Converging on DKK1 to drive metastatic HCC
Hepatocellular carcinoma (HCC) is a common form of liver cancer. The Wnt signaling protein DKK1 and the growth factor receptor EGFR are abundant in HCC and are associated with metastatic progression and poor prognosis in patients. Niu et al. found that these molecular markers are linked. The activation of EGFR in HCC cells induced DKK1 expression through parallel pathways: MEK-ERK pathway–dependent nuclear translocation of the kinase PKM2 and PI3K-AKT pathway–dependent activation of the acetyltransferase p300. These pathways converged on modifying histone H3 at the DKK1 promoter to activate gene transcription. The findings reveal previously unidentified targets that could be therapeutically targeted in patients with HCC.
Abstract
The protein Dickkopf-1 (DKK1) is frequently overexpressed at the transcript level in hepatocellular carcinoma (HCC) and promotes metastatic progression through the induction of β-catenin, a Wnt signaling effector. We investigated how DKK1 expression is induced in HCC and found that activation of the epidermal growth factor receptor (EGFR) promoted parallel MEK-ERK and PI3K-Akt pathway signaling that converged to epigenetically stimulate DKK1 transcription. In HCC cell lines stimulated with EGF, EGFR-activated ERK phosphorylated the kinase PKM2 at Ser37, which promoted its nuclear translocation. Also in these cells, EGFR-activated Akt phosphorylated the acetyltransferase p300 at Ser1834. Subsequently, PKM2 and p300 mediated the phosphorylation and acetylation, respectively, of histone H3 at the DKK1 promoter, which synergistically enhanced DKK1 transcription. The mechanism was supported with mutational analyses in cells and in a chemically induced HCC model in rats. The findings suggest that dual inhibition of the MEK and PI3K pathways might suppress the expression of DKK1 and, consequently, tumor metastasis in patients with HCC.
INTRODUCTION
Hepatocellular carcinoma (HCC) generally has a poor prognosis and is ranked as the second leading cause of cancer-related deaths worldwide (1). The main cause of death in patients with HCC is the high incidence of metastasis before surgical resection of the tumor (2, 3). Therefore, elucidating the mechanisms driving metastasis is urgently needed to develop more effective and timely therapeutic interventions in patients.
Dickkopf-1 (DKK1) is a secreted glycoprotein and an antagonist of the canonical Wnt signaling pathway that functions by binding to low-density lipoprotein receptor–related proteins 5 and 6 (LRP5/6) (4, 5). Although DKK1 was first described as a head inducer during vertebrate development (6), it was later found that DKK1 can act as an important regulator in a variety of human cancers (7), although its role appears to differ by tumor type. For example, DKK1 reportedly suppresses metastasis in breast cancer (8), whereas its overexpression promotes tumor progression and metastasis in esophageal adenocarcinoma and prostate cancer (9, 10). Notably, DKK1 is overexpressed at the transcript level in HCC, wherein it promotes cell migration and invasion (11, 12). However, the molecular mechanism through which DKK1 is transcriptionally regulated in HCC cells is unclear.
Histone acetylation was found to control DKK1 expression at the transcriptional level in breast cancer (13), and evidence suggests that histone H3 modifications—including acetylation and phosphorylation—regulate tumor growth and metastasis in HCC by affecting chromatin structure and gene expression (14, 15). The epidermal growth factor receptor (EGFR) is highly expressed and hyperactivated in HCC and positively correlates with poor prognosis in patients (16, 17). Moreover, EGFR activation was found to result in rapid and sequential phosphorylation and acetylation of H3 in mammalian cells (18, 19). Thus, we explored the possibility that EGFR-induced histone H3 modification may be involved in regulating DKK1 transcription in HCC.
RESULTS
Overexpression of DKK1 promotes tumor growth and metastasis in HCC
We initiated our study by establishing a model of HCC in rats. Tumors were detectable in the liver 16 weeks after intraperitoneal injection of diethylnitrosamine (DEN) (Fig. 1A). Enzyme-linked immunosorbent assays (ELISAs) revealed that the content of DKK1, a secreted protein, in the plasma of DEN-treated rats was increased in a statistically significant manner (Fig. 1B). Real-time polymerase chain reaction (PCR) and Western blotting also revealed increased expression of DKK1 at both the mRNA and protein levels in the livers of DEN-treated rats (Fig. 1, C and D). Immunohistochemical (IHC) analyses further confirmed that DKK1 abundance was increased in the livers of DEN-treated rats relative to controls (Fig. 1E). In addition, IHC staining of human HCC and adjacent normal tissues revealed that the DKK1 abundance was increased in the tumors (Fig. 1F). To investigate the role of DKK1 in the growth and metastasis of HCC in vivo, we injected DKK1-overexpressing and DKK1-knockdown Huh-7 cells (fig. S1, A and B) subcutaneously into the left dorsal flank or into the lateral tail vein of athymic nude mice. The results showed that overexpression of DKK1 promoted tumor growth and metastatic seeding in vivo, whereas knockdown of DKK1 inhibited both, relative to controls (Fig. 1, G and H). Collectively, these results suggest that DKK1 is overexpressed and may promote primary and metastatic growth in HCC.
(A) Representative photos of livers and images of H&E-stained livers from untreated and DEN-treated rats. Scale bar, 50 μm. (B) The level of DKK1 in the serum from untreated and DEN-treated rats. *P < 0.05, two-tailed unpaired t test, n = 8 rats per group. (C) The mRNA expression of DKK1 in the livers of untreated and DEN-treated rats. ***P < 0.001, two-tailed unpaired t test, n = 6 rats per group. (D) The protein expression of DKK1 in the livers of untreated and DEN-treated rats. ***P < 0.001, two-tailed unpaired t test, n = 6 rats per group. (E) Immunohistochemical staining of DKK1 in the livers of untreated and DEN-treated rats. ***P < 0.001, a two-tailed unpaired t test, n = 8 rats per group. Scale bar, 50 μm. (F) Immunohistochemical staining of DKK1 in adjacent nontumor tissues and tumor tissues. **P < 0.01, two-tailed unpaired t test, n = 39 samples per group. (G) Representative images of tumors and quantification of tumor weight collected from vector, DKK1-overexpression (DKK1-OE), sh-Vector, and DKK1-knockdown (DKK1-KD) groups. *P < 0.05, two-tailed unpaired t test, n = 8 mice per group. (H) Representative photos of metastatic lung nodules and quantification of the lung nodules from vector, DKK1-overexpression (DKK1-OE), sh-Vector, and DKK1-knockdown (DKK1-KD) groups. ***P < 0.001, two-tailed unpaired t test, n = 8 mice per group.
