Báo cáo khóa học: Nerve growth factor mediates activation of the Smad pathway in PC12 cells doc

Nerve growth factor mediates activation of the Smad pathwayin PC12 cells Marion Lutz1, Kerstin Krieglstein2, Simone Schmitt1, Peter ten Dijke3, Walter Sebald1, Andrea Wizenmann4 and Petr

Nerve growth factor mediates activation of the Smad pathway

in PC12 cells

Marion Lutz1, Kerstin Krieglstein2, Simone Schmitt1, Peter ten Dijke3, Walter Sebald1, Andrea Wizenmann4 and Petra Knaus1

1

Department of Physiological Chemistry II, Biocenter, University of Wu¨rzburg, Germany;2Department of Anatomy, University

of Go¨ttingen, Germany;3Division of Cellular Biochemistry, the Netherlands Cancer Institute, Amsterdam, the Netherlands;

4

JRG (of) Developmental Neurobiology, Biocenter, University of Wu¨rzburg, Germany

Ligand-induced oligomerization of receptors is a key step

in initiating growth factor signaling Nevertheless, complex

biological responses often require additional trans-signaling

mechanisms involving two or more signaling cascades For

cells of neuronal origin, it was shown that neurotrophic

effects evoked by nerve growth factor or other

neurotro-phins depend highly on the cooperativity with cytokines that

belong to the transforming growth factor b (TGF-b)

superfamily We found that rat pheochromocytoma cells,

which represent a model system for neuronal differentiation,

are unresponsive to TGF-b1 due to limiting levels of its

receptor, TbRII However, stimulation with nerve growth

factor leads to activation of the Smad pathway independent

of TGF-b In contrast to TGF-b signaling, activation of

Smad3 by nerve growth factor does not occur via

phospho-rylation of the C-terminal SSXS-motif, but leads to

hetero-meric complexformation with Smad4, nuclear translocation

of Smad3 and transcriptional activation of Smad-dependent reporter genes This response is direct and does not require

de novoprotein synthesis, as shown by cycloheximide treat-ment This initiation of transcription is dependent on func-tional tyrosine kinase receptors and can be blocked by Smad7 These data provide further evidence that the Smad proteins are not exclusively activated by the classical TGF-b triggered mechanism The potential of NGF to activate the Smad pathway independent of TGF-b represents an import-ant regulatory mechanism with special relevance for the development and function of neuronal cells or of other NGF-sensitive cells, in particular those that are TGF-b-resistant Keywords: PC12 cells; Smads; crosstalk; nerve growth factor; transforming growth factor-b

Proteins of the transforming growth factor b (TGF-b)

family are multifunctional cytokines that display a very

broad range of biological activities including cell

prolifer-ation, differentiation and apoptosis [1] TGF-bs are

ubi-quitously expressed and act on virtually all tissues, thereby

causing distinct cell-specific effects depending on the present

composition of receptors, Smad proteins and DNA-binding

partners [2,3] Referring to cell populations of neuronal

origin, TGF-bs are described to possess neurotrophic effects

when acting in concert with other cytokines or

neurotro-phins [4,5] Signals mediated by TGF-b are propagated by

two receptor serine/threonine kinases designated as TGF-b

type I (TbRI) and type II (TbRII) receptors [6,7] The type

II receptors comprise TbRII [8] and its alternative splice variant TbRII-B [9] The initial binding of TGF-b1 to TbRII is followed by recruitment and activation of TbRI [10] Receptor-associated Smads (R-Smads) involved in TGF-b signaling (Smad2 and Smad3) are phosphorylated

at the C-terminal SSXS-motif [11,12], interact with the common mediator Smad4 [13] and translocate to the nucleus to mediate specific transcriptional responses [14,15] Although Smad2 and Smad3 share 92% amino acid identity, they are functionally distinct A short amino acid sequence in the MAD homology 1 (MH1) domain of Smad2 is responsible for its inability to bind DNA [16,17] However, Smad3 can directly bind to a specific DNA sequence termed the Smad binding element (SBE) These distinct properties account for activation of different subsets

of target genes by either Smad2 or Smad3

Various proteins have been identified that negatively influence TGF-b signaling at different levels [18] One of these proteins, the inhibitory Smad7, mediates its antagon-istic effects by stable interaction with TbRI, thus preventing the transient contact of R-Smads with the receptor and blocking the proceeding cascade [19,20] In addition, Smad7 was shown to recruit the E3 ubiquitin ligases Smurf1 and Smurf2 to TbRI, thereby triggering degradation of the TGF-b receptor complex[21,22] Interestingly, expression

of Smad7 is rapidly induced in response to TGF-b1 [23] and therefore plays a crucial role in regulating TGF-b signaling

Correspondence to P Knaus, Physiological Chemistry II, Biocenter,

University of Wu¨rzburg, 97074 Wu¨rzburg, Germany.

Fax: + 49 931 888 4113, Tel.: + 49 931 888 4127,

E-mail: pknaus@biozentrum.uni-wuerzburg.de

Abbreviations: TGF-b, transforming growth factor-b; TbR, TGF-b

receptor type; NGF, nerve growth factor; R-Smads,

receptor-associ-ated Smads; MH, Mad homology domain; SBE, Smad binding

element; TrkA, tyrosine kinase receptor; ECD, extracellular domain;

MAPK, mitogen activated protein kinase; RSK, receptor serine/

threonine kinase; BMP, bone morphogenetic protein; GFP, green

fluorescent protein.

(Received 5 November 2003, accepted 15 January 2004)

family [26].

