Advertisement
Research ArticleDrug Discovery and Translational Medicine

Differential Binding Activity of TGF-β Family Proteins to Select TGF-β Receptors

Ashraf M. Khalil, Hyna Dotimas, Julius Kahn, Jane E. Lamerdin, David B. Hayes, Priyanka Gupta and Michael Franti
Journal of Pharmacology and Experimental Therapeutics September 2016, 358 (3) 423-430; DOI: https://doi.org/10.1124/jpet.116.232322

Abstract

Growth differentiation factor-11 (GDF11) and myostatin (MSTN) are highly related transforming growth factor-β (TGF-β) ligands with 89% amino acid sequence homology. They have different biologic activities and diverse tissue distribution patterns. However, the activities of these ligands are indistinguishable in in vitro assays. SMAD2/3 signaling has been identified as the canonical pathway for GDF11 and MSTN, However, it remains unclear which receptor heterodimer and which antagonists preferentially mediate and regulate signaling. In this study, we investigated the initiation and regulation of GDF11 and MSTN signaling at the receptor level using a novel receptor dimerization detection technology. We used the dimerization platform to link early receptor binding events to intracellular downstream signaling. This approach was instrumental in revealing differential receptor binding activity within the TGF-β family. We verified the ActR2b/ALK5 heterodimer as the predominant receptor for GDF11- and MSTN-induced SMAD2/3 signaling. We also showed ALK7 specifically mediates activin-B signaling. We verified follistatin as a potent antagonist to neutralize both SMAD2/3 signaling and receptor dimerization. More remarkably, we showed that the two related antagonists, growth and differentiation factor–associated serum protein (GASP)-1 and GASP2, differentially regulate GDF11 (and MSTN) signaling. GASP1 blocks both receptor dimerization and downstream signaling. However, GASP2 blocks only downstream signaling without interference from receptor dimerization. Our data strongly suggest that physical binding of GDF11 (and MSTN) to both ActR2b and ALK5 receptors is required for initiation of signaling.

Introduction

The transforming growth factor-β (TGF-β) superfamily proteins are potent regulatory cytokines. In addition to their vital role in early development and homeostasis (Wu and Hill, 2009), they also play important roles in self-tolerance and autoimmunity (Li et al., 2006). TGF-β proteins selectively signal through multiple and variable cell surface serine/threonine kinase receptors (de Caestecker, 2004). The Mad gene from Drosophila 1 and the related Sma genes from Caenorhabditis elegans 2 (Smad) have been genetically implicated in signaling by members of the bone morphogenetic protein (BMP) group (Liu et al., 1996). The biologically active dimeric mature form of growth differentiation factor-11 (GDF11), myostatin (MSTN), and activin (Huang et al., 2011) can induce signal by initially binding to predominantly ActR2b, a type II receptor. This, in turn, forms a heterodimeric complex with one or more type I receptors, mainly, ALK4, ALK5, and ALK7. Phosphorylation of type I receptors by the ActR2b receptor will recruit and phosphorylate receptor-regulated Smad2 and receptor-regulated Smad3. The activated receptor-regulated Smad2/3 forms a complex with Smad4, which translocates to the nucleus to regulate expression of downstream genes (Dennler et al., 1998). The nature of the ligand: the receptor binding that drives ActR2b to preferentially pair with multiple type I receptors has not been fully elucidated.

In this study, we focused on the two closely related TGF-β family members, GDF11 and MSTN, which have been reported in multiple age-related disorders (Loffredo et al., 2013; Katsimpardi et al., 2014; Sinha et al., 2014; Egerman et al., 2015). They share the same signaling pathway and 89% sequence structure homology (McPherron et al., 1997; Nakashima et al., 1999). The similarity between the two ligands led to an indistinguishable in vitro functional activity (Egerman et al., 2015). However, studies of mRNA expression have revealed distinct in vivo tissue distribution patterns indicating diverse biologic functions. Expression of MSTN is mostly limited to skeletal muscles and a high level is correlated with inhibition of muscle growth (McPherron et al., 1997; Zimmers et al., 2002). GDF11 is widely expressed and involved in the development of multiple tissues (Wu et al., 2003; Harmon et al., 2004; Kim et al., 2005). Levels of circulating GDF11 decline in aged mice leading to multiple age-related disorders (Loffredo et al., 2013; Katsimpardi et al., 2014; Sinha et al., 2014; Egerman et al., 2015). However, a conflicting report showed that GDF11 levels increase with age in rats and humans, and correlate with inhibited myogenesis (Egerman et al., 2015). Moreover, a recent study has shown that GDF11 does not rescue age-related cardiac hypertrophy (Smith et al., 2015).

Multiple proteins have been identified in the regulation of GDF11 and MSTN. They bind these ligands and inhibit their activities (Schneyer et al., 1994; Amthor et al., 2004; Lee and Lee, 2013). One of these binding proteins, follistatin, a secreted glycoprotein, antagonizes numerous members of the TGF-β superfamily including MSTN, GDF11, activin, and BMPs (Hemmati-Brivanlou et al., 1994; Iemura et al., 1998; Amthor et al., 2002; Zimmers et al., 2002; Schneyer et al., 2008). Two related antagonists; growth and differentiation factor–associated serum protein (GASP)-1 and GASP2 were isolated from mouse and human serum (Hill et al., 2003). Both GASP1 and GASP2 have been shown to block MSTN and GDF11 activity in vitro (Hill et al., 2003; Kondás et al., 2008; Szláma et al., 2010), suggesting important roles for GASP1 and GASP2 in regulating the activities of these ligands (Lee and Lee, 2013). GASP1 and GASP2 specifically neutralize GDF11 and MSTN without an effect on activin or TGF-β1 (Lee and Lee, 2013). However, the differential role of GASP1 and GASP2 in regulating GDF11 (and MSTN) signaling has not been investigated.

