Tissue plasminogen activator promotes axonal outgrowth on CNS myelin after conditioned injury

Authors

  • Kenneth Minor,

    1. Department of Biochemistry & Molecular Genetics and Neuroscience Program, University of Colorado Denver – School of Medicine, Aurora, Colorado, USA
    Search for more papers by this author
  • Jacob Phillips,

    1. Department of Biochemistry & Molecular Genetics and Neuroscience Program, University of Colorado Denver – School of Medicine, Aurora, Colorado, USA
    Search for more papers by this author
  • Nicholas W. Seeds

    1. Department of Biochemistry & Molecular Genetics and Neuroscience Program, University of Colorado Denver – School of Medicine, Aurora, Colorado, USA
    Search for more papers by this author

Address correspondence and reprint requests to Nicholas W. Seeds, University Colorado Denver – Anschutz Medical Campus, MS-8315, Neuroscience, Aurora, CO 80045, USA.
E-mail: nicholas.seeds@ucdenver.edu

Abstract

J. Neurochem. (2009) 109, 706–715.

Abstract

Following CNS injury, myelin-associated inhibitors represent major obstacles to axonal regeneration and functional recovery. The following study suggests that the proteolytic enzyme tissue plasminogen activator (tPA) plays a major function in ‘conditioning-injury induced’ axon regeneration. In this paradigm, prior peripheral nerve injury leads to an enhanced ability of sensory neurons to regenerate their central axons in the presence of the CNS inhibitory microenvironment. tPA is widely expressed by CNS and PNS neurons and plays major roles in synaptic reorganization and plasticity. This study shows that cultured neurons from mice deficient in tPA, in contrast to wild-type mice, fail to undergo conditioning-injury induced axonal regeneration in the presence of purified myelin membranes. Interestingly, neurons from mice deficient in plasminogen, the best known substrate for tPA, showed active axon regeneration. These results suggest a novel plasminogen-independent role for tPA in promoting axonal regeneration on CNS myelin.

Abbreviations used
DMEM

Dulbecco’s Modified Eagles Medium

DRG

dorsal root ganglia

ECM

extracellular matrix

HGF

hepatocyte growth factor

LRP1

LDL-like receptor

MAG

myelin-associated glycoprotein

MMP-9

matrix metalloprotease 9

NGF

nerve growth factor

NgR

Nogo-66 receptor

OMgp

oligodendrocyte myelin glycoprotein

PAI-1

plasminogen activator inhibitor-1

PBS

phosphate-buffered saline

PirB

paired immunoglobulin-like receptor B

Plgn−/−

plasminogen-deficient mice

tPA

tissue plasminogen activator

tPA−/−

tPA-deficient mice

Trauma to the adult CNS results in devastating clinical consequences because of the failure of neurons to regenerate their injured axons. This contrasts with what is observed in the adult PNS where regeneration is possible, and even in more extreme cases where surgical intervention is required, much functional recovery can occur and clinical outcomes are more positive. Studies done over 25 years ago demonstrated that adult CNS neurons show a similar intrinsic capacity for axon outgrowth following injury as do those in the PNS, and ultimately implicated factors present in the CNS microenvironment responsible for regeneration failure (Aguayo et al. 1981, 1983). Since then, many factors, including those associated with CNS myelin, have been identified as being inhibitory to axon outgrowth and as potential targets for therapies designed to promote recovery via regeneration and/or plasticity of the injured CNS.

