Wallerian degeneration and axonal regeneration after ... - Springer Link

6 downloads 0 Views 509KB Size Report
Aug 6, 2009 - nerve crush are altered in ICAM-1-deficient mice. Matthias Kirsch & Marianella Campos Friz &. Vassilios I. Vougioukas & Hans-Dieter Hofmann.
Cell Tissue Res (2009) 338:19–28 DOI 10.1007/s00441-009-0837-3

REGULAR ARTICLE

Wallerian degeneration and axonal regeneration after sciatic nerve crush are altered in ICAM-1-deficient mice Matthias Kirsch & Marianella Campos Friz & Vassilios I. Vougioukas & Hans-Dieter Hofmann

Received: 29 April 2009 / Accepted: 26 June 2009 / Published online: 6 August 2009 # Springer-Verlag 2009

Abstract The intercellular cell adhesion molecule-1 (ICAM-1) has been implicated in the recruitment of immune cells during inflammatory processes. Previous studies investigating its involvement in the process of Wallerian degeneration and focusing on its potential role in macrophage recruitement have come to controversial conclusions. To examine whether Wallerian degeneration is altered in the absence of ICAM-1, we have analyzed changes in the expression of axonal and Schwann cell markers following sciatic nerve crush in wildtype and ICAM-1-deficient mice. We report that the lack of ICAM-1 leads to impaired axonal degeneration and regeneration and to alterations in Schwann cell responses following sciatic nerve crush. Degradation of neurofilament protein, the collapse of axonal profiles, and the re-expression of neurofilament proteins are substantially delayed in the distal nerve segment of ICAM-1-/- mice. In contrast, the degradation of myelin, as determined by immunostaining for myelin protein zero, is unaltered in the mutants. Upregulation of GAP-43 and p75 neurotrophin receptor (p75NTR) expression, characteristic for Schwann cells dedifferentiating in M.K. and M.C.F. contributed equally to this work. M. Kirsch : M. Campos Friz : H.-D. Hofmann Institute of Anatomy and Cell Biology, University of Freiburg, Albertstrasse 23, 79104 Freiburg, Germany M. Campos Friz : V. I. Vougioukas Department of Neurosurgery, University Hospital Freiburg, Breisacher Strasse 64, 79106 Freiburg, Germany H.-D. Hofmann (*) Institute of Anatomy and Cell Biology, P.O. Box 111, 79001 Freiburg, Germany e-mail: [email protected]

response to nerve injury, is differentially altered in the mutant animals. These results indicate that ICAM-1 is essential for the normal progression of axonal degeneration and regeneration in distal segments of injured peripheral nerves. Keywords ICAM-1 . Sciatic nerve . Crush-lesion . Neurofilament . GAP-43 . Mouse (ICAM-1-/-; C57BL/6)

Introduction Traumatic injury of a peripheral nerve evokes a stereotypical reaction called Wallerian degeneration (WD), which provides the appropriate cellular and molecular environment for subsequent regenerative axon growth in the distal nerve stump. The distal axon segments degenerate rapidly within 1–2 days in an active process of self-destruction (Raff et al. 2002; Coleman 2005; Saxena and Caroni 2007). In response to axonal degeneration, Schwann cells dedifferentiate and start to proliferate massively; this is accompanied by the disintegration of the myelin sheath, which is initially incorporated by Schwann cells (Stoll et al. 1989; Hirata et al. 1999; Hirata and Kawabuchi 2002). The cells rapidly downregulate myelin protein synthesis (LeBlanc and Poduslo 1990) and upregulate the expression of proteins, such as growth factors, neurotrophins, cyokines, and cell adhesion molecules that promote axon regeneration (Stoll and Müller 1999; Araki et al. 2001). Signals from degenerating axons and Schwann cells lead to the activation and recruitment of macrophages, which, in turn, produce factors that are thought to promote the proliferation and dedifferentiation of the Schwann cells (Gillen et al. 1997; Stoll et al. 2002). The final degradation of axonal remnants and myelin fragments by phagocytic cells including