We investigated differential binding activity of GDF11 and MSTN to their specific receptors and endogenous antagonists. A potential mechanism to mediate the diverse in vivo biologic activities of GDF11 and MSTN could be through differential receptor binding. Identifying differential signal mediators and antagonists at the TGF-β receptor level could accelerate development of therapeutics that target a specific TGF-β member for treatment of age-related disorders.

Materials and Methods

SMAD2/3 Reporter Cells and Assay

Briefly, human liver cancer cell line (HepG2) parental cells (American Type Culture Collection, Manassas, VA) were transfected with Cignal Lentiviral Reporter, SMAD Luciferase (Qiagen Inc., Valencia, CA). Viral particles expressed an inducible firefly luciferase reporter under the control of a SMAD2/3-specific transcriptional response element (CAGA12). SureENTRY transduction reagent (Qiagen) was prepared at 8 µg/ml in transfection media, with 10, 25, and 50 multiplicity of infection in growth transfection media: Dulbecco’s modified Eagle’s medium (Life Technologies Corporation, Grand Island, NY) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Hyclone Logan, Utah), 1X minimal essential medium nonessential amino acids (Life Technologies), and 1X sodium pyruvate (Life Technologies). Transduced pools were selected in HepG2 SMAD selective complete media: Dulbecco’s modified Eagle’s medium supplemented with 10% FBS, 1% penicillin/streptomycin (Life Technologies Corporation), and 10µg/ml puromycin (Sigma Aldrich, St. Louis, MO) to ensure stable integration of the reporter. Puromycin-resistant cells were maintained in selective media and used in experiments as follows. HepG2 SMAD2/3 luciferase reporter cells were harvested, washed, and resuspended at a concentration of 1 × 106 cells per ml in Opti-MEM Assay Media (Life Technologies) supplemented with 1% penicillin/streptomycin and 1X minimal essential medium nonessential amino acids. Reporter cells were incubated in 96-well plates at 50,000 cells per well with serial dilutions of TGF-β ligands. After 24-hour incubation, samples were treated with 100 µl STEADY-Glo reagent (Promega, Madison, WI) and assayed for luciferase expression. Relative luminescence units were plotted versus log10 nanomolar concentrations of the test GDF11, where the EC50 and 90% effective concentration values were calculated using a 4-parameter logistic model, supported by the Excel add-in XL-fit (ID Business Solutions Limited, Boston, MA) and GraphPad Prism, version 6 (GraphPad Software, Inc., La Jolla CA).

PathHunter Dimerization Assays

PathHunter ActR2b/ALK5, ALK4, or ALK7 dimerization assays were developed in collaboration with DiscoverX (Fremont, CA). The goal of this project was to create ActR2b heterodimer pairs with ALK4, ALK5, and ALK7 and generate clones from receptor pairs that respond to GDF11 or other TGF-β ligands. To generate a system to monitor this interaction, DiscoverX used its proprietary β-galactosidase enzyme fragment complementation assay system. The PathHunter dimerization assay detects ligand-induced dimerization of two subunits of a receptor-dimer pair constrained in the cell membrane, since the receptors would be in vivo. Human osteosarcoma cell line (U-2 OS) cells were engineered to coexpress the extracellular domain through the first 8aa of the cytosolic domain of the type II receptor subunit (ActR2b) fused to the small enzyme fragment, Prolink (PK), and the corresponding domain of a type I receptor (ALK4, ALK5, or ALK7) fused to the larger enzyme fragment, the enzyme acceptor (EA). Binding of an agonist to one receptor subunit (e.g., ActR2b-PK) induces it to interact with its dimer partner (e.g., ALK4-EA), forcing complementation of the two enzyme fragments. This results in the formation of a functional enzyme that hydrolyzes its substrate to generate a chemiluminescent signal. Pools were generated by transducing U-2 OS parental cells with vectors expressing the ActR2b-PK-tagged fusions. These pools were selected in appropriate antibiotics to ensure stable integration of the fusion protein. The expression level of the ActR2b-PK fusion protein was evaluated in the single stable pool using in vitro complementation assay. Stable clones with high, medium, and low expression of ActR2b-PK were obtained via limiting dilution and transduced with vectors expressing the EA-tagged fusions, followed by selection in antibiotic to generate double stable pools. Relative expression of each receptor in the double stable pools was evaluated using in vitro complementation in the presence of exogenous enzyme acceptor (EA) or enzyme donor (ED). The results indicated that ActR2b-PK was present in slight excess relative to ALK5-EA (1.5:1), but was expressed in equimolar amounts with ALK7 (1:1.2), and slightly lower relative to ALK4 (1: 1.6).