Several investigators have confirmed the inhibitor potential of isolated myelin membrane preparations on axonal outgrowth in vitro. Major inhibitory molecules identified in CNS myelin include myelin-associated glycoprotein (MAG), Nogo, and oligodendrocyte myelin glycoprotein (OMgp) (for reviews see Filbin 2003; Liu et al. 2006; Schwab 1996). These molecules were thought to inhibit axon growth by binding to a common cell-surface receptor, Nogo-66 receptor (NgR), localized on the growth cones of injured axons (Yiu and He 2006), and acting primarily on growth cone collapse (Chivatakarn et al. 2007). Through the action of NgR co-receptors, which include p75, TROY, and LINGO1, this binding leads to the activation of downstream neural signaling pathways that inhibit axon growth (Yiu and He 2006). The best characterized of these pathways is the RhoA/RhoA-associated kinase pathway that when activated leads to stabilization in the growth cone cytoskeleton which restricts outgrowth of injured axons (He and Koprivica 2004). However, mutant mice deficient in these myelin proteins or NgR show only limited axon regeneration and functional recovery (Zheng et al.2006); suggesting that other myelin components and receptors on neurons play role in this inhibition. In this regard, the paired immunoglobulin-like receptor B (PirB) that also binds Nogo, MAG, and OMgp has been shown recently to inhibit axon regrowth, and blocking both PirB and NgR leads to axon regeneration on myelin (Atwal et al. 2008). Another study has shown that the mammalian target of rapamycin (mTOR) signaling pathway is down-regulated following axonal injury by the tumor suppressor PTEN (phosphatase and tensin homolog), whose deletion allows mTOR activity on downstream targets including ribosomal S6 kinase and eukaryotic initiation factor4E to promote axon regeneration in an inhibitory environment (Park et al. 2008) Further understanding of the complex mechanisms by which myelin-associated inhibitors impede axon outgrowth will aid in developing treatments to counter the inhibitory CNS environment and facilitate regeneration.

One potential means by which neurons might counteract the inhibitory factors present in CNS myelin would be to degrade these factors or interfere with their receptors in ways that neutralize their inhibitory effects on axon growth. Extracellular proteolysis is a strategic mechanism employed by many cell types to disrupt or modify both cell–cell and cell–extracellular matrix (ECM) contacts thereby altering cellular responses to environmental cues (Werb 1997). During many physiological and pathological events, this function is achieved through the expression of several different proteolytic enzymes, including members of the plasminogen activator/plasmin system and matrix metalloproteases (MMPs). Tissue plasminogen activator (tPA) (EC# 3.4.21.68) is a serine protease that converts the inactive precursor plasminogen into its active counterpart, plasmin (EC# 3.4.21.7). Plasmin is a broad-spectrum protease with a long list of extracellular targets including cellular receptors, growth factors, and other ECM components (Tsirka 2002). In the circulatory system, the primary function of plasmin is to degrade fibrin clots, and as such tPA is most recognized for its role in thrombolysis and vascular homeostasis. However, recent years have seen a growing appreciation for tPAs potential roles in the nervous system, with the enzyme being linked to neural mechanisms such as plasticity, synaptic remodeling, and axonal regeneration (for review see Tsirka 2002).

Tissue plasminogen activator is widely expressed in the nervous system by neurons and some glial cells. tPA activity is up-regulated by neurons during both normal and pathological events that lead to neural plasticity. tPA is focally concentrated at axonal growth cones, where it can facilitate axonal growth in vivo and in 3-D cell cultures (Krystosek and Seeds 1981a,b; Pittman and DiBenedetto 1995). In vivo, tPA is up-regulated in the brain during motor learning and promotes a form of plasticity known as long-term potentiation (LTP) (Seeds et al. 1995, 2003; Baranes et al. 1998). tPA has been shown to play a role in experience-dependent plasticity of the visual cortex during the developmental critical period (Mataga et al. 2002, 2004), and tPA induces mossy fiber sprouting and reorganization during seizures (Wu et al. 2000). Peripheral nerve injury leads to an up-regulation of tPA by injured neurons where tPA promotes axon regeneration and functional recovery (Siconolfi and Seeds 2001a,b).

The following study was designed to test the potential role that tPA plays in promoting nerve regeneration in the presence of myelin-associated molecules which represent key barriers to axonal outgrowth following CNS injury. As injured neurons show limited axon growth in the presence of myelin inhibitors, we utilized an experimental model known as the conditioning-injury paradigm that enhances the ability of dorsal root ganglia (DRG) sensory neurons to grow axons in the presence of these inhibitors in culture and more importantly to regenerate CNS axons following a dorsal column spinal cord injury. In this paradigm, axonal growth is stimulated via a pre-conditioning lesion made to the peripheral branch of DRG neurons (Richardson and Issa 1984; Richardson and Verge 1986). Following a conditioning period of 1–2 weeks, injured neurons display a greater propensity for neurite outgrowth in vitro when neurons are explanted and cultured on either growth-permissive substrates or inhibitory myelin (Neumann and Woolf 1999; Qiu et al. 2002). Interestingly, our prior study showed that the peak of tPA up-regulation in the DRG and regenerating axons was also at 1 week following sciatic nerve injury (Siconolfi and Seeds 2001a); suggesting that tPA may play a role in the conditioning-injury paradigm. To test this possibility, mice deficient for either tPA or plasminogen genes were used in the conditioning-injury model of axonal growth on myelin. Our experiments show that tPA is important for conditioned axonal outgrowth and may represent a potential treatment in promoting regenerative growth and plasticity in the injured CNS.

