Combination of engineered neural cell adhesion molecules and GDF-5 for improved neurite extension in nerve guide concepts

ARTICLE IN PRESS

Biomaterials 27 (2006) 3432–3440 www.elsevier.com/locate/biomaterials

Combination of engineered neural cell adhesion molecules and GDF-5 for improved neurite extension in nerve guide concepts Marc Niere, Bettina Braun, Rea Gass, Sabine Sturany, Hansju¨rgen Volkmer Naturwissenschaftliches und Medizinisches Institut an der Universita¨t Tu¨bingen, Markwiesenstr. 55, D-72770 Reutlingen, Germany Received 11 October 2005; accepted 24 January 2006 Available online 23 February 2006

Abstract Current therapeutical approaches for the treatment of severe lesions in the peripheral nervous system rely on the use of autologous tissue or the body’s own Schwann cells. However, these approaches are limited and alternative strategies for peripheral nerve regeneration are required. Here we evaluate combinations of a variety of neuronal regeneration factors including engineered cell adhesion molecules and growth factors in embryonic model neurons to test the possible improvement of artificial nerve guides by cooperative mechanisms. Cell adhesion molecules L1 and neurofascin synergistically promote neurite elongation. The outgrowth promoting properties of both proteins can be combined and further increased within one chimeric protein. Addition of growth and differentiation factor 5 (GDF-5) further enhances neurite outgrowth in a substrate-independent manner. This effect is not due to a protective mode of action of GDF-5 against pro-apoptotic stimuli. Consequently, the study supports the idea that different modes of action of pro-regenerative factors may contribute synergistically to neurite outgrowth and emphasizes the applicability of combinations of proteins specifically involved in development of the nervous system for therapeutical approaches. r 2006 Elsevier Ltd. All rights reserved. Keywords: Cell adhesion; Nerve guide; Nerve tissue engineering; Recombinant protein

1. Introduction In contrast to the situation in the central nervous system, neurons of the peripheral nervous system have an intrinsic potential to regenerate. Axons can regenerate across smaller lesions after suturing the ends of the proximal and the distal stump [1]. However, after severe lesion of the peripheral nervous system other therapeutical approaches to promote regeneration are required [1]. A well-established method to bridge larger gaps is the use of nerve conduits composed of autologous living transplants preferentially isolated from sensory nerves like the Nervus suralis [2]. But suchlike autologous transplants require a second lesion to obtain the required material and the availability is limited [1,3]. Therefore, strategies are to be developed to replace autologous transplants by implantable nerve guides. Cell Corresponding author. Tel.: +49 7121 5153044; fax: +49 7121 5153016. E-mail address: [email protected] (H. Volkmer).

0142-9612/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2006.01.037

adhesion molecules (CAMs) and/or neurotrophic factors embedded in biocompatible scaffolds are currently discussed as promising concepts [4–6]. CAMs of the immunoglobulin superfamily (IgSF) including L1/NgCAM, neurofascin, axonin-1/TAG1 and F11/contactin are interesting candidates. These proteins are expressed in the course of the development of the nervous system and are found on neurons or Schwann cells [7–10]. During embryogenesis they serve as guidance cues for outgrowing neurons, mediate neuronal differentiation, and promote neuronal survival. Their ectodomains, which can readily be expressed as secreted IgFc-fusion proteins in eukaryotic cells, act as permissive substrates for neurite outgrowth [11]. Fc-fusion protein of L1 has already been evaluated in vitro [12,13]. In an animal model, L1-Fc enhanced axonal regeneration and myelination of transected optic nerve [4]. Amongst soluble recombinant factors, nerve growth factor (NGF), glial cell line-derived neurotrophic factor (GDNF), brain-derived neurotrophic factor (BDNF), fibroblast growth factor 1 (FGF-1), ciliary neurotrophic factor (CNTF) and growth and differentiation factor 5

