- Email: [email protected]
Discovery of a Small-Molecule BMP Sensitizer for Human Embryonic Stem Cell Differentiation
Article
Discovery of a Small-Molecule BMP Sensitizer for Human Embryonic Stem Cell Differentiation Graphical Abstract
Authors Lingling Feng, Brandoch Cook, Su-Yi Tsai, Ting Zhou, Brooke LaFlamme, Todd Evans, Shuibing Chen
Correspondence [email protected] (T.E.), [email protected] (S.C.)
In Brief Feng et al. show that PD407824, a CHK1 inhibitor, synergizes with BMP4 to induce human embryonic stem cell differentiation toward mesoderm or cytotrophoblast lineages. Mechanistic studies indicate that CHK1 inhibition depletes p21, leading to a decrease in SMAD2/3 and an increase in levels of nuclear SMAD1.
Highlights d
PD407824 increases the sensitivity of cells to sub-threshold levels of BMP4
d
PD407824 synergizes with BMP4 to reprogram C2C12 myoblasts toward an osteoblast fate
d
PD407824 sensitizes human embryonic stem cells to BMP4 to induce differentiation
d
CHK1 inhibition induces P21 downregulation and SMAD2/3 degradation
Feng et al., 2016, Cell Reports 15, 2063–2075 May 31, 2016 ª 2016 The Author(s). http://dx.doi.org/10.1016/j.celrep.2016.04.066
Cell Reports
Article Discovery of a Small-Molecule BMP Sensitizer for Human Embryonic Stem Cell Differentiation Lingling Feng,1,2,3 Brandoch Cook,2,3 Su-Yi Tsai,2,3 Ting Zhou,2,3 Brooke LaFlamme,2 Todd Evans,2,* and Shuibing Chen2,* 1Key Laboratory of Pesticide and Chemical Biology (CCNU), Ministry of Education, College of Chemistry, Central China Normal University, Wuhan, Hubei 430079, China 2Department of Surgery, Weill Cornell Medical College, New York, NY 10065, USA 3Co-first author *Correspondence: [email protected] (T.E.), [email protected] (S.C.) http://dx.doi.org/10.1016/j.celrep.2016.04.066
SUMMARY
Sorely missing from the ‘‘toolkit’’ for directed differentiation of stem/progenitor cells are agonists of the BMP-signaling pathway. Using a high-throughput chemical screen, we discovered that PD407824, a checkpoint kinase 1 (CHK1) inhibitor, increases the sensitivity of cells to sub-threshold amounts of BMP4. We show utility of the compound in the directed differentiation of human embryonic stem cells toward mesoderm or cytotrophoblast stem cells. Blocking CHK1 activity using pharmacological compounds or CHK1 knockout using single guide RNA (sgRNA) confirmed that CHK1 inhibition increases the sensitivity to BMP4 treatment. Additional mechanistic studies indicate that CHK1 inhibition depletes p21 levels, thereby activating CDK8/9, which then phosphorylates the SMAD2/3 linker region, leading to decreased levels of SMAD2/3 protein and enhanced levels of nuclear SMAD1. This study provides insight into mechanisms controlling the BMP/ transforming growth factor beta (TGF-b) signaling pathways and a useful pharmacological reagent for directed differentiation of stem cells. INTRODUCTION Bone morphogenetic proteins (BMPs) are important regulators of embryonic development and stem/progenitor cell fate decisions. BMPs are required for the establishment of the body plan (Heisenberg and Solnica-Krezel, 2008), limb bud patterning (Robert, 2007), and early hemato-vascular (Larsson and Karlsson, 2005) and neural development (Liu and Niswander, 2005). BMPs regulate the development of multiple organs, including heart (Kruithof et al., 2012), kidney (Cain et al., 2008), liver (Zaret, 2001), and the CNS (Fukuda and Taga, 2006). In adult tissues, BMPs provide signals for differentiation in niches for the hair follicle (Blanpain and Fuchs, 2009), intestinal stem cells (Takashima and Hartenstein, 2012), and germ cells (Knight and Glister, 2006). Due to their important role in embryonic development, BMPs have been used in both maintenance and directed differentiation
of embryonic stem cells (ESCs) to various cell fates. For mouse ESCs, BMP4 is required, together with leukocyte inhibitory factor, to maintain the pluripotent self-renewal state (Li et al., 2012; Ying et al., 2003). In contrast, in human ESCs, BMP4 promotes differentiation, so that inhibition of BMP signaling is required to maintain human ESC self-renewal (James et al., 2005; Wang et al., 2005). Once committed to differentiate, BMPs promote the commitment of ESCs to the mesendoderm germ layer, and these BMP-induced mesendoderm cells can further differentiate into multiple cell lineages, including cardiac, hematopoietic, and hepatic cells. For example, BMP4 has been used to direct differentiation from mesendoderm to Flk1+ hematopoietic progenitor cells and then to blood cells (Lengerke et al., 2008; Nostro et al., 2008). BMP2 and BMP4 direct definitive endoderm cells to a hepatic lineage (Gouon-Evans et al., 2006). BMP7 has been used for differentiation toward brown adipocytes (Nishio et al., 2012). In addition, BMP4 initiates trophoblast differentiation from human ESCs (Xu et al., 2002). Finally, BMP4, 7, and 8b induce germ cell differentiation from both mouse and human ESCs (Kee et al., 2006; Wei et al., 2008). Synthetic small molecules have been widely used to control developmental signaling pathways, as functional agonists or antagonists. Compared to recombinant proteins, synthetic small molecules can be more stable, easier to quantify for reproducible activity and dose response, and far less expensive to produce, which is particularly relevant for scaling cell production. To date, most of the small molecules discovered to regulate BMP signaling are BMP antagonists. A phenotypic screen using zebrafish embryos identified dorsomorphin, which inhibits BMP signaling by targeting BMP type 1 receptors (ALK2, 3, and 6) (Yu et al., 2008). A structure-activity relationship study found a dorsomorphin analog, LDN193189, which demonstrates moderate pharmacokinetic characteristics in mice (Cuny et al., 2008). The in vivo structure-activity relationship study of dorsomorphin analogs identified a specific BMP inhibitor, DMH1 (Hao et al., 2010). Recently, several small molecules have been identified to either activate or synergize with the BMP pathway. For example, SVAK-3 (Okada et al., 2009), SVAK-12 (Kato et al., 2011), KM11073 (Baek et al., 2015), and A1 and A17 (Cao et al., 2014) enhance BMP2-induced early osteoblast marker expression. Small molecules of the flavonoid family have been shown to upregulate Id1, Id2 expression in a human cervical carcinoma cell line (Vrijens et al., 2013). In addition, FK506 activates
Cell Reports 15, 2063–2075, May 31, 2016 ª 2016 The Author(s). 2063 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
A
B
E
C
D
F
G
Figure 1. PD407824 Sensitizes C2C12 Myoblasts to BMP4-Induced Activation of BMP Signaling (A) Scheme of the high-throughput screening. (B) Chemical structure of PD407824. (C) Dose curve of PD407824 on C2C12 myoblasts transfected with pId2-Luc-GFP reporter. (D) PD407824 sensitizes C2C12 myoblasts to BMP4-induced upregulation of endogenous Id2 expression measured by qRT-PCR. (E) PD407824 sensitizes C2C12 myoblasts to BMP4-induced SMAD1/5/9 phosphorylation indicated by western blotting. (F) Quantification of relative SMAD1/5/9 activity. The phosphorylation levels were normalized to total SMAD1 protein. (G) PD407824 sensitizes C2C12 myoblasts to BMP4-induced upregulation of Id1 expression measured by qRT-PCR. See also Figure S1.
BMPR2 and rescues endothelial dysfunction (Spiekerkoetter et al., 2013). However, most of the identified compounds show relatively low activity and fail to induce the generation of mature osteoblasts, which limits their application to modulate BMP signaling. Thus, there is still a strong need to identify efficient BMP activators or sensitizers that can be used in stem cell differentiation. From a high-throughput screen of more than 4000 small molecules we identified PD407824 as a chemical BMP sensitizer. PD407824 sensitizes cells to BMP4-induced upregulation of BMP pathway target genes, including Id1 and Id2. More importantly, we found that PD407824 cooperates with limiting BMP4 protein when added to ESCs during differentiation toward mesoderm or cytotrophoblast stem cells. This provides a useful chemical agent to control human ESC (hESC) differentiation and can facilitate scaled production of defined cell lineages from hESCs. RESULTS Identification of PD407824 as a BMP4 Sensitizer To identify chemical agonists/sensitizers of BMP signaling, we established an assay to monitor BMP-signaling activity through
2064 Cell Reports 15, 2063–2075, May 31, 2016
the expression of a downstream target, Id2 (Figure 1A). We designed a reporter construct containing the Id2 promoter, including two previously validated BMP-responsive regulatory regions from 3,000 to 2,729 and 350 to +80 (Nakahiro et al., 2010) controlling expression of a luciferase-GFP reporter (pId2-LucGFP). The reporter activity was first validated as responsive to BMP4. C2C12 myoblasts transfected with the pId2-LucGFP reporter construct showed increased luciferase activity when treated with 10 ng/ml BMP4 (Figure S1A). To perform high-throughput screening, C2C12 myoblasts were first seeded onto 10-cm2 plates in standard medium (DMEM supplemented with 10% fetal bovine serum [FBS]). On the second day, these cells were transfected with the reporter plasmid pId2LucGFP. 24 hr later, the transfected cells were transferred to 384-well plates at 5,000 cells/100 ml medium per well. After overnight incubation, the culture medium was changed to serum-free (DMEM lacking FBS) to decrease the background luciferase activity induced by serum. After 24 hr, each well was treated with a single compound from a library containing 4,000 chemicals, including kinase inhibitors, signaling pathway regulators, natural products, and US Food and Drug Administration (FDA)approved drugs, which were arrayed similarly in 384-well
format as 10 mM stocks in DMSO. Detailed library information is described in Table S1. The library was screened at 1:1,000 (0.1% DMSO) and 1:10,000 (0.01% DMSO) dilutions, providing 10 mM and 1 mM final concentrations, respectively. After 24-hr chemical treatment, the cells were lysed and analyzed for luciferase activity using the Bright-Glo luciferase substrate (Promega). The effect of compound on luciferase activity was quantified by calculating the fold-induction compared to samples treated with DMSO alone, which served as negative controls. BMP4 at 10 ng/ml was used as a positive control. The compounds showing more than 5-fold induction were chosen as the primary hits for further analysis (Figure S1B; Table S2). Although the cells were cultured in serum-free medium for 24 hr in the primary assay, there might remain some baseline level of BMP signaling (as demonstrated below). In this case, we might expect to identify both BMP agonists and BMP sensitizers in the primary screen. The primary hit compounds were titered and three of them showed reporter induction at a relatively low effective concentration (1–15 mM). One hit compound, isoliquiritigenin, has been previously reported to upregulate alkaline phosphatase (ALK) expression (Vrijens et al., 2013). Since ALK is an early marker of osteoblast differentiation, we further tested the three hit compounds for their capacity to induce bone formation. We could detect ALK expression. However, the hit compounds were unable on their own to induce the generation of mature osteoblasts. C2C12 myoblasts were treated with hit compounds for 21 days and stained with alizarin red S to examine the mineralization of mature osteoblasts. We did not find cells staining positive by alizarin red S in compound-treated conditions (Figure S1C). Since hit compounds themselves did not significantly induce a functional BMP cellular output, we tested whether hit compounds make the cells more sensitive to BMP4 treatment, considering that the primary screen might not be entirely depleted of BMP signals. Using the same platform as for primary screening (Figure 1A), after overnight starvation, C2C12 myoblasts expressing pId2-LucGFP were treated with hit compounds together with 1 ng/ml BMP4. One hit compound, PD407824 (Figure 1B), sensitized C2C12 myoblasts to 1 ng/ml BMP4 or 0.1 ng/ml BMP4 treatment to increase luciferase activity (Figure 1C). Two criteria were used to define the sensitizing effect. First, activity of the reporter in cells treated with PD407824 and BMP4 simultaneously is higher than the sum of the activities in cells treated with PD407824 or BMP4 separately. More importantly, the half maximal effective concentration (EC50) of PD407824 decreases as the BMP4 concentration increases (EC50 = 12.3 mM in the absence of BMP4; EC50 = 0.75 mM in the presence of 0.1 ng/ml BMP4; EC50 = 0.12 mM in the presence of 1 ng/ml BMP4). Neither of the two other primary screen hits, isoliquiritigenin or lindane, showed the same synergistic effect with BMP4. To determine whether PD407824 + BMP4 activates endogenous Id2 expression, C2C12 myoblasts were treated with 10 mM or 1 mM PD407824 plus different doses of BMP4 for 24 hr and analyzed for Id2 expression with quantitative RT-PCR. Consistent with the pId2-LucGFP assay, PD407824 sensitizes C2C12 myoblasts to BMP4 treatment to increase transcription of the endogenous Id2 gene. DMSO + 10 ng/ml BMP4
increased Id2 transcript levels 3.7 ± 1.0-fold; while 1 or 10 mM PD407824 + 10 ng/ml BMP4 treatment increased the Id2 transcript levels 9.6 ± 2.2-fold, and 11.5 ± 0.7-fold, respectively (Figure 1D). Together, these data suggest that PD407824 can sensitize C2C12 myoblasts to sub-threshold concentrations of BMP4 treatment to upregulate expression of Id2, a known target of the BMP-signaling pathway. To examine whether PD407824 does in fact affect BMPsignaling components, we used western blotting experiments to measure the phosphorylation of SMAD1/5/9 at Ser 463/ 465 (Figure 1E). As shown by the DMSO control, we confirmed a low basal level of activated SMAD1/5/9. However, PD407824+BMP4 quantitatively increased SMAD1/5/9 phosphorylation. Interestingly, PD407824 used alone at the same concentration showed a trend to slightly decrease phosphorylated SMAD1/5/9, which was however not statistically significant (Figure 1F). In addition, qRT-PCR was used to measure the expression of Id1, another known key target gene of BMP signaling. PD407824 sensitizes C2C12 myoblasts to subthreshold concentrations BMP4 treatment to upregulate Id1 transcript levels in a dose-dependent manner (Figure 1G). Together, these data suggest that PD407824 sensitizes C2C12 myoblasts to sub-threshold concentrations of BMP4 to activate BMP signaling. PD407824 Synergizes with BMP4 to Reprogram C2C12 Myoblasts toward a Mature Osteoblast Fate BMPs are key drivers of bone formation, and previous studies indicated that BMP treatment can reprogram C2C12 myoblasts to osteoblasts (Katagiri et al., 1994). To determine the impact of PD407824 in mediating BMP-dependent bone formation, C2C12 myoblasts were treated with DMSO as a control, PD407824, 3 ng/ml BMP4, PD407824 + 3 ng/ml BMP4, or 10 ng/ml BMP4. After 6 days of culture, cells were fixed and stained with ALK substrate. In the DMSO control, C2C12 myoblasts spontaneously differentiated into myotubes, but there were no ALK-expressing cells (which would display a purple color from the stain). There were a very limited number of ALK-expressing cells detected in the sample treated with 3 ng/ml BMP4, suggesting this is near or below a bone forming threshold. Indeed, the sample containing 10 ng/ml BMP4 (with DMSO but no compound added) showed a strong induction of ALK expression, although there remained numerous myotubes, suggesting that DMSO + 10 ng/ml BMP4 is not sufficient to block myogenesis. PD407824 treatment alone was not sufficient to block myogenesis or induce ALK expression. However, cells treated with PD407824 + 3 ng/ml BMP4 expressed high levels of ALK and entirely lost the capacity to form myotubes (Figure 2A). Next, C2C12 myoblasts were treated with different doses of PD407824 in the presence or absence of 3 ng/ml BMP4. Six days later, the cell lysates were analyzed for ALK activity using an absorbance-based ALK substrate. The quantitative results confirmed that PD407824 itself does not induce ALK activity; however, PD407824 synergized with 3 ng/ml BMP4 to reprogram C2C12 myoblasts to ALK-expressing cells in a dosedependent manner. The ALK activity generated by cells treated with 1 mM PD407824 + 3 ng/ml BMP4 was equal to that of cells treated with 10 ng/ml BMP4 alone (Figure 2B).
Cell Reports 15, 2063–2075, May 31, 2016 2065
A
B
C
D
Figure 2. PD407824 Synergizes with BMP4 to Reprogram C2C12 Myoblasts into Mature Osteoblasts (A) C2C12 myoblasts were reprogrammed into cells expressing osteogenic marker alkaline phosphatase (ALK). (B) Quantification of ALK activity by fluorescence-based substrate. (C) qRT-PCR to detect the expression of mature osteoblast markers, including collagen I (collagen), osteopontin (OPN), and osteocalcin (OC). (D) Alizarin red S staining of the mature osteoblasts. PD: PD407824. Scale bar, 50 mm.