Cells from the U-2 OS heterodimer clones (ActR2b/ALK4, ActR2b/ALK5, and ActR2b/ALK7 β-galactosidase reporter assays) were maintained in U-2 OS complete selection media: Eagle’s minimum essential medium (American Type Culture Collection, Manassas, VA) supplemented with 10% FBS, 1% penicillin/streptomycin, 10 µg/ml hygromycin B and geneticin (Life Technologies). To establish full-dose response, reporter cells were harvested and plated at 10,000 cells/well in U-2 OS plating media: Eagle’s minimum essential medium supplemented with 2% FBS and 1% penicillin/streptomycin. Plated cells were treated with serial dilutions of TGF-β ligands in 100 µl reactions in 96-well plates. After 16-hour incubation, samples were treated with 100 µl substrate and PathHunter FLASH detection reagent (DiscoverX) and assayed for β-galactosidase expression. Relative luminescence units were plotted versus log10 nanomolar concentrations of the test agonists, where EC50 and 90% effective concentration values were calculated using a 4-parameter logistic model, supported by Excel add-in XL-fit.(ID Business Solutions Limited) and GraphPad Prism, version 6 (GraphPad Software, Inc.).

Protein Ligands and Antagonists

Recombinant human TGF-β ligands used in both SMAD and the dimerization assays were the following: GDF11, MSTN, TGF-β1, BMP2, BMP4 and BMP7 (R&D Systems Inc., Minneapolis, MN). Activin A and Activin B (PeproTech, Rocky Hill, NJ). Antagonists used in both SMAD and the dimerization assays were the following: GASP1 (PeproTech), GASP2 and Follistatin 315 (R&D Systems).

Statistical Analysis.

The EC and IC values were determined using GraphPad Prism, version 6 (GraphPad Software, Inc., La Jolla CA) or XL-fit, version 4 (ID Business Solutions Limited, Boston, MA).

Results

We investigated initiation and downstream signaling of multiple TGF-β ligands by their ability to bind select TGF-β receptors and induce SMAD2/3 signaling, respectively.

SMAD2/3 Signaling.

Functional activity of multiple TGF-β ligands was determined by their ability to activate a SMAD2/3 transcriptional reporter-driving expression of luciferase in HepG2 cells (Fig. 1). This involves binding of the ligands to specific endogenous receptor heterodimers, which lead to coupling to downstream effectors. Activated SMADs bind to a specific transcription element (CAGA) previously identified in the target gene PAI-1 promoter (Dennler et al., 1998). GDF11, MSTN, activin, and TGF-β1 all induced robust SMAD2/3-luciferase activity with low EC50 (Fig. 1; Table 1). However, activin A and activin B showed lower efficacious responses with 63% and 47% maximum efficacy, respectively, compared with GDF11 (Table 1). BMP2 and BMP4 were included as negative controls. Both ligands have been reported to activate SMAD1/5/8 via ALK1, ALK2, and ALK3 receptors (Korchynskyi and ten Dijke, 2002).

Fig. 1.
Fig. 1.

TGF-β-induced SMAD2/3 signaling. SMAD2/3 luciferase reporter HepG2 cells were treated with serial dilutions of TGF-β ligands starting at 200 nM concentration to establish EC50 and 90% effective concentration (EC90) full-dose responses. Both BMP2 and BMP4 were included as negative controls to validate specificity of the assay. After 24-hour incubation, samples were treated with 100 µl STEADY-Glo reagent (Promega, Madison, WI) and assayed for luciferase expression. Each data point is the mean and S.E.M. of luciferase expression, reported as relative luminescence units (RLU). The RLU values were plotted versus log10 nanomolar concentrations of the test agonists. EC50 and EC90 values were calculated using a 4-parameter logistic model, supported by the Excel add-in XLfit. GDF11 and other TGF-β ligands induced SMAD2/3 activity with EC50, EC90 ± S.E.M., and efficacy maximum percentage are presented in Table 1.

View this table:
TABLE 1

Potency of TGF-β ligands in SMAD2/3 signaling

Calculated potency [EC50 (nM) ± S.E.M.] and efficacy (maximum percentage of efficacy) of TGF-β ligands indicate their affinity to specific receptors and ability to induce downstream SMAD2/3 signaling, respectively. EC50 values ± S.E.M. were calculated from the full-dose response of SMAD2/3 luciferase expression to TGF-β ligands using XL-fit and Prism. The maximum efficacy percentage was calculated from value on the y-axis, which corresponds to the EC90 value on the x-axis and was normalized to GDF11.

Receptor Dimerization.

Binding activity of GDF11 and MSTN to specific TGF-β receptors was determined by their ability to induce receptor heterodimers using the receptor dimerization detection technology (PathHunter, DiscoverX), which has been used previously to study the effects of therapeutic antibodies on ErbB2 heterodimer partners (Wehrman et al., 2006). For the TGF-β receptor assays, β-galactosidase (PK)-tagged ActR2b was coexpressed with each of the β-galactosidase (EA)-tagged ALK4, ALK5, or ALK7 in U-2 OS cells. When stimulated with the appropriate ligand, the two enzyme fragments of the receptors interact and reconstitute a functional β-galactosidase enzyme, which produces a luminescent signal when incubated with substrate and detection reagent. To determine the receptor heterodimer that predominately mediates SMAD2/3 signaling, three reporter cell lines were generated to coexpress ActR2b with each of the type I receptors (ALK4, ALK5, or ALK7). We examined select TGF-β ligands for ability to bind ActR2b and preferentially induce receptor dimerization (Fig. 2).

Fig. 2.
Fig. 2.