ARTICLE IN PRESS M. Niere et al. / Biomaterials 27 (2006) 3432–3440

(GDF-5) have already been successfully applied for regeneration in vivo [14–16]. Within the superfamily of transforming growth factor b (TGF-b)-like molecules, GDF-5 is a member of the subfamily of bone morphogenetic proteins (BMP) [17,18]. GDF-5 was characterized in several reports as neurotrophic factor, which promotes survival of dopaminergic neurons and astroglial cells of the central as well as dorsal root ganglia (DRG) cells of the peripheral nerve system [19–21]. Furthermore, administration of GDF-5 was shown to rescue dopaminergic neurons in a rat lesion model for Parkinson’s disease [22]. Although recombinant factors have yet been successfully employed for peripheral nerve regeneration, several problems are still prevailing, which require conceptual improvements. For example, the simulation of a complex interaction of factors inducing neurite outgrowth as it is found in a natural environment might be an important step towards the replacement of grafted nerve tissue. To establish such a strategy, optimal mixtures of different recombinant neural factors need to be evaluated. In this study we have determined the efficiency of combinations of neural CAMs along with GDF-5 to promote neurite outgrowth using NGF-dependent neurons of embryonal chick DRG as a model system. Adding soluble GDF-5 to neurons grown on immobilized CAMs as substrate enhances neurite elongation in a substrateindependent manner. This outgrowth-promoting effect is not due to a protective function of GDF-5 in response to apoptotic stimuli. A synergistic effect on neurite outgrowth is observed on a substrate mixture of L1 and neurofascin, whose outgrowth-promoting capabilities can be combined within one molecule and further enhanced by addition of GDF-5. 2. Material and methods Vector construction: The generation of a vector encoding soluble neurofascin (isoform NF15) fused to IgFc is described elsewhere [11]. For expression of soluble axonin, L1 and chimeric L1/neurofascin proteins as IgFc-fusion proteins the cDNA sequences encoding the ectodomains of the proteins (Fig. 1A) were amplified and cloned into pIG2 vector. Production and purification of recombinant proteins: Stably transfected, CAM expressing monoclonal HEK293 cell lines (293-axonin-1-Fc; 293L1-Fc; 293-NF15-Fc) were generated for the production of Fc-fusion proteins in miniaturized bioreactors (CELLine CL 350, Integra Biosciences). Expression of the chimeric L1/neurofascin fusion molecules was done with the help of the FreeStyleTM 293 Expression System (Invitrogen). Proteins were purified as described previously [11]. Isolation of DRG neurons: DRG were isolated from embryonic chick (day 9) and collected in Ca2+ and Mg2+ free Hank’s balanced salt solution (HBSS) following incubation in 0.08% trypsin/PBS for 15 min. Cells were dissociated and washed once with HBSS and afterwards resuspended in culture medium (DMEM containing 1
N2 supplements [Invitrogen], 2 mM glutamine [PAA Laboratories], 100 U/ml penicillin and 100 mg/ml streptomycin [PAA Laboratories]). Neurite outgrowth in the presence of NGF (25 ng/ml) was quantified as described previously [11]. Cell viability assays: Dissociated DRG neurons were seeded on polyethyleneimine (PEI)-coated coverslips or 96 well plates in NGF (25 ng/ml, Roche Biochemicals) supplemented culture medium for 24 h before treatment with 400 nM staurosporine. For quantification of intact nuclei, neurons were stained with Hoechst 33342 (1 mg/ml) for micro-

3433

scopical evaluation of intact nuclei. DNA fragmentation after 8 h of staurosporine treatment was analysed by TUNEL-assay using In situ Cell Death Detection Kit, TMR red (Roche Biochemicals). Determination of cell viability was done after addition of alamar BlueTM (BioSource International). For quantification of caspase-3 activity in staurosporinetreated (5 h) DRG neurons, cells were permeabilized in 50 ml lysis buffer (20 mM Hepes pH 7.4; 84 mM KCl; 10 mM MgCl2; 0.2 mM EDTA; 0.2 mM EGTA; 0.5% Triton X-100; 100 M DTT; protease inhibitor cocktail [Roche Biochemicals]) following addition of 50 mM caspase substrate AcDEVD-amc (Biomol) dissolved in 150 ml reaction buffer (50 mM Hepes pH 7.4; 2 mM EDTA; 0.2 mM EGTA; 10% sucrose; 0.1% Triton X-100; 3 mM DTT). Increase in fluorescence intensity at 460 nm was measured over a period of 2 h using a microplate reader.

3. Results 3.1. Induction of neurite elongation by individual substrate CAMs If applied as a neuronal substrate, CAMs including L1, neurofascin, axonin-1 and F11 were shown to promote neurite extension. While the extracellular domains of individual CAMs have already been tested, a systematic comparison of different CAMs for induction of neurite outgrowth remained elusive. We therefore examined the potency of a panel of CAMs to stimulate neurite outgrowth in vitro. CAMs were expressed as recombinant IgFc-fusion molecules. Beyond correct protein folding, eukaryotic expression ensures the formation of the highly complex glycosylation pattern of the extracellular domains of neural CAMs. Proteins isolated from cell culture supernatants of stably transfected HEK293 cells were submitted to SDS–PAGE. Silver staining of the gels confirmed a high degree of purity and the correct molecular weight of the CAMs (Fig. 1B, lanes 1–3). In Western blot analyses the proteins were recognized by antibodies to the Fc fragment of human IgG1 and by CAM-specific antibodies, respectively (Fig. 1C, lanes 1, 2, 5, 8, and 9). In a first set of experiments, CAMs L1, neurofascin, axonin-1, and F11 were supplied as immobilized substrates for outgrowing sensory neurons and compared with neurite outgrowth induced by laminin (Fig. 1D). With the exception of F11 (data not shown) all substrates induced neurite outgrowth. Inspection of neurite elongation induced by the CAMs revealed comparable outgrowth promoting capabilities of L1 and neurofascin (median neurite lengths: 123.1 and 136.2 mm, respectively) when coated at a concentration of 25 mg/ml, while of axonin-1induced neurite outgrowth to a lesser extent even at higher concentrations of this protein (median neurite length: 95.7 mm after coating of 50 mg/ml). Therefore, out of the neural CAMs tested so far, L1 and neurofascin were the most potent promoters of neurite elongation. 3.2. Neurite induction by combinations of different CAMs Neurite outgrowth presumably relies on the parallel action of many different environmental factors encountered

