All the experiments were performed in 2-months-old mice and approved by the Italian regulations and San Raffaele Institutional Animal Care and Use Committee (IACUC 487).

For humans samples, written informed consent was received and accepted from all patients, and the study approved by the San Raffaele Ethical Committee.

uPAR−/− mice, previously described [11], were maintained in congenic C57BL/6 strain.

Ten uPAR null mice and 10 control littermates (2 months old) were analyzed for sciatic nerve conduction velocity (NCV) and compound motor action potential (cMAP) in intact nerves, or 45 days post crush as reported [12]. Mice were anesthetized with avertin and placed under a heating lamp to avoid hypothermia. Sciatic NCVs were obtained by stimulating the nerve with steel monopolar needle electrodes. A pair of stimulating electrodes was inserted subcutaneously near the nerve at the ankle. A second pair of electrodes was placed at the sciatic notch, to obtain two distinct sites of stimulation, proximal and distal along the nerve. The muscular response to the electrical nerve stimulation (cMAP) was recorded with a pair of needle electrodes; the active electrode was inserted in muscles in the middle of the paw, while the reference was placed in the skin between the first and second digit.

Histological analysis of sciatic nerves from mice and sural nerves from human biopsies were performed as described [13]. For mouse morphometry, digitalized images of fiber cross sections were obtained from corresponding levels of the sciatic nerve with a digital camera (Leica DFC300F) using a 100× objective of light microscope (Olympus BX51). At least 5 images from 4 different animals per genotype at each evaluation were acquired (25×10³ µm² of sciatic nerve per each animal). The morphometry on semithin sections was analyzed with the Leica QWin software [12]. The ratio between the mean diameter of an axon and the mean diameter of the fiber including myelin (g-ratio), was determined on approximately 200 randomly chosen fibers per animal (3 animals per genotype).

These methods have been previously described [12]. Antibodies used for immunohistochemistry/Western blotting included: anti-calnexin and beta-tubulin TUB2.1 (Sigma); collagen IV, fibronectin, Fibrin, Mac1, Neurofilament-160, laminin-beta1, laminin-beta2 (Millipore); Laminin-alpha2 (4H8-2; Alexis); laminin-alpha4 (Santa Cruz); MBP (SMI-94R; Covance); vitronectin (BioTrend).

Adult mice were anesthetized with avertin (trichloroethanol, 0.4 mg/gr of body weight and crush injury performed as described [14]. After skin incision, the sciatic nerve was exposed and crushed distal to the sciatic notch for 20 seconds with fine forceps previously cooled in dry ice. To identify the site of injury, forceps were previously dropped into vital carbon. The nerve was replaced under the muscle and the incision sutured. All the experiments have been performed on the distal portion of the crushed nerve.

Mouse DRGs and cocultures were established as described [15]. Briefly, mouse DRGs were isolated from E13.5 embryos and placed in C-media containing MEM, 10% fetal calf serum, 2%glucose, 2 mM glutamine, purchased from Life Science and 50 ng/ml Nerve Growth Factor (NGF, from Serotec) on collagen coated glass coverslips. The day after the C-media was removed and replaced for 4–5 days with Neurobasal medium and B27 (vol 1∶1, Life Sciences) without serum and supplemented with NGF, glucose and glutamine as above. Glass coverslips were pre-coated with PLL (final concentration 0.1 mg/ml) and subsequently coated with collagen (rat collagen 0.5 mg/ml from Trevigen). Myelination was induced by treatment for 14 days with C-media supplemented by ascorbic acid (final concentration 50 mg/ml, Sigma).