EGF promotes DKK1 expression through the MEK-ERK and PI3K-Akt pathways
To determine whether activation of EGFR contributes to the up-regulation of DKK1 in HCC, we incubated Huh-7, HepG2, SNU368, and SNU739 cells with epidermal growth factor (EGF) and detected DKK1 expression. The results showed that the mRNA and protein expression of DKK1 was markedly increased in HCC cells after treatment with EGF (100 ng/ml; Fig. 2, A, and B). Considering that DKK1 is a secreted glycoprotein, we then measured the content of DKK1 in cell culture medium. As expected, EGF can promote the secretion of DKK1 from HCC cells into cell culture medium (Fig. 2C). We then assessed whether knockdown of EGFR with short hairpin RNA (shRNA) could block EGF-enhanced DKK1 mRNA and protein expression. As we expected, 24 hours after transfection with EGFR shRNA, the protein expression of EGFR was reduced in Huh-7 and HepG2 cells (Fig. 2D and fig. S2A). Furthermore, our results suggested that knockdown of EGFR could reverse EGF-enhanced DKK1 mRNA and protein expression in Huh-7 and HepG2 cells (Fig. 2, E and F, and fig. S2, B and C). To further determine whether EGF-induced DKK1 overexpression depended on the activation of EGFR, we examined the effect of gefitinib, an inhibitor of the tyrosine kinase activity of EGFR, on DKK1 expression in HCC cells treated with EGF. Using real-time PCR and Western blot assays, we found that gefitinib abrogated EGF-induced up-regulation of DKK1 mRNA and protein expression in Huh-7 and HepG2 cells (Fig. 2, G and H, and fig. S2, D and E).
(A and B) Effects of EGF on the expression of DKK1 mRNA (A) and protein (B) in Huh-7, HepG2, SNU368, and SNU739 cells. *P < 0.05, **P < 0.01, ***P < 0.001 compared with control group, one-way ANOVA, n = 6 independent experiments per group. (C) Effects of EGF on the amount of DKK1 in the cell culture medium from Huh-7, HepG2, SNU368, and SNU739 cells. ***P < 0.001 compared with control group, one-way ANOVA, n = 6 independent experiments per group. (D) Western blotting analysis of the abundance of EGFR protein level in Huh-7 cells after transfection with EGFR shRNA for 24 hours. ***P < 0.001, two-tailed unpaired t test, n = 3 independent experiments per group. (E and F) Effects of EGFR shRNA on the EGF-induced abundance of DKK1 mRNA (E) and protein (F) expression in Huh-7 cells. V, vector; E, EGF; sh, shEGFR; ***P < 0.001, one-way ANOVA, n = 4 independent experiments per group. (G and H) Effects of pretreatment with gefitinib (20 μM; “G”), AG490 (20 μM; “A”), PD98059 (20 μM; “P”), or LY294002 (20 μM; “L”) on the EGF-induced abundance of DKK1 mRNA (G) and protein (H) in Huh-7 cells. ***P < 0.001, one-way ANOVA, n = 6 independent experiments per group.
The biological functions of EGFR are mediated through the activation of the Src–Janus kinase (JAK)–signal transducer and activator of transcription 3 (STAT3), Ras–Raf–mitogen-activated protein kinase kinase (MEK)–extracellular signal–regulated kinase (ERK), and phosphatidylinositol 3-kinase (PI3K)–Akt–mammalian target of rapamycin (mTOR) pathways (20, 21). To identify which of these is involved in EGFR-mediated up-regulation of DKK1 in HCC cells, we examined the effects of the JAK inhibitor AG490, the MEK inhibitor PD98059, and the PI3K inhibitor LY294002 (each applied to cells at 20 μM) on EGF-induced expression of DKK1 mRNA and protein in HCC cells. As shown, both PD98059 and LY294002 blocked EGF-induced DKK1 mRNA and protein expression, whereas AG490 did not affect either, in Huh-7 and HepG2 cells (Fig. 2, G and H, and fig. S2, D and E). These results indicate that the MEK-ERK and PI3K-Akt pathways are required for EGF-induced DKK1 expression.
ERK1/2-dependent phosphorylation of PKM2 promotes DKK1 transcription
ERK1/2-dependent phosphorylation of pyruvate kinase M2 (PKM2) at Ser37 upon EGFR activation is required for the nuclear translocation of PKM2, which can act as a transcriptional coactivator to induce a series of gene expression changes in brain tumors (22). Therefore, we wondered whether ERK1/2-dependent phosphorylation and nuclear translocation of PKM2 were involved in EGFR-mediated DKK1 transcription in HCC. After incubation with EGF (100 ng/ml) for 0.5 or 1 hour, the phosphorylation of PKM2 at Ser37 was enhanced in Huh-7 and HepG2 cells (Fig. 3A and fig. S3A). Pretreatment of Huh-7 and HepG2 cells with the EGFR tyrosine kinase inhibitor gefitinib and the MEK inhibitor PD98059 reversed EGF-induced the phosphorylation of PKM2 Ser37 in these cells (Fig. 3B and fig. S3B). Using a Western blot assay, we also observed increased PKM2 accumulation in the nucleus in Huh-7 and HepG2 cells after treatment with EGF (100 ng/ml) for 1 hour, whereas gefitinib or PD98059 blocked EGF-induced PKM2 nuclear accumulation (Fig. 3C and fig. S3C). Immunofluorescence analysis further confirmed these results (Fig. 3D and fig. S3D).