In addition to ALK, we also examined by qRT-PCR the expression of other osteoblast markers, including collagen I, osteopontin and osteocalcium. Consistent with the ALK substrate assay, PD407824 or 3 ng/ml BMP4 itself was not sufficient to induce bone formation. In contrast, PD407824 + 3 ng/ml BMP4 treatment was able to effectively induce the expression of osteoblast markers. The expression levels of these markers in cells treated with 1 mM PD407824 + 3 ng/ml BMP4 was comparable to those in the cells treated with 10 ng/ml BMP4 (Figure 2C). We also examined the presence of mineralization to evaluate the formation of mature osteoblasts. C2C12 myoblasts were treated with DMSO as control, PD407824 alone, 3 ng/ml BMP4, PD407824 + 3 ng/ml BMP4, or 10 ng/ml BMP4 for 22 days and then stained with alizarin red S solution. In the cells treated with DMSO, 3 ng/ml BMP4, or 1 mM PD407824, no alizarin staining was detected. The cells treated with PD407824 + 3 ng/ml BMP4 showed strong mineralization, comparable to the cells treated with 10 ng/ml BMP4 (Figure 2D). Altogether, these data suggest that PD407824 synergizes with BMP4 to reprogram C2C12 myoblasts into mature osteoblasts. PD407824 as a BMP Sensitizer to Induce hESC Differentiation to Mesoderm BMP4 has been widely used to induce hESC differentiation. The effect of BMP4 in hESC differentiation is context dependent. When cells are allowed to differentiate as embryoid bodies (EBs), the addition of BMP4 directs mouse and human ESC differ-
2066 Cell Reports 15, 2063–2075, May 31, 2016
entiation to mesoderm, which subsequently generates beating cardiomyocytes (Cao et al., 2013; Kattman et al., 2011; Yang et al., 2008). In order to determine whether PD407824 increases hESC sensitivity to BMP treatment in a mesodermal differentiation protocol, hESC H9 cells were treated with DMSO or three different concentrations (0.1 mM, 0.3 mM, and 1 mM) of PD407824 combined with 3 ng/ml BMP4 for 4 days and analyzed with qRTPCR. Consistent with the effect on C2C12 cells, PD407824 sensitizes H9 cells to BMP4 treatment to increase transcription of the endogenous Id2 gene. 1 mM PD407824 + 3 ng/ml BMP4 treatment increased the Id2 transcript levels 7.1 ± 0.3-fold higher than DMSO + 3 ng/ml BMP4 (Figure 3A). Strikingly, PD407824 + 3 ng/ml BMP4 treatment increases transcript levels of Brachyury T, a mesoderm marker, similar as when treated with 10 ng/ml BMP4, suggesting PD407824 sensitizes hESCs to BMP4 treatment during mesoderm differentiation. To further validate that PD407824 + BMP4 induces the generation of functional mesoderm, the chemically treated cells were examined for their capacities to form beating cardiomyocytes, which are mesoderm derivatives. After 4 days treatment with PD407824 + BMP4, EBs were subsequently cultured in cardiac differentiation medium initially containing 20% FBS, which after 10 days was reduced to 2.5%. We used DMSO + 10 ng/ml BMP4 as a positive control, and after 10 days counted the number of beating colonies. In wells treated with 1 mM PD407824 + 3 ng/ml BMP4 the number of beating colonies was comparable to wells treated with 10 ng/ml BMP4 (Figure 3B); notably, treatment with 3 ng/ml BMP4 alone was
A
B
C
D
E
Figure 3. PD407824 Sensitizes Cells to BMP4 Treatment during hESC Mesodermal Differentiation (A) qRT-PCR analysis of ID2 and Brachyury T expression. EBs were treated with either DMSO or three different concentrations (0.1, 0.3, or 1 mM) of PD407824 combined with two concentrations (3 or 10 ng/ml) of BMP4 for 4 days and analyzed with qRT-PCR. DMSO + 3 ng/ml BMP4 was used as a control to normalize data. DMSO + 10 ng/ml BMP4 was used as a positive control. (B) Quantification of beating colonies on day 10. (C) qRT-PCR analysis of the expression levels of human cardiomyocyte markers TNNT2, MHY6, and MLC2v. Levels were normalized to DMSO + 3 ng/ml BMP4. RNA was isolated on day 12 for qRT-PCR analysis. (D) Immunofluorescence staining for TNNT2 expression. Scale bar, 100 mm. Cells were fixed on day 12 and stained for the cardiomyocyte marker TNNT2. (E) Intracellular flow cytometry and quantification of the percentage of TNNT2+ cells.
not sufficient to generate beating colonies. Importantly, the cells treated with 1 mM PD407824 + 3 ng/ml BMP4 expressed high levels of cardiomyocyte differentiation makers, including troponin T type 2 (TNNT2), myosin heavy chain 6 (MHY6), and myosin light chain 2v (MLC2v), at day 12 of differentiation, shown by qRT-PCR (Figure 3C). The expression levels of these genes in the cells treated with 1 mM PD407824 + 3 ng/ml BMP4 treatment were comparable to cells treated with 10 ng/ml BMP4. Moreover, we used immunocytochemistry to examine TNNT2 protein expression. Multiple concentrations (0.1, 0.3, and 1 mM) of PD407824 + 3 ng/ml BMP4 efficiently enhanced TNNT2 expression (Figure 3D). Intracellular flow cytometry was used to quantify the differentiation efficiency. The percentage of TNNT2+ cells in the samples treated with 1 mM PD407824 + 3 ng/ml BMP4 (42.6% ± 4.3%) was significantly enhanced compared to 3 ng/ml BMP4 alone (5.5% ± 0.7%), achieving the level found in cells treated with 10 ng/ml BMP4 (47.5% ± 6.7%) (Figure 3E). Thus, PD407824 sensitizes cells to sub-threshold concentrations of BMP4 during hESC mesodermal differentiation.
PD407824 Sensitizes hESCs to BMP4 during Differentiation toward Cytotrophoblast Stem Cell Lineage The effect of BMP4 on hESC differentiation is stage and context dependent. We next sought to determine if the action of PD407824 is specific to BMP-dependent mesodermal differentiation or can be more generally applied to BMP-dependent differentiation programs. When hESCs are grown in the presence of mouse embryonic fibroblast (MEF) conditional medium (CM), BMP4 functions as a key driver for differentiation toward KERATIN7 (KRT7)+/TP63+ cytotrophoblast stem cells (Amita et al., 2013; Das et al., 2007; Li et al., 2013; Xu et al., 2002). To determine if PD407824 functions as a BMP sensitizer in trophoblast differentiation, hESC H1 cells were treated with DMSO or 2.5 mM PD407824 combined with different concentrations of recombinant BMP4 protein (0, 4, 6, 9, 12, or 30 ng/ml) in MEF-CM for 7 days (Figure 4A). The cells were fixed and stained with antiKRT7 antibody (Figure 4B). More KRT7+ cells were detected in PD407824 + BMP4 treatment condition than the corresponding
Cell Reports 15, 2063–2075, May 31, 2016 2067
A
B
C
D
E
Figure 4. PD407824 Sensitizes Cells to BMP4 Treatment during Differentiation of hESCs toward Cytotrophoblast Stem Cell Fate (A) Scheme of human cytotrophoblast stem cell differentiation. hESC H1 cells were treated with DMSO or 2.5 mM PD407824 combined with different concentrations of recombinant BMP4 protein (0, 4, 6, 9, 12, or 30 ng/ml) in MEF-CM for 7 days. (B) Immunohistochemistry analysis of KRT7 expression. Scale bar, 100 mm. (C and D) Intracellular flow cytometry analysis (C) and quantification (D) of the percentage of KRT7+ cells. (E) qRT-PCR analysis of the expression of human cytotrophoblast stem cell markers KRT7, TP63, and GCM1. Levels were normalized to undifferentiated hESCs. See also Figure S2.
2068 Cell Reports 15, 2063–2075, May 31, 2016
A
B
C
BMP4 treatment condition. Intracellular flow cytometry was used to quantify the differentiation efficiency (Figure 4C). PD407824 treatment significantly increased the percentage of KRT7+ cells in different doses of BMP4 treatment. (Figure 4D). Moreover, qRT-PCR analysis showed that cells treated with 2.5 mM PD407824 + 9 ng/ml BMP4 express other cytotrophoblast stem cells markers, such as TP63 (Li et al., 2013) and GCM1 (Figure 4E). The expression levels of these genes when treated with 2.5 mM PD407824 + 9 ng/ml BMP4 are significantly higher than the condition treated with 9 ng/ml BMP4 and were comparable to the condition treated with 30 ng/ml BMP4. We further examined the proliferation and DNA damage in PD407824 treated hESCs. First, we found that there is no significant difference in the number of Ki67+ cells detected comparing DMSO or PD407824 treated hESCs (Figure S2A). To examine whether CHK1 inhibition affects DNA repair, PD407824-treated hESCs were stained using g-H2A.X antibody. Cells treated with 10 mM etoposide were used as a staining control. No g-H2A.X+ cells were detected in 2.5 mM PD407824 or 2.5 mM PD407824+ 9 ng/ml BMP4-treated conditions (Figure S2B), suggesting that PD407824 treatment does not affect DNA repair during hESC differentiation. Together, these data show that PD407824 increases the sensitivity of hESCs to BMP4 treatment by more than 3-fold during differentiation toward cytotrophoblast stem cell fate without significantly affecting cell proliferation, cell cycle, or DNA repair. PD407824 Functions through Inhibiting Checkpoint Kinase 1 A previous study identified PD407824 as a checkpoint kinase 1 (CHK1) inhibitor (Palmer et al., 2006). To determine whether PD407824 functions through inhibiting CHK1, we tested another CHK1 chemical inhibitor, CHIR124, for its ability to sensitize C2C12 myoblasts to BMP4 treatment. C2C12 myoblasts were treated with different concentrations of CHIR124 in the presence or absence of 1 ng/ml BMP4. After 24 hr treatment, qRT-PCR
Figure 5. PD407824 Inhibiting CHK1
Functions
through
(A and B) CHIR124, a CHK1 inhibitor, mimics the effect of PD407824 to sensitize BMP4-induced upregulation of Id2 expression (A) and increased SMAD1/5/9 phosphorylation (B) in C2C12 myoblasts. The relative Id2 transcript level was normalized to DMSO treatment conditions. The relative SMAD1/5/9 phosphorylation was normalized to total SMAD1. (C) Relative Id2 transcript expression in CHK1+/ and CHK1/ cells. Untransfected cells (wild-type [wt]) and cells infected with sgRNA targeting GFP were used as controls. See also Figure S3.
was used to monitor the transcript levels of Id2. Similar to PD407824, CHIR124 synergized with BMP4 to upregulate Id2 in a dose-dependent manner (Figure 5A). Consistently, 10 mM CHIR124 + BMP4 increases the relative phosphorylation of SMAD1/5/9 (Figures 5B and S3A), showing the similar trend as PD407824 + BMP4 treatment (Figure 1E). Moreover, we complemented the chemical approach using a genetic strategy. C2C12 cells were infected with lentivirus expressing Cas9 and single guide RNA (sgRNA) targeting CHK1 or, as a control, sgRNA targeting GFP. Mutant cells were subcloned. Western blotting experiments confirmed that the monoallelic and biallelic knockout of CHK1 protein in CHK1+/ and CHK1/ clones, respectively (Figure S3B). After overnight starvation, the CHK1+/ and CHK1/ C2C12 cells were treated with 1 ng/ml or 3 ng/ml BMP4. After 24 hr treatment, Id2 transcript levels were analyzed with qRT-PCR. Compared to uninfected or GFP-targeted cells, CHK1+/ and CHK1/ cells show increased expression of Id2 in the absence of added BMP4 and, similar to PD407824-treated cells, in the presence of 1 ng/ml BMP4 (Figure 5C). In the presence of 3 ng/ml BMP4, CHK1/ cells show higher Id2 expression than CHK1+/ cells or cells treated with PD407824, suggesting that CHK1 might play a dose dependent effect. In addition to CHK1, PD407824 also inhibits Wee1 and PKC (Palmer et al., 2006). To determine whether Wee1 or PKC inhibition is involved in the sensitizing effect of PD407824, C2C12 cells were treated with different concentrations of MK1775, a Wee1 inhibitor (Hirai et al., 2009), or bisindolylmaleimide II, a PKC inhibitor, neither of which inhibit CHK1 activity, in the presence or absence of 1 ng/ml BMP4. After 24-hr treatment, qRT-PCR was used to monitor the transcript levels of Id2. In contrast to CHK1 inhibitors, neither MK1775 nor bisindolylmaleimide II sensitizes C2C12 cells to BMP4 treatment (Figures S2C and S2D). Together, these data support a mechanism by which PD407824 impacts BMP activity through inhibition of CHK1, but not Wee1 or PKC. Mechanism of Action: Increasing the Sensitivity to BMP through Depletion of SMAD2/3 In the canonical BMP pathway, the ligands exert their effects through binding to type II serine/threonine kinase receptors (BMPR2 for BMPs), forming an oligomeric complex. The type II
Cell Reports 15, 2063–2075, May 31, 2016 2069
receptors are consequently activated and phosphorylated and then activate the type I receptors (ALK3 and ALK6 for BMPs; Bandyopadhyay et al., 2013). While there are SMADindependent pathways (Bandyopadhyay et al., 2013), a major mode of signal transduction occurs when activated type I receptors phosphorylate intracellular receptor-regulated R-SMADs (SMAD1/5/9 for BMPs), which associate with the co-SMAD (SMAD4) and translocate into the nucleus. The SMAD complexes interact with additional transcription factors to bind promoters of target genes, such as Id2, to control their expression. Since PD407824 + BMP4 treatment increases the Id2 expression levels in both C2C12 cells (Figure 1D) and hESCs (Figure 3A), we first used co-immunoprecipitation experiments to determine whether PD407824 + BMP4 treatment leads to an increased level of SMAD1/5/9 binding with SMAD4. Cells were treated with 3 ng/ml BMP4 in the presence or absence of 10 mM PD407824 for 24 hr. Proteins from the cell lysate were precipitated with anti-SMAD4 antibody and subsequently probed by blotting using anti-SMAD1 antibody. SMAD1 protein bound to SMAD4 was most abundant in the presence of PD407824 + BMP4 compared to other conditions (Figure 6A). Since both SMAD2/3 (R-SMAD for transforming growth factor beta [TGF-b]) and SMAD1/5/9 are capable of binding with SMAD4, we used western blotting to examine the expression of SMAD2/3 at the protein level. Strikingly, western blotting experiments showed that PD407824+BMP4 treatment significantly decreased the total protein levels of SMAD2/3 protein (Figure 6B). In addition, CHIR124, another CHK1 inhibitor, also decreased the total levels of SMAD2/3 protein (Figure S4A). This might due to either reduced levels of SMAD2/3 gene expression or increased degradation of SMAD2/3 protein. To distinguish these possibilities, C2C12 were treated with PD407824 or PD407824 + BMP4 for 24 hr and analyzed with qRT-PCR. Neither PD407824 nor PD407824+BMP4 treatment significantly changed the transcript levels of SMAD2/3 (Figure S3B), suggesting that PD407824 might affect SMAD2/3 protein stability. Previous studies suggested that activated CDK8/9 phosphorylates SMAD2/3 at a linker region (pT220 for SMAD2 and pT179 for SMAD3), which is ultimately recognized by ubiquitin ligase Nedd4L, leading to proteasome-mediated turnover of SMAD2/3 (Alarco´n et al., 2009; Gao et al., 2009). Western blotting experiments were used to measure the phosphorylation of SMAD2/3 at the linker region. In conditions using PD407824 or PD407824+BMP4, the relative phosphorylation (normalized to total SMAD2/3 protein) of linker region (T179 for SMAD3) was significantly elevated compared to BMP4 treatment alone (Figure 6C). The hyper-phosphorylation of the linker region leads to the decreased levels of SMAD2/3 proteins (Figure 6B). Since CDK9 phosphorylates the SMAD2/3 linker region, both chemical and genetic approaches were used to determine whether CDK9 activation is required for the sensitizing effect of PD407824 to upregulate Id2. The cells were treated with or without 3 ng/ml BMP4 plus PD407824 in the presence or absence of 300 nM flavopiridol, a chemical CDK inhibitor, for 24 hr and analyzed with qRT-PCR experiments to measure endogenous Id2 expression levels. Flavopiridol does not affect the activity of BMP4 alone. However, flavopiridol blocks the sensitizing effect of PD407824 with BMP4 (Figure 6D). In
2070 Cell Reports 15, 2063–2075, May 31, 2016
addition, flavopiridol blocks PD407824-induced elevation of SMAD3 T179 phosphorylation, as well as the decrease of total SMAD2/3 protein (Figures 6E and S4C). Moreover, flavopiridol blocks PD407824-induced enhanced binding between SMAD1 and SMAD4 proteins (Figure 6F). Furthermore, C2C12 cells were infected with lentivirus expressing Cas9 and sgRNA targeting CDK9. C2C12 cells infected with lentivirus expressing sgRNA targeting GFP were used as a negative control. After subcloning, the monoallelic and biallelic knockout of CDK9 protein in CDK9+/ and CDK9/ cells was confirmed by western blotting, respectively (Figure S3D). Consistent with the flavopiridol treatment results, CDK9+/ and CDK9/ cells lose the ability to respond to PD407824 treatment, which further confirms that CDK9 activation is required for the ability of PD407824 to sensitize cells to BMP (Figure 6G). It has been shown that p21 inhibits enzymatic activity of CDK9 (Chen et al., 2011). We used western blotting experiments to measure p21 expression levels and found that either PD407824 alone or PD407824 + BMP4 treatment significantly decreased p21 protein levels (p < 0.01) (Figure 6H). In addition, CHIR124 also decreased p21 protein expression levels (Figure S4E). To determine whether p21 depletion is sufficient to increase the sensitivity to BMP4 treatment, C2C12 cells were infected with lentivirus expressing Cas9 and sgRNA targeting p21 and mutant cells were subcloned. C2C12 cells infected with lentivirus expressing sgRNA targeting GFP were used as a negative control. Western blotting confirmed the monoallelic and biallelic knockout of p21 protein expression in p21+/ and p21/ cells, respectively (Figure S3F). The p21+/ and p21/ cells showed increased expression of Id2 in the presence of BMP4, comparable to PD407824-treated cells in the presence of 3 ng/ml BMP4 (Figure 6I). In addition, flavopiridol treatment does not block PD407824 induced decreased expression of p21 protein (Figure S3G), suggesting that CDK9 functions as the downstream of p21. These data suggest that PD407824 impacts BMP activity through p21 inhibition. Overall, the data suggest that PD407824 is sufficient to block CHK1, leading to a significant downregulation of p21, causing activation of CDK8/9, which in turn causes depletion of SMAD2/3. SMAD2/3 competes with SMAD1/5/9 for limiting SMAD4, suggesting that PD407824 effectively sensitizes with sub-threshold BMP by relieving this limitation, freeing SMAD4 to interact productively with the BMP effector SMAD1 (Figure 7A). To further validate this proposed mechanism of action, we created CHK1//p21/ and CHK1//CDK9/ doubleknockout C2C12 cell lines, based on the parental p21/ (Figure S4E) and CDK9/ cells (Figure S4D), respectively. The knockout of CHK1 protein in CHK1//p21/ and CHK1// CDK9/ cells was validated by western blotting experiments (Figure S5A). The double-knockout cells and single-knockout cells were cultured in the absence or presence of 1 ng/ml or 3 ng/ml BMP4. Consistent with the proposed mechanism that p21 is downstream of CHK1, there is no distinguishable difference in the levels of Id2 transcripts comparing CHK1// p21/, CHK1/ and p21/ cells (Figure 7B). In addition, there is no distinguishable difference in the transcript levels of osteoblast markers, including ALP, collagen, OPN, and OC, comparing CHK1//p21/, CHK1/ and p21/ cells after
A
B
D C
G
E
F
I H
Figure 6. PD407824 Increases BMP Sensitivity by Depleting p21 and Activating CDK9, which Then Phosphorylates the SMAD2/3 Linker Region, Leading to Decreased Levels of SMAD2/3 Protein and Enhanced Levels of Nuclear SMAD1 (A) PD407824 treatment increases the binding of SMAD1 and SMAD4 proteins indicated by co-immunoprecipitation. (B) PD407824 causes depletion of SMAD2/3 protein levels, measured by western blotting; SMAD2/3 protein levels are normalized to beta-ACTIN protein levels. (C) PD407824 induces the phosphorylation of the SMAD2/3 linker region (T179 for SMAD3). The phosphorylation level was normalized to total levels of SMAD2 protein. (D–F) Flavopiridol, a CDK inhibitor, blocks the PD407824 induced increased Id2 transcript expression (D), upregulated T179 phosphorylation of SMAD3 and decreased SMAD2/3 total protein (E), and increased binding of SMAD1 and SMAD4 proteins indicated by co-immunoprecipitation (F). (G) Knockout of CDK9 blocks the synergistic effect of PD407824. (H) PD407824 depletes p21 protein levels, shown by western blotting. The p21 protein levels are normalized to b-actin protein levels. (I) Relative Id2 transcript expression in p21+/ and p21/ cells. Uninfected cells (wild-type [wt]) and cells infected with sgRNA targeting GFP were used as negative controls. See also Figure S4.