20

Schwann cells and resident and hematogenous macrophages (Brück et al. 1996; Hirata and Kawabuchi 2002) is thought to be important for facilitating the regrowth of regenerating axons into the distal nerve stump. Whereas the histological features of WD have been described extensively, the underlying cellular and molecular mechanisms remain largely unknown. The surface glycoprotein, intercellular cell adhesion molecule-1 (ICAM-1), a member of the immunoglobulin superfamily, is expressed primarily by endothelial cells and interacts with receptors expressed by lymphocytes and macrophages (Rothlein et al. 1986). It is involved in immune processes such as lymphocyte extravasation and macrophage recruitment (Springer 1990; Butcher 1991). Evidence has been presented that ICAM-1 plays a role during WD, but the findings are ambiguous. Upregulation of intraneural ICAM-1 protein expression has been described to occur following peripheral nerve lesion (Castano et al. 1996; Avellino et al. 2004) but has not been observed in another study (Stoll et al. 1993). Interestingly, Schwann cells have been shown to be induced to express ICAM-1 by pro-inflammatory cytokines, both in vitro (Lisak and Bealmear 1997; Lilje and Armati 1997; Constantin et al. 1999) and in vivo (Shen et al. 2008). Interference with ICAM-1 function by application of antibodies against ICAM-1 or against CR3/Mac-1 has been reported to impair (Lunn et al. 1989; Brück and Friede 1990) or to enhance macrophage invasion into the injured peripheral nerve (Brown et al. 1997; Avellino et al. 2004). Similarly, the number of macrophages recruited during WD has been described to be reduced (Vougioukas et al. 1998) or slightly increased (Avellino et al. 2004) in ICAM-deficient mouse mutants. Morphological analysis has indicated that axon degeneration and myelin degradation are delayed in these animals (Vougioukas et al. 1998; Siebert and Brück 2003). Thus, many open questions remain regarding the role of the ICAM-1 molecule following nerve lesion. To investigate whether ICAM-1 is of importance during the processes of WD and peripheral nerve regeneration, we have analyzed the changes in the level and location of marker proteins for axons and myelin and for Schwann cell responses following sciatic nerve crush in ICAM-/- mice. The time course of axon degeneration and regrowth in the distal nerve segment has been investigated by immunoblotting and immunocytochemistry for neurofilament (NF) proteins and growth-associated protein 43 (GAP-43). In the same way, myelin protein zero (P0) was used as a marker for myelin degradation, and the expression of p75 neurotrophin receptor and GAP-43, which are both induced in Schwann cells after nerve injury (Heumann et al. 1987; Sensenbrenner et al. 1997), for the differentiation state of Schwann cells. Our results indicate that the lack of ICAM-1 does not influence the clearance of myelin but leads to the

Cell Tissue Res (2009) 338:19–28

retardation of both axon degradation and regrowth of NFcontaining axons and to altered Schwann cell responses in the injured peripheral nerve.

Materials and methods Animals and surgery Knockout mice for ICAM-1 (strain B6.129S4-Icam1tmJcgr/J) and C57BL/6 mice recommended as an appropriate control strain (The Jackson Laboratory) were compared. Surgery and subsequent treatment of the animals were performed in accordance with German guidelines for animal experiments under surveillance of the local authorities. Adult animals, 2–3 months of age, were anesthetized with a mixture of 25% Ketamin (Parke-Davis; 100 mg/ml), 6% Xylazine (Bayer; 20 mg/ml), and 2.5% Acepromazine (Albrecht; 10 mg/ml) in physiological saline at a dose of 2.5 ml/kg body weight. The right sciatic nerve was exposed above the femur and crushed at mid-thigh level for 1 min with forceps of 1-mm tip width. The lesion site was marked by loosely ligating it with a thin thread, and the wound was closed by suturing the muscles and skin. Immunoblot analysis At 4, 6, 10, 14, and 28 days after nerve crush, animals were killed by placing them in a CO2 atmosphere, and the lesioned sciatic nerve was dissected. A 2-mm section containing the lesion site was removed, and accurately measured 2-mm nerve pieces containing adjacent proximal and distal parts, respectively, were dissected and immediately frozen in liquid nitrogen. For each time point, at least three animals were examined and their nerve sections were analyzed separately. Equally sized pieces from contralateral unoperated sciatic nerves served as controls. Tissue pieces were lyzed in sample buffer containing protease inhibitors (Complete Mini, Roche) and 0.1 mM Na3VO4, frozen and thawed on ice three times, and homogenized by trituration with a Hamilton syringe fitted with a 26-gauge needle. Samples were cleared by centrifugation, and equal aliquots of the nerve extracts were separated by SDS-polyacrylamide gel electrophoresis on 5%–20% gradient gels and transferred to polyvinylidene difluoride membranes (Millipore). The protocol for immunostaining and detection by chemiluminescence was as described in detail previously (Martin et al. 2003). Primary antibodies used were as follows: monoclonal mouse antibody against 68-kDa NF protein (1:1000; clone NR4, Sigma Aldrich), monoclonal mouse antibody against phosphorylated 160-kDa/200-kDa NF protein (1:1000; Abcam), rabbit anti-p75NTR (1:1000; Promega), rabbit