(A) Effects of EGF on the level of PKM2 Ser37 phosphorylation in Huh-7 cells. *P < 0.05, ***P < 0.001, one-way ANOVA, n = 4 independent experiments per group. (B) Effects of pretreatment with 20 μM gefitinib or 20 μM PD98059 on EGF-induced increased phosphorylation levels of PKM2 at Ser37 in Huh-7 cells after 1 hour of coincubation. **P < 0.01, ***P < 0.001, one-way ANOVA, n = 4 independent experiments per group. (C) Effects of pretreatment with 20 μM gefitinib or 20 μM PD98059 on EGF-induced the expression of PKM2 nuclear protein in Huh-7 cells after incubation with EGF for 1 hour. **P < 0.01, ***P < 0.001, one-way ANOVA, n = 4 independent experiments per group. (D) Immunofluorescence staining of PKM2 in Huh-7 cells. The cells were pretreated with 20 μM gefitinib or 20 μM PD98059 for 30 min before EGF for 1 hour. n = 3 independent experiments per group. Scale bar, 50 μm. (E) Western blotting analysis of the abundance of PKM2 protein level in Huh-7 cells after transfection with PKM2 shRNA for 24 hours. ***P < 0.001, two-tailed unpaired t test, n = 3 independent experiments per group. (F and G) Effects of PKM2 shRNA on the EGF-induced abundance of DKK1 mRNA (F) and protein (G) expression in Huh-7 cells. Twenty-four hours after transfection, cells were incubated in the presence or absence of EGF (100 ng/ml) for 12 hours. V, vector; E, EGF; sh, PKM2 shRNA. ***P < 0.001, one-way ANOVA, n = 4 independent experiments per group. (H) Immunofluorescence staining of PKM2 in Huh-7 cells transiently expressing Flag-tagged WT PKM2, PKM2 S37A mutant, or PKM2 S37D mutant for 24 hours, n = 3 independent experiments per group. Scale bar, 50 μm. (I and J) The expression of DKK1 mRNA (I) and protein (J) in Huh-7 cells expressing a Flag-tagged WT PKM2, PKM2 S37A mutant, or PKM2 S37D mutant for 24 hours. *P < 0.05, ***P < 0.001, one-way ANOVA, n = 4 independent experiments per group.
We next examined whether PKM2 plays a role in EGF-dependent expression of DKK1. Through protein analysis, we found that the protein expression of PKM2 was significantly reduced in Huh-7 and HepG2 cells after 24-hour transfection with specific PKM2 shRNA (Fig. 3E and fig. S3E). Moreover, knockdown of PKM2 with shRNA could reverse EGF-enhanced DKK1 mRNA and protein expression (Fig. 3, F and G, and fig. S3, F and G). Then, we transfected Huh-7 and HepG2 cells with Flag-tagged wild-type (WT) PKM2, a phosphorylation-defective PKM2 S37A mutant, or a phosphorylation-mimic PKM2 S37D mutant and examined the mRNA and protein expression of DKK1. The results revealed that the phosphorylation-mimic PKM2 S37D mutant induced significant PKM2 nuclear accumulation in Huh-7 and HepG2 cells (Fig. 3H and fig. S3H). Meanwhile, the PKM2 S37D mutant also induced a higher expression of DKK1 mRNA and protein compared with that of the WT PKM2 or the S37A mutant in Huh-7 and HepG2 cells (Fig. 3, I and J, and fig. S3, I and J). The results indicate that ERK1/2-dependent PKM2 phosphorylation and nuclear translocation are required for EGF-induced DKK1 transcription.
Akt-dependent p300 phosphorylation mediates the EGF-induced DKK1 transcription
EGFR stimulation can induce the phosphorylation of p300 at Ser1834 through the PI3K-Akt pathway, which, in turn, promotes target gene expression (23). We therefore speculated that Akt-induced phosphorylation of p300 may also be involved in EGFR-mediated DKK1 transcription. Stimulating Huh-7 and HepG2 cells with EGF increased the level of p300 phosphorylation at the Ser1834 residue (Fig. 4A and fig. S4A), and this effect was blocked by pretreatment with the Akt inhibitor MK2206 (Fig. 4B and fig. S4B). Knocking down p300 with shRNA (Fig. 4C and fig. S4C) attenuated the EGF-enhanced expression of DKK1 at the mRNA and protein levels (Fig. 4, D and E, and fig. S4, D and E). Furthermore, in transfecting Huh-7 and HepG2 cells with Flag-tagged phosphorylation-mimic p300 mutant (S1834D), but not a WT or a phosphorylation-defective mutant (S1834A), increased the expression of DKK1 at the mRNA and protein levels (Fig. 4, F and G, and fig. S4, F and G). Thus, Akt-dependent p300 phosphorylation is also required for EGF-induced DKK1 transcription.
(A) Effects of EGF on the level of p300 Ser1834 phosphorylation in Huh-7 cells. ***P < 0.001, one-way ANOVA, n = 7 independent experiments per group. (B) Effects of pretreatment with 20 μM gefitinib or 20 μM MK2206 on the EGF-induced increased phosphorylation levels of p300 at Ser1834 in Huh-7 cells after 1 hour of coincubation. **P < 0.01, ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group. (C) Western blotting analysis of the abundance of p300 protein level in Huh-7 cells after transfection with two specific p300 shRNA for 24 hours. ***P < 0.001, two-tailed unpaired t test, n = 3 independent experiments per group. (D and E) Effects of two specific p300 shRNAs on the abundance of EGF-induced DKK1 mRNA (D) and protein (E) expression in Huh-7 cells. At 24 hours after transfection, cells were incubated in the presence or absence of EGF (100 ng/ml) for 12 hours.*P < 0.05, **P < 0.01, ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group. (F and G) The expression of DKK1 mRNA (F) and protein (G) in Huh-7 cells expressing a Flag-tagged WT p300, p300 S1834A mutant, or p300 S1834D mutant for 24 hours. ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group.
Phosphorylation of PKM2 at Ser37 and p300 at Ser1834 synergistically promotes DKK1 transcription
Subsequent to our results above, we explored whether cotransfection of the PKM2 S37D mutant with the p300 S1834D mutant would produce synergistic effects on DKK1 expression in HCC cells. Using real-time PCR and Western blotting, we found that expressing the PKM2 S37D mutant with the p300 S1834D mutant promoted the expression of DKK1 at both the mRNA and protein levels more effectively than did the expression of WT p300 or the p300 S1834A mutant (Fig. 5, A and B, and fig. S5, A and B). Similarly, cotransfection of the p300 S1834D mutant with the PKM2 S37D mutant, but not with the WT PKM2 or the PKM2 S37A mutant, yielded a synergistic effect on DKK1 mRNA and protein expression (Fig. 5, C and D, and fig. S5, C and D). These results indicate that phosphorylation of PKM2 at Ser37 and p300 at Ser1834 synergistically promotes DKK1 transcription in HCC cells.