7 days of osteogenic differentiation (Figure S5C). Moreover, CDK9 knockout in CHK1/ cells ablates the ability of CHK1/ cells to respond to BMP4 treatment regarding Id2 transcript upregulation (Figure 7C) and capacity to differentiate toward oste-
oblasts (Figure S5D). Furthermore, SB431542, a TGF-b type 1 receptor kinase inhibitor mimics the ability of PD407824 to sensitize C2C12 cells to BMP4 treatment (Figure S5B). Finally, the proposed mechanism was confirmed using hESCs. After
Cell Reports 15, 2063–2075, May 31, 2016 2071
CDK9
C P SMAD4
Linker Region
P
P SMAD2/3
Tail Region
P SMAD1/5/9
2 0
1 ng/ml BMP4
3 ng/ml BMP4
**
14
pSMAD1/5/9
12 10 8
*
6
*
sgGFP CHK1-/CHK1-/-; CDK9-/-
4
p21 pSMAD3 (T179) SMAD2/3
**
16
lB MP 4 3n g/m 9n lB g/m MP 4 lB MP PD 4 +9 ng /m lB MP 4
4
10
sgGFP CHK1-/p21-/CHK1-/-; p21-/-
PD +
6
g/m
* **
3n
n.s.
8
SO
10
PD
12
μM
D
** ** *
14
none
p21
Relative Id2 mRNA expression
ALK1/2/3/6
ALK4/5
TGFβR1
CHK
BMPR1
BMP
TGFβ
n.s.
16
DM
B Relative Id2 mRNA expression
A
SMAD1 beta-ACTIN
2 0
ubiquitination
Id1, Id2 etc
none
1 ng/ml BMP4
3 ng/ml BMP4
Figure 7. Suggested Mechanism of Action (A) Suggested mechanism of action. (B) Relative Id2 transcript expression in CHK1/, p21/, and CHK1/p21/ cells treated with or without BMP4. (C) Relative Id2 transcript expression in CHK1/ and CHK1/CDK9/ cells treated with or without BMP4. (D) Western blotting analysis of hESCs treated with DMSO, 10 mM PD407824, 3 ng/ml BMP4, 10 mM PD407824 + 3 ng/ml BMP4, 9 ng/ml BMP4, or 10 mM PD407824 + 9 ng/ml BMP4. H1 cells were starved in serum-free E6 medium for 24 hr, treated with different conditions for 24 hr, and then analyzed using western blotting. See also Figure S5.
24-hr treatment with DMSO, 10 mM PD407824, 3 ng/ml BMP4, 3 ng/ml BMP4 + 10 mM PD407824, 9 ng/ml BMP4, or 9 ng/ml BMP4 + 10 mM PD407824, hESC lysates were analyzed with western blotting experiments. PD407824, 3 ng/ml BMP4 + 10 mM PD407824, 9 ng/ml BMP4, and 9 ng/ml BMP4 + 10 mM PD407824 treatment decreases p21 protein expression levels, increases the relative phosphorylation of SMAD3 (T179), decreases the total levels of SMAD2/3 proteins, increases the relative phosphorylation of SMAD1/5/9 (Figures 7D and S5E), and leads to enhanced Id2 transcript expression (Figure 3A). DISCUSSION We discovered a previously unreported mechanism by which inhibiting CHK1 can indirectly sensitize the activity of BMP. It does so by downregulating p21 levels, which activates CHK8/9, leading to phosphorylation of the SMAD2/3 linker region and, finally, decreased levels of SMAD2/3 proteins (Figure 6). When BMP ligand is present, SMAD1/5/9 may therefore associate with the increased available SMAD4, and more effectively activate BMP downstream targets, such as Id1 and Id2. This study suggests that SMAD4 is a limiting factor for BMP activation, consistent with our previous observations of competition among SMADs (Cook et al., 2011). Thus, through a screen for BMP agonists, we found a previously unrecognized indirect mechanism to regulate BMP signaling by targeted depletion of SMAD2/3. While PD407824 causes increased levels of phospho-SMAD1/5/9, this may occur through an indirect effect by increasing steadystate levels of active complexes, rather than, for example, enhanced ALK2/3/6 activity. Interestingly, PD407824 itself slightly decreased phospho-SMAD1/5/9 levels (Figure 1E), although it does not reach significance. Previous studies have shown that CHK1 directly induces Smurf1 phosphorylation and
2072 Cell Reports 15, 2063–2075, May 31, 2016
mediates Smurf1 degradation (Wang et al., 2014), which could impact Smad1/5 proteasomal degradation (Zhu et al., 1999). It is possible that CHK1 inhibition reduces both SMAD1/5/9 and SMAD2/3, but much more specifically SMAD2/3. This dynamic competition results in increased binding between SMAD1 and SMAD4 (Figure 6A) at the expense of SMAD2/3 (Figure 6B) and sensitizes the cells to BMP4 treatment (Figures 1D and 1G). While BMP pathway activation leads to phosphorylation of SMAD1/5/9, all the other TGF-b superfamily members, including TGF-b, Activin, and Nodal, cause phosphorylation of SMAD2/3. Many TGF-b inhibitors have been identified, most of which target type I receptors, such as ALK4 or ALK5. Here, we identified a class of compounds, checkpoint kinase inhibitors, which block TGF-b signaling by causing SMAD2/3 depletion. These compounds can therefore be used as useful tools to regulate TGF-b signaling. Checkpoint kinase inhibitors, including MK-8776, LY2606368, and LY2603618, have been used in phase 1 and 2 clinical trials for the treatment of acute myeloid leukemia or breast or ovarian cancers, by themselves or in combination therapies (Lapenna and Giordano, 2009). Here, we found a previously unreported function for checkpoint kinase inhibitors, which sensitize C2C12 cells hESCs with sub-threshold BMP to activate BMP signaling. In addition, recombinant BMPs are widely used for ESC maintenance and for directed differentiation, such as inducing mesoderm, cardiomyocytes, hepatocytes, trophoblasts, germ cells, and brown adipocytes. Recent studies suggested that BMP4 and BMP7 induce the white adipocyte to brown adipocyte transition (Elsen et al., 2013). However, recombinant BMPs are relatively unstable and costly, which limits their application. In our studies, we showed that small-molecule CHK1 inhibitors synergize with BMP4 to induce hESC differentiation toward a beating cardiomyocyte or cytotrophoblast stem
cell fate. Considering the broad application of BMPs in stem cell biology, CHK1 inhibitors can be a useful reagent to control stem cell or progenitor fate. Due to cellular context, the concentrations of CHK1 inhibitors may need be optimized, depending on the particular differentiation protocol and medium employed. While recombinant BMPs have been promising in preclinical studies and even progressed to clinical trials, the clinical outcomes have been less dramatic than might have been hoped. These CHK1 inhibitors, which have themselves been used in phase 1 and 2 clinical trials, might be applicable for therapeutic value, being both stable and cost-efficient, to partially or fully replace recombinant BMPs to treat relevant disorders, such as longbone non-unions, acute fractures, and spinal fusions. EXPERIMENTAL PROCEDURES Cell Culture C2C12 myoblasts were purchased from the American Type Culture Collection (ATCC). C2C12 myoblasts were maintained in complete DMEM supplemented with 10% fetal calf serum, 100 U penicillin, and 100 mg streptomycin at 37 C in 5% CO2. Human ESC H1 and H9 cells were purchased from WiCell Research Institute. Human BMP4 was purchased from R&D Systems. Tissue culture media and sera were purchased from VWR. Bright-Glo luciferase assay reagent was purchased from Promega. Chemicals PD407824 and flavopiridol were purchased from Sigma-Aldrich. Purity of the compounds was higher than 98% high-performance liquid chromatography (HPLC) based on the manufacturer’s information. CHIR124 was purchased from Tocris Bioscience. Unless indicated, all other chemicals were purchased from Sigma-Aldrich. High-Content Small-Molecule Screening Using the Id2 Reporter C2C12 myoblasts were seeded onto 10-cm2 plates in DMEM medium supplemented with 10% FBS. On the second day, the reporter plasmid pId2-LucGFP was transfected into C2C12 myoblasts according to the manufacturer’s protocol for X-treme GENE HP DNA Transfection Reagent (Roche) in serum-free media. 24 hr after transfection, cells were transferred to one 384-well plate per 10-cm dish with 5 3 105 cells/ml in DMEM supplemented with 10% FBS. After another 24 hr, cells were cultured in serum-free conditions for an additional 24 hr and reporter gene expression was induced upon the addition of compounds or BMP4. Cells were then incubated for 24 hr, and luciferase activity was measured using the Bright-Glo luciferase substrate (Promega) following the manufacturer’s recommendations. The effect of compound on luciferase activity was quantified by calculating the fold-induction, determined using the ratio A/B, where A represents the activity in the presence of compound or BMP4 and B represents the activity in wells with only DMSO added. The compounds with more than 5-fold induction were chosen as the primary hit compounds for further analysis. Human ESC Culture and Differentiation Human ESC lines H1 and H9 were cultured on irradiated MEFs and maintained in hESC medium, which contains 80% DMEM/F12, 20% knockout serum replacement, 10 ng/ml human recombinant basic fibroblast growth factor, 2 mM GlutaMax, 0.1 mM nonessential amino acids, and 0.1 mM b-mercaptoethanol. The medium was changed daily. For cardiac differentiation, H9 cells were detached using 0.25%Trysin-EDTA for 1 min at room temperature to form EBs. The EBs were cultured in suspension with hESC medium for 24 hr and then moved to differentiation medium containing 80% knockout DMEM, 20% FBS (Hyclone), 2 mM GlutaMax, 0.1 mM nonessential amino acids, and 0.1 mM b-mercaptoethanol with DMSO or PD407824 combined with 3 or 10 ng/ml BMP4 on ultralow cluster plates (Costar) for 4 days. On day 4, EBs were plated on 0.1% gelatin-coated plates. The following day, differentiated cells were treated with WNT antagonist I (IWR-1, Stemgent) for 3 days.
Beating colonies were counted at day 10. Cells were then dissociated with Accutase for 20 min and replated on gelatin-coated plates at a 1:1 ratio. After day 10, the concentration of FBS was reduced from 20% to 2.5%. RNA was isolated for qRT- PCR analysis or cells were fixed with 4% paraformaldehyde (PFA) for 10 min, followed by staining with anti-TNNT2 antibody for image analysis on day 12. To differentiate hESCs to cytotrophoblast stem cells, hESC H1 cells were treated with DMSO or 2.5 mM PD407824 combined with different concentrations of recombinant BMP4 protein (0, 4, 6, 9, 12, or 30 ng/ml) in MEF-CM for 7 days. RNA was isolated for qRT-PCR analysis or cells were fixed with 4% PFA for 10 min, followed by staining with anti-KRT7 antibody for image analysis. Western Blotting Analysis At the indicated time points, cells were collected in complete lysis buffer (20 mM Tris [pH 7.5], 150 mM NaCl, 50 mM NaF, 1% NP-40 substitute, and Thermo Scientific HALT protease inhibitor cocktail 1:100) and lysates loaded on 10% NuPage Bis-Tris pre-cast gels (Invitrogen). After being resolved by electrophoresis, proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Bio-Rad). Membranes were blocked with 5% BSA in TBS + 0.05% Tween, and then incubated with primary antibody overnight. The antibodies were rabbit anti-phospho-SMAD1/5/9 (Cell Signaling Technology, 9511), rabbit anti-SMAD1 XP (Cell Signal, 6944), rabbit anti-SMAD2/3 (Cell Signal, 3102), rabbit anti-p21 (C-19 antibody; Santa Cruz Biotechnology, sc379), rabbit anti-phospho-SMAD3 T179 (Rockland, 600-401-C48S), and mouse anti-b-actin (Sigma-Aldrich, A1978). The membranes were washed and incubated with secondary antibody for 1 hr at room temperature in 5% milk-TBS-0.05% Tween and developed using SuperSignal West Pico ECL substrate (Thermo Fisher Scientific). Construction of Lentiviral CRISPR Gene Expression Knockdown Plasmids sgRNA sequences were designed using the website http://www.genomeengineering.org/. The target sequences were as follows: sgRNA-eGFP, 50 GGGCGAGGAGCTGTTCACCG-30 ; sgRNA-CHK1, 50 -TAGTGAAATTCTATGG CCAC-30 ; sgRNA-p21, 50 -CTTCGTCACGGAGACGCCGC-30 ; and sgRNACDK9, 50 -ATTGAGATTTGTCGGACCAA-30 . Each target sequence was cloned into the lentiCRISPRv2 vector (Addgene, plasmid # 52961; Sanjana et al., 2014; Shalem et al., 2014) to make knockdown plasmids. C2C12 cells were infected with lentivirus expressing sgRNAs. Two days after infection, cells were treated with 2 mg/ml puromycin for 48 hr. The cells after selection were treated with compounds and used for qRT-PCR analysis. Intracellular Fluorescence-Activated Cell Sorting Analysis The hESC-derived cells were dissociated by Accutase (STEMCELL Technologies), fixed and stained with primary antibody at 4C overnight. Primary antibodies were anti-TNNT2 (1:500) and anti-KRT7 (1:500). After washing three times, cells were incubated with secondary antibodies for 1 hr at room temperature. These were Alexa-488-conjugated donkey anti-mouse or AlexaFluor-647-conjugated donkey anti-guinea pig secondary antibody (1:500) using cytofix/cytoperm (BD Bioscience) and wash buffer (BD Biosciences) according to instructions by the manufacturer. Flow cytometry data were obtained using an Accuri C6 flow cytometry instrument and analyzed with FlowJo. Statistical Analysis Data represent n = 3 independent biological replicates if not specifically indicated. n.s. indicates a non-significant difference. p values were calculated by unpaired two-tailed Student’s t test if not specifically indicated (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001). SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures and five figures and can be found with this article online at http://dx.doi.org/ 10.1016/j.celrep.2016.04.066.