Materials and methods

Animals

Animals used in this study were adult 3-month-old male C57Bl/6 mice including wild-type and those deficient for either the tPA (tPA−/−) or the plasminogen (Plgn−/−) genes generated on the C57Bl/6 genetic background. Mice were house in the Center for Comparative Medicine at the University of Colorado Denver – Health Sciences Center. Animals were cared for in accordance with the guidelines published in the NIH Guide for the Care and Use of Laboratory Animals and the principles presented in the ‘Guidelines for the Use of Animals in Neuroscience Research’ by the Society for Neuroscience. All animal procedures were approved by the UCDHSC Institutional and Animal Care and Use Committee.

Surgical procedures

Pre-conditioning sciatic nerve injuries were carried out to enhance the growth potential of adult sensory DRG neurons. Briefly, animals were anesthetized using a mixture of Ketamine (80 mg/kg) and Xylazine (12 mg/kg). The left hind limb of each animal was then shaved and sterilized with Betadine. A small incision was made in the skin that ran parallel with but slightly caudal to the femur. Careful teasing of muscle with surgical probes was carried out to expose the underlying sciatic nerve. The sciatic nerve was injured using a crush method as described (Siconolfi and Seeds 2001a), where the nerve was placed into a 1-mm-wide needle holder and crushed for 20 s. The holder was then rotated 90° and the crush repeated at the same site. Both the muscle and skin were then sutured. Animals were allowed to recover and given an oral dose of 0.9 mg Ibuprofen per day for analgesia.

Neuronal culture

Seven days following sciatic nerve injury, animals were killed with an overdose injection of sodium pentobarbital. Left lumbar (L4–L6) DRG were carefully isolated under aseptic conditions from both uninjured and injured mice and trimmed to remove nerve roots. DRG were then treated with 1 mL 0.25% collagenase from Clostridium histolyticum (Sigma, St Louis, MO, USA) at 37°C for 1 h. Collagenase was then removed and DRG were treated with 1 mL of 10% heat-inactivated fetal bovine serum (Hyclone, Logan, UT, USA) in Dulbecco’s Modified Eagles Medium (DMEM) (Invitrogen, Carlsbad, CA, USA) to inhibit any remaining collagenase activity. Fetal bovine serum was removed and DRG were washed three times with 1.5 mL volumes of DMEM. DRG were then carefully triturated using polished glass pipettes that were pre-treated with SigmaCote (Sigma). Centrifugation steps were carried out at 33 g for 4 min, and 80 g for 4 min and 125 g for 1 min to remove debris and enrich neurons by the removal of non-neuronal cells such as Schwann and satellite cells. Isolated neurons were then re-suspended in a modified Sato media (Doherty et al. 1990; Cai et al. 1999) containing DMEM plus 5 mg/mL insulin, 29 mg/mL l-glutamine, 4 mg/mL bovine serum albumin, 100 μM putrescine HCl, 0.1 μg/mL l-thyroxine, 0.01 μg/mL 3,3,5,5-triiodo-l-thyronine, 0.08 μg/mL progesterone, 30 nM selenium, 50 ng/mL nerve growth factor (NGF, Sigma) and 125 U/mL penicillin–streptomycin (Invitrogen). Multi-chambered glass microscope slides or 12 mm glass coverslips were pre-coated with 20 μg/mL poly-d-lysine (Sigma) in phosphate-buffered saline (PBS) for 45 min at 37°C. Slide wells were rinsed with PBS, followed by coating overnight at 4°C with either 10 μg/mL fibronectin (Sigma) or 10 μg/mL myelin membranes. Myelin membranes were isolated from adult mouse brain and spinal cord as described (Norton and Poduslo 1973). Before plating neurons, the slide wells were rinsed with PBS and DMEM. Neurons were cultured for 40 h at 37°C in 7% CO2/93% air. In some cases, neuron cultures were supplemented with 140 nM recombinant tPA (Genentech, San Francisco, CA, USA) 1 h following plating.