TGF-β-induced receptor hetrodimerization. U-2 OS cells expressing B Gal reporter and the receptor dimers ActR2b/ALK4 (A), ActR2b/ALK5 (B), or ActR2b/ALK7 (C) were treated with serial dilutions of TGF-β ligands starting at 10 nM concentration to establish EC50 and 90% effective concentration full-dose responses. After 16-hour incubation, samples were treated with 100 µl substrate and PathHunter FLASH detection reagent and assayed for β-galactosidase activity. Each data point is the mean and S.E.M. of β-galactosidase activity, reported as relative luminescence units (RLU). RLU values were plotted versus log10 nanomolar concentrations of the test agonists. EC50 values were calculated using a 4-parameter logistic model, supported by the Excel add-in XLfit. Both EC50 ± S.E.M. and efficacy maximum percentage values are presented in Table 2.

ActR2b/ALK4.

All ligands (GDF11, MSTN, activin A, and activin B) induced ActR2b/ALK4 dimerization with comparable potency (EC50) and efficacy (Fig. 2A; Table 2). TGF-β1 was included as the negative control.

View this table:
TABLE 2

Receptor dimerization by TGF-β ligands

Calculated Potency [EC50 (nM) ± S.E.M.] and efficacy (maximum percentage of efficacy) of TGF-β ligands indicate their affinity to specific receptor dimers (ActR2b/ALK4, ActR2b/ALK5, and ActR2b/ALK7) and their ability to induce receptor dimerization, respectively. The maximum efficacy percentage was calculated from the value on the y-axis, which corresponds to the EC90 value on the x-axis and was normalized to activin A (ALK4), MSTN (ALK5), and activin B (ALK7).

ActR2b/ALK5.

GDF11, MSTN, activin A, and activin B effectively induced ActR2b/ALK5 dimerization with low EC50 (Fig. 2B; Table 2). However, activin A and activin B showed lower efficacious responses with 30% and 42% maximum efficacy, respectively, compared with GDF11 (Table 2). TGF-β1, BMP2, and BMP7 were included as negative controls to validate the specificity of the dimerization assay as demonstrated for the SMAD2/3 signaling assay. TGF-β1 is reported to signal through TGF-B RII/ALK1, ALK2, ALK5, and ALK7 heterodimers, whereas BMP2 and BMP7 signal through BMP RII/ALK3 and ALK6 heterodimers (de Caestecker, 2004). All ligands induced dimerization with higher potency (lower EC50) for ActR2b/ALK5 than ActR2b/ALK4 (Table 2), suggesting dual binding activity of TGF-β ligands to type II and I receptors rather than only type II receptors as previously reported (Andersson et al., 2006).

ActR2b/ALK7.

Only activin B induced ActR2b/ALK7 dimerization (Fig. 2C; Table 2), whereas GDF11 and MSTN showed lower efficacious responses with 26% and 31% maximum efficacy, respectively, compared with activin B. Activin A did not induce ALK7 response. Our data are consistent with previous reports (Tsuchida et al., 2004) indicating that ALK7 is the activin B receptor, but does not mediate activin A signaling.

Antagonists Induce Differential Blocking Activity.

In the previous sections, we showed that both GDF11 and MSTN induced comparable SMAD2/3 (Table 1) and dimerization responses (Table 2). We next sought to determine if endogenous antagonists involved in blocking GDF11 binding were distinct from those involved with blocking MSTN. Therefore, GASP1, GASP2, and follistatin were tested for their ability to neutralize SMAD2/3 signaling (Fig. 3, A and B; Table 3) and to block dimerization of ActR2b/ALK5 (Fig. 3, C and D; Table 4), ActR2b/ALK4 (Fig. 3E), and ActR2b/ALK7 (Fig. 3F; Table 5). The results showed variable blocking ability of GASP1, GASP2, and follistatin. While follistatin, GASP1, and GASP2 effectively neutralized GDF11- and MSTN-induced SMAD2/3 response (Fig. 3, A and B; Table 3), only follistatin and GASP1 blocked GDF11- and MSTN-induced ActR2b/ALK5 dimerization (Fig. 3, C and D; Table 4). Consistent with its effect on ActR2b/ALK5, follistatin effectively blocked GDF11-induced dimerization of ActR2b/ALK4 (Fig. 3E) and activin B induced ActR2b/ALK7 dimerization (Fig. 3F; Table 5), suggesting follistatin targets ActR2b-specific epitope.

Fig. 3.
Fig. 3.

Inhibition of SMAD2/3 and receptor dimerization by antagonists. Serial dilutions of GASP1, GASP2, and follistatin starting at 60 nM concentration preincubated for 30 minutes at 37°C with a constant concentration of 1 nM GDF11 (A) and MSTN (B) in SMAD2/3 luciferase assay or 0.4 nM of GDF11 (C) and MSTN (D) in ActR2b/ALK5 PathHunter β-galactosidase assay. Also, serial dilutions of GASP1, GASP2, and follistatin run against a constant concentration of 6 nM GDF11 (E) in ActR2b/ALK4 β-galactosidase assay or 2 nM activin B (F) in ActR2b/ALK7 β-galactosidase assay. Each data point is the mean and S.E.M. of reporter activity, reported as relative luminescence units (RLU). RLU values were plotted versus log10 nanomolar concentrations of the test antagonists. The IC50 and 90% inhibitory concentration values were calculated using a 4-parameter logistic model, supported by the Excel add-in XLfit. EC50 ± S.E.M. values are presented in Tables 3–5.

View this table:
TABLE 3

Antagonist inhibition of GDF11- or MSTN-induced SMAD2/3 signaling

Potency [IC50 and IC90 (nM) ± S.E.M.] and efficacy (percentage of inhibition) of GASP1, GASP2, and follistatin indicate their ability to inhibit the SMAD2/3 signal and block receptor dimerization. Antagonists effect on GDF11 and MSTN-induced SMAD2/3 luciferase activity. All the antagonists effectively inhibited GDF11- and MSTN-induced SMAD2/3 signaling.