Cell Reports 15, 2063–2075, May 31, 2016 2073
AUTHOR CONTRIBUTIONS S.C. and T.E. designed research; L.F., B.C., B.L., S.-Y.T., and S.C. performed research; L.F., B.C., B.L., S.-Y.T., T.E., and S.C. analyzed data; and L.F., B.L., S.C., and T.E. wrote the manuscript. CONFLICTS OF INTEREST The authors have filed a patent titled ‘‘Chemical BMP sensitizers that replace or partially replace BMP.’’ ACKNOWLEDGMENTS L.F. is funded by the Natural Science Foundation of China (grants 21272089, 21072073, and 20872044). S.C. is funded by The New York Stem Cell Foundation (R-103), NIDDK (DP2 DK098093-01), the American Diabetes Association (1-12-JF-06), and the Cystic Fibrosis Foundation (CHEN15XX0). S.C. is a New York Stem Cell Foundation-Robertson Investigator. T.E. is funded by the NIH (HL056182 and HL111400). B.C. is funded by the NIH (K01 DK096031). T.Z. is funded by a NYSTEM postdoctoral fellowship. This study was also supported by a Shared Facility contract to T.E. and S.C. from the New York State Department of Health (NYSTEM C029156). LentiCRISPRv2 was a gift from Dr. Feng Zhang. We thank Ting-Chu Liu for help with making constructs. We are also very grateful for technical support and advice provided by Harold S. Ralph in the Cell Screening Core Facility at Weill Cornell Medical College. Received: September 3, 2015 Revised: February 6, 2016 Accepted: April 18, 2016 Published: May 19, 2016 REFERENCES
Cook, B.D., Liu, S., and Evans, T. (2011). Smad1 signaling restricts hematopoietic potential after promoting hemangioblast commitment. Blood 117, 6489– 6497. Cuny, G.D., Yu, P.B., Laha, J.K., Xing, X., Liu, J.F., Lai, C.S., Deng, D.Y., Sachidanandan, C., Bloch, K.D., and Peterson, R.T. (2008). Structure-activity relationship study of bone morphogenetic protein (BMP) signaling inhibitors. Bioorg. Med. Chem. Lett. 18, 4388–4392. Das, P., Ezashi, T., Schulz, L.C., Westfall, S.D., Livingston, K.A., and Roberts, R.M. (2007). Effects of fgf2 and oxygen in the bmp4-driven differentiation of trophoblast from human embryonic stem cells. Stem Cell Res. (Amst.) 1, 61–74. Elsen, M., Raschke, S., Tennagels, N., Schwahn, U., Jelenik, T., Roden, M., Romacho, T., and Eckel, J. (2013). BMP4 and BMP7 induce the white-tobrown transition of primary human adipose stem cells. Am. J. Physiol. Cell Physiol. 306, C431–C440. Fukuda, S., and Taga, T. (2006). [Roles of BMP in the development of the central nervous system]. Clin. Calcium 16, 781–785. Gao, S., Alarco´n, C., Sapkota, G., Rahman, S., Chen, P.Y., Goerner, N., Macias, M.J., Erdjument-Bromage, H., Tempst, P., and Massague´, J. (2009). Ubiquitin ligase Nedd4L targets activated Smad2/3 to limit TGF-beta signaling. Mol. Cell 36, 457–468. Gouon-Evans, V., Boussemart, L., Gadue, P., Nierhoff, D., Koehler, C.I., Kubo, A., Shafritz, D.A., and Keller, G. (2006). BMP-4 is required for hepatic specification of mouse embryonic stem cell-derived definitive endoderm. Nat. Biotechnol. 24, 1402–1411. Hao, J., Ho, J.N., Lewis, J.A., Karim, K.A., Daniels, R.N., Gentry, P.R., Hopkins, C.R., Lindsley, C.W., and Hong, C.C. (2010). In vivo structure-activity relationship study of dorsomorphin analogues identifies selective VEGF and BMP inhibitors. ACS Chem. Biol. 5, 245–253. Heisenberg, C.P., and Solnica-Krezel, L. (2008). Back and forth between cell fate specification and movement during vertebrate gastrulation. Curr. Opin. Genet. Dev. 18, 311–316.
Alarco´n, C., Zaromytidou, A.I., Xi, Q., Gao, S., Yu, J., Fujisawa, S., Barlas, A., Miller, A.N., Manova-Todorova, K., Macias, M.J., et al. (2009). Nuclear CDKs drive Smad transcriptional activation and turnover in BMP and TGF-beta pathways. Cell 139, 757–769.
Hirai, H., Iwasawa, Y., Okada, M., Arai, T., Nishibata, T., Kobayashi, M., Kimura, T., Kaneko, N., Ohtani, J., Yamanaka, K., et al. (2009). Small-molecule inhibition of Wee1 kinase by MK-1775 selectively sensitizes p53-deficient tumor cells to DNA-damaging agents. Mol. Cancer Ther. 8, 2992–3000.
Amita, M., Adachi, K., Alexenko, A.P., Sinha, S., Schust, D.J., Schulz, L.C., Roberts, R.M., and Ezashi, T. (2013). Complete and unidirectional conversion of human embryonic stem cells to trophoblast by BMP4. Proc. Natl. Acad. Sci. USA 110, E1212–E1221.
James, D., Levine, A.J., Besser, D., and Hemmati-Brivanlou, A. (2005). TGFbeta/activin/nodal signaling is necessary for the maintenance of pluripotency in human embryonic stem cells. Development 132, 1273–1282.
Baek, S.H., Choi, S.W., Park, S.J., Lee, S.H., Chun, H.S., and Kim, S.H. (2015). Quinoline compound KM11073 enhances BMP-2-dependent osteogenic differentiation of C2C12 cells via activation of p38 signaling and exhibits in vivo bone forming activity. PLoS ONE 10, e0120150. Bandyopadhyay, A., Yadav, P.S., and Prashar, P. (2013). BMP signaling in development and diseases: a pharmacological perspective. Biochem. Pharmacol. 85, 857–864. Blanpain, C., and Fuchs, E. (2009). Epidermal homeostasis: a balancing act of stem cells in the skin. Nat. Rev. Mol. Cell Biol. 10, 207–217. Cain, J.E., Hartwig, S., Bertram, J.F., and Rosenblum, N.D. (2008). Bone morphogenetic protein signaling in the developing kidney: present and future. Differentiation 76, 831–842. Cao, N., Liang, H., Huang, J., Wang, J., Chen, Y., Chen, Z., and Yang, H.T. (2013). Highly efficient induction and long-term maintenance of multipotent cardiovascular progenitors from human pluripotent stem cells under defined conditions. Cell Res. 23, 1119–1132.
Katagiri, T., Yamaguchi, A., Komaki, M., Abe, E., Takahashi, N., Ikeda, T., Rosen, V., Wozney, J.M., Fujisawa-Sehara, A., and Suda, T. (1994). Bone morphogenetic protein-2 converts the differentiation pathway of C2C12 myoblasts into the osteoblast lineage. J. Cell Biol. 127, 1755–1766. Kato, S., Sangadala, S., Tomita, K., Titus, L., and Boden, S.D. (2011). A synthetic compound that potentiates bone morphogenetic protein-2-induced transdifferentiation of myoblasts into the osteoblastic phenotype. Mol. Cell. Biochem. 349, 97–106. Kattman, S.J., Witty, A.D., Gagliardi, M., Dubois, N.C., Niapour, M., Hotta, A., Ellis, J., and Keller, G. (2011). Stage-specific optimization of activin/nodal and BMP signaling promotes cardiac differentiation of mouse and human pluripotent stem cell lines. Cell Stem Cell 8, 228–240. Kee, K., Gonsalves, J.M., Clark, A.T., and Pera, R.A. (2006). Bone morphogenetic proteins induce germ cell differentiation from human embryonic stem cells. Stem Cells Dev. 15, 831–837. Knight, P.G., and Glister, C. (2006). TGF-beta superfamily members and ovarian follicle development. Reproduction 132, 191–206.
Cao, Y., Wang, C., Zhang, X., Xing, G., Lu, K., Gu, Y., He, F., and Zhang, L. (2014). Selective small molecule compounds increase BMP-2 responsiveness by inhibiting Smurf1-mediated Smad1/5 degradation. Sci. Rep. 4, 4965.
Kruithof, B.P., Duim, S.N., Moerkamp, A.T., and Goumans, M.J. (2012). TGFbeta and BMP signaling in cardiac cushion formation: lessons from mice and chicken. Differentiation 84, 89–102.
Chen, H., Li, C., Huang, J., Cung, T., Seiss, K., Beamon, J., Carrington, M.F., Porter, L.C., Burke, P.S., Yang, Y., et al. (2011). CD4+ T cells from elite controllers resist HIV-1 infection by selective upregulation of p21. J. Clin. Invest. 121, 1549–1560.
Lapenna, S., and Giordano, A. (2009). Cell cycle kinases as therapeutic targets for cancer. Nat. Rev. Drug Discov. 8, 547–566.
2074 Cell Reports 15, 2063–2075, May 31, 2016
Larsson, J., and Karlsson, S. (2005). The role of Smad signaling in hematopoiesis. Oncogene 24, 5676–5692.
Lengerke, C., Schmitt, S., Bowman, T.V., Jang, I.H., Maouche-Chretien, L., McKinney-Freeman, S., Davidson, A.J., Hammerschmidt, M., Rentzsch, F., Green, J.B., et al. (2008). BMP and Wnt specify hematopoietic fate by activation of the Cdx-Hox pathway. Cell Stem Cell 2, 72–82.
Spiekerkoetter, E., Tian, X., Cai, J., Hopper, R.K., Sudheendra, D., Li, C.G., El-Bizri, N., Sawada, H., Haghighat, R., Chan, R., et al. (2013). FK506 activates BMPR2, rescues endothelial dysfunction, and reverses pulmonary hypertension. J. Clin. Invest. 123, 3600–3613.
Li, Z., Fei, T., Zhang, J., Zhu, G., Wang, L., Lu, D., Chi, X., Teng, Y., Hou, N., Yang, X., et al. (2012). BMP4 Signaling Acts via dual-specificity phosphatase 9 to control ERK activity in mouse embryonic stem cells. Cell Stem Cell 10, 171–182.
Takashima, S., and Hartenstein, V. (2012). Genetic control of intestinal stem cell specification and development: a comparative view. Stem Cell Rev. 8, 597–608.
Li, Y., Moretto-Zita, M., Soncin, F., Wakeland, A., Wolfe, L., Leon-Garcia, S., Pandian, R., Pizzo, D., Cui, L., Nazor, K., et al. (2013). BMP4-directed trophoblast differentiation of human embryonic stem cells is mediated through a DNp63+ cytotrophoblast stem cell state. Development 140, 3965–3976. Liu, A., and Niswander, L.A. (2005). Bone morphogenetic protein signalling and vertebrate nervous system development. Nat. Rev. Neurosci. 6, 945–954. Nakahiro, T., Kurooka, H., Mori, K., Sano, K., and Yokota, Y. (2010). Identification of BMP-responsive elements in the mouse Id2 gene. Biochem. Biophys. Res. Commun. 399, 416–421. Nishio, M., Yoneshiro, T., Nakahara, M., Suzuki, S., Saeki, K., Hasegawa, M., Kawai, Y., Akutsu, H., Umezawa, A., Yasuda, K., et al. (2012). Production of functional classical brown adipocytes from human pluripotent stem cells using specific hemopoietin cocktail without gene transfer. Cell Metab. 16, 394–406.
Vrijens, K., Lin, W., Cui, J., Farmer, D., Low, J., Pronier, E., Zeng, F.Y., Shelat, A.A., Guy, K., Taylor, M.R., et al. (2013). Identification of small molecule activators of BMP signaling. PLoS ONE 8, e59045. Wang, G., Zhang, H., Zhao, Y., Li, J., Cai, J., Wang, P., Meng, S., Feng, J., Miao, C., Ding, M., et al. (2005). Noggin and bFGF cooperate to maintain the pluripotency of human embryonic stem cells in the absence of feeder layers. Biochem. Biophys. Res. Commun. 330, 934–942. Wang, M., Guo, L., Wu, Q., Zeng, T., Lin, Q., Qiao, Y., Wang, Q., Liu, M., Zhang, X., Ren, L., et al. (2014). ATR/Chk1/Smurf1 pathway determines cell fate after DNA damage by controlling RhoB abundance. Nat. Commun. 5, 4901. Wei, W., Qing, T., Ye, X., Liu, H., Zhang, D., Yang, W., and Deng, H. (2008). Primordial germ cell specification from embryonic stem cells. PLoS ONE 3, e4013.
Nostro, M.C., Cheng, X., Keller, G.M., and Gadue, P. (2008). Wnt, activin, and BMP signaling regulate distinct stages in the developmental pathway from embryonic stem cells to blood. Cell Stem Cell 2, 60–71.
Xu, R.H., Chen, X., Li, D.S., Li, R., Addicks, G.C., Glennon, C., Zwaka, T.P., and Thomson, J.A. (2002). BMP4 initiates human embryonic stem cell differentiation to trophoblast. Nat. Biotechnol. 20, 1261–1264.
Okada, M., Sangadala, S., Liu, Y., Yoshida, M., Reddy, B.V., Titus, L., and Boden, S.D. (2009). Development and optimization of a cell-based assay for the selection of synthetic compounds that potentiate bone morphogenetic protein-2 activity. Cell Biochem. Funct. 27, 526–534.
Yang, L., Soonpaa, M.H., Adler, E.D., Roepke, T.K., Kattman, S.J., Kennedy, M., Henckaerts, E., Bonham, K., Abbott, G.W., Linden, R.M., et al. (2008). Human cardiovascular progenitor cells develop from a KDR+ embryonicstem-cell-derived population. Nature 453, 524–528.
Palmer, B.D., Thompson, A.M., Booth, R.J., Dobrusin, E.M., Kraker, A.J., Lee, H.H., Lunney, E.A., Mitchell, L.H., Ortwine, D.F., Smaill, J.B., et al. (2006). 4-Phenylpyrrolo[3,4-c]carbazole-1,3(2H,6H)-dione inhibitors of the checkpoint kinase Wee1. Structure-activity relationships for chromophore modification and phenyl ring substitution. J. Med. Chem. 49, 4896–4911. Robert, B. (2007). Bone morphogenetic protein signaling in limb outgrowth and patterning. Dev. Growth Differ. 49, 455–468. Sanjana, N.E., Shalem, O., and Zhang, F. (2014). Improved vectors and genome-wide libraries for CRISPR screening. Nat. Methods 11, 783–784. Shalem, O., Sanjana, N.E., Hartenian, E., Shi, X., Scott, D.A., Mikkelsen, T.S., Heckl, D., Ebert, B.L., Root, D.E., Doench, J.G., and Zhang, F. (2014). Genome-scale CRISPR-Cas9 knockout screening in human cells. Science 343, 84–87.
Ying, Q.L., Nichols, J., Chambers, I., and Smith, A. (2003). BMP induction of Id proteins suppresses differentiation and sustains embryonic stem cell selfrenewal in collaboration with STAT3. Cell 115, 281–292. Yu, P.B., Hong, C.C., Sachidanandan, C., Babitt, J.L., Deng, D.Y., Hoyng, S.A., Lin, H.Y., Bloch, K.D., and Peterson, R.T. (2008). Dorsomorphin inhibits BMP signals required for embryogenesis and iron metabolism. Nat. Chem. Biol. 4, 33–41. Zaret, K.S. (2001). Hepatocyte differentiation: from the endoderm and beyond. Curr. Opin. Genet. Dev. 11, 568–574. Zhu, H., Kavsak, P., Abdollah, S., Wrana, J.L., and Thomsen, G.H. (1999). A SMAD ubiquitin ligase targets the BMP pathway and affects embryonic pattern formation. Nature 400, 687–693.