Neurite outgrowth analyses

Following 40 h in culture, neurons were carefully rinsed with PBS and then fixed using 4% paraformaldehyde in PBS for 5 min. Fixed cultures were washed three times with PBS and then blocked using 0.1% bovine serum albumin in PBS containing 0.1% triton X-100. Following blocking, primary antibody incubation was carried out using rabbit anti-β-tubulin-III (Sigma) (1 : 300 in blocking solution) overnight at 4°C. Secondary detection was performed using a VECTASTAIN ABC peroxidase system (Vector Labs, Burlingame, CA, USA) and 3,3′-diaminobenzidine. Digital photomicrographs of neurons and their neurite outgrowth patterns were taken using taken using a SPOT RT camera system (Sterling Heights, MI, USA). For outgrowth analysis, two parameters were measured. First, all neurons were counted and identified as being non-neurite bearing or those displaying at least one neurite > 25 μm in length. This analysis was used to obtain percentages of neurite initiation. Lastly, elongation analysis was made by measuring all neurites for length. All studies were performed a minimum of three times, and more than 300 cells or neurites assessed in each study. Statistical analysis using a one-way anova was done to compare results for both neurite initiation and elongation between experimental groups, where p < 0.01 were considered significant; however, all p values indicated by an asterisk in these studies were < 0.01.

tPA zymography

Plasminogen activator zymography was performed as described previously (Heussen and Dowdle 1980). Non-reducing sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels were polymerized with protein substrates, casein and plasminogen. Equal amounts of cell culture media from DRG neurons isolated from non-injured and injury-conditioned wild-type mice and recombinant human tPA (Genentech) and mouse urine (primarily urokinase plasminogen activator) standards were electrophoresed in 10% separating acrylamide gels. Sodium dodecyl sulfate was extracted from the gels using a 2.5% Triton X-100 wash for 30 min. before incubating the gel for 18 h in 0.1 M Tris–HCl pH 8.1 at 37°C to allow degradation of the casein. Then gels were stained with Coomassie blue (0.125%), and destained in 50% methanol and 10% acetic acid until clear zones of PA mediated caseinolysis appeared in the blue background. The gels were analyzed using a Chemilmager 5500 system (San Leandro, CA, USA).

Results

To assess the potential role of tPA in the conditioning-injury model of axonal growth on CNS myelin, sciatic nerves of adult wild-type, tPA−/−, and Plgn−/− mice were injured as described in the Materials and methods. One week after injury, L4/L5 DRG neurons were isolated and cultured on a substrate of either growth-permissive fibronectin or inhibitory myelin membranes purified from mouse brain and spinal cord. DRG neurons were cultured over a period of 40 h. As shown in Fig. 1(a and c), neurite outgrowth by non-conditioned wild-type or tPA−/− DRG neurons is inhibited in the presence of myelin membranes. Pre-conditioned wild-type neurons (Figs 1b and 2c) display robust neurite outgrowth with complex arborization patterns on myelin membranes demonstrating the conditioning-injury response of axonal regeneration. In contrast, conditioned DRG neurons from tPA−/− mice (Fig. 1d) fail to display extensive neurite outgrowth and growth is visibly more like that of non-conditioned neurons. As seen in Fig. 2(a), many tPA−/− neurons fail to produce neurites and show signs of abortive outgrowth with broad lamellapodia-like structures on the myelin substrate, in marked contrast to the dramatic neurite outgrowth seen with conditioned wild-type neurons on myelin (Fig. 2c). Fewer tPA−/− neurons are able to initiate outgrowth, but many of those display abnormal processes that are in some cases swollen and not well-defined, while others appear to be blebbing (Fig. 2b), a morphological indicator of neurite degeneration (Ivins et al. 1998).

Figure 1.

 Mice deficient for tPA fail to display conditioned-injury induced neurite outgrowth on myelin membranes. Photomicrographs of (a) non-conditioned wild-type neurons, (b) conditioned wild-type neurons, (c) non-conditioned tPA−/− neurons, and (d) conditioned tPA−/− neurons cultured on myelin membranes. Neurons were stained using rabbit anti-β tubulin III. All bars are 50 μm.

