NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Alzate O, editor. Neuroproteomics. Boca Raton (FL): CRC Press/Taylor & Francis; 2010.

Cover of Neuroproteomics

Neuroproteomics.

Show details

Chapter 15Applications of Proteomics to Nerve Regeneration Research

, , , , and .

15.1. PERIPHERAL NERVE INJURY: BACKGROUND AND CLINICAL SIGNIFICANCE

Peripheral nerve injury is a major clinical and public health challenge. Although a common and increasingly prevalent wartime condition (1), injury to peripheral nerves, plexuses, and roots is present in 5% of patients seen in civilian trauma centers (2). In one study, almost half of peripheral nerve injuries at trauma centers were due to motor vehicle accidents and about half required surgery (3). Peripheral nerve injuries can substantially impact quality of life through loss of function and increased risk of secondary disabilities from falls, fractures, and other injuries (2).

Neurons are connected in intricate communication networks established during development to convey sensory information from peripheral receptors of sensory neurons to the central nervous system (the brain and spinal cord), and to convey commands from the central nervous system to effector organs such as skeletal muscle innervated by motor neurons. The peripheral nerve environment is quite complex, consisting of axonal projections from neurons, supporting cells such as Schwann cells and fibroblasts, and the blood supply to the nerve.

Connective tissue known as endoneurium surrounds peripheral nerve axons. Within peripheral nerves, axons are grouped into fascicles surrounded by connective tissue known as perineurium. Between and surrounding groups of fascicles is the epineurium. Microvessel plexuses course longitudinally through the epineurium and send branches through the perineurium to form a vascular network of capillaries in the endoneurium (4).

The primary supporting cell for peripheral nerves is the Schwann cell. Schwann cells wrap around axons in a spiral fashion multiple times and their plasma membranes form a lipid-rich tubular cover around the axon known as the myelin sheath or the neurilemma. Schwann cells and the myelin sheath support and maintain axons and help to guide axons during axonal regeneration following nerve injury (5).

It has been known for quite some time that regenerating axons exhibit a strong preference for growing along the inside portion of remaining basal lamina tubes in the distal nerve stump, the well-characterized “bands of Bungner” (6–10). Schwann cells originally associated with myelinated axons form such bands all the way from the transection site to the distal end organ target. The critical concept here is that the eventual distal destination of regenerating axons is largely determined by the Schwann cell tubes as they enter at the nerve transection site (9,11,12).

Recent elegant work with transgenic mice expressing fluorescent proteins in their axons has verified that most (but not all) regenerating axons distal to either a crush or a transection injury remain within a single Schwann cell band as they grow within the distal nerve stump (13,14). Within the motor system it has even been shown that the same axon predominantly reinnervates the same neuromuscular junctions, and that Schwann cell bands act as mechanical barriers to direct axon outgrowth (13).

The neuronal cell body is the site of synthesis of virtually all proteins and organelles in the cell. A complex process known as anterograde transport continuously moves materials from the neuronal cell body via the axon to its terminal synapse. These transported substances include neurotransmitters that facilitate communications between the neuron and end organ tissue across a narrow extracellular space known as the synaptic cleft (5) or, as in the case of the motor neuron innervation of muscle, the neuromuscular junction (15).

Conversely, end organs such as muscle produce substances that act as nerve growth factors. These make their way across the neuromuscular junction to the innervating motor neuron axon (16). Some of these substances, or chemical messengers induced by them, are packaged and conveyed by retrograde transport from the synapse via the axon to the neuronal cell body. In this manner, the neuron and its end organ are continuously informed about the status of the connection between them. It has been suggested that information from end organs takes the form of factors that sustain existing nerve cell connections and promote the regeneration of damaged nerve cells. For instance, it has long been known that muscle exerts a strong influence on developing and regenerating motor neurons, and we have recently shown that even within an individual muscle there are factors that can influence the accuracy of reinnervation (17). Recent work has elegantly shown that if a single muscle fiber is selectively lesioned, the motor neuron axon terminal making up the proximal side of the neuromuscular junction rapidly atrophies and withdraws from the muscle postsynaptic sites within a matter of hours (18).