Cell Reports 15, 2063–2075, May 31, 2016 2075
Discovery of a Small-Molecule BMP Sensitizer for Human Embryonic Stem Cell Differentiation Graphical Abstract
Authors Lingling Feng, Brandoch Cook, Su-Yi Tsai, Ting Zhou, Brooke LaFlamme, Todd Evans, Shuibing Chen
Correspondence [email protected] (T.E.), [email protected] (S.C.)
In Brief Feng et al. show that PD407824, a CHK1 inhibitor, synergizes with BMP4 to induce human embryonic stem cell differentiation toward mesoderm or cytotrophoblast lineages. Mechanistic studies indicate that CHK1 inhibition depletes p21, leading to a decrease in SMAD2/3 and an increase in levels of nuclear SMAD1.
Highlights d
PD407824 increases the sensitivity of cells to sub-threshold levels of BMP4
d
PD407824 synergizes with BMP4 to reprogram C2C12 myoblasts toward an osteoblast fate
d
PD407824 sensitizes human embryonic stem cells to BMP4 to induce differentiation
d
CHK1 inhibition induces P21 downregulation and SMAD2/3 degradation
Feng et al., 2016, Cell Reports 15, 2063–2075 May 31, 2016 ª 2016 The Author(s). http://dx.doi.org/10.1016/j.celrep.2016.04.066
Cell Reports
Article Discovery of a Small-Molecule BMP Sensitizer for Human Embryonic Stem Cell Differentiation Lingling Feng,1,2,3 Brandoch Cook,2,3 Su-Yi Tsai,2,3 Ting Zhou,2,3 Brooke LaFlamme,2 Todd Evans,2,* and Shuibing Chen2,* 1Key Laboratory of Pesticide and Chemical Biology (CCNU), Ministry of Education, College of Chemistry, Central China Normal University, Wuhan, Hubei 430079, China 2Department of Surgery, Weill Cornell Medical College, New York, NY 10065, USA 3Co-first author *Correspondence: [email protected] (T.E.), [email protected] (S.C.) http://dx.doi.org/10.1016/j.celrep.2016.04.066
SUMMARY
Sorely missing from the ‘‘toolkit’’ for directed differentiation of stem/progenitor cells are agonists of the BMP-signaling pathway. Using a high-throughput chemical screen, we discovered that PD407824, a checkpoint kinase 1 (CHK1) inhibitor, increases the sensitivity of cells to sub-threshold amounts of BMP4. We show utility of the compound in the directed differentiation of human embryonic stem cells toward mesoderm or cytotrophoblast stem cells. Blocking CHK1 activity using pharmacological compounds or CHK1 knockout using single guide RNA (sgRNA) confirmed that CHK1 inhibition increases the sensitivity to BMP4 treatment. Additional mechanistic studies indicate that CHK1 inhibition depletes p21 levels, thereby activating CDK8/9, which then phosphorylates the SMAD2/3 linker region, leading to decreased levels of SMAD2/3 protein and enhanced levels of nuclear SMAD1. This study provides insight into mechanisms controlling the BMP/ transforming growth factor beta (TGF-b) signaling pathways and a useful pharmacological reagent for directed differentiation of stem cells. INTRODUCTION Bone morphogenetic proteins (BMPs) are important regulators of embryonic development and stem/progenitor cell fate decisions. BMPs are required for the establishment of the body plan (Heisenberg and Solnica-Krezel, 2008), limb bud patterning (Robert, 2007), and early hemato-vascular (Larsson and Karlsson, 2005) and neural development (Liu and Niswander, 2005). BMPs regulate the development of multiple organs, including heart (Kruithof et al., 2012), kidney (Cain et al., 2008), liver (Zaret, 2001), and the CNS (Fukuda and Taga, 2006). In adult tissues, BMPs provide signals for differentiation in niches for the hair follicle (Blanpain and Fuchs, 2009), intestinal stem cells (Takashima and Hartenstein, 2012), and germ cells (Knight and Glister, 2006). Due to their important role in embryonic development, BMPs have been used in both maintenance and directed differentiation
of embryonic stem cells (ESCs) to various cell fates. For mouse ESCs, BMP4 is required, together with leukocyte inhibitory factor, to maintain the pluripotent self-renewal state (Li et al., 2012; Ying et al., 2003). In contrast, in human ESCs, BMP4 promotes differentiation, so that inhibition of BMP signaling is required to maintain human ESC self-renewal (James et al., 2005; Wang et al., 2005). Once committed to differentiate, BMPs promote the commitment of ESCs to the mesendoderm germ layer, and these BMP-induced mesendoderm cells can further differentiate into multiple cell lineages, including cardiac, hematopoietic, and hepatic cells. For example, BMP4 has been used to direct differentiation from mesendoderm to Flk1+ hematopoietic progenitor cells and then to blood cells (Lengerke et al., 2008; Nostro et al., 2008). BMP2 and BMP4 direct definitive endoderm cells to a hepatic lineage (Gouon-Evans et al., 2006). BMP7 has been used for differentiation toward brown adipocytes (Nishio et al., 2012). In addition, BMP4 initiates trophoblast differentiation from human ESCs (Xu et al., 2002). Finally, BMP4, 7, and 8b induce germ cell differentiation from both mouse and human ESCs (Kee et al., 2006; Wei et al., 2008). Synthetic small molecules have been widely used to control developmental signaling pathways, as functional agonists or antagonists. Compared to recombinant proteins, synthetic small molecules can be more stable, easier to quantify for reproducible activity and dose response, and far less expensive to produce, which is particularly relevant for scaling cell production. To date, most of the small molecules discovered to regulate BMP signaling are BMP antagonists. A phenotypic screen using zebrafish embryos identified dorsomorphin, which inhibits BMP signaling by targeting BMP type 1 receptors (ALK2, 3, and 6) (Yu et al., 2008). A structure-activity relationship study found a dorsomorphin analog, LDN193189, which demonstrates moderate pharmacokinetic characteristics in mice (Cuny et al., 2008). The in vivo structure-activity relationship study of dorsomorphin analogs identified a specific BMP inhibitor, DMH1 (Hao et al., 2010). Recently, several small molecules have been identified to either activate or synergize with the BMP pathway. For example, SVAK-3 (Okada et al., 2009), SVAK-12 (Kato et al., 2011), KM11073 (Baek et al., 2015), and A1 and A17 (Cao et al., 2014) enhance BMP2-induced early osteoblast marker expression. Small molecules of the flavonoid family have been shown to upregulate Id1, Id2 expression in a human cervical carcinoma cell line (Vrijens et al., 2013). In addition, FK506 activates
Cell Reports 15, 2063–2075, May 31, 2016 ª 2016 The Author(s). 2063 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
A
B
E
C
D
F
G
Figure 1. PD407824 Sensitizes C2C12 Myoblasts to BMP4-Induced Activation of BMP Signaling (A) Scheme of the high-throughput screening. (B) Chemical structure of PD407824. (C) Dose curve of PD407824 on C2C12 myoblasts transfected with pId2-Luc-GFP reporter. (D) PD407824 sensitizes C2C12 myoblasts to BMP4-induced upregulation of endogenous Id2 expression measured by qRT-PCR. (E) PD407824 sensitizes C2C12 myoblasts to BMP4-induced SMAD1/5/9 phosphorylation indicated by western blotting. (F) Quantification of relative SMAD1/5/9 activity. The phosphorylation levels were normalized to total SMAD1 protein. (G) PD407824 sensitizes C2C12 myoblasts to BMP4-induced upregulation of Id1 expression measured by qRT-PCR. See also Figure S1.
BMPR2 and rescues endothelial dysfunction (Spiekerkoetter et al., 2013). However, most of the identified compounds show relatively low activity and fail to induce the generation of mature osteoblasts, which limits their application to modulate BMP signaling. Thus, there is still a strong need to identify efficient BMP activators or sensitizers that can be used in stem cell differentiation. From a high-throughput screen of more than 4000 small molecules we identified PD407824 as a chemical BMP sensitizer. PD407824 sensitizes cells to BMP4-induced upregulation of BMP pathway target genes, including Id1 and Id2. More importantly, we found that PD407824 cooperates with limiting BMP4 protein when added to ESCs during differentiation toward mesoderm or cytotrophoblast stem cells. This provides a useful chemical agent to control human ESC (hESC) differentiation and can facilitate scaled production of defined cell lineages from hESCs. RESULTS Identification of PD407824 as a BMP4 Sensitizer To identify chemical agonists/sensitizers of BMP signaling, we established an assay to monitor BMP-signaling activity through
2064 Cell Reports 15, 2063–2075, May 31, 2016
the expression of a downstream target, Id2 (Figure 1A). We designed a reporter construct containing the Id2 promoter, including two previously validated BMP-responsive regulatory regions from 3,000 to 2,729 and 350 to +80 (Nakahiro et al., 2010) controlling expression of a luciferase-GFP reporter (pId2-LucGFP). The reporter activity was first validated as responsive to BMP4. C2C12 myoblasts transfected with the pId2-LucGFP reporter construct showed increased luciferase activity when treated with 10 ng/ml BMP4 (Figure S1A). To perform high-throughput screening, C2C12 myoblasts were first seeded onto 10-cm2 plates in standard medium (DMEM supplemented with 10% fetal bovine serum [FBS]). On the second day, these cells were transfected with the reporter plasmid pId2LucGFP. 24 hr later, the transfected cells were transferred to 384-well plates at 5,000 cells/100 ml medium per well. After overnight incubation, the culture medium was changed to serum-free (DMEM lacking FBS) to decrease the background luciferase activity induced by serum. After 24 hr, each well was treated with a single compound from a library containing 4,000 chemicals, including kinase inhibitors, signaling pathway regulators, natural products, and US Food and Drug Administration (FDA)approved drugs, which were arrayed similarly in 384-well
format as 10 mM stocks in DMSO. Detailed library information is described in Table S1. The library was screened at 1:1,000 (0.1% DMSO) and 1:10,000 (0.01% DMSO) dilutions, providing 10 mM and 1 mM final concentrations, respectively. After 24-hr chemical treatment, the cells were lysed and analyzed for luciferase activity using the Bright-Glo luciferase substrate (Promega). The effect of compound on luciferase activity was quantified by calculating the fold-induction compared to samples treated with DMSO alone, which served as negative controls. BMP4 at 10 ng/ml was used as a positive control. The compounds showing more than 5-fold induction were chosen as the primary hits for further analysis (Figure S1B; Table S2). Although the cells were cultured in serum-free medium for 24 hr in the primary assay, there might remain some baseline level of BMP signaling (as demonstrated below). In this case, we might expect to identify both BMP agonists and BMP sensitizers in the primary screen. The primary hit compounds were titered and three of them showed reporter induction at a relatively low effective concentration (1–15 mM). One hit compound, isoliquiritigenin, has been previously reported to upregulate alkaline phosphatase (ALK) expression (Vrijens et al., 2013). Since ALK is an early marker of osteoblast differentiation, we further tested the three hit compounds for their capacity to induce bone formation. We could detect ALK expression. However, the hit compounds were unable on their own to induce the generation of mature osteoblasts. C2C12 myoblasts were treated with hit compounds for 21 days and stained with alizarin red S to examine the mineralization of mature osteoblasts. We did not find cells staining positive by alizarin red S in compound-treated conditions (Figure S1C). Since hit compounds themselves did not significantly induce a functional BMP cellular output, we tested whether hit compounds make the cells more sensitive to BMP4 treatment, considering that the primary screen might not be entirely depleted of BMP signals. Using the same platform as for primary screening (Figure 1A), after overnight starvation, C2C12 myoblasts expressing pId2-LucGFP were treated with hit compounds together with 1 ng/ml BMP4. One hit compound, PD407824 (Figure 1B), sensitized C2C12 myoblasts to 1 ng/ml BMP4 or 0.1 ng/ml BMP4 treatment to increase luciferase activity (Figure 1C). Two criteria were used to define the sensitizing effect. First, activity of the reporter in cells treated with PD407824 and BMP4 simultaneously is higher than the sum of the activities in cells treated with PD407824 or BMP4 separately. More importantly, the half maximal effective concentration (EC50) of PD407824 decreases as the BMP4 concentration increases (EC50 = 12.3 mM in the absence of BMP4; EC50 = 0.75 mM in the presence of 0.1 ng/ml BMP4; EC50 = 0.12 mM in the presence of 1 ng/ml BMP4). Neither of the two other primary screen hits, isoliquiritigenin or lindane, showed the same synergistic effect with BMP4. To determine whether PD407824 + BMP4 activates endogenous Id2 expression, C2C12 myoblasts were treated with 10 mM or 1 mM PD407824 plus different doses of BMP4 for 24 hr and analyzed for Id2 expression with quantitative RT-PCR. Consistent with the pId2-LucGFP assay, PD407824 sensitizes C2C12 myoblasts to BMP4 treatment to increase transcription of the endogenous Id2 gene. DMSO + 10 ng/ml BMP4
increased Id2 transcript levels 3.7 ± 1.0-fold; while 1 or 10 mM PD407824 + 10 ng/ml BMP4 treatment increased the Id2 transcript levels 9.6 ± 2.2-fold, and 11.5 ± 0.7-fold, respectively (Figure 1D). Together, these data suggest that PD407824 can sensitize C2C12 myoblasts to sub-threshold concentrations of BMP4 treatment to upregulate expression of Id2, a known target of the BMP-signaling pathway. To examine whether PD407824 does in fact affect BMPsignaling components, we used western blotting experiments to measure the phosphorylation of SMAD1/5/9 at Ser 463/ 465 (Figure 1E). As shown by the DMSO control, we confirmed a low basal level of activated SMAD1/5/9. However, PD407824+BMP4 quantitatively increased SMAD1/5/9 phosphorylation. Interestingly, PD407824 used alone at the same concentration showed a trend to slightly decrease phosphorylated SMAD1/5/9, which was however not statistically significant (Figure 1F). In addition, qRT-PCR was used to measure the expression of Id1, another known key target gene of BMP signaling. PD407824 sensitizes C2C12 myoblasts to subthreshold concentrations BMP4 treatment to upregulate Id1 transcript levels in a dose-dependent manner (Figure 1G). Together, these data suggest that PD407824 sensitizes C2C12 myoblasts to sub-threshold concentrations of BMP4 to activate BMP signaling. PD407824 Synergizes with BMP4 to Reprogram C2C12 Myoblasts toward a Mature Osteoblast Fate BMPs are key drivers of bone formation, and previous studies indicated that BMP treatment can reprogram C2C12 myoblasts to osteoblasts (Katagiri et al., 1994). To determine the impact of PD407824 in mediating BMP-dependent bone formation, C2C12 myoblasts were treated with DMSO as a control, PD407824, 3 ng/ml BMP4, PD407824 + 3 ng/ml BMP4, or 10 ng/ml BMP4. After 6 days of culture, cells were fixed and stained with ALK substrate. In the DMSO control, C2C12 myoblasts spontaneously differentiated into myotubes, but there were no ALK-expressing cells (which would display a purple color from the stain). There were a very limited number of ALK-expressing cells detected in the sample treated with 3 ng/ml BMP4, suggesting this is near or below a bone forming threshold. Indeed, the sample containing 10 ng/ml BMP4 (with DMSO but no compound added) showed a strong induction of ALK expression, although there remained numerous myotubes, suggesting that DMSO + 10 ng/ml BMP4 is not sufficient to block myogenesis. PD407824 treatment alone was not sufficient to block myogenesis or induce ALK expression. However, cells treated with PD407824 + 3 ng/ml BMP4 expressed high levels of ALK and entirely lost the capacity to form myotubes (Figure 2A). Next, C2C12 myoblasts were treated with different doses of PD407824 in the presence or absence of 3 ng/ml BMP4. Six days later, the cell lysates were analyzed for ALK activity using an absorbance-based ALK substrate. The quantitative results confirmed that PD407824 itself does not induce ALK activity; however, PD407824 synergized with 3 ng/ml BMP4 to reprogram C2C12 myoblasts to ALK-expressing cells in a dosedependent manner. The ALK activity generated by cells treated with 1 mM PD407824 + 3 ng/ml BMP4 was equal to that of cells treated with 10 ng/ml BMP4 alone (Figure 2B).
Cell Reports 15, 2063–2075, May 31, 2016 2065
A
B
C
D
Figure 2. PD407824 Synergizes with BMP4 to Reprogram C2C12 Myoblasts into Mature Osteoblasts (A) C2C12 myoblasts were reprogrammed into cells expressing osteogenic marker alkaline phosphatase (ALK). (B) Quantification of ALK activity by fluorescence-based substrate. (C) qRT-PCR to detect the expression of mature osteoblast markers, including collagen I (collagen), osteopontin (OPN), and osteocalcin (OC). (D) Alizarin red S staining of the mature osteoblasts. PD: PD407824. Scale bar, 50 mm.