Figure 2.

 Sensory DRG neurons from conditioned-injury tPA−/− mice show signs of abortive neurite outgrowth and blebbing on myelin when compared to conditioned wild-type neurons. Photomicrographs of representative neurons from tPA−/− mice grown on myelin show signs of (a) abortive neurite outgrowth and broad lamellapodia-like structures and (b) swelling and blebbing of neurites, compared to the conditioned-injury wild-type neurons (c). All bars equal 25 μm.

Quantitative analysis of neurite outgrowth was carried out to obtain measurements of both neurite initiation and elongation. Initiation analysis was done by counting the number of neurons possessing at least one neurite > 25 μm in length and comparing this value to the total number of neurons randomly sampled. Results are presented as % total number neurons assayed and provided as a bar graph in Fig. 3(a). As expected, injury-conditioned wild-type neurons display enhanced neurite initiation over that of non-conditioned neurons (35.0 ± 3.1% vs. 21.7 ± 2.0%). Injury-conditioned neurons from tPA−/− mice show no increase in neurite initiation over their non-injured counterparts, and display significantly fewer (18.8 ± 3.3%) neurite-bearing cells comparable to that observed for non-conditioned wild-type neurons. Interestingly, addition of 140 nM recombinant tPA to the culture medium of tPA−/− neurons was able to induce neurite initiation in these neurons to levels (30.0 ± 5%) that are on average not significantly different from those of conditioned wild-type neurons. However, addition of tPA to non-conditioned wild-type neurons has no enhancement effect. These results demonstrate the importance of tPA as one of the initiating factors for conditioning-induced neurite outgrowth on the inhibitory myelin membrane substrate.

Figure 3.

 Quantitatively tPA−/− neurons fail to display conditioning-injury induced neurite initiation and elongation on myelin. (a) Neurite initiation was measured by calculating the average number of neurons that displayed at least one neurite of 25 μm or longer in length. Results are represented as the percent of neurite-bearing neurons out of the total number of neurons counted. (b) Average lengths for each type of neuron assayed. Gray bars indicate neurons from wild-type mice, and Black bars indicated neurons from tPA−/− mice. Error bars represent standard errors, and asterisks indicate p < 0.01 compared to other cultures from the same genotype mice.

For quantitative analysis of neurite elongation, all observed neurites for each neuron type or condition were measured for length and averaged. Results are presented in Fig. 3(b). As expected conditioned wild-type neurons display on average significantly greater (3.1-fold increase) neurite lengths than that of non-conditioned neurons (187.4 ± 12.4 μm vs. 60.3 ± 10.0 μm), again confirming the conditioning-injury effect on axon regeneration. Conditioned neurons from tPA−/− mice show only a slight increase in neurite length over neurons from non-conditioned tPA−/− mice, but display a 45% reduction in average neurite length (102.9 ± 6.4 μm) compared to conditioned wild-type neurons. However, tPA−/− neurons still are able to grow neurites on average that are longer than those of non-conditioned wild-type neurons, demonstrating the possibility of compensatory molecules that can partially recover the effects of a tPA deficiency. Addition of recombinant tPA to conditioned tPA−/− cultures did induce neurite elongation (140.1 ± 9.1 μm) that is statistically significant to the non-conditioned tPA−/− neurons, but not quite to the levels observed for conditioned wild-type neurons; however, tPA addition had no effect on neurite length when added to non-conditioned wild-type neurons.