The clinical significance, prognosis, and treatment of peripheral nerve injury depend on the site and extent of the injury. Despite regeneration, extensive peripheral nerve injuries can result in the effective paralysis of the entire limb or distal portions of the limb. Two peripheral nerve injury classification schemes, the Seddon (19) and the Sunderland (20), are in common use. These classify nerve injury according to whether the injury was confined to demyelination only or a more severe disruption of axons and supporting connective tissue. According to Seddon, the most severe injuries are classified as axonotmesis and neurotmesis. Axonotmesis is a nerve injury characterized by axon disruption rather than destruction of the connective tissue framework. The connective tissue and Schwann tubes are relatively intact. This is typical of stretch injuries common in falls and motor vehicle accidents. In contrast, neurotmesis involves the disruption of the nerve trunk and the connective tissue structure. This would occur in injuries where the nerve has been completely severed or badly crushed.

Prognosis is good in peripheral nerve injuries where endoneurial Schwann cell tubes remain intact. Disruption of the Schwann cell tubes results in the loss of established pathways that regenerating axons follow. For extensive injuries, surgery is usually necessary to remove damaged nerve tissue and join viable nerve ends by direct anastomosis or by a nerve tissue graft (1). Refinement of microsurgical techniques involving the introduction of the surgical microscope and microsutures has increased the accuracy of this mechanical process, yet only 10% of adults will recover normal nerve function using state-of-the-art current techniques (21–23). The limits of microsurgical techniques have been reached; this is not surprising given that the finest suture material and needles (18–22 and 50–75 microns, respectively) are still quite a bit larger than the smallest axons that need to be repaired. The major key to recovery of function following peripheral nerve lesions is the accurate regeneration of axons to their original target end organs. A recognized leader of clinical nerve repair once stated, “The core of the problem is not promoting axon regeneration, but in getting them back to where they belong” (Sunderland, 1991) (23).

At the level of a mixed peripheral nerve where motor and sensory axons are intermixed, correct discriminatory choices for appropriate terminal nerve branches at the lesion site are necessary prerequisites for the subsequent successful reinnervation of appropriate end-organ targets. Motor axons previously innervating muscle may be misdirected to sensory organs, and sensory axons typically innervating skin can be misdirected to muscle. Misdirected regeneration is a major barrier to functional recovery.

In order to understand axonal regeneration and the mechanisms that axons use to navigate to target tissues, our laboratory has conducted a series of studies that are now culminating in proteomic investigations to identify specific biochemical mediators that may be the underlying mechanisms that direct accurate axon regeneration. We describe our work and that of others in the development of a model of axonal regeneration in the rodent femoral nerve and what we have learned from it. Then we will lay out our current research direction illustrating how approaches in proteomics such as two-dimensional differential gel electrophoresis (2D-DIGE) and mass spectrometry can be used to identify the underlying mediators that may lead to new therapies for peripheral nerve injury.

15.2. FEMORAL NERVE REGENERATION MODEL

15.2.1. Motor Axon Regeneration Accuracy

The rodent femoral nerve is an elegant model to examine motor neuron reinnervation accuracy. Two terminal nerve pathways roughly equal in size, one to the skin and the other to the quadriceps muscle, are intermixed at the level of the parent nerve, but bifurcate distally into distinct terminal nerve branches: a sensory cutaneous branch (continuing as the saphenous nerve and providing local cutaneous innervation) and a muscle branch to the quadriceps.

Weiss and Edds originally introduced the rat femoral nerve in 1945 as a model system to study the fate of axons that originally innervated muscle or skin when they were forcibly misdirected into the inappropriate nerve branch (24). When motor axons were forced into the cutaneous branch (and vice versa) via nerve crosses, axons were able to grow and survive in the foreign territory for extended periods of time. This led these researchers to conclude that neurotropism within the context of distal nerve stumps “has been ruled out conclusively” [(24), p. 173]. The original femoral nerve model has more recently been modified to look at the fate of regenerating axons when they are given equal access to the terminal cutaneous and muscle branches, rather than being forced into foreign nerve stumps. In contrast to the conclusions of Weiss and Edds (1945), the outcome of these equal access experiments showed a preference for regenerating motor axons to reinnervate their original terminal nerve branch, a process that has been termed “preferential” motor reinnervation (PMR) (Figure 15.1) (25,26).