Cell Tissue Res (2009) 338:19–28

anti-GAP-43 (1:500; Chemicon), and monoclonal mouse anti-P0 (1:2000; Society to Support Research, Diagnosis and Treatment of Neurological Disorders, Austria). Alkaline-phosphatase-coupled goat anti-mouse or antirabbit IgG antibodies (1:10,000; Applied Biosystems) were used as secondary antibodies, and detection was performed with the CPD-Star chemiluminescence kit (Applied Biosystems). Chemiluminescence was recorded by using the Fujifilm LAS-3000 imaging system. Each blotting membrane was probed sequentially for the four marker proteins. Between the various immunostainings, membranes were stripped by incubating them in 0.1 M TRIS buffer (pH 6.7) containing 2% SDS for 45 min at 60°C. Densitometric quantification was performed with the AIDA 1D software package (Raytest, Straubenhardt, Germany). Endogenous standard proteins as usually used for the normalization of blot data are not applicable in the injured nerve where the degradation of axons and myelin sheaths occur in parallel with changes in the differentiation state of Schwann cells and other cells. Therefore, to compare the time course of NF and P0 degradation in wildtype and ICAM-/- mice, values obtained from injured nerve segments were normalized to pooled extracts from unlesioned contralateral nerves (28 days postlesion), which were obtained from wildtype and ICAM-1-/- mice, respectively. Statistical analysis was performed by two-way analysis of variance (ANOVA) followed by the Bonferroni post-test by using GraphPad Prism (GraphPad Software, San Diego).

21

and brightness by applying the same correction factors for images from wildtype and ICAM-1-/- animals.

Results

Immunocytochemistry

The early phase of WD is characterized by the rapid disintegration of the distal axon segment within 2 days and the degradation of NF proteins, which is completed between 3 and 7 days after peripheral nerve lesion (Terada et al. 1998; Zhai et al. 2003). This was confirmed when we analyzed extracts from the sciatic nerve of wildtype mice 10 days after nerve crush by immunoblotting with two antibodies recognizing the 68-kDa low-molecular-weight NF protein (NF-L) and phosphorylated NFs (mainly the 200-kDa subunit), respectively (Fig. 1). In wildtype animals, signals for the NF proteins had decreased to undetectable levels distal to the lesion site, whereas protein levels in proximal segments were not changed compared with uncrushed control nerves. In contrast, NF protein levels had decreased less markedly in the distal nerve segment of ICAM-1-/- mice, NF degradation not being completed by 10 days postlesion. This difference between wildtype and ICAM-1-/- mice was observed with both NF antibodies used, indicating that axonal degeneration was retarded in these mutants. GAP-43 has been described as a marker for both growing axons and non-myelinating Schwann cells (Sensenbrenner et al. 1997). Expression of GAP-43 mRNA has been shown to be upregulated in dedifferentiating Schwann cells in the