(A and B) Effects of cotransfection of the PKM2 S37D mutant with the Flag-tagged WT p300, p300 S1834A mutant, or p300 S1834D mutant for 24 hours on the expression of DKK1 mRNA (A) and protein (B) in Huh-7 cells. *P < 0.05, ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group. (C and D) Effects of cotransfection of the p300 S1834D mutant with Flag-tagged WT PKM2, PKM2 S37A mutant, or PKM2 S37D mutant for 24 hours on the expression of DKK1 mRNA (C) and protein (D) expression in Huh-7 cells. *P < 0.05, **P < 0.01, ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group.
Phosphorylation and acetylation of histone H3 participate in EGF-induced DKK1 expression
PKM2 reportedly phosphorylates histone H3 at the Thr11 site, which is required for histone H3 acetylation at the Lys9 site and subsequent gene expression (19). Given that acetyltransferase p300 can acetylate histone H3 in mammalian cells, it raised the possibility that PKM2 phosphorylates histone H3 at Thr11 upon EGFR activation, and p300 then acetylates histone H3 Lys9 at the promoter region of DKK1, leading to the activation of DKK1 transcription in HCC. To test this idea, we first examined whether EGF induces H3-Thr11 phosphorylation and H3-Lys9 acetylation in Huh-7 and HepG2 cells; as expected, it did (Fig. 6, A and B, and fig. S6, A and B). Furthermore, depletion of PKM2 substantially reduced the EGF-induced Thr11 phosphorylation and Lys9 acetylation of H3 in Huh-7 and HepG2 cells (Fig. 6, C and D, and fig. S6, C and D). The data may suggest that PKM2-dependent phosphorylation of histone H3 at Thr11 is required for the acetylation of histone H3 at Lys9 upon EGFR activation in HCC. To test this further, we first used the EGFR tyrosine kinase inhibitor gefitinib and the MEK/ERK inhibitor PD98059 to block EGF-induced phosphorylation of PKM2 at Ser37 and examined the level of H3-Thr11 phosphorylation in Huh-7 and HepG2 cells. Both inhibitors attenuated EGF-induced phosphorylation of H3-Thr11 (Fig. 6E and fig. S6E). Moreover, expression of the PKM2 S37D mutant in Huh-7 and HepG2 cells also enhanced the level of phosphorylation of H3-Thr11 (Fig. 6F and fig. S6F). Together, these results demonstrated that H3-Thr11 phosphorylation was dependent on the phosphorylation of PKM2 at Ser37.
(A and B) Effects of EGF on the level of H3-Thr11 phosphorylation (A) and H3-Lys9 acetylation (B) in Huh-7 cells. *P < 0.05, **P < 0.01, ***P < 0.001, one-way ANOVA, n = 6 independent experiments per group. (C and D) Effects of PKM2 shRNA on EGF-induced H3-Thr11 phosphorylation (C) and H3-Lys9 acetylation (D) in Huh-7 cells. Twenty-four hours after transfection, cells were incubated in the presence or absence of EGF (100 ng/ml) for 1 hour. V, vector; E, EGF; Sh, PKM2 shRNA. *P < 0.05, ***P < 0.001, one-way ANOVA, n = 4 independent experiments per group. (E) Effects of gefitinib (20 μM; Gefi) or PD98059 (20 μM; PD) on the level of EGF-induced H3-Thr11 phosphorylation in Huh-7 cells after 1 hour of coincubation. **P < 0.01, ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group. (F) The level of H3-Thr11 phosphorylation in Huh-7 cells transiently expressing a Flag-tagged WT PKM2, PKM2 S37A mutant, or PKM2 S37D mutant. *P < 0.05, **P < 0.01, one-way ANOVA, n = 5 independent experiments per group.(G) Western blotting analysis of the abundance of histone H3 protein level in Huh-7 cells after transfection with H3 shRNA for 24 hours. ***P < 0.001, two-tailed unpaired t test, n = 3 independent experiments per group. (H and I) Effects of reconstituted expression of RNAi-resistant histone rH3-T11A on EGF-induced the mRNA (H) and protein (I) abundance of DKK1in endogenous H3-knockdown Huh-7 cells. Twenty-four hours after transfection, cells were incubated in the presence or absence of EGF (100 ng/ml) for 12 hours. *P < 0.05, **P < 0.01, ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group. (J) Effects of EGF on the binding of phospho–H3-Thr11 to DKK1 promoter in Huh-7 cells. **P < 0.01, a two-tailed unpaired t test, n = 5 independent experiments per group. (K) Protein expression of phospho–H3-Thr11 in the livers of untreated and DEN-treated rats. ***P < 0.001, a two-tailed unpaired t test, n = 6 rats per group. (L) Immunohistochemical staining of phospho–H3-Thr11 in the livers of untreated and DEN-treated rats, **P < 0.01, two-tailed unpaired t test, n = 8 rats per group. Scale bars, 100 μm. (M) ChIP analysis of phospho–H3-Thr11 binding to DKK1 promoter in the livers of untreated and DEN-treated rats. **P < 0.01, two-tailed unpaired t test, n = 5 rats per group. (N) Immunohistochemical staining of phospho–H3-Thr11 and DKK1 in 39 human HCC specimens. Left: Representative photos of two tumors are shown. Scale bar, 100 μm. Right: Semiquantitative scoring was performed, Pearson’s product moment correlation test; note that some of the dots represent more than one specimen.