Sural nerve samples from 24 patients undergoing nerve biopsy for diagnostic reasons were included in the study. Patient's characteristics are reported in Table 1. Patients were classified as regenerating or non-regenerating on the basis of their histological characteristics at the time of the diagnosis of axonal neuropathy as described [5]. In brief, we considered regenerating nerves those with abundant signs of regeneration (Index of regenerating axons was calculated as the number of regenerating clusters per mm² of sural nerve as follow: 0<40 clusters, 1 = 41–100, 2 = 101–200, 3>201). Severity of the neuropathy was assigned counting large/medium-sized (5–12 µm) myelinated axons in 6–8 representative areas of sural nerve (0.22–0.28 mm²); each sural nerve was classified accordingly to the mean number of large/medium-sized myelinated axons per mm² as follow: very severe <200 axons, severe = 201–500, moderate = 501–1400, mild >1401. The neuropathy was diagnosed by the presence of distal sensory disturbances with progressive weakness at four limbs, reduced/absent deep tendon reflexes and neurophysiology showing reduction of motor/sensory nerve action potentials with variable slowing of motor/sensory nerve conduction velocities. Patients underwent routine laboratory tests including screening for disimmune/inflammatory disorders, liver and kidney function, glucose intolerance, muscle enzymes, and cerebrospinal fluid examination. Two independent physicians performed clinical and histopathological evaluations blindly.

We determined the activation of the uPA/plasminogen system in frozen sections by in situ caseinolytic zymography using the EnzChek® protease green fluorescence assay kit (Invitrogen), following the protocol previously described [16]. This assay is based on the release of fluorescent BODIPY FL®-labeled peptides by plasmin-dependent caseinolysis. Sural nerve sections were incubated with BODIPY-FL® casein solution (10 µg/ml in Tris-HCl; Molecular Probes) for 3 h at 37°C. Caseinolytic activity in nerve sections generated fluorescein peptides that are viewed at confocal microscopy. To test specificity for uPA, sections were pre-incubated with a serine protease inhibitor such as 0.5 µM tPA-STOP (American Diagnostica) or amiloride (0.1 mM; Sigma) at 37°C for 1 h. Semiquantitative analysis of fluorescence staining was measured by ImageJ (NIH) on images of 0.22–0.26 mm² of sural nerve endoneurium viewed using confocal microscope (Leica SP5). For grading, we measured the mean gray value corresponding to the sum of the gray value of all pixels in the selected area divided by the number of pixels, defined as follows: 0 if mean <1, 0.5 if = 2–10, 1 if = 11–20, 2 if = 21–40, 3 if >41.

The uPA and tPA activity was evaluated in cell homogenate and conditioned (C- medium, serum free) media from DRG neuron-Schwann cell cocultures grown for 5, 10, 15 and 20 days with (or without) ascorbic acid (50 mg/ml), or sciatic nerve homogenate from 3 Wt and 3 uPAR−/− mice. Conditioned medium from DRG neuron-Schwann cell cocultures was collected after 6 h of incubation and centrifuged at 14,000 rpm for 5 min to remove detached cells and debris. Proteins in the medium were precipitated by the acetone/trichloroacetic acid procedure and were resuspended in lysis buffer (0.1M Tris-HCl pH 8.0, 0.2% Triton-X100). Cell monolayers were washed twice with ice-cold PBS and lysed on the tissue culture dish by addition of ice-cold lysis buffer. Frozen sciatic nerves dissected from uPAR−/− and wild type littermates were pulverized, sonicated in lysis buffer (0,1M Tris-HCl pH 8,0, 1% Triton-X100, 150 mM NaCl), boiled for 5 minutes, and spun at 14,000 rpm for 5 min at room temperature to eliminate insoluble material. Proteins precipitated from media, cells and nerve homogenate were normalized for protein concentration using Bio-Rad Bradford kit. Forty micrograms of Schwann cell/neuron coculture homogenate, or proteins precipitated from medium, or 25 µg of nerve extract were separated on non-reducing SDS/PAGE (10%), after the gels were washed in 2.5% Triton X100 for 45 minutes and subsequently placed in contact with a casein gel containing 2% non-fat dry milk, 0.01M Tris-HCl, 1.25% agarose, and plasminogen (6.25 ng/ml; Roche), at 37°C. After 16 hours, proteolytic activity areas appear as clear bands on casein gel. The polyacrylamyde gel containing nerve homogenate was stained in 0.125% (w/v) coomassie brilliant blue (G-250 Bio-Rad) in 34% methanol, 17% ammonium sulphate, 3,57% phosphoric acid and destained in 1% acetic acid solution as loading control and were analyzed by densitometry. As positive controls, human uPA (0.04 mM, bands 33 and 54 kDa) and rtPA (Actilyse® 0.01 mM, band 65 kDa) were used.

Statistical analyses were evaluated by two tail Student's t test.