It is increasingly evident that signal transduction in

general does not only occur in a linear fashion but rather

comprises a complexnetwork of signaling pathways that

mutually influence their activity [27] TGF-b family

members are implicated in multiple interdependent signals

between pathways originating from receptor

serine/threo-nine kinases and receptor tyrosine kinases Cellular

responses induced by bone morphogenetic protein

(BMP) for example, can be impaired by epidermal growth

factor and hepatocyte growth factor, which lead to the

phosphorylation of Smad1 in the linker region and thus

prevent Smad1 nuclear translocation [28] Direct effects

of signaling intermediates were shown for the c-Jun

N-terminal kinase and protein kinase C c-Jun N-terminal

kinase phosphorylates Smad3 outside the SSXS-motif,

thus supporting nuclear transport of Smad3 [29] Protein

kinase C however, abrogates direct DNA binding of

Smad3 by serine phosphorylation in the MH1 domain

[30] This indicates that Smads are not restricted to

TGF-b/BMP pathways; rather, they represent a point of

convergence of various signals and their activation is a

precise contextually regulated process

Here we demonstrate that in rat pheochromocytoma

cells (PC12), NGF stimulation results in activation of the

Smad cascade This Smad activation is independent of

TGF-b1 and occurs by a mechanism which is different

from that induced by TGF-b1 in that it does not lead to

C-terminal phosphorylation of R-Smads However, NGF

rapidly triggers association between Smad3 and Smad4,

translocation to the nucleus and gene expression

NGF-mediated transcriptional activation of TGF-b responsive

reporter constructs requires the presence of functional

TrkA receptors and can be impaired by the inhibitory

Smad7

Materials and methods

Antibodies and reagents

The monoclonal antibody against TGF-b1, -b2 and -b3

(clone #1D11) was purchased from R&D Systems Details

of polyclonal antisera against Smad2 (anti-S2), Smad3

(anti-S3) and C-terminally phosphorylated forms of

Smad1 (anti-PS1) and Smad2 (anti-PS2) were published

previously [31,32] The P-Smad1 antibody shows

cross-reactivity to phosphorylated Smad3 and can thus be used

as an
anti-P-Smad3 (anti-PS3) [33] The monoclonal

antibody Smad2/3 was purchased from BD Biosciences

and the monoclonal antibody Smad4 (B-8) was obtained

from Santa Cruz Peroxidase-coupled goat anti-(rabbit

IgG) Ig was obtained from Dianova Doxycycline was

purchased from Sigma Human TGF-b1 was from R&D

Systems and mouse NGF (2.5S) from Alomone labs

(Jerusalem, Israel)

solution of 38 lgÆmL)1collagen in 0.1% (v/v) acetic acid for at least 1 h followed by thorough washing with sterile

dH2O and medium without supplements

Neutralization of TGF-b All TGF-b isoforms were neutralized by the addition of either monoclonal antibodies against TGF-b1, -b2, -b3 (20 lgÆmL)1) or a 100- or 1000-fold molar excess of the soluble extracellular domain (ECD) of TbRII-B (TbRII-B-ECD), kindly provided by J Nickel (Biocenter, University

of Wu¨rzburg, Germany)

DNA constructs Smad2 and Smad7 constructs were published previously [20,34] The Smad3 construct was kindly provided by

R Derynck (University of California at San Francisco, CA, USA) Smad1, Smad4 and Smad4 (DSAD) constructs were

a gift from M de Caestecker (National Cancer Institute, Bethesda, MD, USA) [35] The NGF receptor constructs (TrkA and TrkA-K538R) were obtained from M Chao (Skirball Institute of Biomolecular Medicine, New York,

NY, USA)

Cell culture and transient transfection PC12 cells [36] were cultured in RPMI supplemented with 10% (v/v) horse serum, 5% (v/v) fetal bovine serum and antibiotics

100 lgÆmL)1) Transient transfection of PC12 cells was performed using LipofectAMINETM (Life Technologies Inc.) according to the manufacturer’s protocol 293T cells were maintained in minimum Eagle’s medium (MEM) containing 10% (v/v) fetal bovine serum and antibiotics Transfection of 293T cells was performed using calcium phosphate–DNA coprecipitation [37] L6 rat myoblasts were cultured in DMEM with 10% (v/v) fetal bovine serum MLEC cells [38] were cultured in DMEM supplemented with 10% (v/v) fetal bovine serum and 250 lgÆmL)1 geneticin

Retroviral constructs and transduction of PC12 cells

by retroviral transfer The retroviral construct pMX-GFP-Smad3 was a kind gift from Y Henis (Tel Aviv University, Israel) [39] The construct for N-terminally HA-tagged TbRII-wt was subcloned into the retroviral vector pczCFG-EGIRT (D Lindemann, unpublished results)

tetracycline-inducible cytomegalovirus (CMV) minimal promoter [40] Cells were transduced by infection with helper-free VSV-G pseudotyped retroviruses as described previously [37] Briefly, 293T cells were cotransfected with the retroviral construct and plasmids for gag-pol and

VSV-G Twenty-four hours post-transfection, cells were

treated with 10 mMsodium-butyrate for 10 h Infection of

the target cells was performed 48 h and 72 h after

transfec-tion Because the retroviral sequences contain the gfp gene,

infected cells could be selected by FACS sorting

Luciferase reporter gene assays

For reporter gene assays, different TGF-b responsive

elements were used The p3TP-luc(+) reporter is derived

from the p3TP-luc construct [7] but contains a modified luc

gene (pSP-luc(+), Promega) The pSBE reporter [41] serves

as a readout for TGF-b as well as BMP signaling whereas

p3TP-luc(+) and (CAGA)12-luc constructs respond

speci-fically to TGF-b mediated signals [42,43]

PC12 cells were plated on collagen-coated 6-well plates

prior to cotransfection with reporter constructs, pRL-TK

(Renilla luciferase, Promega) and with the indicated receptor

or Smad constructs The total amount of DNA was kept

constant by the addition of empty vector (pcDNA3)