To investigate the relevance of H3-Thr11 phosphorylation in EGF-induced DKK1 expression, we expressed RNA interference (RNAi)–resistant WT histone H3 or histone H3-T11A in endogenous histone H3-knockdown Huh-7 and HepG2 cells. As expected, the protein expression of endogenous histone H3 was reduced in Huh-7 and HepG2 cells 24 hours after transfection with H3 shRNA (Fig. 6G and fig. S6G), and the expression of the T11A mutant abrogated EGF-induced expression of DKK1 at the mRNA and protein levels (Fig. 6, H and I, and fig. S6, H and I). Chromatin immunoprecipitation (ChIP) assays also suggested that EGF treatment resulted in enhanced H3-Thr11 phosphorylation at the DKK1 promoter (Fig. 6J and fig. S6J). In support of these in vitro results, we also observed in vivo an increase in phospho–H3-Thr11 levels in DEN-treated rats using Western blotting and IHC (Fig. 6, K and L). ChIP analyses with a phospho–H3-Thr11 antibody demonstrated that DEN administration resulted in an enhanced H3-Thr11 phosphorylation at the DKK1 promoter in rats (Fig. 6M). Furthermore, IHC analysis revealed that the amount of H3-Thr11 phosphorylation correlated with the expression levels of DKK1 in a statistically significant manner across 39 human HCC tumor specimens (Fig. 6N). These results indicate that PKM2-induced phosphorylation of histone H3 at Thr11 is required for EGFR-mediated histone H3-Lys9 acetylation and DKK1 transcription in HCC.
Next, we explored whether p300 mediates the acetylation of histone H3 at Lys9 in response to EGFR activation in HCC. Western blotting revealed that knockdown of p300 significantly diminished EGF-induced Lys9 acetylation of histone H3 in Huh-7 and HepG2 cells (Fig. 7A and fig. S7A). Blocking EGF-induced phosphorylation of p300 at Ser1834 with the EGFR tyrosine kinase inhibitor gefitinib or the Akt inhibitor MK2206 abrogated EGF-induced acetylation of H3-Lys9 in Huh-7 and HepG2 cells (Fig. 7B and fig. S7B). Furthermore, expression of the p300 S1834D mutant in Huh-7 and HepG2 cells also enhanced the level of acetylation of H3-Lys9 (Fig. 7C and fig. S7C). Collectively, these results indicated that the phosphorylation of p300 Ser1834 was required for H3-Lys9 acetylation in HCC cells.
(A) Effects of p300 shRNA on EGF-induced H3-Lys9 acetylation in Huh-7 cells. Twenty-four hours after transfection, cells were incubated in the presence or absence of EGF (100 ng/ml) for 1 hour. C, control; E, EGF; #1, #2, p300 shRNA construct. ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group. (B) Effects of pretreatment with gefitinib (20 μM; Gefi) and MK2206 (20 μM; MK) on EGF-induced H3-Lys9 acetylation after 1 hour of coincubation in Huh-7 cells. **P < 0.01, ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group. (C) The level of H3-Lys9 acetylation in Huh-7 cells transiently expressing Flag-tagged WT p300, p300 S1834A mutant, or p300 S1834D mutant for 24 hours. *P < 0.05, **P < 0.01, one-way ANOVA, n = 5 per group. (D and E) Effects of reconstituted expression of RNAi-resistant histone H3-K9R on EGF-induced the mRNA (D) and protein (E) expression of DKK1 in endogenous H3-knockdown Huh-7 cells. Twenty-four hours after transfection, cells were incubated in the presence or absence of EGF (100 ng/ml) for 12 hours. **P < 0.01, ***P < 0.001, one-way ANOVA, n = 5 independent experiments per group. (F) Effects of EGF on the binding of acetyl-H3-Lys9 to DKK1 promoter in Huh-7 cells. **P < 0.01, two-tailed unpaired t test, n = 5 independent experiments per group. (G) The expression of acetyl-H3-Lys9 in the livers of untreated and DEN-treated rats. *P < 0.05, two-tailed unpaired t test, n = 6 rats per group. (H) Immunohistochemical staining of acetyl-H3-Lys9 in the livers of untreated and DEN-treated rats. **P < 0.001, two-tailed unpaired t test, n = 8 rats per group. Scale bar, 100 μm. (I) ChIP analysis of acetyl-H3-Lys9 binding to DKK1 promoter in the livers of untreated and DEN-treated rats. **P < 0.01, two-tailed unpaired t test, n = 5 rats per group. (J) Immunohistochemical staining of acetyl-H3-Lys9 in 39 human HCC specimens. DKK1 staining is shown in Fig. 6N. Left: Representative photos of two tumors are shown. Scale bar, 100 μm. Right: Semiquantitative scoring was performed, Pearson’s product moment correlation test; note that some of the dots represent more than one specimen.
Subsequently, we generated an RNAi-resistant histone H3 mutant in which Lys9 was replaced with nonacetylatable arginine (H3K9R) and transfected this into Huh-7 cells that had been depleted of endogenous histone H3. In these cells, EGF-induced DKK1 expression, at both the mRNA and protein levels, was significantly decreased in H3K9R mutant–expressing cells compared with that in WT H3-expressing cells (Fig. 7, D and E, and fig. S7, D and E). Furthermore, ChIP assays also suggested that EGF treatment resulted in enhanced H3-Lys9 acetylation at the DKK1 promoter (Fig. 7F and fig. S7F). These results suggest that H3-Lys9 acetylation is required for DKK1 expression in response to EGFR activation.
To expand our findings in vivo, we examined the levels of H3-Lys9 acetylation in the DEN-induced rodent model of HCC. Western blotting and IHC data indicate that DEN administration significantly increased the level of H3-Lys9 acetylation (Fig. 7, G and H), specifically at the DKK1 promoter as shown by ChIP data (Fig. 7I). In support of the clinical relevance of this finding that H3-Lys9 acetylation regulates DKK1 transcription, IHC analyses of 39 human HCC tumor specimens revealed that the amount of H3-Lys9 acetylation correlated significantly with that of DKK1 (Fig. 7J), thus adding p300-dependent acetylation of histone H3 to the mechanism mediating EGFR-induced DKK1 transcription in HCC.
DISCUSSION
DKK1 expression, which is rarely expressed in normal human adult tissues outside of placental and embryonic tissues (24), has been associated with a high rate of recurrence and metastasis in patients with HCC (25, 26) and with the migration and invasive behavior of cultured HCC cells by inducing β-catenin nuclear accumulation (11, 27). Here, we found that EGF stimulation induced DKK1 gene expression in HCC. Data from cell lines and a rat model of HCC indicated that EGFR regulates DKK1 expression at the transcriptional level through activation of the PI3K-Akt and MEK-ERK pathways in HCC (Fig. 8).