View this table:
TABLE 4

Antagonist inhibition of GDF11- or MSTN-induced ActR2b/ALK5 dimerization

Potency [IC50 and IC90 (nM) ± S.E.M.] and efficacy (percentage of inhibition) of GASP1, GASP2, and follistatin indicate their ability to block receptor dimerization. Antagonists effect on GDF11- and MSTN-induced ActR2b/ALK5 dimerization. Follistatin, and to a lesser extent GASP1, effectively blocked GDF11- and MSTN-induced ActR2b/ALK5 dimerization. GASP2 had no effect.

View this table:
TABLE 5

Antagonist inhibition of activin B–mediated ActR2b/ALK7 dimerization

Potency [IC50 and IC90 (nM) ± S.E.M.] and efficacy (percentage of inhibition) of GASP1, GASP2, and follistatin indicate their ability to block receptor dimerization. Antagonists effect on ActR2b/ALK7 dimerization. Only follistatin blocked activin B–induced ActR2b/ALK7 dimerization. GASP1 and GASP2 have no effect.

Discussion

The canonical SMAD2/3 signaling induced by the activin branch of TGF-β superfamily proteins (GDF11, GDF8, and activin) is initiated by binding to type II and I serine/threonine receptors and is regulated by multiple antagonists including GASP1, GASP2, and follistatin. It remains not fully determined which receptor heterodimers and antagonists preferentially mediate and regulate signal, respectively. The endogenous antagonists GASP1, GASP2, and follistatin bind and neutralize members of the TGF-β superfamily, including GDF11 and activin. However, possible divergent roles of these antagonists have not been fully explored. In this study, we used two complementary assay systems, SMAD signaling and receptor dimerization, to provide novel information about how the specificity of signaling is achieved by given TGF-β family ligands. In one approach (the luciferase reporter assay), signaling through endogenous receptors was evaluated using a readout that integrates signals emanating from multiple receptors. Therefore, an important element of our approach was to address the specificity of ligand binding by reconstructing the heteromeric receptors using an assay that quantitatively measures the first step of activation of the receptors, which is ligand-induced dimerization. The latter approach was previously used to determine the mechanism of action of therapeutic molecules for several different types of receptors (receptor tyrosine kinases and cytokine receptors) (Wehrman et al., 2006; Moraga et al., 2015). Importantly, this approach models the context dependent dimerization of heteromeric receptors, topologically constrained in the plasma membrane in a physiologically relevant manner.

Using these models, we investigated the initiation and regulation of GDF11 and MSTN signaling at the receptor level. Assigning differential binding activity between GDF11 and MSTN either to their receptors or to endogenous antagonists can be useful if a specific modulator needs to be developed. Our data revealed multiple TGF-β ligands with differential responses on SMAD2/3 signaling and receptor dimerization. The robust SMAD2/3 signals induced by GDF11 and MSTN are predominantly mediated by the ActR2b/ALK5 heterodimer and to a lesser extent by ActR2b/ALK4. However, our data indicate that ALK7 does not mediate GDF11 or MSTN signaling. While ALK4 equally mediates signaling of GDF11, MSTN, and activin A (but not activin B), ALK7 is the predominant receptor for activin B only, consistent with reports showing ALK7 as a crucial receptor for activin B (Tsuchida et al., 2004). The highly potent (low EC50 values), yet submaximal, efficacy of the activins, as revealed in both SMAD signaling and ActR2b/ALK5 dimerization, indicates partial agonism of these ligands. This may reflect intrinsic efficacy of activin or low receptor occupancy. Activin predominantly binds to ActR2A/ALK1 and ALK2 receptors in addition to ActR2b/ALK4 receptors (de Caestecker, 2004). TGF-β1 was included as a negative control for the dimerization assays because it does not bind to ActR2b; rather, it is reported to signal through TGF-B RII/ALK1, ALK2, ALK5, and ALK7 heterodimers. Two other negative controls included in the study, BMP2 and BMP7, signal through BMP RII/ALK3 and ALK6 heterodimers (de Caestecker, 2004) and were not active in each of the three ActR2b heterodimers tested.

This differential binding activity could contribute to diverse in vivo biologic activities between GDF11, MSTN, activin A, and TGF-β1. Considering their extensive biophysical and structural similarity (McPherron et al., 1997; Nakashima et al., 1999), the indistinguishable responses of GDF11 and MSTN are in agreement with published data (Schneyer et al., 2008). Similar data have been reported for ActR2b-mediated interaction of GDF11 with ALK4 and ALK5 receptors using pull-down assays (Andersson et al., 2006).