Twenty-four hours post-transfection, cells were starved in

medium containing 0.2% (v/v) horse serum for 4 h,

followed by stimulation with either 2 nMNGF or 200 pM

TGF-b1 for an additional 24 h Cell lysis and luciferase

measurement were performed according to manufacturer’s

instructions (Dual Luciferase Assay system, Promega) and

data were normalized using Renilla luminescence The value

of unstimulated transfected cells was set to one and all other

values were calculated accordingly

Smad-Western blotting

To investigate Smad phosphorylation, cells were seeded on

Petri dishes, starved for 4 h in medium containing 0.2%

(v/v) serum and stimulated with 2 nM NGF or 200 pM

TGF-b1 for 30 min Cells were lysed in TNE lysis-buffer

[20 mMTris/HCl pH 7.4, 150 mMNaCl, 1% (v/v) Triton

X-100, 1 mMEDTA, 1 mM phenylmethanesulfonyl

fluor-ide] containing protease inhibitors (CompleteTM, Roche)

and phosphatase inhibitors (50 mMNaF, 10 mMNa4P2O7

and 1 mM Na3VO4) Equal amounts of cell lysates were

analyzed by immunoblotting using anti-PS2, anti-Smad2,

anti-PS3 or anti-Smad3 Immunoreactive proteins were

visualized by enhanced chemiluminescence

Co-immunoprecipitation of Smad proteins

PC12 cells (5· 106) were starved in medium containing

0.2% (v/v) horse serum for 4 h Ligand stimulation was

carried out for the indicated periods of time using 2 nM

NGF Cells were lysed in a buffer containing NaCl/Pi

pH 7.4, 0.5% (v/v) Triton X-100, 1 mMEDTA,

phospha-tase inhibitors and protease inhibitors Cell lysates were

incubated with monoclonal antibody Smad2/3 or Smad4

for 2 h, followed by incubation with protein-A sepharose

beads overnight at 4C The beads were washed twice with

lysis buffer and twice with NaCl/Pi before

immunocom-plexes were eluted by boiling in SDS sample buffer for

5 min Following separation by SDS/PAGE and

electro-transfer to a nitrocellulose membrane, proteins were

immunoblotted with Smad4 or Smad2/3 antibodies as

appropriate

Nuclear and cytoplasmic fractionation PC12 cells and L6 cells were starved in low serum medium for 4 h Ligand stimulation was performed for the indicated periods of time with 2 nM NGF or 200 pM TGF-b1, respectively Control cells were stimulated with 2 nMNGF for 1 h following a 1 h treatment with cycloheximide (5 lgÆmL)1in dimethylsulfoxide) or dimethylsulfoxide only Cells were washed with NaCl/Pi, centrifuged (1000 g, 4C, 10min) and the cell pellet was then resuspended in hypotonic buffer (10 mMHepes pH 7.9, 1.5 mMMgCl2, 10 mMKCl, protease inhibitors) Cells were vortexed thoroughly and cell lysis was followed by microscopy until 90% of the cells were lysed Following centrifugation (1000 g, 4C, 10 min), the supernatant was referred to as the cytoplasmic fraction The pellet containing the nuclei was resuspended in high salt buffer [20 mMHepes pH 7.9, 25% (v/v) glycerol, 420 mM NaCl2, 1.5 mMMgCl2, 0.2 mMEDTA, protease inhibitors] Extraction of nuclear proteins was achieved by vortexing this solution thoroughly, incubating for 30 min on ice and subsequent centrifugation (25 000 g, 4C, 20 min) The supernatant was collected and represents the nuclear fraction

Nuclear translocation PC12 cells were plated on collagen-coated dishes Starvation

in a low serum medium for 4 h was followed by stimulation with 2 nMNGF for the indicated times Control cells were stimulated with 2 nM NGF for 30 min following a 1 h treatment with cycloheximide (5 lgÆmL)1in dimethylsulf-oxide) or dimethylsulfoxide only The cells were then washed with NaCl/Piand 3% (v/v) BSA, fixed in 4% (v/v) paraformaldehyde and 0.2% (v/v) TX-100 for 10 min at room temperature After washing with NaCl/Picontaining 3% (v/v) BSA, Smad2/3 staining was performed with antibodies from BD Biosciences Nuclei were stained by the addition of 1 lgÆmL)1Hoechst 33342 for 2 min The subcellular distribution of Smad2/3 was then analyzed by confocal microscopy

Results

NGF mediates the activation of Smad-dependent reporter genes independently of TGF-b

Survival of neuronal cells is described to be synergistically promoted by TGF-bs and neurotrophic factors (e.g NGF) [44,45] To examine whether NGF has the potential to modulate the Smad pathway in PC12 cells, we performed reporter gene assays using luciferase constructs containing promoter elements that are responsive to proteins of the TGF-b superfamily, i.e pSBE-luc, p3TP-luc(+) and (CAGA)12-luc [7,41,42] As indicated in Fig 1A, PC12 cells show a significant increase of transcriptional activation after stimulation with NGF on all tested Smad-dependent reporter constructs In contrast, TGF-b1 is not able to induce transcription from these reporters in PC12 cells Different approaches were chosen to exclude that this NGF-mediated transcriptional response is a secondary effect, caused for instance by NGF-triggered secretion of TGF-b1

as published previously [46] First, the reporter gene assay

was carried out in the presence of monoclonal antibodies

against TGF-b1, -b2 and -b3 to neutralize all TGF-b

isoforms (Fig 1B) Stimulation with NGF leads to a

significant increase in transcriptional activity which is not impaired by the presence of neutralizing TGF-b antibodies TGF-b, however, does not induce luciferase activity, neither

Fig 1 NGF mediates transcription from

TGF-b responsive reporter genes (A) PC12

cells were plated on collagen-coated dishes and

cotransfected with pRL-TK and pSBE-luc,

p3TP-luc(+) or (CAGA) 12 -luc Following

starvation for 4 h in medium containing 0.2%

(v/v) horse serum, cells were stimulated for

24 h with either 2 n M (50 ngÆmL)1) NGF

(black bars) or 200 p M TGF-b1 (gray bars), or

were left untreated (white bars) Cell lysates

were prepared and the luciferase activity was

measured Data were normalized as described

in Materials and methods, and error bars

represent the SD evaluated from three

inde-pendent experiments (B) Transfection and

starvation of PC12 cells was carried out as

described in (A) using pSBE-luc as the

repor-ter construct Subsequently, cells were treated

with 2 n M NGF (black bars) or 200 p M

TGF-b1 (gray bars) either in presence or in absence

of 20 lgÆmL)1TGF-b1, -b2, -b3 antibody.