ARTICLE IN PRESS M. Niere et al. / Biomaterials 27 (2006) 3432–3440

3434 Ig-like domains FNIII-like domains IgG1-Fc

alternatively spliced exons

1

2

3

4

5

207 kD

117 kD 95 kD

1

3

4

5

6

L1(Ig1-6) NF15ecto-IgFc

L1-IgFc

NF15-IgFc 2

L1(Ig1-6) NF(Ig1-6)-IgFc

(A)

axonin-1-IgFc

49 kD

7

8

(B) 100

9

80

207 kD

laminin neurofascin L1 axonin-1

117 kD

%

60 40 20

95 kD 0 0

(C)

(D)

100 200 300 400 500 600 700 800 900 1000

neurite length [µm]

Fig. 1. Molecular architecture, purity, specificity, and functionality of recombinant CAM-Fc-fusion proteins used throughout this study. (A) Schematic representation of CAM-Fc-fusion proteins. (B) Silver-stained 7% SDS–PAG after electrophoresis of each 1 mg of purified protein (lane 1, axonin-1; lane 2, L1; lane 3, NF15; lane 4, L1[Ig1-6]-NF15ecto; lane 5, L1[Ig1-6]-NF[Ig1-6]). (C) Western blot analysis of purified CAM-Fc-fusion proteins after electrophoresis in a 7% PAG (lane 1, axonin-1; lane 2, NF15; lane 3, L1[Ig1-6]-NF15ecto; lane 4, L1[Ig1-6]-NF[Ig1-6]; lane 5, L1; lane 6, L1[Ig1-6]NF15ecto; lane 7, L1[Ig1-6]-NF[Ig1-6]; lane 8, NF15; lane 9, axonin-1). Detection was done using polyclonal antibody specific for axonin-1 (lane 1), polyclonal antibody specific for neurofascin (lanes 2–4), and polyclonal antibody specific for human IgFc (lanes 5–9), and alkaline phosphatase-conjugated secondary antibody. (D) Neurite outgrowth on purified CAMs in comparison to laminin. Chick DRG neurons were cultivated on immobilized axonin-1 (50 mg/ml), L1 (25 mg/ml), NF15 (25 mg/ml), and laminin (25 mg/ml). After 24 h cells were immunostained using monoclonal antibody A2B5 and subjected to microscopy. The length of the longest neurite of process bearing cells was determined. Only weak neurite extension was detected on an axonin-1 substrate, while the ability of L1 and neurofascin to induce neurite outgrowth was comparably strong. The longest neurites were measured on a laminin substrate.

by regrowing axons. It is therefore conceivable that the presence of a combination of different CAMs may enhance neurite outgrowth when compared to single molecules. We therefore applied DRG neurons to surfaces coated with combinations of different CAMs to analyze neuritepromoting activity. A mixture of L1 and neurofascin turned out to be most efficient in the promotion of neurite outgrowth (median neurite length 190.6 mm, Fig. 2A). A mixture of L1 and

axonin-1 was less efficient (Fig. 2B, median neurite length: 144.0 mm, po0:0001, Mann–Whitney), while a combination of neurofascin and axonin-1 (Fig. 2C, median neurite length: 88.8 mm) even significantly inhibited neurite outgrowth in comparison to outgrowth observed on a pure neurofascin substrate (median neurite length: 123.1 mm, po0:0001, Mann–Whitney). This finding may be explained by the observation that axonin-1 acts as cellular receptor of sensory neurons grown on substrate-bound neurofascin

ARTICLE IN PRESS M. Niere et al. / Biomaterials 27 (2006) 3432–3440

100

100

80

80

L1 + neurofascin L1 neurofascin

40

20

0

0 100

200

300

400

500

600

700

neurite length [µm]

(A)

0

100

80

80

neurofascin + axonin-1 neurofascin axonin-1

40

20

20

0

0 0

100

200

300

400

500

600

700

neurite length [µm]

0

(D)

200

300

400

500

600

700

L1 + neurofascin + axonin-1 L1 + neurofascin + BSA L1 + axonin-1 + BSA neurofascin + axonin-1 + BSA

60

%

% 40

100

neurite length [µm]

(B)

100

60

(C)

40

20

0

L1 + axonin-1 L1 axonin-1

60

%

%

60

3435

100 200 300 400 500 600 700 800 900 1000

neurite length [µm]

Fig. 2. L1 and neurofascin synergistically enhance neurite outgrowth (A–C). Chick DRG neurons were seeded on immobilized co-substrates of each 12.5 mg/ml of two different substrate molecules. Neurite extension was compared with outgrowth on 25 mg/ml of the mono-substrate. Neurite extension on a L1/NF15 co-substrate was significantly increased in comparison to cells grown on mono-substrates (A). Neurite outgrowth on a co-substrate of L1 and axonin-1 did not enhance neurite outgrowth (B). A mixture of NF15 and axonin-1 inhibited neurite outgrowth (C). (D). Neurite outgrowth on a L1-NF15 co-substrate is not enhanced by addition of axonin-1. Chick DRG neurons were cultivated as in (A–C) on a co-substrate composed of axonin-1 (50 mg/ml), NF15 (25 mg/ml), and L1 (12.5 mg/ml). In controls, one of each substrate molecule was replaced by BSA at corresponding concentrations.