In agreement with previous findings, where follistatin was shown to bind and neutralize activin and MSTN (Schneyer et al., 1994), we similarly verified that follistatin neutralizes SMAD2/3 signaling and, furthermore, it blocks ActR2b/type I receptor heterodimerization. We also verified that GASP1 inhibition of GDF11- and MSTN-induced SMAD2/3 signaling (Hill et al., 2003; Kondás et al., 2008; Szláma et al., 2010) was mediated through blocking ActR2b/ALK5 dimerization. However, relative to follistatin, GASP1 partially inhibits GDF11- and MSTN-induced ActR2b/ALK5 dimerization (Fig. 3, C and D). The variable activity of the antagonists could be related to the target epitope on the ligand or an intrinsic property of the antagonists. Additionally, GASP2 inhibited GDF11- and MSTN-induced SMAD2/3 activity (Fig. 3, A and B) but surprisingly failed to block GDF11- and MSTN-induced ActR2b/ALK5 receptor dimerization (Fig. 3, C and D), indicating GASP2 targets a GDF11 epitope important for binding to ALK5 rather than ActR2b. Conversely, GASP1 neutralized SMAD2/3 signaling and blocked GDF11-induced ActR2b/ALK5 dimerization, indicating GASP1 targets a GDF11 epitope crucial for binding to ActR2b. Furthermore, the same ligands induced differential binding activity to ActR2b/ALK5 compared with ActR2b/ALK4 (see the EC50 values in Table 2). In summary, our data strongly suggest that TGF-β ligands (GDF11 and MSTN) bind to both type II and I receptors, not just type II receptors as previously reported (Andersson et al., 2006). GDF11 has been reported to only bind directly to ActR2b but not to any of the type I receptors alone, possibly due to the low binding affinity of GDF11 to type I receptors or sensitivity of their assay.

Currently, direct structural evidence is not available. It is possible that dual GDF11 binding to type II and I receptors occurs, where follistatin and GASP1 interfere with the high affinity GDF11 binding to ActR2b while GASP2 blocks the low affinity GDF11 binding to ALK5. However, our study does not rule out the possibility of direct interaction of GASP1 and follistatin with type I receptors as well. Follistatin was reported to interfere with ligand binding to both type II and I receptors (Thompson et al., 2005), which explains its higher potency in the SMAD and dimerization assays compared with GASP1 and GASP2 (Fig. 3). Our data suggest a model where GDF11 (or MSTN) has dual binding activity to both type II and I receptors, where GASP1 and GASP2 differentially block the downstream signal. Our dual binding hypothesis agrees with a previous report (Thompson et al., 2005), where follistatin co-crystalized with activin A reveals an interaction of follistatin with both type II and I receptors. Furthermore, the nature of the GDF11 binding that drives ActR2b heterodimerization with multiple type I receptors has not been discussed. We show that GASP2 neutralization of SMAD2/3 signaling is not due to blocking the binding of GDF11 to ActR2b but likely to blocking GDF11-mediated crosslinking of ActR2b and ALK5. This indicates that the GASP2-specific domain on GDF11 is not required for ActR2b binding and dimerization.

In summary, our approach revealed differential receptor binding activity of GDF11 and MSTN compared with other TGF-β ligands. We showed that the robust SMAD2/3 signal induced by GDF11 (similar to MSTN) was predominantly mediated by dimerization of ActR2b with ALK5 and to a lesser extent with ALK4. However, ALK7 in the context of ActR2b plays no role in GDF11- and MSTN-mediated signaling. The endogenous antagonists GASP1 and follistatin both neutralize GDF11-induced signal by blocking ActR2b/ALK5 dimerization. Interestingly, GASP2 neutralizes SMAD2/3 signaling without blocking receptor dimerization. Our data suggest that GASP2 neutralization of GDF11-induced SMAD2/3 signaling is likely via blocking GDF11 binding to type I receptors but not blocking its binding to ActR2b. Therefore, we hypothesized a dual binding activity of GDF11 to both type I and II receptors, which are targets for GASP1 and GASP2, respectively.

Acknowledgments

The authors thank Nancy Burgh for providing administrative assistance and Dr. Scott DeWire for valuable comments.

Authorship Contributions

Participated in research design: Khalil, Lamerdin, Franti.

Conducted experiments: Khalil, Dotimas.

Contributed new reagents or analytic tools: Khalil, Kahn, Hayes, Gupta.

Performed data analysis: Khalil, Lamerdin, Franti, Kahn, Hayes, Gupta.

Wrote or contributed to the writing of the manuscript: Khalil, Lamerdin, Franti.

Footnotes

Abbreviations

BMP
bone morphogenetic protein
EA
enzyme acceptor
FBS
fetal bovine serum
GASP
growth and differentiation factor–associated serum protein
GDF11
growth differentiation factor-11
HepG2
human liver cancer cell line
MSTN
myostatin
PK
Prolink
TGF-β
transforming growth factor-β
U-2 OS
human osteosarcoma cell line