Further analysis of neurite elongation by conditioned wild-type and tPA−/− neurons in the presence of myelin membranes was carried out to obtain distribution patterns on the percentage of neurites that fell between various length ranges (≤ 50 μm, 51–99 μm, 100–149 μm, 150-199 μm, 200–249 μm, and ≥ 250 μm). Neurite lengths were binned and tabulated for the various length ranges and presented as a bar graph in Fig. 4(a). The majority (67.7%) of conditioned tPA−/− neurites fail to grow to lengths greater that 100 μm. In contrast, only 24.3% of conditioned wild-type neurites fail to grow to lengths greater that 100 μm. In fact, 35.7% of conditioned wild-type neurites grow to lengths equal to or greater that 200 μm, compared to only 11.5% of conditioned tPA−/− neurites falling into this range. Clearly, tPA-deficiency drastically limits the ability of conditioned neurons to grow long neurites in the presence of myelin membranes. Finally, it is important to note that when grown in the presence of the permissive substrate fibronectin, injury-induced conditioning stimulated neurite growth by both wild-type and tPA−/− neurons, but to similar lengths that are not significantly different between genotypes (Fig. 4b). This finding indicates that tPA has a important role in conditioning-induced outgrowth only in the presence of inhibitory myelin membranes, but doesn’t necessarily play a part in the general enhancement of growth potential that occurs as a result of a conditioning injury.

Figure 4.

 The majority of neurons from conditioned-injury tPA−/− mice display shorter processes relative to those from wild-type mice. (a) Distribution patterns on the percentage of neurites that fell between various length ranges (≤ 50, 51–99, 100–149, 150–199, 200–249, and ≥ 250 μm). Neurite lengths were binned and tabulated for the various length ranges and presented as a bar graph. The majority (67.7%) of conditioned tPA−/− neurites fail to grow to lengths > 100 μm. (b) No difference was observed in the conditioning stimulated average neurite lengths for wild-type and tPA−/− neurons grown on fibronectin. Error bars represent standard errors, and asterisks indicate p < 0.01 compared to unconditioned neurons of the same genotype.

Since our prior studies (Siconolfi and Seeds 2001a) showed that sciatic nerve injury led to an induction of tPA in DRG neurons and their regenerating axons, we compared the tPA activity in culture media collected from non-injured and injury-conditioned DRG neurons of wild-type mice. The media were collected at 17hr. when axon outgrowth is in its earliest stages on the myelin substratum. Zymography of equivalent amounts of media from cultures containing the same number of plated neurons shows that the conditioning-injury leads to a 60% greater secretion of tPA by these neurons, when analyzed densitometrically under conditions of linear activity (Fig. 5). There was no difference in the urokinase plasminogen activator (uPA) activity with a conditioning-injury.

Figure 5.

 Tissue plasminogen activator activity is increased in cultures of injury-conditioned wild-type DRG neurons. Culture media was removed from cultures of dissociated DRG neurons from injury-conditioned and non-injured wild-type mice at 17 h, when axon outgrowth is just beginning on myelin. The media was subjected to zymography to assess relative amounts of secreted tPA activity between these different cultures. tPA activity is 60% greater in the media from injury-conditioned wild-type DRG neurons (lane 4) compared to media from DRG neurons of non-injured wild-type mice (lane 3). Plasminogen activator standards for tPA (r-tPA in lane 1) top arrow, and urokinase plasminogen activator (uPA) (mouse urine in lane 2) thin lower arrow were included on this gel.

In a separate analysis for neurite elongation, random DRG neurite-bearing neurons were measured for cellular diameter and were separated into groups defined as small (< 25 μm) or medium-to-large (≥ 25 μm) subtypes of neurons. As shown in Fig. 6, the greatest impact of tPA deficiency on neurite elongation is imparted on small caliper neurons. Injury-conditioned small caliper tPA−/− neurons grow neurites that are 2.5-fold shorter in length than those of corresponding wild-type neurons. For medium- to large-sized neurons, the difference is not so robust; conditioned tPA−/− neurons grow neurites that are only 1.3-fold shorter than those of wild-type neurons. Interestingly, the addition of recombinant tPA only improves the outgrowth of small neurons bringing the average neurite lengths to levels not different from those of wild-type neurons; whereas, addition of recombinant tPA fails to rescue the diminished outgrowth observed for medium–large tPA−/− neurons. This finding reflects a greater dependency by small DRG neurons for tPA in axonal elongation, which coincides with the earlier report from this laboratory showing that small neurons were the primary source of tPA in these DRG cultures (Hayden and Seeds 1996).

Figure 6.