For immunofluorescence staining, 1-cm sections of the sciatic nerves, containing the lesion site of the injured nerves, were dissected, fixed in 4% paraformaldehyde overnight, and cryoprotected in 20% sucrose solution. The nerves were embedded in Tissue-Tek, and 10-μm longitudinal sections were cut on a Leica cryotome. For comparison, sections from wildtype and ICAM-/- animals were processed in parallel for immunofluorescence as described previously (Martin et al. 2003) . Unspecific binding sites were blocked by incubation with 10% appropriate normal serum, 1% Triton X-100 in 0.1 M phosphate buffer for 1 h. The sections were then incubated overnight at 4°C with the following primary antibodies: rabbit pan-anti-NF (1:1000; BioTrend), rabbit anti-p75NTR (1:300; Promega), rabbit anti-GAP-43 (1:300; Chemicon), or monoclonal mouse anti-P0 (1:400; GFN, Graz, Austria). Secondary Cy3coupled donkey anti-rabbit or anti-mouse IgG antibodies (1:400; Dianova) diluted in phosphate-buffered saline/1% normal serum were applied for 1 h. Stained sections were photographed with a Leica DFC 359X digital camera by using identical exposure times for all sections labeled for the same marker. The images were optimized for contrast

Fig. 1 Immunoblot analysis of axonal and Schwann cell markers following crush-lesion of the sciatic nerve. Extracts from nerve segments proximal (p) and distal (d) to the lesion site were analyzed by immunoblotting with antibodies against 68-kDa neurofilament (NF) protein (NF-L), phosphorylated 160-kDa and 200-kDa NF proteins (pNF-H; the 200-kDa band is shown), and growthassociated protein 43 (GAP-43) at 10 days after nerve crush. Three animals each of wildtype (WT) and the ICAM-1-/- genotype (ICAM-1-/-) are documented. Signals for NF proteins remained unchanged in the proximal nerve segments. They had completely disappeared in distal segments of the wildtype mice but were still clearly visible in the mutants. In ICAM-1-/- mice, lesion-induced GAP-43 signals of similar intensity were present in distal and proximal segments, whereas distal segments of wildtype mice showed no or only weak GAP-43 expression (con control)

22

distal segment of injured peripheral nerves with a delay of 1– 2 weeks after lesion (Curtis et al. 1992; Scherer et al. 1994; Awatramani et al. 2002). Analysis of GAP-43 expression by immunoblotting showed that the protein was upregulated to a similar extent in the proximal segment of crushed sciatic nerves from wildtype and ICAM-1-/- mice within 10 days after lesion (Fig. 1). In the distal nerve segments of ICAM-1-/- mice, GAP-43 expression was consistently observed at this time point suggesting that the response of Schwann cells was also altered in the absence of ICAM-1. In wildtype animals, upregulation was observed in a minority of the mice examined but was absent in the others. A similar variability was seen in another set of day-10 wildtype mice; this may have been because GAP-43 upregulation started around this postlesional time point in wildtype mice (see below). The difference in GAP-43 regulation seemed to indicate that Schwann cell responses might also be altered in ICAM-1-/mutants. To compare WD in wildtype and ICAM-1-/- mice in more detail, we studied the time course of NF degradation and the regeneration of NF-positive axons in the nerve segment distal to the site of injury. Immunoblotting demonstrated that NF-L levels were minimal as early as 4 days after crush (Fig. 2a, b). They remained low for at least 10 days in wildtype animals and then increased to reach more than 80% of the level found in unlesioned control nerves after 4 weeks (Fig. 2b). The time course was significantly different in ICAM-deficient mice, the degradation of NF-L protein not being completed before 10– 14 days after lesion. However, at 4 weeks postlesion, NF-L levels were virtually identical to those in wildtype mice indicating that axonal regeneration was retarded but not impaired in the mutants. Immunocytochemistry with antibodies recognizing all NF subunits confirmed the delay of NF degradation in animals lacking ICAM-1. Without lesion, NF immunoreactivity of sciatic axons appeared to be unaltered in ICAM-1-/- mice (Fig. 2c, f). By 6 days after nerve crush, the distal nerve of ICAM-1-/- mice contained significantly more NF-immunoreactive material in the form of brightly labeled aggregates demonstrating that the decomposition of NF proteins was still under way (Fig. 2d, e, g, h). In addition, many NF-positive axonal profiles, which were not seen in the normal nerves at this postlesional stage, were preserved in the mutant nerves. These results indicated that the disintegration of the distal severed axons and the degradation of NF-proteins were delayed by more than 1 week in the absence of ICAM-1 without affecting the extent of regeneration. We also studied the time course of GAP-43 upregulation in the distal nerve segment during a 4-week period following nerve crush. In wildtype mice, GAP-43 protein levels were low in unlesioned control nerves and during the