After 24 h, luciferase activity was recorded

and the data were evaluated as described

above (C) L6 cells stably expressing

GFP-Smad3 were starved for 4 h followed by

treatment with either 20 lgÆmL)1

anti-(TGF-b1, -b2, -b3) (lane 2), 200 p M TGF-b1

(lane 3) or 200 p M TGF-b1 together with

20 lgÆmL)1anti-(TGF-b1, -b2, -b3) (lane 4).

Cell lysates were analysed for C-terminally

phosphorylated Smad3 (upper panel) or for

total amounts of Smad3 (lower panel) (D)

PC12 cells were cotransfected with pSBE and

pRL-TK The day after transfection, PC12

cells were starved as described previously and

stimulated with either 2 n M NGF (black bars),

200 p M TGF-b1 (gray bars) or were left

untreated (white bars) In addition, the

experiment was carried out in the absence or

presence of the soluble extracellular domain of

TbRII-B (TbRII-B-ECD) TbRII-B-ECD

was used at 100-fold and 1000-fold molar

excess as indicated Twenty-four hours after

stimulation, the supernatant of the PC12 cells

was transferred to L6 cells (lower panel) which

were equivalently transfected and starved

prior to addition of the supernatant PC12

cells (upper panel) as well as L6 cells (lower

panel) were lysed and tested for luciferase

activity Error bars represent the SD from two

independent experiments.

in the presence nor in the absence of antibodies The

neutralizing capacity of the TGF-b1, -b2 and -b3 antibodies

was verified in L6 cells stably transduced with a GFP–

Smad3 construct (Fig 1C) In the absence of TGF-b1, -b2

and -b3 antibodies, stimulation with TGF-b1 leads to

C-terminal phosphorylation of Smad3, whereas treatment

with TGF-b1 together with neutralizing antibodies impedes

the phosphorylation of Smad3 Second, we performed a

complementary experiment using the soluble ECD of

TbRII-B, a TGF-b type II receptor splice variant that

binds all three TGF-b isoforms [9] TbRII-B-ECD was

added in two different concentrations (100- and 1000-fold

molar excess) to PC12 cells transfected with the pSBE-luc

reporter in the presence of NGF or TGF-b1 (Fig 1D,

upper panel) The TbRII-B-ECD did not abolish

NGF-mediated Smad activation in PC12 cells Next, we harvested

the supernatant of PC12 cells treated in this way and placed

it on the TGF-b sensitive L6 myoblast cell line that was

equally transfected with pSBE-luc (Fig 1D, lower panel)

As there was no detectable increase in luciferase activity in

TGF-b-sensitive L6 cells that were treated with the

super-natant from NGF stimulated PC12 cells (Fig 1D, lower

panel, bar 2), we can conclude that NGF treatment of PC12

cells does not lead to the production of active TGF-b In

contrast, the TGF-b-treated PC12 cell supernatant results in

reporter activation in L6 cells (Fig 1D, lower panel, bar 3)

This activation is almost completely blocked by a 1000-fold

molar excess of TbRII-B-ECD (Fig 1D, lower panel, bar 9)

Furthermore, we measured the amount of TGF-b that

is produced in response to NGF and checked whether this

TGF-b is present in an active or latent form The

quantification of TGF-b was performed by using the

MLEC cell line, which stably expresses a luciferase

reporter gene under the control of a truncated PAI-1

promoter [38] We found that NGF leads to production

of only marginal amounts ( 1.1 pM) of active TGF-b

(data not shown) This is in accordance with the results

presented above (Fig 1D, lower panel) Taken together,

all approaches clearly demonstrate that NGF mediates the

activation of the Smad pathway independently of the

TGF-b ligand

TGF-b resistance is due to limiting amounts of TbRII

in PC12 cells

The expression of TGF-b receptors, particularly TbRII, in

PC12 cells is controversially discussed in the literature

[47,48] Phosphorylation studies and reporter gene assays

demonstrate that TbRII represents the limiting component

of TGF-b1 signaling in PC12 cells Smad2 phosphorylation

was analysed in parental PC12 cells and in stable PC12 cells

expressing TbRII-wt under the control of a

doxycycline-inducible promoter Figure 2A shows that treatment with

TGF-b1 results in C-terminal phosphorylation only in

PC12 cells that ectopically express TbRII but not in

parental PC12 cells However, in response to NGF, there

is no phosphorylation of Smad2 at the SSXS-motif

Luciferase assays also show that transient transfection of

PC12 cells with TbRII but not TbRI constructs leads to

increased responsiveness to TGF-b1 (Fig 2B), again

indi-cating that TbRII is the limiting component of TGF-b1

signaling in PC12 cells

Mechanism of Smad activation by NGF The mechanism of Smad reporter activation was investigated

by using luciferase constructs that allowed us to distinguish between signals originating from different R-Smads, i.e pSBE-luc, p3TP(+)-luc and (CAGA)12-luc [7,41,42] (see Materials and methods) Using p3TP-luc(+), NGF-induced reporter activation was investigated after ectopic expression

of various R-Smad constructs (Fig 3A) From all R-Smads tested, Smad3 shows the most prominent induction of transcription after NGF stimulation Similar results were obtained with the (CAGA)12-luc reporter (data not shown) Given that in TGF-b signaling, phosphorylation of the C-terminal serine residues (SSXS) is essential for dissoci-ation from the type I receptor and for heteromeric complex formation with Smad4 [34,49], we investigated C-terminal phosphorylation in response to TGF-b1 as well as NGF The phosphorylation pattern of Smad3 resulting from stimulation with either TGF-b or NGF was analysed in