In addition to ALK, we also examined by qRT-PCR the expression of other osteoblast markers, including collagen I, osteopontin and osteocalcium. Consistent with the ALK substrate assay, PD407824 or 3 ng/ml BMP4 itself was not sufficient to induce bone formation. In contrast, PD407824 + 3 ng/ml BMP4 treatment was able to effectively induce the expression of osteoblast markers. The expression levels of these markers in cells treated with 1 mM PD407824 + 3 ng/ml BMP4 was comparable to those in the cells treated with 10 ng/ml BMP4 (Figure 2C). We also examined the presence of mineralization to evaluate the formation of mature osteoblasts. C2C12 myoblasts were treated with DMSO as control, PD407824 alone, 3 ng/ml BMP4, PD407824 + 3 ng/ml BMP4, or 10 ng/ml BMP4 for 22 days and then stained with alizarin red S solution. In the cells treated with DMSO, 3 ng/ml BMP4, or 1 mM PD407824, no alizarin staining was detected. The cells treated with PD407824 + 3 ng/ml BMP4 showed strong mineralization, comparable to the cells treated with 10 ng/ml BMP4 (Figure 2D). Altogether, these data suggest that PD407824 synergizes with BMP4 to reprogram C2C12 myoblasts into mature osteoblasts. PD407824 as a BMP Sensitizer to Induce hESC Differentiation to Mesoderm BMP4 has been widely used to induce hESC differentiation. The effect of BMP4 in hESC differentiation is context dependent. When cells are allowed to differentiate as embryoid bodies (EBs), the addition of BMP4 directs mouse and human ESC differ-
2066 Cell Reports 15, 2063–2075, May 31, 2016
entiation to mesoderm, which subsequently generates beating cardiomyocytes (Cao et al., 2013; Kattman et al., 2011; Yang et al., 2008). In order to determine whether PD407824 increases hESC sensitivity to BMP treatment in a mesodermal differentiation protocol, hESC H9 cells were treated with DMSO or three different concentrations (0.1 mM, 0.3 mM, and 1 mM) of PD407824 combined with 3 ng/ml BMP4 for 4 days and analyzed with qRTPCR. Consistent with the effect on C2C12 cells, PD407824 sensitizes H9 cells to BMP4 treatment to increase transcription of the endogenous Id2 gene. 1 mM PD407824 + 3 ng/ml BMP4 treatment increased the Id2 transcript levels 7.1 ± 0.3-fold higher than DMSO + 3 ng/ml BMP4 (Figure 3A). Strikingly, PD407824 + 3 ng/ml BMP4 treatment increases transcript levels of Brachyury T, a mesoderm marker, similar as when treated with 10 ng/ml BMP4, suggesting PD407824 sensitizes hESCs to BMP4 treatment during mesoderm differentiation. To further validate that PD407824 + BMP4 induces the generation of functional mesoderm, the chemically treated cells were examined for their capacities to form beating cardiomyocytes, which are mesoderm derivatives. After 4 days treatment with PD407824 + BMP4, EBs were subsequently cultured in cardiac differentiation medium initially containing 20% FBS, which after 10 days was reduced to 2.5%. We used DMSO + 10 ng/ml BMP4 as a positive control, and after 10 days counted the number of beating colonies. In wells treated with 1 mM PD407824 + 3 ng/ml BMP4 the number of beating colonies was comparable to wells treated with 10 ng/ml BMP4 (Figure 3B); notably, treatment with 3 ng/ml BMP4 alone was
A
B
C
D
E
Figure 3. PD407824 Sensitizes Cells to BMP4 Treatment during hESC Mesodermal Differentiation (A) qRT-PCR analysis of ID2 and Brachyury T expression. EBs were treated with either DMSO or three different concentrations (0.1, 0.3, or 1 mM) of PD407824 combined with two concentrations (3 or 10 ng/ml) of BMP4 for 4 days and analyzed with qRT-PCR. DMSO + 3 ng/ml BMP4 was used as a control to normalize data. DMSO + 10 ng/ml BMP4 was used as a positive control. (B) Quantification of beating colonies on day 10. (C) qRT-PCR analysis of the expression levels of human cardiomyocyte markers TNNT2, MHY6, and MLC2v. Levels were normalized to DMSO + 3 ng/ml BMP4. RNA was isolated on day 12 for qRT-PCR analysis. (D) Immunofluorescence staining for TNNT2 expression. Scale bar, 100 mm. Cells were fixed on day 12 and stained for the cardiomyocyte marker TNNT2. (E) Intracellular flow cytometry and quantification of the percentage of TNNT2+ cells.
not sufficient to generate beating colonies. Importantly, the cells treated with 1 mM PD407824 + 3 ng/ml BMP4 expressed high levels of cardiomyocyte differentiation makers, including troponin T type 2 (TNNT2), myosin heavy chain 6 (MHY6), and myosin light chain 2v (MLC2v), at day 12 of differentiation, shown by qRT-PCR (Figure 3C). The expression levels of these genes in the cells treated with 1 mM PD407824 + 3 ng/ml BMP4 treatment were comparable to cells treated with 10 ng/ml BMP4. Moreover, we used immunocytochemistry to examine TNNT2 protein expression. Multiple concentrations (0.1, 0.3, and 1 mM) of PD407824 + 3 ng/ml BMP4 efficiently enhanced TNNT2 expression (Figure 3D). Intracellular flow cytometry was used to quantify the differentiation efficiency. The percentage of TNNT2+ cells in the samples treated with 1 mM PD407824 + 3 ng/ml BMP4 (42.6% ± 4.3%) was significantly enhanced compared to 3 ng/ml BMP4 alone (5.5% ± 0.7%), achieving the level found in cells treated with 10 ng/ml BMP4 (47.5% ± 6.7%) (Figure 3E). Thus, PD407824 sensitizes cells to sub-threshold concentrations of BMP4 during hESC mesodermal differentiation.
PD407824 Sensitizes hESCs to BMP4 during Differentiation toward Cytotrophoblast Stem Cell Lineage The effect of BMP4 on hESC differentiation is stage and context dependent. We next sought to determine if the action of PD407824 is specific to BMP-dependent mesodermal differentiation or can be more generally applied to BMP-dependent differentiation programs. When hESCs are grown in the presence of mouse embryonic fibroblast (MEF) conditional medium (CM), BMP4 functions as a key driver for differentiation toward KERATIN7 (KRT7)+/TP63+ cytotrophoblast stem cells (Amita et al., 2013; Das et al., 2007; Li et al., 2013; Xu et al., 2002). To determine if PD407824 functions as a BMP sensitizer in trophoblast differentiation, hESC H1 cells were treated with DMSO or 2.5 mM PD407824 combined with different concentrations of recombinant BMP4 protein (0, 4, 6, 9, 12, or 30 ng/ml) in MEF-CM for 7 days (Figure 4A). The cells were fixed and stained with antiKRT7 antibody (Figure 4B). More KRT7+ cells were detected in PD407824 + BMP4 treatment condition than the corresponding
Cell Reports 15, 2063–2075, May 31, 2016 2067
A
B
C
D
E
Figure 4. PD407824 Sensitizes Cells to BMP4 Treatment during Differentiation of hESCs toward Cytotrophoblast Stem Cell Fate (A) Scheme of human cytotrophoblast stem cell differentiation. hESC H1 cells were treated with DMSO or 2.5 mM PD407824 combined with different concentrations of recombinant BMP4 protein (0, 4, 6, 9, 12, or 30 ng/ml) in MEF-CM for 7 days. (B) Immunohistochemistry analysis of KRT7 expression. Scale bar, 100 mm. (C and D) Intracellular flow cytometry analysis (C) and quantification (D) of the percentage of KRT7+ cells. (E) qRT-PCR analysis of the expression of human cytotrophoblast stem cell markers KRT7, TP63, and GCM1. Levels were normalized to undifferentiated hESCs. See also Figure S2.
2068 Cell Reports 15, 2063–2075, May 31, 2016
A
B
C
BMP4 treatment condition. Intracellular flow cytometry was used to quantify the differentiation efficiency (Figure 4C). PD407824 treatment significantly increased the percentage of KRT7+ cells in different doses of BMP4 treatment. (Figure 4D). Moreover, qRT-PCR analysis showed that cells treated with 2.5 mM PD407824 + 9 ng/ml BMP4 express other cytotrophoblast stem cells markers, such as TP63 (Li et al., 2013) and GCM1 (Figure 4E). The expression levels of these genes when treated with 2.5 mM PD407824 + 9 ng/ml BMP4 are significantly higher than the condition treated with 9 ng/ml BMP4 and were comparable to the condition treated with 30 ng/ml BMP4. We further examined the proliferation and DNA damage in PD407824 treated hESCs. First, we found that there is no significant difference in the number of Ki67+ cells detected comparing DMSO or PD407824 treated hESCs (Figure S2A). To examine whether CHK1 inhibition affects DNA repair, PD407824-treated hESCs were stained using g-H2A.X antibody. Cells treated with 10 mM etoposide were used as a staining control. No g-H2A.X+ cells were detected in 2.5 mM PD407824 or 2.5 mM PD407824+ 9 ng/ml BMP4-treated conditions (Figure S2B), suggesting that PD407824 treatment does not affect DNA repair during hESC differentiation. Together, these data show that PD407824 increases the sensitivity of hESCs to BMP4 treatment by more than 3-fold during differentiation toward cytotrophoblast stem cell fate without significantly affecting cell proliferation, cell cycle, or DNA repair. PD407824 Functions through Inhibiting Checkpoint Kinase 1 A previous study identified PD407824 as a checkpoint kinase 1 (CHK1) inhibitor (Palmer et al., 2006). To determine whether PD407824 functions through inhibiting CHK1, we tested another CHK1 chemical inhibitor, CHIR124, for its ability to sensitize C2C12 myoblasts to BMP4 treatment. C2C12 myoblasts were treated with different concentrations of CHIR124 in the presence or absence of 1 ng/ml BMP4. After 24 hr treatment, qRT-PCR
Figure 5. PD407824 Inhibiting CHK1
Functions
through
(A and B) CHIR124, a CHK1 inhibitor, mimics the effect of PD407824 to sensitize BMP4-induced upregulation of Id2 expression (A) and increased SMAD1/5/9 phosphorylation (B) in C2C12 myoblasts. The relative Id2 transcript level was normalized to DMSO treatment conditions. The relative SMAD1/5/9 phosphorylation was normalized to total SMAD1. (C) Relative Id2 transcript expression in CHK1+/ and CHK1/ cells. Untransfected cells (wild-type [wt]) and cells infected with sgRNA targeting GFP were used as controls. See also Figure S3.
was used to monitor the transcript levels of Id2. Similar to PD407824, CHIR124 synergized with BMP4 to upregulate Id2 in a dose-dependent manner (Figure 5A). Consistently, 10 mM CHIR124 + BMP4 increases the relative phosphorylation of SMAD1/5/9 (Figures 5B and S3A), showing the similar trend as PD407824 + BMP4 treatment (Figure 1E). Moreover, we complemented the chemical approach using a genetic strategy. C2C12 cells were infected with lentivirus expressing Cas9 and single guide RNA (sgRNA) targeting CHK1 or, as a control, sgRNA targeting GFP. Mutant cells were subcloned. Western blotting experiments confirmed that the monoallelic and biallelic knockout of CHK1 protein in CHK1+/ and CHK1/ clones, respectively (Figure S3B). After overnight starvation, the CHK1+/ and CHK1/ C2C12 cells were treated with 1 ng/ml or 3 ng/ml BMP4. After 24 hr treatment, Id2 transcript levels were analyzed with qRT-PCR. Compared to uninfected or GFP-targeted cells, CHK1+/ and CHK1/ cells show increased expression of Id2 in the absence of added BMP4 and, similar to PD407824-treated cells, in the presence of 1 ng/ml BMP4 (Figure 5C). In the presence of 3 ng/ml BMP4, CHK1/ cells show higher Id2 expression than CHK1+/ cells or cells treated with PD407824, suggesting that CHK1 might play a dose dependent effect. In addition to CHK1, PD407824 also inhibits Wee1 and PKC (Palmer et al., 2006). To determine whether Wee1 or PKC inhibition is involved in the sensitizing effect of PD407824, C2C12 cells were treated with different concentrations of MK1775, a Wee1 inhibitor (Hirai et al., 2009), or bisindolylmaleimide II, a PKC inhibitor, neither of which inhibit CHK1 activity, in the presence or absence of 1 ng/ml BMP4. After 24-hr treatment, qRT-PCR was used to monitor the transcript levels of Id2. In contrast to CHK1 inhibitors, neither MK1775 nor bisindolylmaleimide II sensitizes C2C12 cells to BMP4 treatment (Figures S2C and S2D). Together, these data support a mechanism by which PD407824 impacts BMP activity through inhibition of CHK1, but not Wee1 or PKC. Mechanism of Action: Increasing the Sensitivity to BMP through Depletion of SMAD2/3 In the canonical BMP pathway, the ligands exert their effects through binding to type II serine/threonine kinase receptors (BMPR2 for BMPs), forming an oligomeric complex. The type II
Cell Reports 15, 2063–2075, May 31, 2016 2069
receptors are consequently activated and phosphorylated and then activate the type I receptors (ALK3 and ALK6 for BMPs; Bandyopadhyay et al., 2013). While there are SMADindependent pathways (Bandyopadhyay et al., 2013), a major mode of signal transduction occurs when activated type I receptors phosphorylate intracellular receptor-regulated R-SMADs (SMAD1/5/9 for BMPs), which associate with the co-SMAD (SMAD4) and translocate into the nucleus. The SMAD complexes interact with additional transcription factors to bind promoters of target genes, such as Id2, to control their expression. Since PD407824 + BMP4 treatment increases the Id2 expression levels in both C2C12 cells (Figure 1D) and hESCs (Figure 3A), we first used co-immunoprecipitation experiments to determine whether PD407824 + BMP4 treatment leads to an increased level of SMAD1/5/9 binding with SMAD4. Cells were treated with 3 ng/ml BMP4 in the presence or absence of 10 mM PD407824 for 24 hr. Proteins from the cell lysate were precipitated with anti-SMAD4 antibody and subsequently probed by blotting using anti-SMAD1 antibody. SMAD1 protein bound to SMAD4 was most abundant in the presence of PD407824 + BMP4 compared to other conditions (Figure 6A). Since both SMAD2/3 (R-SMAD for transforming growth factor beta [TGF-b]) and SMAD1/5/9 are capable of binding with SMAD4, we used western blotting to examine the expression of SMAD2/3 at the protein level. Strikingly, western blotting experiments showed that PD407824+BMP4 treatment significantly decreased the total protein levels of SMAD2/3 protein (Figure 6B). In addition, CHIR124, another CHK1 inhibitor, also decreased the total levels of SMAD2/3 protein (Figure S4A). This might due to either reduced levels of SMAD2/3 gene expression or increased degradation of SMAD2/3 protein. To distinguish these possibilities, C2C12 were treated with PD407824 or PD407824 + BMP4 for 24 hr and analyzed with qRT-PCR. Neither PD407824 nor PD407824+BMP4 treatment significantly changed the transcript levels of SMAD2/3 (Figure S3B), suggesting that PD407824 might affect SMAD2/3 protein stability. Previous studies suggested that activated CDK8/9 phosphorylates SMAD2/3 at a linker region (pT220 for SMAD2 and pT179 for SMAD3), which is ultimately recognized by ubiquitin ligase Nedd4L, leading to proteasome-mediated turnover of SMAD2/3 (Alarco´n et al., 2009; Gao et al., 2009). Western blotting experiments were used to measure the phosphorylation of SMAD2/3 at the linker region. In conditions using PD407824 or PD407824+BMP4, the relative phosphorylation (normalized to total SMAD2/3 protein) of linker region (T179 for SMAD3) was significantly elevated compared to BMP4 treatment alone (Figure 6C). The hyper-phosphorylation of the linker region leads to the decreased levels of SMAD2/3 proteins (Figure 6B). Since CDK9 phosphorylates the SMAD2/3 linker region, both chemical and genetic approaches were used to determine whether CDK9 activation is required for the sensitizing effect of PD407824 to upregulate Id2. The cells were treated with or without 3 ng/ml BMP4 plus PD407824 in the presence or absence of 300 nM flavopiridol, a chemical CDK inhibitor, for 24 hr and analyzed with qRT-PCR experiments to measure endogenous Id2 expression levels. Flavopiridol does not affect the activity of BMP4 alone. However, flavopiridol blocks the sensitizing effect of PD407824 with BMP4 (Figure 6D). In
2070 Cell Reports 15, 2063–2075, May 31, 2016
addition, flavopiridol blocks PD407824-induced elevation of SMAD3 T179 phosphorylation, as well as the decrease of total SMAD2/3 protein (Figures 6E and S4C). Moreover, flavopiridol blocks PD407824-induced enhanced binding between SMAD1 and SMAD4 proteins (Figure 6F). Furthermore, C2C12 cells were infected with lentivirus expressing Cas9 and sgRNA targeting CDK9. C2C12 cells infected with lentivirus expressing sgRNA targeting GFP were used as a negative control. After subcloning, the monoallelic and biallelic knockout of CDK9 protein in CDK9+/ and CDK9/ cells was confirmed by western blotting, respectively (Figure S3D). Consistent with the flavopiridol treatment results, CDK9+/ and CDK9/ cells lose the ability to respond to PD407824 treatment, which further confirms that CDK9 activation is required for the ability of PD407824 to sensitize cells to BMP (Figure 6G). It has been shown that p21 inhibits enzymatic activity of CDK9 (Chen et al., 2011). We used western blotting experiments to measure p21 expression levels and found that either PD407824 alone or PD407824 + BMP4 treatment significantly decreased p21 protein levels (p < 0.01) (Figure 6H). In addition, CHIR124 also decreased p21 protein expression levels (Figure S4E). To determine whether p21 depletion is sufficient to increase the sensitivity to BMP4 treatment, C2C12 cells were infected with lentivirus expressing Cas9 and sgRNA targeting p21 and mutant cells were subcloned. C2C12 cells infected with lentivirus expressing sgRNA targeting GFP were used as a negative control. Western blotting confirmed the monoallelic and biallelic knockout of p21 protein expression in p21+/ and p21/ cells, respectively (Figure S3F). The p21+/ and p21/ cells showed increased expression of Id2 in the presence of BMP4, comparable to PD407824-treated cells in the presence of 3 ng/ml BMP4 (Figure 6I). In addition, flavopiridol treatment does not block PD407824 induced decreased expression of p21 protein (Figure S3G), suggesting that CDK9 functions as the downstream of p21. These data suggest that PD407824 impacts BMP activity through p21 inhibition. Overall, the data suggest that PD407824 is sufficient to block CHK1, leading to a significant downregulation of p21, causing activation of CDK8/9, which in turn causes depletion of SMAD2/3. SMAD2/3 competes with SMAD1/5/9 for limiting SMAD4, suggesting that PD407824 effectively sensitizes with sub-threshold BMP by relieving this limitation, freeing SMAD4 to interact productively with the BMP effector SMAD1 (Figure 7A). To further validate this proposed mechanism of action, we created CHK1//p21/ and CHK1//CDK9/ doubleknockout C2C12 cell lines, based on the parental p21/ (Figure S4E) and CDK9/ cells (Figure S4D), respectively. The knockout of CHK1 protein in CHK1//p21/ and CHK1// CDK9/ cells was validated by western blotting experiments (Figure S5A). The double-knockout cells and single-knockout cells were cultured in the absence or presence of 1 ng/ml or 3 ng/ml BMP4. Consistent with the proposed mechanism that p21 is downstream of CHK1, there is no distinguishable difference in the levels of Id2 transcripts comparing CHK1// p21/, CHK1/ and p21/ cells (Figure 7B). In addition, there is no distinguishable difference in the transcript levels of osteoblast markers, including ALP, collagen, OPN, and OC, comparing CHK1//p21/, CHK1/ and p21/ cells after
A
B
D C
G
E
F
I H
Figure 6. PD407824 Increases BMP Sensitivity by Depleting p21 and Activating CDK9, which Then Phosphorylates the SMAD2/3 Linker Region, Leading to Decreased Levels of SMAD2/3 Protein and Enhanced Levels of Nuclear SMAD1 (A) PD407824 treatment increases the binding of SMAD1 and SMAD4 proteins indicated by co-immunoprecipitation. (B) PD407824 causes depletion of SMAD2/3 protein levels, measured by western blotting; SMAD2/3 protein levels are normalized to beta-ACTIN protein levels. (C) PD407824 induces the phosphorylation of the SMAD2/3 linker region (T179 for SMAD3). The phosphorylation level was normalized to total levels of SMAD2 protein. (D–F) Flavopiridol, a CDK inhibitor, blocks the PD407824 induced increased Id2 transcript expression (D), upregulated T179 phosphorylation of SMAD3 and decreased SMAD2/3 total protein (E), and increased binding of SMAD1 and SMAD4 proteins indicated by co-immunoprecipitation (F). (G) Knockout of CDK9 blocks the synergistic effect of PD407824. (H) PD407824 depletes p21 protein levels, shown by western blotting. The p21 protein levels are normalized to b-actin protein levels. (I) Relative Id2 transcript expression in p21+/ and p21/ cells. Uninfected cells (wild-type [wt]) and cells infected with sgRNA targeting GFP were used as negative controls. See also Figure S4.