Cell Tissue Res (2009) 338:19–28

first 6 days after lesion (Fig. 3a, b). They had increased significantly by day 10 and were highest after 4 weeks, the latest time point studied. GAP-43 expression in ICAM-/mice was significantly higher at early postlesional stages (Fig. 3a, b). They increased to almost maximum levels within 4 days after nerve injury and remained relatively constant thereafter (Fig. 3a, b). At late postlesional time points (28 days postlesion), upregulation was more pronounced in wildtype mice. Immunocytochemical observations were in agreement with the immunoblot data. In both mouse strains, GAP-43 immunoreactivity was not detectable in control nerves (Fig. 3c) but was induced proximal to the lesion site (Fig. 3d, f; compare with Fig. 1). The labeling pattern corresponded to that demonstrated previously for non-myelinating Schwann cells of the sciatic nerve by double-labeling with glial fibrillary acidic protein (Curtis et al. 1992). GAP-43-immunoreactive regenerating axons that were expected to be present were not discernible on the background of labeled Schwann cells. Immunostaining of sections from distal nerve segments reflected the quantitative differences between wildtype and ICAM-/mice as determined by immunoblotting at 6 days after nerve lesion (Fig. 3e, g). In the mutants, GAP-43 labeling was strong and located in Schwann cells that frequently formed bands of Büngner, whereas Schwann cell labeling was weak in wildtype nerves. Immunoreactive axonal profiles indicating nerve regeneration were identifiable in wildtype nerves (Fig. 3e) but were not discernible in nerves from mutants. This agreed with the delayed axonal regeneration in ICAM-/- mice indicated by the NF expression data (Fig. 2), although axonal GAP-43 immunoreactivity might have been masked by the brightly labeled Schwann cells in these animals. The differences in injury-induced GAP-43 upregulation indicated that Schwann cell reactions during WD are altered in the absence of ICAM-1. To investigate this further, we studied the expression of the low affinity neurotrophin receptor p75NTR, which is not expressed in cells of the normal adult peripheral nerve but is induced in dedifferentiating Schwann cells following nerve injury (Taniuchi et al. 1986). Since available antibodies to p75NTR turned out not to be suited for quantitative immunoblot analysis, we focused on immunocytochemical visualization of p75NTR expression. Immunoreactivity for p75NTR was weak in proximal segments of crushed sciatic nerves both from wildtype and ICAM-/- mice, and a few spindle-shaped cells probably representing Schwann cells (Taniuchi et al. 1986) were recognizable (Fig. 4a, d). In segments distal to the crush site, labeling intensity was markedly enhanced 6 days after lesion in wildtype mice. The labeled cells had the characteristic appearance of proliferating Schwann cells forming bands of Büngner (Fig. 4b, c). Injury-induced increase of p75NTR immunoreactivity was also seen in

Cell Tissue Res (2009) 338:19–28

23

Fig. 2 Time course of degradation and re-expression of NF proteins following sciatic nerve crush in wildtype and ICAM-1-/- mice. a Immunoblots of distal nerve segments from crush-lesioned sciatic nerves stained with antibodies against 68-kDa NF protein at 4, 6, 10, and 14 days postlesion (dpl); nerves from two to three animals are shown at each time point (WT wildtype, con control). b Densitometric quantification of 68-kDa NF protein levels in wildtype (open bars) and ICAM-1-/- (filled bars) mice analyzed by immunoblotting. Values were normalized to signals measured in unlesioned control nerves and are given as mean±SD (n=3; not all animals evaluated are shown in a).

Statistical analysis by two-way ANOVA indicated significant differences between wildtype and ICAM-1-/- mice (P