EGFR activation in HCC cells induces PKM2 phosphorylation at the Ser37 residue through the ERK1/2 signaling pathway. Phosphorylated PKM2 then translocates into the nucleus where it phosphorylates histone H3 at Thr11. Meanwhile, EGF binding to EGFR also promotes p300 phosphorylation at the Ser1834 residue through the Akt signaling pathway. Subsequently, p300 acetylates histone H3 at the Lys9 site. Histone H3-Thr11 phosphorylation and Lys9 acetylation cooperatively promote DKK1 gene expression in HCC.
Himburg et al. (28) reported that DKK1 treatment induced EGF secretion in bone marrow cells; with our findings here, this raises the possibility that DKK1 and EGF may establish a positive feedback loop in HCC and accelerate tumor metastasis. DKK1 has been reported to perform its biological functions by binding to its receptors cytoskeleton-associated membrane protein 4 (CKAP4) or LRP5/6, which are significantly elevated in HCC tumor tissues (29–32). Although aberrantly increased CKAP4 or LRP5/6 is positively correlated with tumor progression (31, 33), it remains to be investigated whether EGF has an effect on the expression of CKAP4 or LRP5/6 in HCC cells in our future work. We think this will help us understand the complexity of possible positive loop interactions involving EGF and DKK1 signaling.
Persistent activation of the EGFR signaling system has been implicated in liver responses to injury and carcinogenesis (16, 34). EGFR facilitates HCC progression and metastasis through the activation of several major intracellular signaling cascades such as the JAK2-STAT3, Ras-Raf-MEK-ERK, and PI3K-Akt-mTOR pathways (35, 36). These signaling pathways often activate transcriptional programs in the cell nucleus, leading to the expression of genes involved in cell proliferation, survival, adhesion, or migration in cancer cells (37). However, there is no evidence showing that these downstream signaling pathways of EGFR participate in the transcriptional regulation of DKK1 expression, although DKK1 does activate the PI3K-Akt signaling pathways and promote tumor progression in esophageal adenocarcinoma (9). In support of our findings on the role of ERK in regulating DKK1 expression, Zhou et al. (38) found that inhibition of the MEK-ERK pathway abolished the increase in DKK1 mRNA during amino acid limitation in a human colon cancer cell line. In addition, it is currently unclear whether other tyrosine kinase receptors, including vascular EGFR (VEGFR) and insulin-like growth factor receptor (IGFR), have a regulatory role in DKK1 expression in HCC; overexpression of VEGFR and IGFR has been detected in HCC cell lines and tissues, and these receptors also activate the PI3K-Akt and the RAF-MEK-ERK pathways (39, 40).
PKM2, one of the key glycolytic enzymes in tumor cells, normally localizes to the cytoplasm and acts as a prominent driver of the Warburg effect (19, 41). Under growth factor stimulation, PKM2 also translocates into the nucleus and acts as a transcriptional coactivator to promote targeted gene expression in a variety of cancers (42, 43). Consistent with these reports, we documented that nuclear PKM2 could serve as a transcriptional cofactor to promote DKK1 transcription upon EGF treatment in HCC cells. Accumulating evidence suggested that posttranslational modifications such as phosphorylation and acetylation are critical for the protein localization and functional modulation of PKM2 (44, 45). Consistent with these findings, we also observed increased PKM2 nuclear accumulation when its Ser37 residue was phosphorylated by ERK1/2 in HCC cells. Acting as both a serine/threonine protein kinase and a tyrosine protein kinase, PKM2 binds to and phosphorylates multiple transcription factors such as β-catenin and STAT3 in the nucleus to promote tumor-related gene transcription (46, 47). Notably, Yang et al. (19) found that PKM2-induced phosphorylation of histone H3 at Thr11 is critical for EGFR-mediated MYC and CCND1 transcription in glioblastoma. Here, we also demonstrated that PKM2 can phosphorylate histone H3 at residue Thr11 upon EGFR activation, leading to DKK1 transcription in HCC.
Phosphorylation of histone H3 is one of the most frequent epigenetic modifications that affect chromatin structure and gene expression (14, 15). For example, phosphorylation of H3 at the Thr3, Thr6, Ser10, or Thr11 residue is critical for transcriptional activation of its target genes in mammalian cells (48). Accumulating evidence suggests that histone H3 phosphorylation usually acts in combination with other modifications (such as methylation or acetylation) to participate in the regulation of gene transcription (18). Notably, Yang et al. (19) found that H3-Thr11 phosphorylation is required for H3-Lys9 acetylation. In line with their report, our results suggest that both histone H3 phosphorylation at Thr11 and acetylation at Lys9 are required for EGFR-mediated DKK1 transcription in HCC. Also, according to Yang et al. (19), no posttranslational modifications (namely, phosphorylation or acetylation) occurred at the residues of Thr3, Lys4, Thr6, and Ser10 in histone H3 in response to EGF stimulation; thus, although we cannot yet rule it out, we supposed these sites are not essential for EGF-induced DKK1 transcription in HCC.
Histone acetylation can be regulated by two groups of enzymes with opposing functions: histone acetyltransferases, which add acetyl moieties to lysine residues, and histone deacetylases (HDACs), which catalyze the removal of these acetyl groups (14, 49). Here, our results suggest that histone H3 acetylation at Lys9 upon EGFR activation depends on the acetyltransferase activity of p300 in HCC. Certainly, we cannot exclude that histone H3-Lys9 acetylation is mediated by HDAC3 dissociation from DKK1 promoter region, because Yang et al. (19) suggested that PKM2-dependent H3-Thr11 phosphorylation promotes HDAC3 removal from histone H3 and facilitates subsequent H3-Lys9 acetylation. Besides, one possibility that p300 was recruited to the DKK1 promoter region to acetylate histone H3 at Lys9 when HDAC3 was away from this region may exist. These speculations need to be further explored. As an acetyltransferase, p300 is involved in several cellular processes such as cell proliferation, survival, and differentiation (50). Overexpression of p300 has been identified in HCC and was associated with enhanced epithelial-mesenchymal transition of HCC cells (51). Histone acetyltransferase p300 promotes tumor progression by enhancing the acetylation of some transcriptional factors or histones, which results in the activation or repression of target gene expression (52). Consistent with these reports, we demonstrated that p300 could acetylate histone H3 at Lys9 and promote DKK1 transcription in HCC. In support of our results, Kim et al. (13) found that DKK1 expression was modulated epigenetically by p300 in breast cancer; however, there, to the effect, was suppression of tumorigenicity (13), indicating that the role of DKK1 is likely context dependent, possibly attributed to different tumor microenvironments or genetic backgrounds. The activity of p300 can be regulated by several protein kinases, such as protein kinase A (PKA), PKB, PKC, and mitogen-activated protein kinases (MAPKs) (23, 53, 54). For instance, PKC could repress p300 acetyltransferase activity and transcriptional function through enhancing its phosphorylation at the Ser89 residue (53). In contrast, Akt-mediated phosphorylation of p300 at Ser1834 enhances acetyltransferase activity (23). In conclusion, our findings that activation of EGFR can induce DKK1 transcription through the MEK-ERK and PI3K-Akt signaling pathways in HCC suggest that pathway inhibitors might be explored for clinical use in patients to slow the progression of the cancer.