 Mainly smaller neurons from tPA−/− mice show a failure for conditioning-injury induced neurite outgrowth on myelin. In a separate analysis for neurite elongation, random DRG neurite-bearing neurons were measured for cellular diameter and were separated into groups defined as small (< 25 μm) or medium-to-large (≥ 25 μm) subtypes of neurons. For each neuron type and condition, the longest neurite displayed was measured and these values were used to generate overall average lengths. Gray bars, conditioned wild-type neurons; black bars, conditioned tPA−/− neurons; striped bars, conditioned tPA−/− neurons with added rtPA (140 nM). Error bars represent standard errors, and asterisk indicates p < 0.01 for tPA−/− small neurons compared to wild-type small neurons.

To begin investigating the mechanism(s) by which tPA can promote conditioning-induced neurite outgrowth in the presence of inhibitory myelin, we evaluated the importance of tPA action on the presence of its principal substrate plasminogen/plasmin. To test the potential role that this conversion might play in conditioning-induced outgrowth on myelin, injury-conditioned neurons were also generated from Plgn−/− mice. As shown in Fig. 7(a), conditioned Plgn−/− neurons were able to grow neurites that were visually comparable to those of conditioned wild-type neurons shown in Fig. 1(b). Quantitative analysis of neurite elongation demonstrated that conditioned Plgn−/− neurons grow neurites of lengths not different from corresponding wild-type neurons (194.3 ± 19.6 vs. 187.4 ± 12.4, respectively) (Fig. 7b); nor was any difference seen in neurite initiation. This finding points to a plasminogen/plasmin-independent role of tPA in conditioning-injury induced neurite outgrowth on myelin.

Figure 7.

 Neurons from Plgn−/− mice show robust conditioning-injury induced neurite outgrowth on myelin. (a) Photomicrograph of sensory neurons from Plgn−/− mice grown on myelin (bar: 50 μm). (b) Quantitative length analysis showed no difference in the conditioning response displayed by wild-type and Plgn−/− neurons. Error bars represent standard errors.

Discussion

Earlier we showed that peripheral lesion of DRG sensory nerves led to induction of tPA and its transport down the regenerating peripheral nerve root; furthermore, the induced tPA expression was necessary for regeneration and timely recovery of sensory function (Siconolfi and Seeds 2001a,b). Similarly our continuing studies (manuscript in preparation) have shown that there is also a dramatic increase in tPA transport in the central nerve root within the dorsal column of the spinal cord 1 week following peripheral injury of the sciatic nerve; suggesting that tPA may facilitate axonal regrowth following a subsequent dorsal column injury as seen with the conditioning injury model (Neumann and Woolf 1999; Qiu et al. 2002). In the current study, our findings demonstrate a role for tPA in conditioning injury induced axonal growth by DRG neurons cultured in the presence of inhibitory CNS myelin. Here, sensory DRG neurons from mice deficient for tPA fail to display this conditioning response following peripheral injury. Interestingly, this effect was only observed when neurons were cultured on myelin membranes and not on the growth-permissive substrate fibronectin, indicating a specific role for tPA in facilitating myelin-impeded axon growth in the conditioning injury model.

What are the potential mechanisms by which tPA facilitates axon growth on myelin?

It is well documented that tPA plays a role in neurite outgrowth. Early experiments from this laboratory demonstrated that neurons utilize tPA expression and secretion localized to their growth cones to proteolytically break-down ECM molecules and facilitate migration and axonal elongation (Krystosek and Seeds 1984; McGuire and Seeds 1990). Many of the related effects of tPA in the nervous system are regulated by inhibitors, including plasminogen activator inhibitor-1 (PAI-1), neuroserpin and protease nexin-1, which inhibit the proteolytic activity of tPA. It is intriguing to speculate the possibility that myelin-associated inhibitors such as MAG, Nogo, and OMgp or their receptors including PirB represent proteolytic substrates for tPA and that they play a role in the results described here. We are currently investigating this possibility; in this regard preliminary studies of simply pre-treating myelin-coated coverslips with r-tPA did not promote axonal growth from conditioned tPA−/− neurons cultured without r-tPA. Alternative substrates for tPA are important to consider, especially as the results reported here show that mice deficient for plasminogen were able to display conditioned-injury axon outgrowth on myelin, implicating a plasminogen-independent role for tPA in regulating this neuritogenic effect on CNS myelin.