FIGURE 15.1. Axon regeneration from motor neurons illustrating preferential motor reinnervation (PMR) in the femoral nerve.

FIGURE 15.1

Axon regeneration from motor neurons illustrating preferential motor reinnervation (PMR) in the femoral nerve. This nerve divides distally into two terminal nerve branches: a cutaneous branch and a motor branch to the quadriceps muscle. Normally, as shown (more...)

In the normal femoral nerve no motor neurons project into the cutaneous branch. Thus reinnervation of this distal branch by regenerating motor neurons represents a failure of specificity, which can be quantified by retrograde tracing. Previous work with this model system in both rats and mice when muscle and skin contact are maintained has revealed that motor axons initially grow equally into both nerve branches but over time are preferentially retained in the muscle branch, thus resulting in PMR (27–33). Conversely, we have recently shown that when muscle contact is denied but the cutaneous branch remains intact to skin, the cutaneous branch now becomes the preferred terminal nerve branch for motor neuron projections (33,34), suggesting that regeneration accuracy is highly dependent on the accessibility options allowed by a particular surgical preparation.

15.2.2. Investigating Relative Roles of Pathways and End Organs in Preferential Motor Reinnervation

The above studies indicate that both the end organ and the pathway can influence PMR. We have conducted a series of studies to assess the relative roles of these two sources of influence by developing surgical models that manipulate relative levels of trophic support in the muscle and cutaneous nerve branches. In all of these studies, the measurement of regeneration accuracy is the number of motor neurons that are retrogradely labeled from just one of the two terminal pathways, or simultaneously from both pathways. This is determined as follows. After the predetermined survival period following the initial proximal nerve repair (8 weeks in Figure 15.2), the two terminal nerve branches (which project either to muscle or to skin) are re-exposed and separated from each other by food-grade silicone grease dams, trimmed to ~3 mm distal to the normal parent nerve bifurcation, and randomly assigned to receive crystals of either fluorescein dextran (D-1820, Molecular Probes, Eugene, OR, USA) or tetramethylrhodamine dextran (D-1817, Molecular Probes). After crystal application, each branch is blotted and sealed with silicone grease, and the surgical site is closed in layers. Three days later, animals receive an overdose of anesthetic and are perfused through the heart with 0.1 M phosphate-buffered saline (PBS, pH 7.4) followed by 4% paraformaldehyde in PBS. The lumbar spinal cord is removed, post-fixed for several hours, and sucrose protected overnight. The cord is frozen on dry ice and stored at −80°C until being sectioned with a cryostat. Serial 25-μm frozen sections are thawed in PBS, mounted onto glass slides, briefly air-dried, and coverslipped with Prolong according to the manufacturer’s instructions (P-7481, Molecular Probes).

FIGURE 15.2. Data obtained eight weeks after femoral nerve repair in adult rats.

FIGURE 15.2

Data obtained eight weeks after femoral nerve repair in adult rats. (A) All animals received parent femoral nerve transection and repair. In the long muscle branch and short cutaneous branch model (LM-SC; left-most column) the cutaneous branch was also (more...)

Retrogradely labeled motor neurons containing a nucleus are identified using a composite filter set that allows simultaneous visualization of both labels (#51006, Chroma Technology, Brattleboro, VT, USA) in a fluorescence-equipped Zeiss Axiophot microscope, at 250× magnification. Labeling with one or both fluorophores is verified by single filters tailored to appropriate absorption and emission spectra (Chroma Technology). Motor neuron counts are carried out by blinded independent observers and scored as either single labeled (fluorescein or tetramethylrhodamine only) or double labeled (both fluorescein and tetramethylrhodamine). Neurons are tabulated based on staining as projecting to the muscle or cutaneous pathway only, or as projecting to both pathways. Counting variation among observers is ~2%. Neuron counts are corrected for split cells in microtome sections (35). Although there are certainly newer methods to correct for split cell counts, we use the Abercrombie method in order to relate our work to previous works in this field by other laboratories. Previously published work from our laboratory has documented the validity and reliability of these methods (e.g., lack of cross-contamination due to tracer leakage, intra- and inter-counter reliability, etc. (34,36).