(unpublished results). Both axonin-1 and neurofascin may form a complex with each other that prevents interaction of cellular axonin-1 with substrate-bound neurofascin. Finally, combining all three molecules did not further promote neurite elongation in comparison to the L1/neurofascin cosubstrate (Fig. 2D). Consequently, in our study neurofascin and L1 was the optimum combination to induce neurite growth in sensory neurons. 3.3. Enhanced neurite elongation by neurofascin-L1 chimeric molecules Synergy between neurofascin and L1 with respect to neurite elongation may rely on the induction of separate signalling pathways, which may converge at an unknown specific level further downstream of L1 and neurofascin. We therefore assumed that forced aggregation of L1 and neurofascin signalling components may enhance the synergistic effects. Therefore, we combined cDNA se-

quences encoding L1 and neurofascin ectodomains to obtain protein chimera comprising parts of both molecules within one single polypeptide chain and investigated the functionality and potential of these chimeric proteins as substrate for outgrowing neurites. Since the six Ig-like domains of L1 were previously shown to be sufficient for L1-dependent neurite induction, they were linked in one chimera to the amino terminus of the complete ectodomains of neurofascin (construct L1[Ig1-6]-NF15ecto). A second chimeric Fc-fusion protein was constructed composed exclusively of L1 and neurofascin Ig-like domains (construct L1[Ig1-6]NF[Ig1-6]). Correct size and specificity of purified fusion proteins was analysed by SDS–PAGE and Western blot (Fig. 1B, lanes 4 and 5; Fig. 1C, lanes 3, 4, 6, and 7). DRG neurons were grown on immobilized L1 and neurite induction compared to outgrowth on L1-neurofascin chimera. Both chimera turned out to induce outgrowth of significantly longer neurites than L1 alone (Fig. 3A).

ARTICLE IN PRESS M. Niere et al. / Biomaterials 27 (2006) 3432–3440

3436

3.4. Enhanced neurite elongation by the application of GDF-5

100 80 60

L1(Ig1-6)NF15ecto

40

L1(Ig1-6)NF(Ig1-6)

%

L1

20 0 0

200

400

600

800

1000

1200

neurite length [µm]

(A) 100 80 60

L1(Ig1-6)NF15ecto

40

L1(Ig1-6)NF(Ig1-6)

%

L1 + neurofascin

20 0 0

(B)

200

400

600

800

1000

1200

neurite length [µm]

Fig. 3. The outgrowth-promoting properties of L1 and neurofascin can be combined within chimeric molecules. Chick DRG neurons (E9, 5000 cells/cm2) were cultivated on 25 mg/ml of immobilized chimeric CAM L1(Ig1-6)-NF15ecto (25 mg/ml), L1(Ig1-6)-NF(Ig1-6) (25 mg/ml). Neurite elongation was compared to outgrowth on a L1 substrate (A, 25 mg/ml) and on a L1-NF15 co-substrate (B, 12.5 mg/ml each). Both chimeric molecules induced significantly longer neurites in comparison to outgrowth on a L1 substrate. Beyond this, outgrowth on chimeric substrate L1(Ig1-6)-NF15ecto was significantly enhanced in comparison to a L1-NF15 co-substrate.

Therefore, the synergistic effect of L1 and neurofascin was recapitulated by the application of the L1-neurofascin chimera. In addition, the chimeric protein L1(Ig1-6)NF15ecto-induced neurite outgrowth even better than a combined substrate of separated neurofascin and L1 polypeptides (Fig. 3B, median neurite lengths: 278.5 and 187.8 mm, po0:0001, Mann–Whitney). Finally, only Ig-like domains of L1 and neurofascin expressed in the chimera (L1[Ig1-6]NF[Ig1-6]) was not as efficient in inducing neurite elongation as a chimera composed of L1 Ig-like domains in conjunction with the complete ectodomains of neurofascin (median neurite lengths: 203.0 and 278.5 mm, po0:0001, Mann–Whitney). Therefore, fibronectin type III-like repeats may contribute to full neurite outgrowth promoting activity. In summary, the application of chimeric proteins, composed of partners acting cooperatively, may further improve outgrowthpromoting activity.