MATERIALS AND METHODS
Reagents and antibodies
All the inhibitors used in the present study were purchased from Cell Signaling Technology (Beverly, MA, USA). Rabbit monoclonal anti-DKK1 antibody (ab109416), rabbit polyclonal anti-histone H3 antibody (ab1791), rabbit polyclonal anti-histone H3 (acetyl Lys9) antibody (ab10812), and rabbit polyclonal anti-histone H3 (phosphor Thr11) antibody (ab5168) were obtained from Abcam (Cambridge, MA, USA). Rabbit monoclonal anti-PKM2 antibody (4053) was purchased from Cell Signaling Technology (Danvers, MA, USA). Rabbit polyclonal anti-PKM2 (phosphor Ser37) antibody (PA5-37684) and rabbit polyclonal anti-p300 (phosphor Ser1834) antibody (PA5-64531) were obtained from Thermo Fisher Scientific (Waltham, MA, USA). Mouse monoclonal anti–β-actin antibody (sc-47778), horseradish peroxidase (HRP)–conjugated goat anti-rabbit immunoglobulin G (IgG) (sc-2054), and goat anti-mouse IgG (sc-2973) were all purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). EGF was purchased from R&D systems (Minneapolis, MN, USA).
Cell lines and cell culture
All cell lines used in the experiments were obtained from the Cell Bank of the Chinese Academy of Science (Shanghai, China). Cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (HyClone, Logan, UT, USA) at 37°C in a humidified incubator containing 5% CO2.
ELISA assay
Levels of rat in the plasma and human DKK1 in the cell culture supernatant were measured using ELISA kits (R&D Systems, Minneapolis, MN) according to the instructions provided by the manufacturer. Absorbance was measured at 450 nm by using a Vmax Kinetic microplate reader (Molecular Devices, Sunnyvale, CA).
Western blot analysis
-Western blot analysis was conducted as previously described (55). Briefly, the total proteins were extracted from cancer cells and rat livers using ice-chilled radioimmunoprecipitation assay lysis buffer containing 50 mM tris-HCl (pH 8.0), 150 mM NaCl, 0.5% sodium deoxycholate, 0.1% SDS, 1% NP-40, 5 mM EDTA, 0.25 mM phenylmethylsulfonyl fluoride, and protease inhibitor cocktail. The nuclear proteins were isolated from cancer cells according to previous report (8). In brief, the harvested cells were washed three times with cold phosphate-buffered saline (PBS) buffer and then resuspended gently in hypotonic buffer containing 20 mM tris-HCl (pH 7.4), 10 mM NaCl, and 3 mM MgCl2. After incubation on ice for 15 min, these cells were lysed with a Dounce homogenizer. The homogenate was centrifuged to remove intact cells, followed by centrifugation at 800g to collect the nuclei. The supernatant contains the cytoplasmic fraction. The nuclear pellets were washed three times with PBS buffer and lysed through sonication. The concentration of total or nuclear protein was determined using a BCA assay kit (Pierce, Rockford, IL). Protein samples were denatured and separated by SDS–polyacrylamide gel electrophoresis. After separation, the protein was transferred onto a polyvinylidene difluoride membrane. The membranes were then blocked with 5% nonfat milk or 3% bovine serum albumin followed by the incubation of the indicated primary and HRP-conjugated secondary antibodies. The blots were detected using the enhanced chemiluminescence plus reagents and visualized using a FluorChem E Imager (Protein Simple, San Jose, CA, USA).
Immunofluorescence
Cells were washed with ice-cold PBS and fixed for 10 min in 4% paraformaldehyde at room temperature. After being blocked with 5% fetal calf serum for 1 hour at room temperature, the cells were stained with the corresponding primary antibody overnight at 4°C, followed by incubation with Alexa Fluor 488–conjugated secondary antibodies (Invitrogen, Carlsbad, CA, USA) for 1 hour at room temperature. Nuclei were labeled with Hoechst 33342 stain for 5 min at room temperature. Images were acquired using an LSM 510 M (Carl Zeiss) confocal microscope.
Real-time PCR
RNA from cancer cells was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Complementary DNA was synthesized by using random hexamers and MMLV reverse transcriptase according to the manufacturer’s instructions (Takara, Tokyo, Japan). Real-time PCR was performed using 2× SYBR Green PCR Master Mix (Promega) on an ABI 7500 sequence detection system (Applied Biosystems). The reaction conditions were as follows: 95°C, predenaturation for 5 min, 15 s at 95°C, and 1 min at 60°C for a total of 40 cycles. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control. The relative expression level of the genes was calculated by the 2−ΔΔCt method. Specific primers for human HCC cells were as follows: DKK1, 5′-TGT GCT AGA CAC TTC TGG TCC AA-3′ and 5′-TGA TCT TTC TGT ATC CGG CAAG-3′; GAPDH, 5′-GAC ACC CAC TCC TCC ACC TTT-3′ and 5′-TTG CTG TAG CCA AAT TCG TTGT-3′. Specific primers for rat were as follows: DKK1, 5′-ATG CCC TCT GAC CAC AGC CATT-3′ and 5′-CAC CGT GGT CAT TGC CAA GGT-3′; GAPDH, 5′-AGC CAT GTA CGT AGC CAT CC-3′ and 5′-GCC ATC TCT TGC TCG AAG TC-3′.