One of our working hypotheses regarding the development of PMR in the femoral nerve is that there is a hierarchy of trophic support for regenerating motor neurons, with muscle contact being the highest, followed by the length of the terminal nerve branch and/or contact with skin. Because we hypothesized that muscle contact by regenerating motor axons results in the greatest level of trophic support, we examined the effect on PMR when muscle contact was allowed, but skin contact was prevented by capping the cutaneous branch; a preparation referred to as long-muscle:short cutaneous. This preparation allows trophic support from muscle and the muscle branch without the competing influence of contact with skin or a long cutaneous nerve branch. The results were an extremely robust bias for regenerating motor neurons to project to the muscle branch, with approximately a 9:1 ratio (Figure 15.2, far left column). Conversely, when we allowed skin contact but prevented muscle contact (a preparation referred to as short-muscle:long cutaneous) the ratio of regenerating motor neuron projections was reversed, with approximately a 1:3 ratio of motor neurons projecting to the muscle versus the cutaneous branch (far right column of Figure 15.2). The only surgical model that did not result in a significant difference in the number of motor neurons projecting to the two terminal nerve branches was when contact with both end organs was prevented, a preparation expected to make the level of trophic support in both terminal branches the most similar (referred to as short-muscle:short cutaneous).

These findings are consistent with the hypotheses that motor neurons compare relative levels of trophic support from various distal targets, and that there is a hierarchy of trophic support for regenerating motor axons with muscle contact being the highest, followed by the length of the terminal nerve branch and/or contact with skin.

These studies suggest several important aspects of PMR. Since initial axonal growth into the two terminal branches appears to be random, it is not likely that mechanical guidance plays a major role in PMR. The finding that the muscle branch becomes a more hospitable environment for motor axons over time suggests a role for attractive factors in the muscle branch or repulsive factors in the cutaneous branch.

We next focused on the time course of the development of preferential projections in the long-muscle:short cutaneous surgical model, where muscle contact is allowed, but skin contact is prevented by capping the cutaneous branch. We found that preferential projections developed as early as two weeks following the initial surgery [Figure 15.3, from (36)].

FIGURE 15.3. Development of pathway choice in adult rats after femoral nerve repair.

FIGURE 15.3

Development of pathway choice in adult rats after femoral nerve repair. When the influence of the cutaneous pathway/skin was removed, preferential reinnervation of the muscle pathway was more pronounced and was seen beginning at 2 weeks (mean ± (more...)

Given the rapidity with which preferential projections were established in the muscle branch (i.e., by two weeks), we decided to investigate the importance of maintaining the connection to muscle during this early regeneration time point. Given the fact that retrograde transport is known to continue within denervated rodent nerves for approximately 48 hours (37–42), and that the rate of retrograde transport is at least several millimeters/hour [ibid, see also (43)], muscle-derived signals could easily reach the proximal femoral nerve repair site if they were moved via retrograde transport. We therefore decided to investigate the role of retrograde transport distal to the proximal femoral nerve using the long-muscle:short-cutaneous model that demonstrated PMR at two weeks. The two-week time point is also important because we wanted to minimize the possible influence of regenerating motor neurons physically contacting muscle. By two weeks motor axons are just beginning to reach the quadriceps muscle and motor endplate reinnervation is very rare (27).

Colchicine disrupts microtubule function and has previously been used by many laboratories as a means of disrupting retrograde transport; it is typically used at a concentration of 1–50 niM [e.g., (44–47)]. At the same time as the proximal femoral nerve repair, we injected 3 µL of colchicine (25 mM, in saline) into the distal muscle branch using a pulled glass micro-pipette and a Hamilton syringe. As a control for possible damage due to the injection procedure another group of animals received injection of saline alone. A third group of animals received a crush of the distal muscle branch. The crush was carried out using fine needle holders with a 0.5-mm tip. Pressure was applied for 15 seconds, the orientation of the needle holder reversed, and pressure applied for another 15 seconds. The completeness of the crush procedure was assessed visually; the nerve crush site became completely translucent in all animals. Animals survived for two weeks and then were processed for retrogradely labeled motor neurons as described above.