Beside CAMs further soluble component like neurotrophins or cytokins may promote neurite elongation. Application of neurotrophins is obligatory in our culture system because survival and differentiation of embryonic sensory neurons require the presence of NGF. On the other hand, GDF-5 was shown to enhance the activity of NGF for neuronal differentiation synergistically. Following the concept that addressing different signalling pathways may synergistically promote neurite outgrowth, we investigated the effect of GDF-5 on neurite elongation induced by a panel of neural CAMs. As depicted in Fig. 4, adding GDF-5 to the medium significantly enhanced outgrowth of DRG neurons on axonin-1, neurofascin, L1, and laminin substrates in the presence of NGF (Fig. 4A–D). After 24 h of cultivation of DRG neurons in presence of GDF-5, median neurite length increased from 95.7 mm to 113.0 mm (on axonin-1, po0:013, Mann–Whitney, Fig. 4A), from 136.2 to 172.6 mm (on neurofascin, po0:0001, Mann–Whitney, Fig. 4B) from 123.1 to 177.6 mm (on L1, po0:0001, Mann–Whitney, Fig. 4C), and from 297.0 to 423.2 mm (on laminin, po0:0001, Mann–Whitney, Fig. 4D). Adding GDF-5 also significantly enhanced neurite elongation on a substrate composed of both neurofascin and L1 (Fig. 4E, po0:0001, Mann–Whitney). Therefore, among the substrates applied neurite length varied from 18.1% on a substrate composed of axonin-1 up to 44.3% on an L1 substrate. These findings indicate that GDF-5 promotes neurite elongation on all substrates, but to a different extent. Therefore, inclusion of GDF-5 in a mixture of CAMs and neurotrophins may further enhance neurite elongation. 3.5. GDF-5 does not induce anti-apoptotic pathways In a previous report, the characterization of GDF-5 as a survival factor for chick DRG neurons was based on the quantification of the number of process bearing cells kept in culture [20]. In fact, the question remains open whether the observed increasing percentage of neurons in the presence of GDF-5 relied on enhanced neurite induction or on better survival. DRG neurons were seeded in absence and presence of GDF-5. NGF was included in all experimental conditions because it is basically required for survival and neurite extension of sensory neurons [20]. As NGF was obligatory, the effects of GDF-5 on survival could not be examined by the removal of NGF. Instead, the kinase inhibitor staurosporine was added to induce apoptosis. After this treatment cells were analysed for hallmarks of cell death and apoptosis including quantification of pyknotic nuclei, DNA fragmentation, overall changes in cellular redox state, and caspase-3 activity (Fig. 5). Exposing DRG neurons to staurosporine at a

ARTICLE IN PRESS M. Niere et al. / Biomaterials 27 (2006) 3432–3440 100

100

80

80

60

60 40

20

20

0

0 0

50

(A)

100

150

200

250

300

350

0

100

80

80 60 40

20

20

0

0 0

100

200

300

400

500

600

700

neurite length [µm]

0

(D)

300

400

500

600

700

1200

1400

laminin + GDF-5 laminin - GDF-5

%

%

L1 + GDF-5 L1 - GDF-5

(C)

200

neurite length [µm]

100

40

100

(B)

neurite length [µm]

60

neurofascin + GDF-5 neurofascin - GDF-5

%

%

axonin-1 + GDF-5 axonin-1 - GDF-5

40

3437

200

400

600

800

1000

neurite length [µm]

100 80 60

%

L1 + neurofascin + GDF-5 L1 + neurofascin - GDF-5

40 20 0 0

(E)

100 200 300 400 500 600 700 800 900 1000

neurite length [µm]

Fig. 4. Soluble GDF-5 enhances neurite outgrowth independently from the immobilized substrate molecule(s). Chick DRG neurons (E9, 5000 cells/cm2) were seeded on immobilized axonin-1 (A, 50 mg/ml), NF15 (B, 25 mg/ml), L1 (C, 25 mg/ml), laminin (D, 25 mg/ml), and (E) on a substrate composed of L1 (12,5 mg/ml) as well as NF15 (25 mg/ml). Cells were cultivated in NGF-supplemented medium (25 ng/ml) in the absence and presence of 10 ng/ml GDF-5. Except for the situation on an axonin-1 substrate (po0; 013, Mann–Whitney), GDF-5 significantly enhanced neurite outgrowth independently from the substrate molecule (po0:0001, Mann–Whitney).

concentration of 400 nM resulted in a decrease in intact nuclei in comparison to pyknotic nuclei, increase in TUNEL positive nuclei, reduction in metabolic activity, and activation of caspase 3. Consequently, under these experimental conditions staurosporine treatment induced apoptotic pathways initiated by caspase 3 activation in accordance with previously published results [23,24]. Adding GDF-5 to the cell culture did not result in

significant changes in protecting cells from staurosporine-induced cell death as revealed throughout all assays. These data indicate that promotion of neurite outgrowth of DRG neurons by GDF-5 is not accompanied by an increased protection against proapoptotic stimuli. Therefore, the function of GDF-5 is presumably the induction of neurite extension in the presence of NGF.