We assessed whether uPAR loss affects nerve function and structure. Neurophysiological analysis of uPAR−/− mice showed normal nerve conduction velocities and compound motor action potentials (Figure 1D). Consistently, morphological analysis of sciatic nerves did not reveal differences in total fiber number (Wt 18452±270 fibers/mm² vs uPAR−/− 17750±679 fibers/mm²), fiber size distribution and myelination (Figure 1A–C and 1E).

As uPAR-uPA complex plays a role in ECM remodeling, we then investigated the endoneurial ECM composition. By immunohistochemistry we did not observe differences in the expression of laminins (α2, β1, β2, γ1), fibronectin, and collagen IV (data not shown). Instead, we observed increased deposition of fibrin and vitronectin in uPAR−/− mice (Figure 1F–K, and not shown). Western blot analysis confirmed increased amount of fibrin and vitronectin in uPAR−/− mice (Figure 1L). Thus, uPAR is dispensable for nerve development.

Since in vitro system may unravel defects masked by compensatory or redundant molecules during development in vivo, we exploited uPAR−/− Schwann cell-DRG neuron cocultures. As a preliminary experiment, we evaluated whether fibrinolytic activity is present and modulated in cocultures. Thus we evaluated tPA and uPA activity by zymography in organotypic DRG explants from wild type mice, in which myelination was induced by ascorbic acid. We found that both tPA and uPA activity are induced when myelination starts, and that uPA activity increased proportionally with myelination (Figure 2A). We then evaluated myelination of uPAR−/− DRG explants, measured as number of MBP positive segments. Myelination was not significantly different between uPAR−/− and Wt DRG explants seven days after ascorbic acid treatment (Figure 2B–F).

Overall, although uPA and tPA activity parallels myelination, suggesting a role in this process, uPAR is dispensable also in vitro for myelination during development.

As abnormal fibrin and vitronectin deposition affects nerve regeneration, we evaluated the response to nerve injury in uPAR−/− mice.

We performed sciatic crush injury in 12 uPAR−/− and 12 littermate control mice and analyzed peripheral nerve regeneration in the nerve stump distally to the site of the lesion at 15 and 45 days post crush (dpc). Fifteen dpc, both wild type and uPAR−/− nerves showed axonal degeneration, invading macrophages and few regenerating fibers. The number of regenerating fibers was significantly reduced (34%) in uPAR−/− nerves (Wt 5850±430/mm² vs uPAR−/− 3900±250/mm²; p = 0.02), and consistent for all fiber diameters (Figure 3A–C). Forty-five dpc, when Wt nerves are usually regenerated, uPAR−/− nerves still showed a significant reduced number (22%) of regenerating fibers (Wt 17800±889 fibers/mm² versus uPAR−/− 14007±260 fibers/mm²; p = 0.05); this was primarily evident for small diameter fibers (Figure 3D–F). Consistent with defective regeneration, we also observed that regenerating fibers in uPAR−/− mice had reduced myelin thickness, as measured by g-ratio (Figure 3G; n. 26, p = 0.01). Finally, neurophysiological analysis confirmed impaired regeneration as uPAR−/− mice displayed a significant reduction in NCV as compared to Wt mice at 45 dpc (Figure 3H; n. 8, p = 0.001).

Overall our results suggest that nerve repair is impaired in uPAR−/− mice. As uPAR has been associated to leukocyte migration [7], [17], especially into the nervous system during experimental autoimmune encephalomyelitis [18], we evaluated if impaired nerve repair in uPAR−/− mice was due to reduced macrophage recruitment in the nerve. In fact, macrophage activity is crucial for myelin clearance in damaged nerve to promote efficient nerve repair [1]. However, staining for macrophages using anti-Mac1 antibody did not show a reduced number of macrophages in uPAR−/− nerves (Figure 3I–J). The number of macrophages was similar at 15dpc, while significantly increased in uPAR−/− mice at 21dpc (p = 0.04) and 45dpc (p = 0.008; Figure 3I–K). An increased number of macrophages in uPAR−/− nerve would suggest a delay in clearance of nerve debris, thus in accordance with a delay in nerve regeneration.