Chromatin immunoprecipitation
ChIP assays were performed using a SimpleChIP Plus Enzymatic Chromatin IP Kit (Cell Signaling Technology) following the manufacturer’s instructions. In brief, cells were cross-linked with 1% formaldehyde for 10 min at 37°C. Then, the chromatin was digested with micrococcal nuclease to generate 150 to 900 base pair DNA/protein fragments. The DNA-protein complexes were immunoprecipitated with the ChIP-grade antibodies and appropriate protein G–agarose beads. Normal rabbit or mouse IgG was used as a negative control. The ChIP samples were verified by quantitative PCR (qPCR) to evaluate histone modification status on the DKK1 promoter (promoter region: −409 to −273). The following primers were used for real-time qPCR after ChIP assay: 5′-GCA CAG TCA GCG AGT ATT GG-3′ and 5′-GAA CTT GGG TGC CCT TGC CTG-3′.
Animals and experimental design
Male Sprague-Dawley rats (6 weeks old) and male athymic BALB/c nu/nu mice (5 weeks old) were purchased from Beijing Weitong Lihua Animal Co. All animal procedures were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of Henan University. All animals were housed in 12-hour light/dark cycles at a temperature of 20° to 24°C and a relative humidity of 50 ± 10%. During the period of the study, the rats were housed three per cage, and the BALB/c nude mice were housed five per cage, with free access to food and water. All animals were allowed 1 week of acclimatization before the experimental procedures.
To induce HCC, the rats were intraperitoneally injected with DEN at a dose of 50 mg/kg once a week for 16 weeks. Rats in the control group were intraperitoneally injected with normal saline solution on the same days as the DEN-treated group. At the end of the study, all animals were euthanized, and the livers were removed. After photographing with a camera, some livers were fixed in 4% paraformaldehyde for histopathology and IHC examinations, and the remaining portions were kept at −80°C for Western blotting. Blood samples from all the rats were also collected in tubes, and the serum was separated and analyzed for DKK1 expression by ELISA.
For the in vivo tumorigenesis experiment, nude BALB/c mice were divided into four groups: Huh-7/vector, Huh-7/DKK1-OE (DKK1 overexpression), Huh-7/sh-vector, and Huh-7/DKK1-KD. A total of 2 × 106 cells were subcutaneously injected into the left dorsal flank of each nude mouse. Eight mice per group in each experiment were included. The tumors were allowed to grow for 2 weeks, and then, the mice were euthanized by cervical dislocation.
For the experimental metastasis assay, nude BALB/c mice were divided into four groups: Huh-7/vector, Huh-7/DKK1-OE (DKK1 overexpression), Huh-7/sh-vector, and Huh-7/DKK1-KD. Eight mice per group were included in each experiment. A total of 1 × 106 cells were injected into the lateral tail vein of nude BALB/c mice. To ensure that all mice bore actively growing lung tumors, pulmonary metastasis was allowed to develop for 6 weeks. Then, the mice were euthanized, and the lungs were removed. After being fixed with 4% paraformaldehyde for 1 day, the number of metastatic nodules on the lung surface was counted.
Human tissue specimens
Matched pairs of HCC and adjacent nontumor tissues from 39 patients were collected from the First Affiliated Hospital of Henan University between 2014 and 2018. The use of patient tissues was approved by the Institute Research Ethics Committee at the hospital, and informed consent was obtained from all enrolled patients.
Hematoxylin and eosin staining and IHC
The paraformaldehyde-fixed tissues were dehydrated and embedded in paraffin. The paraffin-embedded tissues were then cut into pieces (4 μm) and placed on polylysine-coated slides. The slides were stained with hematoxylin and eosin (H&E) or IHC stains. For IHC staining, the slides were incubated with the DKK1 antibody (1:100), p-histone H3 antibody (1:100), or Ac-histone H3 antibody (1:100) at 4°C overnight. Slides were washed with PBS twice, incubated with streptavidin-conjugated HRP antibody for 30 min, and developed with diaminobenzidine. Immunoreactivity was examined under a light microscope. Two independent pathologists analyzed the expression of the target protein by visualizing the brown-stained section. The staining intensity was scored on a scale of 0 to 3: 0 (negative), 1 (weak), 2 (moderate), or 3 (strong). The staining extent was scored as follows: 0, if no tumor sections were stained; 1, if <1% of sections were stained; 2, if 2 to 10% of sections were stained; 3, if 11 to 30% of sections were stained; 4, if 31 to 70% of sections were stained; or 5, if 71 to 100% of sections were stained. The staining intensity and staining extent scores were then added to calculate the final stain score (0 to 8) for each tissue.
Statistical analysis
Statistical analyses were performed using the GraphPad Prism 5.0 (GraphPad Software Inc., La Jolla, CA, USA). All data are presented as means ± SEM. The statistical significance of the difference between the two groups was tested by an unpaired, two-tailed Student’s t test. One-way analysis of variance (ANOVA) followed by a Tukey or Dunnett’s posttest was used to compare means of multiple experimental groups. All data were subjected to a normal distribution test, and only the data with a normal distribution were subjected to a t test or ANOVA. Pearson’s correlation test was used to examine the correlation of two proteins; P < 0.05 was considered significant.
SUPPLEMENTARY MATERIALS
stke.sciencemag.org/cgi/content/full/13/657/eabb5727/DC1
Fig. S1. DKK1-overexpressing and DKK1-knockdown Huh-7 cells.
Fig. S2. EGFR activation promotes DKK1 expression in HCC through the MEK-ERK and PI3K-Akt pathways.
Fig. S3. ERK1/2-dependent phosphorylation and nuclear translocation of PKM2 promote DKK1 transcription in HepG2 cells.
Fig. S4. Akt-dependent p300 phosphorylation mediates the EGF-induced increase in DKK1 transcription in HepG2 cells.
Fig. S5. Cotransfection of the PKM2 S37D mutant with the p300 S1834D mutant produced a synergistic effect on DKK1 expression in HepG2 cells.
Fig. S6. PKM2-dependent phosphorylation of histone H3 at Thr11 mediates EGFR-induced histone H3-Lys9 acetylation and DKK1 transcription in HepG2 cells.
Fig. S7. p300-dependent acetylation of histone H3 mediates EGFR-induced DKK1 transcription in HepG2 cells.
This is an article distributed under the terms of the Science Journals Default License.