ARTICLE IN PRESS M. Niere et al. / Biomaterials 27 (2006) 3432–3440

TUNEL-positi ve cells [%]

3438

Intact nuclei [%]

100 80 60 40 20 0

NGF

NGF GDF-5

NGF STA

(A)

NGF GDF-5 STA

100 80 60 40 20 0

NGF

NGF GDF-5

NGF STA A

(B)

NGF GDF-5 STA A

100

FU [%]

80 60 40 20 0

NGF

NGF GDF-5

NGF STA

(C)

NGF GDF-5 STA

NGF NGF+STA NGF+GDF-5+STA 35.000

NGF+GDF-5 NGF+zVAD

30.000 25.000

NGF+STA+zVAD NGF+GDF-5+STA+zVAD NGF+GDF-5+zVAD

20.000 15.000 10.000 5.000

Caspase activity [%]

fluorescence intensity [AU]

40.000

100 80 60 40 20 0

0 60 120 180 240 300 360 420 480 540 600 660 720 780

(D)

time [s]

(E)

NGF

NGF GDF-5

NGF STA

NGF GDF-5 STA

Fig. 5. GDF-5 does not protect from staurosporine-induced apoptosis. Isolated NGF (25 ng/ml)-dependent DRG neurons were cultivated on PEI-coated dishes or slides for 24 h in presence or absence of 10 ng/ml GDF-5 before exposure to 400 nM staurosporine (STA). Cell viability was analysed by evaluation of intact nuclei of neurons using Hoechst 33342 (A, n ¼ 4), by DNA fragmentation using TUNEL assay (B, n ¼ 3), by measuring of the cellular redox state using alamar BlueTM reagent (C, n ¼ 3), and by quantification of caspase-3 activity in response to induction of apoptosis (D, one representative experiment, E, n ¼ 5). All experiments revealed no significant protection of DRG neurons from staurosporine-induced apoptosis by GDF-5.

4. Discussion Our results show that the combined application of CAMs neurofascin and L1 is best to enhance neurite outgrowth of NGF-dependent sensory neurons. The outgrowth promoting activity of both factors can be enhanced (1) by combining components of both molecules in one chimeric fusion protein and (2) by adding GDF-5. We have also provided evidence that GDF-5 mainly acts as a neurite outgrowth-promoting factor in sensory neurons rather than a survival factor. The study presented here supports the idea that neurite outgrowth of sensory neurons may be optimized by combining purified recombinant neural CAMs along with further trophic factors like GDF-5 as shown for embryonic

neurons. Neurite outgrowth may be further enhanced by the fusion of different CAMs on one polypeptide chain. Hence, this approach is expected to be more efficient than others, which rely on the application of single factors. For practical reasons, embryonic neurons were applied to enable the test of many different combinations of regenerative factors with a simple cell culture system. Therefore, a re-evaluation of the combinations characterized in this study may be required for a regeneration model in vivo. Nevertheless, our results suggest that our novel strategy may be applicable, in principle. Extracellular matrix components like laminin and fibronectin or extracellular domains of CAM L1 were previously included into scaffold materials [12,13,25,26]. A combination of laminin and NGF turned out to mimic the

ARTICLE IN PRESS M. Niere et al. / Biomaterials 27 (2006) 3432–3440

neurite promoting activity of autologous transplants [3]. Actually, our results show that laminin is superior to CAMs in inducing neurite outgrowth. However, laminin as a ‘‘gold standard’’ with regard to its outgrowth-promoting capability cannot be considered as a suitable factor for a synthetic nerve guide. Laminin and fibronectin are universal substrates for many migrating cell types expressing its ligand, integrins, and may thus favour the immigration of fibroblasts that impede neuronal regeneration [13]. Therefore, neuronal CAMs, which are selective for interactions within the nervous system, provide a specific approach for the establishment of nerve guides. Furthermore, the complex molecular architecture of laminin does not enable its convenient production by using recombinant technologies. This is essential for the development of biomaterials because well-defined and highly purified molecules are required for the application in patients. Finally, neural CAMs fused to the Fc portion of human IgG1 turned out to be a preferred source of highly pure material, which is also appropriate for functional analysis in protein-binding studies, neurite outgrowth assays as well as therapeutic strategies. Our results showed that a combination of neurofascin and L1 is best for neurite elongation of sensory neurons. Both molecules belong to the same subfamily of L1-like CAMs of the IgSF as judged by the analysis of sequence conservation [27] and are characterized by an overlapping repertoire of extracellular binding partners including F11 and axonin-1 [28]. Although similar in structure and binding behaviour, neurofascin and L1 might address different signalling pathways to explain the obvious synergistic effect observed in our experiments. L1 signalling was previously shown to rely on the induction of components of the PLC-g and the MAPK pathways [29–31]. Neurofascin binds to the PDZ protein syntenin, but the mechanisms of downstream signalling remain to be elucidated [32]. The finding that GDF-5 increases neurite outgrowth in the presence of NGF confirms previous results. NGF addresses several pathways including signal transduction by PLC-g as well as MAPK- and PI3-K/Akt-related signalling. Signalling by GDF-5 is reported to be primarily associated with Smad activation [33,34]. Cooperativity between GDF-5 and NGF implies use of a common component in signalling, which is addressed by pathways activated by both factors. In fact, induction of Smad signalling has been found in PC12 cells [35]. Furthermore, observations suggesting a function of GDF-5 as a survival factor imply that GDF-5 might assist NGF in the induction of anti-apoptotic signalling, for instance via induction of Akt. On the other hand, we did not find any evidence that components of classical apoptosis are involved in GDF-5 signalling. Therefore, GDF-5 may represent a factor, which contributes to axon elongation rather than neuronal survival of DRG neurons. With this regard, the observation that GDF-5-related signalling may also be linked to transduction by MAPK-pathways via p38