7 days of osteogenic differentiation (Figure S5C). Moreover, CDK9 knockout in CHK1/ cells ablates the ability of CHK1/ cells to respond to BMP4 treatment regarding Id2 transcript upregulation (Figure 7C) and capacity to differentiate toward oste-
oblasts (Figure S5D). Furthermore, SB431542, a TGF-b type 1 receptor kinase inhibitor mimics the ability of PD407824 to sensitize C2C12 cells to BMP4 treatment (Figure S5B). Finally, the proposed mechanism was confirmed using hESCs. After
Cell Reports 15, 2063–2075, May 31, 2016 2071
CDK9
C P SMAD4
Linker Region
P
P SMAD2/3
Tail Region
P SMAD1/5/9
2 0
1 ng/ml BMP4
3 ng/ml BMP4
**
14
pSMAD1/5/9
12 10 8
*
6
*
sgGFP CHK1-/CHK1-/-; CDK9-/-
4
p21 pSMAD3 (T179) SMAD2/3
**
16
lB MP 4 3n g/m 9n lB g/m MP 4 lB MP PD 4 +9 ng /m lB MP 4
4
10
sgGFP CHK1-/p21-/CHK1-/-; p21-/-
PD +
6
g/m
* **
3n
n.s.
8
SO
10
PD
12
μM
D
** ** *
14
none
p21
Relative Id2 mRNA expression
ALK1/2/3/6
ALK4/5
TGFβR1
CHK
BMPR1
BMP
TGFβ
n.s.
16
DM
B Relative Id2 mRNA expression
A
SMAD1 beta-ACTIN
2 0
ubiquitination
Id1, Id2 etc
none
1 ng/ml BMP4
3 ng/ml BMP4
Figure 7. Suggested Mechanism of Action (A) Suggested mechanism of action. (B) Relative Id2 transcript expression in CHK1/, p21/, and CHK1/p21/ cells treated with or without BMP4. (C) Relative Id2 transcript expression in CHK1/ and CHK1/CDK9/ cells treated with or without BMP4. (D) Western blotting analysis of hESCs treated with DMSO, 10 mM PD407824, 3 ng/ml BMP4, 10 mM PD407824 + 3 ng/ml BMP4, 9 ng/ml BMP4, or 10 mM PD407824 + 9 ng/ml BMP4. H1 cells were starved in serum-free E6 medium for 24 hr, treated with different conditions for 24 hr, and then analyzed using western blotting. See also Figure S5.
24-hr treatment with DMSO, 10 mM PD407824, 3 ng/ml BMP4, 3 ng/ml BMP4 + 10 mM PD407824, 9 ng/ml BMP4, or 9 ng/ml BMP4 + 10 mM PD407824, hESC lysates were analyzed with western blotting experiments. PD407824, 3 ng/ml BMP4 + 10 mM PD407824, 9 ng/ml BMP4, and 9 ng/ml BMP4 + 10 mM PD407824 treatment decreases p21 protein expression levels, increases the relative phosphorylation of SMAD3 (T179), decreases the total levels of SMAD2/3 proteins, increases the relative phosphorylation of SMAD1/5/9 (Figures 7D and S5E), and leads to enhanced Id2 transcript expression (Figure 3A). DISCUSSION We discovered a previously unreported mechanism by which inhibiting CHK1 can indirectly sensitize the activity of BMP. It does so by downregulating p21 levels, which activates CHK8/9, leading to phosphorylation of the SMAD2/3 linker region and, finally, decreased levels of SMAD2/3 proteins (Figure 6). When BMP ligand is present, SMAD1/5/9 may therefore associate with the increased available SMAD4, and more effectively activate BMP downstream targets, such as Id1 and Id2. This study suggests that SMAD4 is a limiting factor for BMP activation, consistent with our previous observations of competition among SMADs (Cook et al., 2011). Thus, through a screen for BMP agonists, we found a previously unrecognized indirect mechanism to regulate BMP signaling by targeted depletion of SMAD2/3. While PD407824 causes increased levels of phospho-SMAD1/5/9, this may occur through an indirect effect by increasing steadystate levels of active complexes, rather than, for example, enhanced ALK2/3/6 activity. Interestingly, PD407824 itself slightly decreased phospho-SMAD1/5/9 levels (Figure 1E), although it does not reach significance. Previous studies have shown that CHK1 directly induces Smurf1 phosphorylation and
2072 Cell Reports 15, 2063–2075, May 31, 2016
mediates Smurf1 degradation (Wang et al., 2014), which could impact Smad1/5 proteasomal degradation (Zhu et al., 1999). It is possible that CHK1 inhibition reduces both SMAD1/5/9 and SMAD2/3, but much more specifically SMAD2/3. This dynamic competition results in increased binding between SMAD1 and SMAD4 (Figure 6A) at the expense of SMAD2/3 (Figure 6B) and sensitizes the cells to BMP4 treatment (Figures 1D and 1G). While BMP pathway activation leads to phosphorylation of SMAD1/5/9, all the other TGF-b superfamily members, including TGF-b, Activin, and Nodal, cause phosphorylation of SMAD2/3. Many TGF-b inhibitors have been identified, most of which target type I receptors, such as ALK4 or ALK5. Here, we identified a class of compounds, checkpoint kinase inhibitors, which block TGF-b signaling by causing SMAD2/3 depletion. These compounds can therefore be used as useful tools to regulate TGF-b signaling. Checkpoint kinase inhibitors, including MK-8776, LY2606368, and LY2603618, have been used in phase 1 and 2 clinical trials for the treatment of acute myeloid leukemia or breast or ovarian cancers, by themselves or in combination therapies (Lapenna and Giordano, 2009). Here, we found a previously unreported function for checkpoint kinase inhibitors, which sensitize C2C12 cells hESCs with sub-threshold BMP to activate BMP signaling. In addition, recombinant BMPs are widely used for ESC maintenance and for directed differentiation, such as inducing mesoderm, cardiomyocytes, hepatocytes, trophoblasts, germ cells, and brown adipocytes. Recent studies suggested that BMP4 and BMP7 induce the white adipocyte to brown adipocyte transition (Elsen et al., 2013). However, recombinant BMPs are relatively unstable and costly, which limits their application. In our studies, we showed that small-molecule CHK1 inhibitors synergize with BMP4 to induce hESC differentiation toward a beating cardiomyocyte or cytotrophoblast stem
cell fate. Considering the broad application of BMPs in stem cell biology, CHK1 inhibitors can be a useful reagent to control stem cell or progenitor fate. Due to cellular context, the concentrations of CHK1 inhibitors may need be optimized, depending on the particular differentiation protocol and medium employed. While recombinant BMPs have been promising in preclinical studies and even progressed to clinical trials, the clinical outcomes have been less dramatic than might have been hoped. These CHK1 inhibitors, which have themselves been used in phase 1 and 2 clinical trials, might be applicable for therapeutic value, being both stable and cost-efficient, to partially or fully replace recombinant BMPs to treat relevant disorders, such as longbone non-unions, acute fractures, and spinal fusions. EXPERIMENTAL PROCEDURES Cell Culture C2C12 myoblasts were purchased from the American Type Culture Collection (ATCC). C2C12 myoblasts were maintained in complete DMEM supplemented with 10% fetal calf serum, 100 U penicillin, and 100 mg streptomycin at 37 C in 5% CO2. Human ESC H1 and H9 cells were purchased from WiCell Research Institute. Human BMP4 was purchased from R&D Systems. Tissue culture media and sera were purchased from VWR. Bright-Glo luciferase assay reagent was purchased from Promega. Chemicals PD407824 and flavopiridol were purchased from Sigma-Aldrich. Purity of the compounds was higher than 98% high-performance liquid chromatography (HPLC) based on the manufacturer’s information. CHIR124 was purchased from Tocris Bioscience. Unless indicated, all other chemicals were purchased from Sigma-Aldrich. High-Content Small-Molecule Screening Using the Id2 Reporter C2C12 myoblasts were seeded onto 10-cm2 plates in DMEM medium supplemented with 10% FBS. On the second day, the reporter plasmid pId2-LucGFP was transfected into C2C12 myoblasts according to the manufacturer’s protocol for X-treme GENE HP DNA Transfection Reagent (Roche) in serum-free media. 24 hr after transfection, cells were transferred to one 384-well plate per 10-cm dish with 5 3 105 cells/ml in DMEM supplemented with 10% FBS. After another 24 hr, cells were cultured in serum-free conditions for an additional 24 hr and reporter gene expression was induced upon the addition of compounds or BMP4. Cells were then incubated for 24 hr, and luciferase activity was measured using the Bright-Glo luciferase substrate (Promega) following the manufacturer’s recommendations. The effect of compound on luciferase activity was quantified by calculating the fold-induction, determined using the ratio A/B, where A represents the activity in the presence of compound or BMP4 and B represents the activity in wells with only DMSO added. The compounds with more than 5-fold induction were chosen as the primary hit compounds for further analysis. Human ESC Culture and Differentiation Human ESC lines H1 and H9 were cultured on irradiated MEFs and maintained in hESC medium, which contains 80% DMEM/F12, 20% knockout serum replacement, 10 ng/ml human recombinant basic fibroblast growth factor, 2 mM GlutaMax, 0.1 mM nonessential amino acids, and 0.1 mM b-mercaptoethanol. The medium was changed daily. For cardiac differentiation, H9 cells were detached using 0.25%Trysin-EDTA for 1 min at room temperature to form EBs. The EBs were cultured in suspension with hESC medium for 24 hr and then moved to differentiation medium containing 80% knockout DMEM, 20% FBS (Hyclone), 2 mM GlutaMax, 0.1 mM nonessential amino acids, and 0.1 mM b-mercaptoethanol with DMSO or PD407824 combined with 3 or 10 ng/ml BMP4 on ultralow cluster plates (Costar) for 4 days. On day 4, EBs were plated on 0.1% gelatin-coated plates. The following day, differentiated cells were treated with WNT antagonist I (IWR-1, Stemgent) for 3 days.
Beating colonies were counted at day 10. Cells were then dissociated with Accutase for 20 min and replated on gelatin-coated plates at a 1:1 ratio. After day 10, the concentration of FBS was reduced from 20% to 2.5%. RNA was isolated for qRT- PCR analysis or cells were fixed with 4% paraformaldehyde (PFA) for 10 min, followed by staining with anti-TNNT2 antibody for image analysis on day 12. To differentiate hESCs to cytotrophoblast stem cells, hESC H1 cells were treated with DMSO or 2.5 mM PD407824 combined with different concentrations of recombinant BMP4 protein (0, 4, 6, 9, 12, or 30 ng/ml) in MEF-CM for 7 days. RNA was isolated for qRT-PCR analysis or cells were fixed with 4% PFA for 10 min, followed by staining with anti-KRT7 antibody for image analysis. Western Blotting Analysis At the indicated time points, cells were collected in complete lysis buffer (20 mM Tris [pH 7.5], 150 mM NaCl, 50 mM NaF, 1% NP-40 substitute, and Thermo Scientific HALT protease inhibitor cocktail 1:100) and lysates loaded on 10% NuPage Bis-Tris pre-cast gels (Invitrogen). After being resolved by electrophoresis, proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Bio-Rad). Membranes were blocked with 5% BSA in TBS + 0.05% Tween, and then incubated with primary antibody overnight. The antibodies were rabbit anti-phospho-SMAD1/5/9 (Cell Signaling Technology, 9511), rabbit anti-SMAD1 XP (Cell Signal, 6944), rabbit anti-SMAD2/3 (Cell Signal, 3102), rabbit anti-p21 (C-19 antibody; Santa Cruz Biotechnology, sc379), rabbit anti-phospho-SMAD3 T179 (Rockland, 600-401-C48S), and mouse anti-b-actin (Sigma-Aldrich, A1978). The membranes were washed and incubated with secondary antibody for 1 hr at room temperature in 5% milk-TBS-0.05% Tween and developed using SuperSignal West Pico ECL substrate (Thermo Fisher Scientific). Construction of Lentiviral CRISPR Gene Expression Knockdown Plasmids sgRNA sequences were designed using the website http://www.genomeengineering.org/. The target sequences were as follows: sgRNA-eGFP, 50 GGGCGAGGAGCTGTTCACCG-30 ; sgRNA-CHK1, 50 -TAGTGAAATTCTATGG CCAC-30 ; sgRNA-p21, 50 -CTTCGTCACGGAGACGCCGC-30 ; and sgRNACDK9, 50 -ATTGAGATTTGTCGGACCAA-30 . Each target sequence was cloned into the lentiCRISPRv2 vector (Addgene, plasmid # 52961; Sanjana et al., 2014; Shalem et al., 2014) to make knockdown plasmids. C2C12 cells were infected with lentivirus expressing sgRNAs. Two days after infection, cells were treated with 2 mg/ml puromycin for 48 hr. The cells after selection were treated with compounds and used for qRT-PCR analysis. Intracellular Fluorescence-Activated Cell Sorting Analysis The hESC-derived cells were dissociated by Accutase (STEMCELL Technologies), fixed and stained with primary antibody at 4C overnight. Primary antibodies were anti-TNNT2 (1:500) and anti-KRT7 (1:500). After washing three times, cells were incubated with secondary antibodies for 1 hr at room temperature. These were Alexa-488-conjugated donkey anti-mouse or AlexaFluor-647-conjugated donkey anti-guinea pig secondary antibody (1:500) using cytofix/cytoperm (BD Bioscience) and wash buffer (BD Biosciences) according to instructions by the manufacturer. Flow cytometry data were obtained using an Accuri C6 flow cytometry instrument and analyzed with FlowJo. Statistical Analysis Data represent n = 3 independent biological replicates if not specifically indicated. n.s. indicates a non-significant difference. p values were calculated by unpaired two-tailed Student’s t test if not specifically indicated (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001). SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures and five figures and can be found with this article online at http://dx.doi.org/ 10.1016/j.celrep.2016.04.066.