Fig 2 PC12 cells express low levels of endogenous TbRII (A) C-terminal phosphorylation of Smad2 was investigated in either par-ental PC12 cells (lanes 1–3) or PC12 cells stably expressing TbRII-wt (lanes 4–6) Cells were kept in media containing 1 lgÆmL)1doxycycline for 3 days to induce expression of TbRII-wt Following starvation, cells were treated either with 2 n M NGF (lanes 2 and 5) or with 200 p M

TGF-b1 (lanes 3 and 6) or were left untreated (lanes 1 and 4) Cell lysates were examined for Smad2 phosphorylation by immunoblotting using an antibody raised against the phosphorylated SSXS-motif of Smad2 (anti-PS2) (B) PC12 cells were plated on collagen-coated dishes and were cotransfected with pRL-TK, pSBE and the indicated expression constructs Luciferase activity was measured after starva-tion and treatment for 24 h with control medium (white bars) or with medium containing 200 p M TGF-b1 (black bars).

TGF-b-responsive L6 rat myoblasts and in PC12 cells,

which were both stably transduced with a retroviral GFP–

Smad3 construct [39] Immunoblotting using an antibody

that specifically recognizes C-terminally phosphorylated

Smad3 [32,33] revealed that phosphorylation of both the

heterologous GFP–Smad3 and the endogenous Smad3

protein occurs in L6 cells after treatment with TGF-b1, but

not with NGF (Fig 3B) In PC12 cells, however, we did not

detect any phosphorylation at the SSXS-motif of Smad3;

neither in response to NGF nor in response to TGF-b1 The

low amounts of TbRII expressed in PC12 cells can account

for the lack of phosphorylation upon TGF-b stimulation

This suggests that NGF-mediated Smad activation occurs

independently from

SSXS-motif

Involvement of Smad4 in NGF-triggered activation

of the Smad signaling cascade

In TGF-b1-induced signaling, R-Smads form heteromeric

complexes with Smad4 following the activation by TbRI

Reporter gene assays using the p3TP(+)-luc construct were performed to determine whether NGF-mediated activation

of Smad response elements is also Smad4-dependent PC12 cells were transfected with Smad3, Smad4 or a functionally inactive Smad4 variant – Smad4(DSAD) – either alone or

in the indicated combinations Smad4(DSAD) lacks amino acids 274–321 which encode the Smad activation domain (SAD) [35,50] Figure 4A demonstrates that ectopic expres-sion of Smad3 results in efficient transcriptional activation

of Smad-dependent reporter genes, whereas neither Smad4 nor the mutant Smad4 show an effect on luciferase induction when expressed alone Coexpression of Smad3 and Smad4, however, enhances the Smad3 effect In

Fig 3 Mode of Smad activation by NGF (A) Induction of specific

R-Smads was investigated using the p3TP-luc(+) reporter PC12 cells

were transiently transfected with the indicated Smad constructs

Fol-lowing starvation, cells were left untreated (white bars) or were

stimu-lated with 2 n M NGF (black bars) Data were normalized as described

in Materials and methods Error bars were calculated from three

independent measurements (B) Phosphorylation of Smad3 was tested

in TGF-b responsive L6 rat myoblasts (lanes 1–4) and PC12 cells (lanes

5–8) that were transduced with pMX-GFP-Smad3 (lanes 2–4 and 6–8)

using retroviral transfer Cells were starved for 4 h, followed by

sti-mulation with 2 n M NGF (lanes 3 and 7) or 200 p M TGF-b1 (lanes 4

and 8) and cell lysates were analysed for C-terminally phosphorylated

Smad3 (upper panel) or for total amounts of Smad3 (lower panel).

Fig 4 Role of Smad4 in NGF-triggered activation of the Smad path-way (A) The role of functional Smad4 was investigated by reporter gene assays using p3TP-luc(+) PC12 cells were cotransfected with the luciferase constructs and with different combinations of Smad3 and Smad4 variants as indicated Following starvation and treatment for

24 h with either 2 n M NGF (black bars) or control medium (white bars), cell lysates were prepared and used for luminescence measure-ment (B) NGF-induced association of Smad3 and Smad4 was assessed by coimmunoprecipitation studies PC12 cells were starved and treated with 2 n M NGF for the indicated periods of time Smad3 was immunoprecipitated from cell lysates using the anti-Smad2/3 Ig and analyzed by Western blotting using anti-Smad4 Igs Additionally, the experiment was repeated with immunoprecipitation of Smad4 followed by detection of Smad3 by Western blotting (upper panels) Total amounts of protein were verified by immunoblotting proteins of total lysates using the appropriate antibodies (lower panels).

contrast, coexpression of the functionally inactive Smad4

variant – Smad4(DSAD) – largely prevents

Smad3-mediated reporter gene activation These results were also

confirmed by using the (CAGA)12-luc reporter (data not

shown)