3439

and ERK [36] suggests a possible involvement of this pathway in GDF-5-mediated neurite elongation, since p38MAPK-related signalling has been shown to contribute to neuronal differentiation in NGF-induced PC12 cells [37]. 5. Conclusions Neurite elongation may be enhanced by the introduction of a mixture of different recombinant factors in artificial nerve guides. Our results show that a combination of CAMs L1 and neurofascin along with NGF and GDF-5 supplies an improved environment of higher complexity for the induction of neurite elongation of embryonic neurons. Application of an artificial fusion protein composed of L1 and neurofascin further enhances neurite outgrowth. Consequently, we provide an improved basic strategy based on a combinatory approach and engineered CAMs to enhance neural regeneration by fully synthetic materials. Acknowledgements We like to thank Uta Rheinweiler for the construction of plasmid expression vectors and Biopharm GmbH, Heidelberg, for generously supplying us with recombinant human GDF-5. We are grateful for an L1 expression vector supplied by Vance Lemmon. The project was funded by BMBF Grant No. 0311568. References [1] Schmidt CE, Leach JB. Neural tissue engineering: strategies for repair and regeneration. Ann Rev Biomed Eng 2003;5:293–347. [2] Mackinnon SE, Dellon AL. A comparison of nerve regeneration across a sural nerve graft and a vascularized pseudosheath. J Hand Surg (Am) 1988;13(6):935–42. [3] Yu X, Bellamkonda RV. Tissue-engineered scaffolds are effective alternatives to autografts for bridging peripheral nerve gaps. Tissue Eng 2003;9(3):421–30. [4] Xu G, Nie DY, Wang WZ, Zhang PH, Shen J, Ang BT, et al. Optic nerve regeneration in polyglycolic acid-chitosan conduits coated with recombinant L1-Fc. Neuroreport 2004;15(14):2167–72. [5] Pittier R, Sauthier F, Hubbell JA, Hall H. Neurite extension and in vitro myelination within three-dimensional modified fibrin matrices. J Neurobiol 2005;63(1):1–14. [6] Yang Y, De Laporte L, Rives CB, Jang JH, Lin WC, Shull KR, et al. Neurotrophin releasing single and multiple lumen nerve conduits. J Control Release 2005;104(3):433–46. [7] Seilheimer B, Persohn E, Schachner M. Neural cell adhesion molecule expression is regulated by Schwann cell–neuron interactions in culture. J Cell Biol 1989;108(5):1909–15. [8] Rathjen FG, Wolff JM, Chang S, Bonhoeffer F, Raper JA. Neurofascin: a novel chick cell-surface glycoprotein involved in neurite–neurite interactions. Cell 1987;51(5):841–9. [9] Sonderegger P. Axonin-1 and NgCAM as ‘‘recognition’’ components of the pathway sensor apparatus of growth cones: a synopsis. Cell Tissue Res 1997;290(2):429–39. [10] Brummendorf T, Rathjen FG. Cell adhesion molecules 1: immunoglobulin superfamily. Protein Profile 1995;2(9):963–1108. [11] Volkmer H, Leuschner R, Zacharias U, Rathjen FG. Neurofascin induces neurites by heterophilic interactions with axonal NrCAM while NrCAM requires F11 on the axonal surface to extend neurites. J Cell Biol 1996;135(4):1059–69.