Cell Reports 15, 2063–2075, May 31, 2016 2073
AUTHOR CONTRIBUTIONS S.C. and T.E. designed research; L.F., B.C., B.L., S.-Y.T., and S.C. performed research; L.F., B.C., B.L., S.-Y.T., T.E., and S.C. analyzed data; and L.F., B.L., S.C., and T.E. wrote the manuscript. CONFLICTS OF INTEREST The authors have filed a patent titled ‘‘Chemical BMP sensitizers that replace or partially replace BMP.’’ ACKNOWLEDGMENTS L.F. is funded by the Natural Science Foundation of China (grants 21272089, 21072073, and 20872044). S.C. is funded by The New York Stem Cell Foundation (R-103), NIDDK (DP2 DK098093-01), the American Diabetes Association (1-12-JF-06), and the Cystic Fibrosis Foundation (CHEN15XX0). S.C. is a New York Stem Cell Foundation-Robertson Investigator. T.E. is funded by the NIH (HL056182 and HL111400). B.C. is funded by the NIH (K01 DK096031). T.Z. is funded by a NYSTEM postdoctoral fellowship. This study was also supported by a Shared Facility contract to T.E. and S.C. from the New York State Department of Health (NYSTEM C029156). LentiCRISPRv2 was a gift from Dr. Feng Zhang. We thank Ting-Chu Liu for help with making constructs. We are also very grateful for technical support and advice provided by Harold S. Ralph in the Cell Screening Core Facility at Weill Cornell Medical College. Received: September 3, 2015 Revised: February 6, 2016 Accepted: April 18, 2016 Published: May 19, 2016 REFERENCES
Cook, B.D., Liu, S., and Evans, T. (2011). Smad1 signaling restricts hematopoietic potential after promoting hemangioblast commitment. Blood 117, 6489– 6497. Cuny, G.D., Yu, P.B., Laha, J.K., Xing, X., Liu, J.F., Lai, C.S., Deng, D.Y., Sachidanandan, C., Bloch, K.D., and Peterson, R.T. (2008). Structure-activity relationship study of bone morphogenetic protein (BMP) signaling inhibitors. Bioorg. Med. Chem. Lett. 18, 4388–4392. Das, P., Ezashi, T., Schulz, L.C., Westfall, S.D., Livingston, K.A., and Roberts, R.M. (2007). Effects of fgf2 and oxygen in the bmp4-driven differentiation of trophoblast from human embryonic stem cells. Stem Cell Res. (Amst.) 1, 61–74. Elsen, M., Raschke, S., Tennagels, N., Schwahn, U., Jelenik, T., Roden, M., Romacho, T., and Eckel, J. (2013). BMP4 and BMP7 induce the white-tobrown transition of primary human adipose stem cells. Am. J. Physiol. Cell Physiol. 306, C431–C440. Fukuda, S., and Taga, T. (2006). [Roles of BMP in the development of the central nervous system]. Clin. Calcium 16, 781–785. Gao, S., Alarco´n, C., Sapkota, G., Rahman, S., Chen, P.Y., Goerner, N., Macias, M.J., Erdjument-Bromage, H., Tempst, P., and Massague´, J. (2009). Ubiquitin ligase Nedd4L targets activated Smad2/3 to limit TGF-beta signaling. Mol. Cell 36, 457–468. Gouon-Evans, V., Boussemart, L., Gadue, P., Nierhoff, D., Koehler, C.I., Kubo, A., Shafritz, D.A., and Keller, G. (2006). BMP-4 is required for hepatic specification of mouse embryonic stem cell-derived definitive endoderm. Nat. Biotechnol. 24, 1402–1411. Hao, J., Ho, J.N., Lewis, J.A., Karim, K.A., Daniels, R.N., Gentry, P.R., Hopkins, C.R., Lindsley, C.W., and Hong, C.C. (2010). In vivo structure-activity relationship study of dorsomorphin analogues identifies selective VEGF and BMP inhibitors. ACS Chem. Biol. 5, 245–253. Heisenberg, C.P., and Solnica-Krezel, L. (2008). Back and forth between cell fate specification and movement during vertebrate gastrulation. Curr. Opin. Genet. Dev. 18, 311–316.
Alarco´n, C., Zaromytidou, A.I., Xi, Q., Gao, S., Yu, J., Fujisawa, S., Barlas, A., Miller, A.N., Manova-Todorova, K., Macias, M.J., et al. (2009). Nuclear CDKs drive Smad transcriptional activation and turnover in BMP and TGF-beta pathways. Cell 139, 757–769.
Hirai, H., Iwasawa, Y., Okada, M., Arai, T., Nishibata, T., Kobayashi, M., Kimura, T., Kaneko, N., Ohtani, J., Yamanaka, K., et al. (2009). Small-molecule inhibition of Wee1 kinase by MK-1775 selectively sensitizes p53-deficient tumor cells to DNA-damaging agents. Mol. Cancer Ther. 8, 2992–3000.
Amita, M., Adachi, K., Alexenko, A.P., Sinha, S., Schust, D.J., Schulz, L.C., Roberts, R.M., and Ezashi, T. (2013). Complete and unidirectional conversion of human embryonic stem cells to trophoblast by BMP4. Proc. Natl. Acad. Sci. USA 110, E1212–E1221.
James, D., Levine, A.J., Besser, D., and Hemmati-Brivanlou, A. (2005). TGFbeta/activin/nodal signaling is necessary for the maintenance of pluripotency in human embryonic stem cells. Development 132, 1273–1282.
Baek, S.H., Choi, S.W., Park, S.J., Lee, S.H., Chun, H.S., and Kim, S.H. (2015). Quinoline compound KM11073 enhances BMP-2-dependent osteogenic differentiation of C2C12 cells via activation of p38 signaling and exhibits in vivo bone forming activity. PLoS ONE 10, e0120150. Bandyopadhyay, A., Yadav, P.S., and Prashar, P. (2013). BMP signaling in development and diseases: a pharmacological perspective. Biochem. Pharmacol. 85, 857–864. Blanpain, C., and Fuchs, E. (2009). Epidermal homeostasis: a balancing act of stem cells in the skin. Nat. Rev. Mol. Cell Biol. 10, 207–217. Cain, J.E., Hartwig, S., Bertram, J.F., and Rosenblum, N.D. (2008). Bone morphogenetic protein signaling in the developing kidney: present and future. Differentiation 76, 831–842. Cao, N., Liang, H., Huang, J., Wang, J., Chen, Y., Chen, Z., and Yang, H.T. (2013). Highly efficient induction and long-term maintenance of multipotent cardiovascular progenitors from human pluripotent stem cells under defined conditions. Cell Res. 23, 1119–1132.
Katagiri, T., Yamaguchi, A., Komaki, M., Abe, E., Takahashi, N., Ikeda, T., Rosen, V., Wozney, J.M., Fujisawa-Sehara, A., and Suda, T. (1994). Bone morphogenetic protein-2 converts the differentiation pathway of C2C12 myoblasts into the osteoblast lineage. J. Cell Biol. 127, 1755–1766. Kato, S., Sangadala, S., Tomita, K., Titus, L., and Boden, S.D. (2011). A synthetic compound that potentiates bone morphogenetic protein-2-induced transdifferentiation of myoblasts into the osteoblastic phenotype. Mol. Cell. Biochem. 349, 97–106. Kattman, S.J., Witty, A.D., Gagliardi, M., Dubois, N.C., Niapour, M., Hotta, A., Ellis, J., and Keller, G. (2011). Stage-specific optimization of activin/nodal and BMP signaling promotes cardiac differentiation of mouse and human pluripotent stem cell lines. Cell Stem Cell 8, 228–240. Kee, K., Gonsalves, J.M., Clark, A.T., and Pera, R.A. (2006). Bone morphogenetic proteins induce germ cell differentiation from human embryonic stem cells. Stem Cells Dev. 15, 831–837. Knight, P.G., and Glister, C. (2006). TGF-beta superfamily members and ovarian follicle development. Reproduction 132, 191–206.
Cao, Y., Wang, C., Zhang, X., Xing, G., Lu, K., Gu, Y., He, F., and Zhang, L. (2014). Selective small molecule compounds increase BMP-2 responsiveness by inhibiting Smurf1-mediated Smad1/5 degradation. Sci. Rep. 4, 4965.
Kruithof, B.P., Duim, S.N., Moerkamp, A.T., and Goumans, M.J. (2012). TGFbeta and BMP signaling in cardiac cushion formation: lessons from mice and chicken. Differentiation 84, 89–102.
Chen, H., Li, C., Huang, J., Cung, T., Seiss, K., Beamon, J., Carrington, M.F., Porter, L.C., Burke, P.S., Yang, Y., et al. (2011). CD4+ T cells from elite controllers resist HIV-1 infection by selective upregulation of p21. J. Clin. Invest. 121, 1549–1560.
Lapenna, S., and Giordano, A. (2009). Cell cycle kinases as therapeutic targets for cancer. Nat. Rev. Drug Discov. 8, 547–566.
2074 Cell Reports 15, 2063–2075, May 31, 2016
Larsson, J., and Karlsson, S. (2005). The role of Smad signaling in hematopoiesis. Oncogene 24, 5676–5692.
Lengerke, C., Schmitt, S., Bowman, T.V., Jang, I.H., Maouche-Chretien, L., McKinney-Freeman, S., Davidson, A.J., Hammerschmidt, M., Rentzsch, F., Green, J.B., et al. (2008). BMP and Wnt specify hematopoietic fate by activation of the Cdx-Hox pathway. Cell Stem Cell 2, 72–82.
Spiekerkoetter, E., Tian, X., Cai, J., Hopper, R.K., Sudheendra, D., Li, C.G., El-Bizri, N., Sawada, H., Haghighat, R., Chan, R., et al. (2013). FK506 activates BMPR2, rescues endothelial dysfunction, and reverses pulmonary hypertension. J. Clin. Invest. 123, 3600–3613.
Li, Z., Fei, T., Zhang, J., Zhu, G., Wang, L., Lu, D., Chi, X., Teng, Y., Hou, N., Yang, X., et al. (2012). BMP4 Signaling Acts via dual-specificity phosphatase 9 to control ERK activity in mouse embryonic stem cells. Cell Stem Cell 10, 171–182.
Takashima, S., and Hartenstein, V. (2012). Genetic control of intestinal stem cell specification and development: a comparative view. Stem Cell Rev. 8, 597–608.
Li, Y., Moretto-Zita, M., Soncin, F., Wakeland, A., Wolfe, L., Leon-Garcia, S., Pandian, R., Pizzo, D., Cui, L., Nazor, K., et al. (2013). BMP4-directed trophoblast differentiation of human embryonic stem cells is mediated through a DNp63+ cytotrophoblast stem cell state. Development 140, 3965–3976. Liu, A., and Niswander, L.A. (2005). Bone morphogenetic protein signalling and vertebrate nervous system development. Nat. Rev. Neurosci. 6, 945–954. Nakahiro, T., Kurooka, H., Mori, K., Sano, K., and Yokota, Y. (2010). Identification of BMP-responsive elements in the mouse Id2 gene. Biochem. Biophys. Res. Commun. 399, 416–421. Nishio, M., Yoneshiro, T., Nakahara, M., Suzuki, S., Saeki, K., Hasegawa, M., Kawai, Y., Akutsu, H., Umezawa, A., Yasuda, K., et al. (2012). Production of functional classical brown adipocytes from human pluripotent stem cells using specific hemopoietin cocktail without gene transfer. Cell Metab. 16, 394–406.
Vrijens, K., Lin, W., Cui, J., Farmer, D., Low, J., Pronier, E., Zeng, F.Y., Shelat, A.A., Guy, K., Taylor, M.R., et al. (2013). Identification of small molecule activators of BMP signaling. PLoS ONE 8, e59045. Wang, G., Zhang, H., Zhao, Y., Li, J., Cai, J., Wang, P., Meng, S., Feng, J., Miao, C., Ding, M., et al. (2005). Noggin and bFGF cooperate to maintain the pluripotency of human embryonic stem cells in the absence of feeder layers. Biochem. Biophys. Res. Commun. 330, 934–942. Wang, M., Guo, L., Wu, Q., Zeng, T., Lin, Q., Qiao, Y., Wang, Q., Liu, M., Zhang, X., Ren, L., et al. (2014). ATR/Chk1/Smurf1 pathway determines cell fate after DNA damage by controlling RhoB abundance. Nat. Commun. 5, 4901. Wei, W., Qing, T., Ye, X., Liu, H., Zhang, D., Yang, W., and Deng, H. (2008). Primordial germ cell specification from embryonic stem cells. PLoS ONE 3, e4013.
Nostro, M.C., Cheng, X., Keller, G.M., and Gadue, P. (2008). Wnt, activin, and BMP signaling regulate distinct stages in the developmental pathway from embryonic stem cells to blood. Cell Stem Cell 2, 60–71.
Xu, R.H., Chen, X., Li, D.S., Li, R., Addicks, G.C., Glennon, C., Zwaka, T.P., and Thomson, J.A. (2002). BMP4 initiates human embryonic stem cell differentiation to trophoblast. Nat. Biotechnol. 20, 1261–1264.
Okada, M., Sangadala, S., Liu, Y., Yoshida, M., Reddy, B.V., Titus, L., and Boden, S.D. (2009). Development and optimization of a cell-based assay for the selection of synthetic compounds that potentiate bone morphogenetic protein-2 activity. Cell Biochem. Funct. 27, 526–534.
Yang, L., Soonpaa, M.H., Adler, E.D., Roepke, T.K., Kattman, S.J., Kennedy, M., Henckaerts, E., Bonham, K., Abbott, G.W., Linden, R.M., et al. (2008). Human cardiovascular progenitor cells develop from a KDR+ embryonicstem-cell-derived population. Nature 453, 524–528.
Palmer, B.D., Thompson, A.M., Booth, R.J., Dobrusin, E.M., Kraker, A.J., Lee, H.H., Lunney, E.A., Mitchell, L.H., Ortwine, D.F., Smaill, J.B., et al. (2006). 4-Phenylpyrrolo[3,4-c]carbazole-1,3(2H,6H)-dione inhibitors of the checkpoint kinase Wee1. Structure-activity relationships for chromophore modification and phenyl ring substitution. J. Med. Chem. 49, 4896–4911. Robert, B. (2007). Bone morphogenetic protein signaling in limb outgrowth and patterning. Dev. Growth Differ. 49, 455–468. Sanjana, N.E., Shalem, O., and Zhang, F. (2014). Improved vectors and genome-wide libraries for CRISPR screening. Nat. Methods 11, 783–784. Shalem, O., Sanjana, N.E., Hartenian, E., Shi, X., Scott, D.A., Mikkelsen, T.S., Heckl, D., Ebert, B.L., Root, D.E., Doench, J.G., and Zhang, F. (2014). Genome-scale CRISPR-Cas9 knockout screening in human cells. Science 343, 84–87.
Ying, Q.L., Nichols, J., Chambers, I., and Smith, A. (2003). BMP induction of Id proteins suppresses differentiation and sustains embryonic stem cell selfrenewal in collaboration with STAT3. Cell 115, 281–292. Yu, P.B., Hong, C.C., Sachidanandan, C., Babitt, J.L., Deng, D.Y., Hoyng, S.A., Lin, H.Y., Bloch, K.D., and Peterson, R.T. (2008). Dorsomorphin inhibits BMP signals required for embryogenesis and iron metabolism. Nat. Chem. Biol. 4, 33–41. Zaret, K.S. (2001). Hepatocyte differentiation: from the endoderm and beyond. Curr. Opin. Genet. Dev. 11, 568–574. Zhu, H., Kavsak, P., Abdollah, S., Wrana, J.L., and Thomsen, G.H. (1999). A SMAD ubiquitin ligase targets the BMP pathway and affects embryonic pattern formation. Nature 400, 687–693.
Cell Reports 15, 2063–2075, May 31, 2016 2075