References

  1. ↵
    1. Amthor H,
    2. Christ B,
    3. Rashid-Doubell F,
    4. Kemp CF,
    5. Lang E, and
    6. Patel K
    (2002) Follistatin regulates bone morphogenetic protein-7 (BMP-7) activity to stimulate embryonic muscle growth. Dev Biol 243:115–127.
    OpenUrlCrossRefPubMed
  2. ↵
    1. Amthor H,
    2. Nicholas G,
    3. McKinnell I,
    4. Kemp CF,
    5. Sharma M,
    6. Kambadur R, and
    7. Patel K
    (2004) Follistatin complexes Myostatin and antagonises Myostatin-mediated inhibition of myogenesis. Dev Biol 270:19–30.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Andersson O,
    2. Reissmann E, and
    3. Ibáñez CF
    (2006) Growth differentiation factor 11 signals through the transforming growth factor-β receptor ALK5 to regionalize the anterior–posterior axis. EMBO Rep 7:831–837.
    OpenUrlAbstract/FREE Full Text
  4. ↵
    1. de Caestecker M
    (2004) The transforming growth factor-β superfamily of receptors. Cytokine Growth Factor Rev 15:1–11.
    OpenUrlCrossRefPubMed
  5. ↵
    1. Dennler S,
    2. Itoh S,
    3. Vivien D,
    4. ten Dijke P,
    5. Huet S, and
    6. Gauthier JM
    (1998) Direct binding of Smad3 and Smad4 to critical TGFβ-inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene. EMBO J 17:3091–3100.
    OpenUrlAbstract/FREE Full Text
  6. ↵
    1. Egerman MA,
    2. Cadena SM,
    3. Gilbert JA,
    4. Meyer A,
    5. Nelson HN,
    6. Swalley SE,
    7. Mallozzi C,
    8. Jacobi C,
    9. Jennings LL,
    10. Clay I,
    11. et al.
    (2015) GDF11 increases with age and inhibits skeletal muscle regeneration. Cell Metab 22:164–174.
    OpenUrlCrossRefPubMed
  7. ↵
    1. Harmon EB,
    2. Apelqvist AA,
    3. Smart NG,
    4. Gu X,
    5. Osborne DH, and
    6. Kim SK
    (2004) GDF11 modulates NGN3+ islet progenitor cell number and promotes β-cell differentiation in pancreas development. Development 131:6163–6174.
    OpenUrlAbstract/FREE Full Text
  8. ↵
    1. Hemmati-Brivanlou A,
    2. Kelly OG, and
    3. Melton DA
    (1994) Follistatin, an antagonist of activin, is expressed in the Spemann organizer and displays direct neuralizing activity. Cell 77:283–295.
    OpenUrlCrossRefPubMed
  9. ↵
    1. Hill JJ,
    2. Qiu Y,
    3. Hewick RM, and
    4. Wolfman NM
    (2003) Regulation of myostatin in vivo by growth and differentiation factor-associated serum protein-1: a novel protein with protease inhibitor and follistatin domains. Mol Endocrinol 17:1144–1154.
    OpenUrlCrossRefPubMed
  10. ↵
    1. Huang Z,
    2. Chen X, and
    3. Chen D
    (2011) Myostatin: a novel insight into its role in metabolism, signal pathways, and expression regulation. Cell Signal 23:1441–1446.
    OpenUrlCrossRefPubMed
  11. ↵
    1. Iemura S,
    2. Yamamoto TS,
    3. Takagi C,
    4. Uchiyama H,
    5. Natsume T,
    6. Shimasaki S,
    7. Sugino H, and
    8. Ueno N
    (1998) Direct binding of follistatin to a complex of bone-morphogenetic protein and its receptor inhibits ventral and epidermal cell fates in early Xenopus embryo. Proc Natl Acad Sci USA 95:9337–9342.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Katsimpardi L,
    2. Litterman NK,
    3. Schein PA,
    4. Miller CM,
    5. Loffredo FS,
    6. Wojtkiewicz GR,
    7. Chen JW,
    8. Lee RT,
    9. Wagers AJ, and
    10. Rubin LL
    (2014) Vascular and neurogenic rejuvenation of the aging mouse brain by young systemic factors. Science 344:630–634.
    OpenUrlAbstract/FREE Full Text
  13. ↵
    1. Kim J,
    2. Wu HH,
    3. Lander AD,
    4. Lyons KM,
    5. Matzuk MM, and
    6. Calof AL
    (2005) GDF11 controls the timing of progenitor cell competence in developing retina. Science 308:1927–1930.
    OpenUrlAbstract/FREE Full Text
  14. ↵
    1. Kondás K,
    2. Szláma G,
    3. Trexler M, and
    4. Patthy L
    (2008) Both WFIKKN1 and WFIKKN2 have high affinity for growth and differentiation factors 8 and 11. J Biol Chem 283:23677–23684.
    OpenUrlAbstract/FREE Full Text
  15. ↵
    1. Korchynskyi O and
    2. ten Dijke P
    (2002) Identification and functional characterization of distinct critically important bone morphogenetic protein-specific response elements in the Id1 promoter. J Biol Chem 277:4883–4891.
    OpenUrlAbstract/FREE Full Text
  16. ↵
    1. Lee YS and
    2. Lee SJ
    (2013) Regulation of GDF-11 and myostatin activity by GASP-1 and GASP-2. Proc Natl Acad Sci USA 110:E3713–E3722.
    OpenUrlAbstract/FREE Full Text
  17. ↵
    1. Li MO,
    2. Wan YY,
    3. Sanjabi S,
    4. Robertson AK, and
    5. Flavell RA
    (2006) Transforming growth factor-β regulation of immune responses. Annu Rev Immunol 24:99–146.
    OpenUrlCrossRefPubMed
  18. ↵
    1. Liu F,
    2. Hata A,
    3. Baker JC,
    4. Doody J,
    5. Cárcamo J,
    6. Harland RM, and
    7. Massagué J