ARTICLE IN PRESS 3440

M. Niere et al. / Biomaterials 27 (2006) 3432–3440

[12] Adcock KH, Brown DJ, Shearer MC, Shewan D, Schachner M, Smith GM, et al. Axon behaviour at Schwann cell—astrocyte boundaries: manipulation of axon signalling pathways and the neural adhesion molecule L1 can enable axons to cross. Eur J Neurosci 2004;20(6):1425–35. [13] Webb K, Budko E, Neuberger TJ, Chen S, Schachner M, Tresco PA. Substrate-bound human recombinant L1 selectively promotes neuronal attachment and outgrowth in the presence of astrocytes and fibroblasts. Biomaterials 2001;22(10):1017–28. [14] Fine EG, Decosterd I, Papaloizos M, Zurn AD, Aebischer P. GDNF and NGF released by synthetic guidance channels support sciatic nerve regeneration across a long gap. Eur J Neurosci 2002;15(4):589–601. [15] Midha R, Munro CA, Dalton PD, Tator CH, Shoichet MS. Growth factor enhancement of peripheral nerve regeneration through a novel synthetic hydrogel tube. J Neurosurg 2003;99(3):555–65. [16] Lang EM, Asan E, Plesnila N, Hofmann GO, Sendtner M. Motoneuron survival after C7 nerve root avulsion and replantation in the adult rabbit: effects of local ciliary neurotrophic factor and brain-derived neurotrophic factor application. Plast Reconstr Surg 2005;115(7):2042–50. [17] Hotten G, Neidhardt H, Jacobowsky B, Pohl J. Cloning and expression of recombinant human growth/differentiation factor 5. Biochem Biophys Res Commun 1994;204(2):646–52. [18] Storm EE, Huynh TV, Copeland NG, Jenkins NA, Kingsley DM, Lee SJ. Limb alterations in brachypodism mice due to mutations in a new member of the TGF beta-superfamily. Nature 1994;368(6472): 639–43. [19] Krieglstein K, Suter-Crazzolara C, Hotten G, Pohl J, Unsicker K. Trophic and protective effects of growth/differentiation factor 5, a member of the transforming growth factor-beta superfamily, on midbrain dopaminergic neurons. J Neurosci Res 1995;42(5):724–32. [20] Farkas LM, Scheuermann S, Pohl J, Unsicker K, Krieglstein K. Characterization of growth/differentiation factor 5 (GDF-5) as a neurotrophic factor for cultured neurons from chicken dorsal root ganglia. Neurosci Lett 1997;236(2):120–2. [21] O’Keeffe GW, Dockery P, Sullivan AM. Effects of growth/ differentiation factor 5 on the survival and morphology of embryonic rat midbrain dopaminergic neurones in vitro. J Neurocytol 2004;33(5):479–88. [22] Hurley FM, Costello DJ, Sullivan AM. Neuroprotective effects of delayed administration of growth/differentiation factor-5 in the partial lesion model of Parkinson’s disease. Exp Neurol 2004;185(2): 281–9. [23] Dawson G. Mechanisms of apoptosis in embryonic cortical neurons (E6 and E7) in culture involve lipid signalling, protein phosphorylation and caspase activation. Int J Dev Neurosci 2000;18(2–3):247–57. [24] Pike BR, Zhao X, Newcomb JK, Wang KK, Posmantur RM, Hayes RL. Temporal relationships between de novo protein synthesis, calpain and caspase 3-like protease activation, and DNA fragmenta-

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33] [34]

[35]

[36]

[37]

tion during apoptosis in septo-hippocampal cultures. J Neurosci Res 1998;52(5):505–20. Mosahebi A, Wiberg M, Terenghi G. Addition of fibronectin to alginate matrix improves peripheral nerve regeneration in tissueengineered conduits. Tissue Eng 2003;9(2):209–18. Toba T, Nakamura T, Lynn AK, Matsumoto K, Fukuda S, Yoshitani M, et al. Evaluation of peripheral nerve regeneration across an 80-mm gap using a polyglycolic acid (PGA)—collagen nerve conduit filled with laminin-soaked collagen sponge in dogs. Int J Artif Organs 2002;25(3):230–7. Volkmer H, Hassel B, Wolff JM, Frank R, Rathjen FG. Structure of the axonal surface recognition molecule neurofascin and its relationship to a neural subgroup of the immunoglobulin superfamily. J Cell Biol 1992;118(1):149–61. Volkmer H, Zacharias U, Norenberg U, Rathjen FG. Dissection of complex molecular interactions of neurofascin with axonin-1, F11, and tenascin-R, which promote attachment and neurite formation of tectal cells. J Cell Biol 1998;142(4):1083–93. Schmid RS, Pruitt WM, Maness PF. A MAP kinase-signaling pathway mediates neurite outgrowth on L1 and requires Srcdependent endocytosis. J Neurosci 2000;20(11):4177–88. Schaefer AW, Kamiguchi H, Wong EV, Beach CM, Landreth G, Lemmon V. Activation of the MAPK signal cascade by the neural cell adhesion molecule L1 requires L1 internalization. J Biol Chem 1999;274(53):37965–73. Hall H, Williams EJ, Moore SE, Walsh FS, Prochiantz A, Doherty P. Inhibition of FGF-stimulated phosphatidylinositol hydrolysis and neurite outgrowth by a cell-membrane permeable phosphopeptide. Curr Biol 1996;6(5):580–7. Koroll M, Rathjen FG, Volkmer H. The neural cell recognition molecule neurofascin interacts with syntenin-1 but not with syntenin2, both of which reveal self-associating activity. J Biol Chem 2001;276(14):10646–54. Patapoutian A, Reichardt LF. Trk receptors: mediators of neurotrophin action. Curr Opin Neurobiol 2001;11(3):272–80. Aoki H, Fujii M, Imamura T, Yagi K, Takehara K, Kato M, et al. Synergistic effects of different bone morphogenetic protein type I receptors on alkaline phosphatase induction. J Cell Sci 2001;114(Pt 8):1483–9. Lutz M, Krieglstein K, Schmitt S, ten Dijke P, Sebald W, Wizenmann A, et al. Nerve growth factor mediates activation of the Smad pathway in PC12 cells. Eur J Biochem 2004;271(5):920–31. Nakamura K, Shirai T, Morishita S, Uchida S, Saeki-Miura K, Makishima F. p38 mitogen-activated protein kinase functionally contributes to chondrogenesis induced by growth/differentiation factor-5 in ATDC5 cells. Exp Cell Res 1999;250(2):351–63. Morooka T, Nishida E. Requirement of p38 mitogen-activated protein kinase for neuronal differentiation in PC12 cells. J Biol Chem 1998;273(38):24285–8.