It is important to keep in mind that the manipulations to the distal muscle branch are carried out immediately after the femoral nerve repair, during the same surgical exposure. Or, in other words, to a nerve that has already had its axons severed more proximally. Both the colchicine and crush groups failed to show PMR at two weeks, whereas the saline injection group did show PMR. The variability of the colchicine group was greater than the crush group as might be expected from a somewhat variable disruption of retrograde transport. The demonstration of PMR in the saline injection group repeats the initial finding above of PMR at two weeks using the long-muscle:short-cutaneous model (Figure 15.4).

FIGURE 15.4. PMR at two weeks in adult rats is dependent on retrograde transport from the denervated muscle branch.

FIGURE 15.4

PMR at two weeks in adult rats is dependent on retrograde transport from the denervated muscle branch. Animals received the long-muscle:short-cutaneous (LM-SC) surgical preparation. In addition, at the same time as the proximal femoral nerve repair the (more...)

Within the confines of this model system, these findings show that when retrograde transport is disrupted between the proximal femoral nerve lesion site and distal muscle, PMR fails to develop. The results of all of the above work suggest that contact with, or influence from, muscle is a major determinant of motor neuron regeneration accuracy.

15.2.3. Target-Derived Trophic Support and Trophomorphism

As briefly discussed above, a generalized concept of neuronal development that has received much interest over the past several decades has been that neurons require target-derived trophic support for survival (48–50). A related concept is that of trophomorphism, which suggests that competition among sibling axonal branches determines which axonal collaterals survive, and this is based on the relative amounts of trophic support to each of the axon collaterals (51,52).

Our findings are consistent with the hypothesis that trophomorphism plays a prominent role in PMR. This is especially the case since, during regeneration, axons from the regenerating motor neurons have access to both terminal nerve branches. We have shown that when muscle contact is allowed, significantly more motor neurons project to the distal muscle branch; however, when muscle end-organ contact is prevented, motor axons preferentially reinnervate the foreign cutaneous terminal pathway (Figure 15.2). This suggests that motor neurons assess the level of trophic support from each of the terminal branches and over time become preferentially located in the one that provides the greater amount of trophic support. This explanation is also consistent with the classical work of Campenot showing that neurons retain their axons in compartments with relatively higher levels of trophic support (53,54). Studies from other laboratories using Y-shaped tubes also offer support for the assertion that motor neurons assess relative levels of trophic support and respond accordingly. Regenerating axons have been shown to distinguish between nerve and non-nerve targets, suggesting tissue specificity with respect to attracting and retaining regenerating axons. These studies have also shown that distal nerve volume influences axonal growth (55).

This strong evidence from surgical experiments is consistent with models where trophomorphism plays a prominent role in PMR, especially for motor neurons. Trophic support implies the presence of soluble factors migrating from end organs in a retrograde fashion. Interestingly, it has been known for almost 30 years that the endocytotic activity of skeletal muscle increases after denervation (56). Subsequent work demonstrated a pairing between the increased endocytotic activity with significant increases in exocytotic activity (57). These observations have led to the suggestion that the high exocytotic activity of denervated muscle is due to the increased secretion of neurotrophic factors by the muscle in an attempt to stimulate reinnervation by regenerating axons (58–62). Consistent with such an interpretation is the finding that the expression levels of several neurotrophic factors are indeed increased following either nerve or muscle damage [e.g., hepatocyte growth factor (HGF), fibroblast growth factor (FGF), brain-derived neurotrophic factor (BDNF) (63–69)]. The increased endocytotic/exocytotic activity occurs predominantly at the endplate region of the muscle (70), where it would be ideally situated to pass muscle-derived factors across the neuromuscular junction to the preterminal motor axon.

A major ongoing component of our work focuses on the identification of mediators of trophomorphism within the femoral nerve model. We are taking advantage of novel techniques developed in the field of proteomics. In the remainder of this chapter we consider the various proteomic methodologies and explore how they are used in our laboratory to understand biochemical mediators of axonal regeneration and PMR.

