver analysis software. For mouse livers, gene expression analysis was conducted with Whole Mouse Genome 4×44 multiplex format oligo arrays (014868) (Agilent Technologies) following the Agilent one-color microarray-based gene expression analysis protocol. Three separate biological replicates of cytoplasmic RNA samples were purified with RNeasy Midi kits (Invitrogen) from control or HESKOL mouse livers treated with dexamethasone (1 mg/kg) or vehicle for 6 hours. Starting with 500 ng of total RNA, Cy3-labeled cRNA was produced according to the manufacturer’s protocol. For each sample, 1.65 μg of Cy3-labeled cRNA was fragmented and hybridized for 17 hours in a rotating hybridization oven. Slides were washed and then scanned with an Agilent Scanner. Data were obtained with the Agilent Feature Extraction software (v9.5) using the one-color defaults for all parameters. The Agilent Feature Extraction software performed error modeling, adjusting for additive and multiplicative noise. To determine differentially expressed probes, we performed an error-weighted ANOVA and Benjamini-Hochberg multiple test correction with a P value <0.01 using Rosetta Resolver System (version 7.0, Rosetta Biosoftware). Significantly regulated genes were analyzed by IPA software (Ingenuity Systems).

Statistical analysis

Student’s t tests, one-way ANOVA with Tukey’s post hoc analysis, or Mann-Whitney tests were used to evaluate whether differences were statistically significant with GraphPad Prism 6 software. Statistical significance was defined as P < 0.05.


Fig. S1. The GR Dim4 DNA binding mutant represses HES1 expression.

Fig. S2. Conservation of the HES1 NFκB site across species.

Fig. S3. Overexpression of HES1 silences various genes.

Fig. S4. Absence of HES1 increases glucocorticoid sensitivity.

Fig. S5. Absence of HES1 elicits a faster response to glucocorticoid signaling.

Fig. S6. Notch signaling inhibits glucocorticoid actions through HES1.

Fig. S7. HES1 and GR do not physically interact.

Fig. S8. Expression of Flag-WT-HES1 and Flag-mut-HES1.

Fig. S9. Glucocorticoids repress HES1 gene expression in multiple tissues.

Fig. S10. Absence of HES1 potentiates GR-mediated induction of G6P and PEPCK.

Fig. S11. Insulin concentrations.


Acknowledgments: We thank A. Jetten, X. Li, and members of the Cidlowski laboratory for critical reading of the manuscript. We thank G. E. Kissling (Biostatistics Branch, NIEHS) for help with statistical analyses. We thank the NIEHS Flow Cytometry Center and the Viral Vector Core for technical support. Funding: This research was supported by the Intramural Research Program of the NIH, NIEHS. Author contributions: J.R.R. designed and performed the experiments, analyzed the data, and wrote the manuscript. R.H.O. designed the experiments, analyzed the data, and wrote the manuscript. N.Z.L. designed and performed the experiments. M.K. designed and performed the experiments and analyzed the data. M.G. performed the experiments. J.A.C. designed the experiments, analyzed the data, and wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The gene expression array data have been deposited in the Gene Expression Omnibus database with the accession number GSE52106.
View Abstract

Navigate This Article