Although plasminogen expression was necessary for recovery of peripheral sensory function in the in vivo studies (Siconolfi and Seeds 2001b), the neural culture system used here to monitor axonal growth on inhibitory substrates is much simpler and was not plasminogen-dependent. There are many potential reasons for a plasminogen requirement during in vivo axonal regeneration; these include initial wound healing events, fibrinolysis required in areas of regrowth, vascular hemostasis supporting the regrowing axon, Schwann cell interaction and re-myelination of the regenerating axon, synapse formation, recovery of sensory terminals, etc.

One of the more important plasminogen-independent roles of tPA is the modulation of growth factor activities via post-translational proteolysis that converts them from an inactive pro-form to an active mature-form. tPA can directly cleave pro-hepatocyte growth factor (HGF) to active mature HGF, and this is thought to occur at or near the HGF receptor c-met (Mars et al. 1993; Thewke and Seeds 1996). HGF is known to be neurotrophic and neurotropic for a variety of neurons in both the CNS and PNS, and to promote axonal regrowth after spinal cord injury (Kitamura et al. 2007). Interestingly HGF promotes axonal outgrowth and survival of the NGF-dependent pain and temperature sensitive DRG neurons, but has no effect on brain-derived neurotrophic factor or neurotrophin-3-dependent sensory neurons (Maina et al. 1997). These HGF-responsive nociceptive and thermoreceptive neurons are likely to be those same small neurons that express tPA (Hayden and Seeds 1996) and that are shown here (Fig. 6) to be tPA-dependent for axonal outgrowth on the myelin membrane substrate.

Another important plasminogen-independent action of tPA centers around its ability to act as a cytokine in cell signaling events. tPA binding to the low density lipoprotein-like receptor (LRP1) either alone or as an inhibited complex with PAI-1 triggers LRP1 tyrosine phosphorylation leading to intracellular signaling (Orth et al. 1994). tPA binds to LRP1 on hippocampal neurons and enhances late-phase LTP by activating cAMP-dependent protein kinase (Zhuo et al. 2000). Binding of tPA to LRP1 can also activate Mek1 and its downstream targets Erk-1 and -2 (Hu et al. 2006, 2008). Interestingly, this activation results in increased expression of another important proteolytic enzyme, MMP-9 (EC# 3.4.24.35). It is possible that tPA exerts its effects on axon growth observed in this study directly in a non-proteolytic manner, but an overall effect that requires the proteolytic degradation of myelin by down-stream effectors such as MMP-9. In vivo studies have shown that tPA regulates levels of MMP-9 expression both in animal models of stroke (Tsuji et al. 2005) and peripheral nerve injury/regeneration (Siconolfi and Seeds 2003).

Two recent reports have shown that independent of plasmin formation tPA binding to an LRP can facilitate NMDA receptor activity and calcium influx into neurons (Martin et al. 2008; Samson et al. 2008). tPA binding to LRP1 leads to phosphorylation of the C-terminal NPxY domain and interaction with PSD95, which bridges LRP1 to the cytoplasmic face of the NMDA receptor (Martin et al. 2008). Somewhat differently, Samson et al. (2008) showed that tPA proteolytic activity is required for binding and activation of LRP1 in a plasminogen-independent activation of the NMDA receptor. However, tPAs proteolytic activity is required and tPA may have to bind a serpin like PAI-1 or protease nexin-1 through its active site prior to binding and activation of LRP1 and NMDA-R.

In summary, this is the first study to define a role for tPA in facilitating axon growth in the presence of inhibitory molecules like CNS myelin that are known to block CNS regeneration. This property fits with the many identified roles of tPA in the nervous system, including neural plasticity and peripheral nerve regeneration. The ability of tPA to promote axon growth of neurons in the presence of myelin inhibitors raises the possibility of tPA as a potential therapy in the treatment of both acute and chronic CNS injuries. Considering this, we are currently investigating whether this ability of tPA to promote axon growth is related to spinal cord regeneration of sensory nerve tracts observed in the conditioning-injury paradigm (Neumann and Woolf 1999; Qiu et al. 2002).

Acknowledgements

The authors would like to thank Susan Haffke for her excellent technical assistance, and help from Shay Fabbro, Katherine Brandao, and Steve Mikesell with these studies. This study was supported in part by grants from NIH: NS-441219 and T32-NS-007083, and the Paralyzed Veterans of America – #2047, and the Paralysis Project of America #28880.

Ancillary