Furthermore, co-immunoprecipitation studies confirmed

that NGF stimulation of PC12 cells leads to interaction

between Smad3 and Smad4 (Fig 4B) Whereas in the

absence of ligand there is no heteromeric

complexforma-tion, NGF treatment triggers association of the Smad

proteins within 30 min, indicating that NGF directly

activates Smad signaling

NGF stimulation rapidly initiates nuclear accumulation

of Smad3

To assess whether Smad3 translocates to the nucleus in

response to NGF treatment, nuclear extracts were

investi-gated for the content of Smad3 protein and the cellular

distribution of Smad3

Nuclear extracts were prepared from L6 rat myoblasts that

were stimulated with TGF-b1 (Fig 5, upper panel) and

from PC12 cells at several time points after NGF treatment

(Fig 5

7 , middle panel) In both cell lines, Smad3 can be

detected in the nuclear fraction after 5 min of ligand

stimulation and the amount of nuclear Smad3 increases

with prolonged growth factor treatment, reaching a

maxi-mum after 15–30 min of stimulation with TGF-b1 in L6 cells and after 1 h of stimulation with NGF in PC12 cells Referring to Smad4, there are significant levels of protein

in the nucleus of both cell lines already in the absence of ligand Cycloheximide treatment for 1 h prior to the addition of NGF indicated that Smad3 was directly stimulated by NGF for nuclear translocation, with no

de novoprotein synthesis required (Fig 5, lower panel) Next, we investigated the nuclear transport of Smad3 in PC12 cells by confocal microscopy Staining with Hoe-chst 33342 was performed to visualize the nuclei Without NGF treatment, Smad3 is distributed throughout the whole cell (Fig 6, first row) Stimulation with NGF for 30 min shows a strong decrease of cytoplasmic Smad3 staining and accumulation of Smad3 in the nucleus, and NGF treatment for 3 h results in a solely nuclear localization of Smad3 (Fig 6, second and third rows, respectively) Cycloheximide treatment indicated that Smad3 was immediately stimulated

by NGF for nuclear translocation; this process does not require de novo protein synthesis (Fig 6, rows 4 and 5) This agrees with the data that we obtained by cellular fraction-ation of parental PC12 cells (Fig 5)

NGF-mediated activation of Smad reporter constructs can be efficiently abrogated by either kinase-dead TrkA receptors or by the inhibitory Smad7 protein

To assess the involvement of TrkA receptors in the activation of the Smad pathway, different TrkA variants were tested in reporter gene assays using the pSBE-luc construct (Fig 7) In PC12 cells transfected with wild-type TrkA (TrkA-wt), the basal level of luciferase activity is elevated already but the signal can be potently enhanced by stimulation with NGF Transfection of the TrkA variant (TrkA–K538A) that carries a mutation resulting in the inactivation of the tyrosine kinase activity causes a signifi-cant reduction of responsiveness An even stronger inhi-bitory effect can be observed after cotransfection of the wild-type TrkA receptor together with Smad7 The antago-nizing impact of Smad7 becomes additionally apparent by the strong inhibitory effect on endogenous signaling that is elicited following expression of ectopic Smad7 (Fig 7, lanes

2 and 4) These results suggest functional TrkA receptors

to be necessary for NGF-mediated activation of Smad-dependent reporter genes and demonstrate the inhibitory role of Smad7 on this NGF-mediated effect

Discussion

Originally, Smad proteins were exclusively attributed to pathways activated by TGF-b family members but it becomes increasingly evident that multiple signaling cas-cades originating from other receptor systems are involved

in modulating Smad signaling [14,27,30,51,52] In the present work, we demonstrate that in PC12 cells NGF-stimulated signaling via the TrkA receptor leads to activa-tion of the Smad pathway NGF-mediated Smad activaactiva-tion

is independent of TGF-b ligand and occurs by a mechanism which is different from that induced by TGF-b

PC12 rat pheochromocytoma cells represent a widely used model system to investigate neuronal differentiation that is initiated following stimulation with NGF [36]

Fig 5.

10 NGF induces nuclear accumulation of Smad3 L6 cells and PC12

cells were starved for 4 h in medium containing 0.2% (v/v) fetal bovine

serum or horse serum, respectively, and nuclear fractions were

pre-pared at various time points after exposure to either 200 p M TGF-b1

or 2 n M NGF as indicated Proteins contained in the nuclear fraction

were subjected to SDS/PAGE, electrotransferred to nitrocellulose and

immunoblotted with monoclonal antibodies to Smad2/3 or Smad4.

The purity of the cytoplasmic and nuclear fractions was confirmed

by immunoblotting with an anti-lamin serum Control cells were

stimulated with 2 n M NGF for 1 h following a 1 h treatment with

cycloheximide (lower panel) Nuclear fractions were probed with

anti-Smad2/3 and subsequently with anti-lamin.

Although TGF-bs do not promote survival or

differenti-ation of neuronal populdifferenti-ations on their own, they elicit a

neurotrophic potential if they are applied together with

other cytokines (GDNF, GDF-5) or neurotrophins (NGF,

NT-3) [4,5], suggesting that the signaling cascades of

TGF-bs and neurotrophins are somehow interdependent However, ectopic expression of inhibitors of the TGF-b/ Smad pathway such as Smad7 or neutralizing TGF-b antibodies did not prevent NGF-induced neurite formation (data not shown), suggesting that the Smad pathway that

Fig 6 Smad3 nuclear translocation in PC12 cells using confocal microscopy Cells were plated on collagen-coated dishes, starved in medium containing 0.2% (v/v) horse serum for 4 h and stimulated for with 2 n M NGF for 30 min (second row), 3 h (third row) or were left untreated (first row) Control cells were stimulated with 2 n M NGF for 30 min following a 1 h treatment with cyclohex imide (CHX; fourth row) or dimethyl-sulfoxide (DMSO; fifth row) Cells were fixed, nuclei were stained with Hoechst 33342 for 2 min and the cells were analysed by confocal microscopy The projection of multiple sections is seen on the left for each panel to visualize the morphology of the cells The middle row shows staining of Smad3 and on the right an overlay of Smad3 and Hoechst staining is seen.