We then assessed whether defect in nerve repair was mediated by abnormal ECM composition. We focused on fibrin, fibronectin and vitronectin, which are targets of the fibrinolytic complex. As expected, we observed increased expression of fibrin, vitronectin and fibronectin after nerve damage in the endoneurium of both Wt and uPAR−/− mice, although this expression was consistently higher in uPAR−/− mice (Figure 4A–F and Figure 5A; values are expressed as average of 3 experiments±SEM). At 45 dpc, when nerve should be almost recovered, Wt nerves showed roughly normal levels of all three ECM molecules, whereas fibrin and vitronectin levels were still increased in mutant nerves (Figure 4G–L and Figure 5A), and fibronectin levels markedly decreased (Figure 5A). Overall these results suggest that normal remodeling of ECM after damage is impaired in uPAR−/− mice, which consistently have higher levels of fibrin and vitronectin.

We therefore evaluated whether uPAR−/− mice have a defective function of the fibrinolytic complex. Nerve tissue from the crush site of 3 Wt and 3 uPAR−/− mice was obtained at 7 and 15 dpc and from non-injured nerves. tPA and uPA activity was determined by zymography (the experiment was performed 3 times). In non-injured nerve, uPA and tPA activity was similar between uPAR−/− and Wt mice. After nerve damage, uPA and tPA activity was modulated in Wt nerves, reaching a peak at 7dpc and decreasing at 15dpc (Figure 5B). In uPAR−/− mice instead, uPA activity was only mildly increased and not modulated (Figure 5B).

These results demonstrate that uPAR−/− mice have a reduced uPA activity after nerve damage, which sustains their inability to degrade fibrin.

To verify whether reduced fibrin deposition would improve nerve repair in uPAR−/− mice, we performed exogenous fibrinolysis by treating mice with recombinant (r)tPA (5 mg/ml; Actilyse®). We first verified that intra-peritoneal administration of rtPA (roughly 1 mg per mouse) was sufficient to remove fibrin in the endoneurium of uPAR−/− crushed sciatic nerves. Five uPAR−/− and 5 Wt mice have been treated with rtPA either other day from 1dpc and the sciatic nerve analyzed at 4dpc. Fibrin that is highly expressed in the endoneurium of non-treated mice was reduced in the endoneurium of rtPA treated mice (Figure 6A–D).

We then performed sciatic crush in 3 uPAR−/− and 3 Wt mice and treated these mice for 3 weeks, 2 times per week from 1dpc, with rtPA (5 mg/100 g). Forty-five dpc, motor nerve conduction velocities and cMAPs were similar in uPAR null mice and controls (Figure 6H). Accordingly, both wild type and uPAR−/− nerves showed quite complete nerve recovery, and we did not observe differences in total number of fibers (Wt 16300±1082 fibers/mm² vs uPAR−/− 17680±1090 fibers/mm²), fiber type distribution (Figure 6E–G) and myelin thickness (Figure 6I). Immunohistochemistry for fibrin expression showed similar pattern in Wt and uPAR−/− mice (Figure 6J–K).

Our previous findings indicate that nerve regeneration in human neuropathies is impaired in the presence of abnormal accumulation of fibrin and vitronectin in the endoneurium [5]. We investigated whether the fibrinolytic activity is involved in the abnormal deposition of fibrin and vitronectin in peripheral nerves of human neuropathies (Table 1). We analyzed 24 nerve biopsies from patients affected by peripheral neuropathy that, on histopathology evaluation, were divided in regenerative (patients #1–13) and non-regenerative (patients #14–24) neuropathies. On these nerve samples, we performed in situ zymography and observed that all the nerves showing higher fibrinolytic activity belong to the group of regenerating neuropathies (Table 1 and Figure 7). This result was independent from the diagnosis of the neuropathy. When an inhibitor of tPA or uPA activity was added before zymography, the fluorescent activity was reduced or abolished, confirming the specificity of the result Figure 7A^I–B^I. Consistently, nerves with signs of regeneration and high fibrinolytic activity showed significantly lower expression of fibrin and vitronectin (Table 1). On the contrary, patients with no signs of regeneration and low fibrinolytic activity showed significantly higher expression of fibrin and vitronectin (Table 1).

Overall these data suggest that efficient fibrinolytic activity remodels the endoneurial ECM in order to favor nerve regeneration.