    (1996) A human Mad protein acting as a BMP-regulated transcriptional activator. Nature 381:620–623.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Loffredo FS,
    2. Steinhauser ML,
    3. Jay SM,
    4. Gannon J,
    5. Pancoast JR,
    6. Yalamanchi P,
    7. Sinha M,
    8. Dall’Osso C,
    9. Khong D,
    10. Shadrach JL,
    11. et al.
    (2013) Growth differentiation factor 11 is a circulating factor that reverses age-related cardiac hypertrophy. Cell 153:828–839.
    OpenUrlCrossRefPubMed
  20. ↵
    1. McPherron AC,
    2. Lawler AM, and
    3. Lee SJ
    (1997) Regulation of skeletal muscle mass in mice by a new TGF-β superfamily member. Nature 387:83–90.
    OpenUrlCrossRefPubMed
  21. ↵
    1. Moraga I,
    2. Wernig G,
    3. Wilmes S,
    4. Gryshkova V,
    5. Richter CP,
    6. Hong WJ,
    7. Sinha R,
    8. Guo F,
    9. Fabionar H,
    10. Wehrman TS,
    11. et al.
    (2015) Tuning cytokine receptor signaling by re-orienting dimer geometry with surrogate ligands. Cell 160:1196–1208.
    OpenUrlCrossRefPubMed
  22. ↵
    1. Nakashima M,
    2. Toyono T,
    3. Akamine A, and
    4. Joyner A
    (1999) Expression of growth/differentiation factor 11, a new member of the BMP/TGFβ superfamily during mouse embryogenesis. Mech Dev 80:185–189.
    OpenUrlCrossRefPubMed
  23. ↵
    1. Schneyer AL,
    2. Rzucidlo DA,
    3. Sluss PM, and
    4. Crowley WF Jr.
    (1994) Characterization of unique binding kinetics of follistatin and activin or inhibin in serum. Endocrinology 135:667–674.
    OpenUrlCrossRefPubMed
  24. ↵
    1. Schneyer AL,
    2. Sidis Y,
    3. Gulati A,
    4. Sun JL,
    5. Keutmann H, and
    6. Krasney PA
    (2008) Differential antagonism of activin, myostatin and growth and differentiation factor 11 by wild-type and mutant follistatin. Endocrinology 149:4589–4595.
    OpenUrlCrossRefPubMed
  25. ↵
    1. Sinha M,
    2. Jang YC,
    3. Oh J,
    4. Khong D,
    5. Wu EY,
    6. Manohar R,
    7. Miller C,
    8. Regalado SG,
    9. Loffredo FS,
    10. Pancoast JR,
    11. et al.
    (2014) Restoring systemic GDF11 levels reverses age-related dysfunction in mouse skeletal muscle. Science 344:649–652.
    OpenUrlAbstract/FREE Full Text
  26. ↵
    1. Smith SC,
    2. Zhang X,
    3. Zhang X,
    4. Gross P,
    5. Starosta T,
    6. Mohsin S,
    7. Franti M,
    8. Gupta P,
    9. Hayes D,
    10. Myzithras M,
    11. et al.
    (2015) GDF11 does not rescue aging-related pathological hypertrophy. Circ Res 117:926–932.
    OpenUrlAbstract/FREE Full Text
  27. ↵
    1. Szláma G,
    2. Kondás K,
    3. Trexler M, and
    4. Patthy L
    (2010) WFIKKN1 and WFIKKN2 bind growth factors TGFβ1, BMP2 and BMP4 but do not inhibit their signalling activity. FEBS J 277:5040–5050.
    OpenUrlCrossRefPubMed
  28. ↵
    1. Thompson TB,
    2. Lerch TF,
    3. Cook RW,
    4. Woodruff TK, and
    5. Jardetzky TS
    (2005) The structure of the follistatin:activin complex reveals antagonism of both type I and type II receptor binding. Dev Cell 9:535–543.
    OpenUrlCrossRefPubMed
  29. ↵
    1. Tsuchida K,
    2. Nakatani M,
    3. Yamakawa N,
    4. Hashimoto O,
    5. Hasegawa Y, and
    6. Sugino H
    (2004) Activin isoforms signal through type I receptor Serine/threonine kinase ALK7. Mol Cell Endocrinol 220:59–65.
    OpenUrlCrossRefPubMed
  30. ↵
    1. Wehrman TS,
    2. Raab WJ,
    3. Casipit CL,
    4. Doyonnas R,
    5. Pomerantz JH, and
    6. Blau HM
    (2006) A system for quantifying dynamic protein interactions defines a role for Herceptin in modulating ErbB2 interactions. Proc Natl Acad Sci USA 103:19063–19068.
    OpenUrlAbstract/FREE Full Text
  31. ↵
    1. Wu HH,
    2. Ivkovic S,
    3. Murray RC,
    4. Jaramillo S,
    5. Lyons KM,
    6. Johnson JE, and
    7. Calof AL
    (2003) Autoregulation of neurogenesis by GDF11. Neuron 37:197–207.
    OpenUrlCrossRefPubMed
  32. ↵
    1. Wu MY and
    2. Hill CS
    (2009) TGF-β superfamily signaling in embryonic development and homeostasis. Dev Cell 16:329–343.
    OpenUrlCrossRefPubMed
  33. ↵
    1. Zimmers TA,
    2. Davies MV,
    3. Koniaris LG,
    4. Haynes P,
    5. Esquela AF,
    6. Tomkinson KN,
    7. McPherron AC,
    8. Wolfman NM, and
    9. Lee SJ
    (2002) Induction of cachexia in mice by systemically administered myostatin. Science 296:1486–1488.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Download PDF
Article Alerts
Email Article
Citation Tools
Research ArticleDrug Discovery and Translational Medicine

Novel Insight on TGF-β Receptors Dimerization

Ashraf M. Khalil, Hyna Dotimas, Julius Kahn, Jane E. Lamerdin, David B. Hayes, Priyanka Gupta and Michael Franti
Journal of Pharmacology and Experimental Therapeutics September 1, 2016, 358 (3) 423-430; DOI: https://doi.org/10.1124/jpet.116.232322
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section

Advertisement