can be activated by NGF is mainly important for other

cellular responses

PC12 cells show transcriptional activation of

TGF-b-responsive reporter genes upon NGF but not TGF-b1

stimulation (Fig 1) Low amounts of TbRII expressed in

these cells can account for TbRII being the limiting factor

for proper TGF-b1 signaling, which is in accordance with

results showing that ectopic expression of TbRII restores

TGF-b responsiveness (Fig 2) As an earlier report shows

upregulation of TGF-b by NGF [46], we investigated

whether NGF-triggered Smad activation is caused by

autocrine action of TGF-b Considering the limiting

amount of TbRII discussed above, PC12 cells are equally

resistant to signals evoked by either exogenous or autocrine

TGF-b Furthermore, even if all TGF-b isoforms are

neutralized by the addition of antibodies or the soluble

extracellular domain of TbRII-B (TbRII-B-ECD), NGF is

still capable of activating Smad-dependent reporter genes

(Fig 1B,D) Concerning the amounts of secreted TGF-b, we

found that besides latent (i.e biologically inactive) TGF-b,

only marginal amounts of active TGF-b can be detected

in the supernatant of NGF-treated PC12 cells TGF-b is

synthesized as a precursor proprotein that is cleaved during

secretion However, the mature TGF-b remains associated

with its propeptide thereby forming a latent complexuntil

activation [53] Because TGF-b activation takes place in the

extracellular compartment, intracellular signaling events

initiated by autocrine TGF-b can be excluded Taken

together, the effects of NGF on the Smad pathway are

independent of TGF-b ligand

To characterize the point of convergence and the mode of

Smad activation, C-terminal phosphorylation, heteromeric

complexformation with Smad4 and nuclear translocation

of R-Smads was investigated Comparison of signaling

through different R-Smad proteins revealed that Smad3

mediates the most potent activation (Fig 3A) As Smad2

contains an additional exon in the MH1 domain that is not

present in Smad3, it lacks the capacity to bind directly to

DNA [16,17], which might explain the different behavior of

the two TGF-b-activated Smads in response to NGF

Although treatment of PC12 cells with NGF results in

the activation of Smad-dependent reporter genes, the

preceeding signaling events are not identical to those that are known from TGF-b signaling as NGF does not induce phosphorylation of the C-terminal SSXS-motif (Fig 3B) Recent publications describe alternative mechanisms of Smad activation which are likewise independent of C-terminal phosphorylation: c-Jun N-terminal kinase was shown to be rapidly activated by TGF-b stimulation in a Smad-independent manner and to cause initial phosphory-lation of Smad3 at sites other than the SSXS motif This modification in turn promotes TbRI-dependent C-terminal phosphorylation of Smad3 [29] The mitogen-activated protein kinase kinase kinase was shown to trigger phosphorylation outside the C-terminal motif, which results in enhanced transcriptional activity of Smad2 in endothelial cells [54] These examples support our findings that activation of Smad proteins can occur independently

of C-terminal phosphorylation Besides direct phosphory-lation events, NGF potentially triggers other modifications

of R-Smads resulting in Smad nuclear translocation and transcriptional activation

As the NGF-initiated processes were shown to be dependent on functional Smad4 proteins (Fig 4A) and to lead to heteromeric complexformation between Smad3 and Smad4 (Fig 4B), we assume that the presence of the SSXS-motif is crucial to allow interaction between R-Smads and Smad4, even if the C-terminal serines are not phosphorylated Recent reports show that phosphory-lation of the SSXS-motif enhances heteromeric complex formation and stabilizes the assembly of the Smad homo-and hetero-oligomers Nevertheless, Smad3 homo-and Smad4 were shown to heterotrimerize in the absence of phos-phorylation [55,56]

Thus it remains to be elucidated whether phosphorylation

of other residues or different modifications causes the same

or even a distinct oligomerization pattern of Smads Nuclear translocation of Smad3 could be confirmed by the appearance of the Smad3 protein in nuclear extracts following NGF stimulation (Fig 5) and by investigation of the cellular distribution of Smad3 by confocal microscopy (Fig 6) The observation, that Smad3 appears in the nuclear fraction of PC12 cells after only 5 min of NGF stimulation and reaches a maximum after 1 h hints of a direct effect of NGF on Smad proteins This is also confirmed by cycloheximide treatment (Figs 5 and 6), demonstrating that no de novo protein synthesis is required for NGF-mediated nuclear translocation of Smad3 Whereas Smad3 is not present in the nucleus in the absence

of NGF, Smad4 can be found in the nucleus regardless of ligand stimulation This is in accordance with the findings that Smad4 continuously shuttles between the cytoplasm and the nucleus [57] R-Smads, however, underlie cytoplas-mic retention in the absence of ligand due to their interaction with Smad anchor for receptor activation

8[58,59] or microtubules [60] Confocal microscopy studies additionally confirm the NGF-induced nuclear transloca-tion of Smad3 within 30 min (Fig 6) While untreated cells reveal Smad3 staining throughout the whole cell, stimula-tion with NGF for 30–60 min provokes complete nuclear accumulation of Smad3

To define the role of the high-affinity NGF receptor, TrkA, we ectopically expressed functionally inactive NGF receptors in PC12 cells and found that the tyrosine kinase

Fig 7 Functional TrkA receptors are essential for the activation of

Smad-dependent reporters by NGFPC12 Cells were transfected with

pSBE-luc and the indicated DNA constructs Total amounts of DNA

were kept constant by the addition of empty vector (pcDNA3) The

experiment was carried out as described in Fig 1 and error bars are

calculated from three independent measurements.

signal abrogation have been previously described Smad7 is

capable of blocking Smad signaling at the receptor level by

interaction with activated TbRI [19] or by recruiting the

E3 ubiquitin ligases Smurf1 and Smurf2 to the receptors,

resulting in enhanced turnover of TGF-b receptors [21,22]

Furthermore, Smad7 was shown to interfere with signal

transduction by interaction with cytoplasmic proteins such

as TAB1 [61] or mitogen-activated protein kinase kinase

kinase [54] These distinct antagonizing mechanisms of

Smad7 open up the question whether Smad7 blocks

NGF-induced Smad signaling at the receptor level or by

interaction with other proteins As dominant-negative

TGF-b receptor mutants did not block NGF-induced

Smad activation (data not shown), they seem to be

dispensible for NGF-mediated signals, and therefore a

mechanism that involves Smad7 interaction with

cytoplas-mic proteins is favored

The Alk7 type I receptor is highly similar in its

intracellular domain to TbRI and the constitutively active

form of Alk7 was shown also to induce Smad2/3

phos-phorylation Studies in PC12 cells have indicated that Alk7

signaling augments differentiation response to NGF [48]

The ligand for this receptor, however, is presently unknown

In conclusion, we describe here that NGF stimulation of

PC12 cells results in activation of the Smad pathway

independently of TGF-b1 This activation is direct and

results in nuclear translocation of Smad3 within only

30 min of NGF treatment Binding of NGF to its

high-affinity receptor TrkA induces activation of Smad3,

heteromeric complexformation with Smad4, nuclear

trans-location and transcriptional activation However, unlike

TGF-b1 signaling, this process does not include

phosphory-lation of the C-terminal SSXS-motif of the R-Smad Based

on the diverse mechanism of Smad activation by either

TGF-b1 or NGF, specific subsets of target genes might be

induced

The potential of NGF to activate the Smad pathway

independently of TGF-b might be of special importance in

regulating the expression of genes that are essential for

the development and function of neuronal cells or other

NGF-sensitive cells, in particular those which are TGF-b

resistant

Acknowledgements

R Derynck, M de Caestecker and M Chao are gratefully

acknow-ledged for expression vectors and D Lindemann, Y Henis and X Liu

for retroviral vector constructs We thank F Neubauer for generating

the p3TP-luc(+) construct and J Nickel for providing the

TbRII-B-ECD We are grateful to J Fey for preparation of collagen and to

Y Kehl for excellent technical assistance We also acknowledge

S Hassel, R Scha¨fer and M Sammar for helpful discussions This

work was supported by the Deutsche Forschungsgemeinschaft (DFG)

grant Kn332/3–2 to P Knaus and EEC, Project 171R

ERB-FMRXCT980216 to P ten Dijke M Lutz was supported by GK 181.

2105–2124.

4 Krieglstein, K., Henheik, P., Farkas, L., Jaszai, J., Galter, D., Krohn, K & Unsicker, K (1998) Glial cell line-derived neuro-trophic factor requires transforming growth factor-b for exerting its full neurotrophic potential on peripheral and CNS neurons.

J Neurosci 18, 9822–9834.

5 Bo¨ttner, M., Krieglstein, K & Unsicker, K (2000) The trans-forming growth factor-bs: Structure, signaling, and roles in nervous system development and functions J Neurochem 75, 2227–2240.

6 ten Dijke, P., Miyazono, K & Heldin, C.-H (1996) Signaling via hetero-oligomeric complexes of type I and type II serine/threonine kinase receptors Curr Opin Cell Biol 8, 139–145.

7 Wrana, J.L., Attisano, L., Carcamo, J., Zentella, A., Doody, J., Laiho, M., Wang, X.F & Massague´, J (1992) TGF-b signals through a heteromeric protein kinase receptor complex Cell 71, 1003–1014.

8 Lin, H.Y., Wang, X.F., Ng-Eaton, E., Weinberg, R.A & Lodish, H.F (1992) Expression cloning of the TGF-b type II receptor, a functional transmembrane serine/threonine kinase Cell 68, 775–785.

9 Rotzer, D., Roth, M., Lutz, M., Lindemann, D., Sebald, W & Knaus, P (2001) Type III TGF-b receptor-independent signalling

of TGF-b2 via TbRII- B, an alternatively spliced TGF-b type II receptor EMBO J 20, 480–490.

10 Wrana, J.L., Attisano, L., Wieser, R., Ventura, F & Massague´, J (1994) Mechanism of activation of the TGF-b receptor Nature

370, 341–347.

11 Heldin, C.-H., Miyazono, K & ten Dijke, P (1997) TGF-b sig-nalling from cell membrane to nucleus through Smad proteins Nature 390, 465–471.

12 Massague´, J & Chen, Y.G (2000) Controlling TGF-b signaling Genes Dev 14, 627–644.

13 Lagna, G., Hata, A., Hemmati-Brivanlou, A & Massague´, J (1996) Partnership between DPC4 and Smad proteins in TGF-b signalling pathways Nature 383, 832–836.

14 Zhang, Y & Derynck, R (1999) Regulation of Smad signalling by protein associations and signalling crosstalk Trends Cell Biol 9, 274–279.

15 Itoh, S., Itoh, F., Goumas, M.-J & ten Dijke, P (2000) Signaling

of transforming growth factor-b family members through Smad proteins Eur J Biochem 267, 6954–6967.

16 Dennler, S., Huet, S & Gauthier, J.M (1999) A short amino-acid sequence in MH1 domain is responsible for functional differences between Smad2 and Smad3 Oncogene 18, 1643–1648.

17 Yagi, K., Goto, D., Hamamoto, T., Takenoshita, S., Kato, M & Miyazono, K (1999) Alternatively spliced variant of Smad2 lacking exon 3 Comparison with wild-type Smad2 and Smad3.

J Biol Chem 274, 703–709.

18 ten Dijke, P., Miyazono, K & Heldin, C.-H (2000) Signaling inputs converge on nuclear effectors in TGF-b signaling Trends Biochem Sci 25, 64–70.

19 Hayashi, H., Abdollah, S., Qiu, Y., Cai, J., Xu, Y.Y., Grinnell, B.W., Richardson, M.A., Topper, J.N., Gimbrone, M.A Jr, Wrana, J.L & Falb, D (1997) The MAD-related protein Smad7 associates with the TGF-b receptor and functions as an antagonist

of TGF-b signaling Cell 89, 1165–1173.

20 Nakao, A., Afrakhte, M., More´n, A., Nakayama, T., Christian, J.L., Heuchel, R., Itoh, S., Kawabata, M., Heldin, N.E., Heldin,