15.3. ROLE OF PROTEOMICS IN IDENTIFYING MEDIATORS OF AXONAL REGENERATION AND PREFERENTIAL MOTOR REINNERVATION

15.3.1. RESEARCH Goals and Relevance of Proteomics

Findings from surgical models and other evidence described previously are consistent with a prominent role for unknown trophic mediators in PMR. It is thought that end organs produce soluble protein trophic factors that either promote or inhibit axonal growth. Levels of these factors are believed to be altered in response to nerve injury, making a terminal nerve branch either more or less attractive to regenerating motor neurons. Trophic factors may be expressed by supporting cells such as Schwann cells, but previously described experimental evidence suggests that muscle tissue has a particularly strong trophic effect.

The objective of our current investigations is to identify the soluble proteins produced by muscle tissue in response to denervation. Specifically, we are looking for proteins that demonstrate changes in expression or are released secondarily to nerve damage. We reason that some subset of these may be involved in axonal regeneration and PMR. This task has only recently become technically feasible with the development of powerful proteomic techniques for biomarker discovery and protein identification.

Biomarker discovery involves the comparison of entire proteomes of sample tissues that differ according to some known process. In our case, we will compare proteomes of normal tissue to those of tissue involved in nerve injury. The objective is to identify protein expression differences that enable us to discover small subsets of proteins somehow involved in nerve regeneration.

15.3.2. Power and Limitations of Proteomics

Fundamental challenges in biomarker discovery for neuroproteomics research include the immense size and variability of the neuroproteome, phenotypic variations as cells interact with their environment, and the wide range of relative protein abundances. The size of the proteome even within a single cell is immense. Lefkovits et al. (71) have estimated that a single B-lymphoblast cell contains 109protein molecules of 4000 different molecular species. Frequently, complex tissues rather than single cell lines are examined, further increasing the number of protein species. The estimated size of the proteome, all proteins potentially expressed from the genome, is far larger and may exceed 100,000 in humans (72). This implies great potential phenotypic variation between tissues and within cell lines under varying conditions. Add to this the variation introduced by post-translational modification and the magnitude of the challenge becomes truly impressive. Proteomic techniques must also manage the wide range of protein abundances where just a few hundred protein species predominate, and species most likely of interest to researchers are relatively rare. The abundance range comparing the most common to the more rare protein species has been estimated to span seven to eight orders of magnitude (71).

Historically protein chemistry focused on one or a few proteins of interest. Laboratory techniques such as chromatography, electrophoresis, and affinity columns were developed to identify, quantify, and characterize individual proteins. But proteomics required the development of laboratory techniques, information systems, and statistical approaches to process hundreds to thousands of proteins simultaneously. The theoretical advantage of proteomics is in the ability to simultaneously examine all proteins in a biological system so that relations among them may be explored under different conditions. This is a theoretical advantage because, in practice, no proteomic technique has the resolution to examine all proteins, especially those in low abundance. These undetectable low-abundance proteins my make up 70% of the expressed proteome (71).

During the early 1960s, Norman G. Anderson proposed the Molecular Anatomy (MAN) program to catalog the human proteome (73,74). Parts of this program were initiated at Oak Ridge National Laboratory and included many of the objectives of proteomic programs today such as the identification of all cellular proteins, their polymorphisms, their structure and functions, their location, and their chemical properties. Limitations in technology, funding, and a public focus on the Human Genome Project delayed the realization of this dream. In many respects we are only now beginning to realize the ambitious goals set out by Anderson almost a half century ago.

The goals of the MAN project were likely ahead of their time. Their realization awaited the development of techniques that enable large numbers of proteins to be processed simultaneously. In the mid-1970s such a breakthrough occurred in the form of two-dimensional gel electrophoresis (2-DE) (75) (for a discussion of 2-DE see Chapters 3 and 4). By the end of that decade, 2-DE was greatly refined and standardized, leading to proposals for its use in the development of a human protein index (HPI) consistent with the basic concepts of the MAN project (71). Proteomics has been described as the “rebellious child of 2-DE” (73), suggesting a prominent role for 2-DE as a foundation technique in the field. Although automated and refined, the basic concept of 2-DE has changed little since it was first described (75). In his seminal manuscript, O’Farrell reported the ability to resolve a maximum of 5000 proteins with 2-DE (75). This was the first technique to simultaneously evaluate large numbers of proteins, and it led to the advances that make modern protein biomarker discovery possible.

Although a breakthrough, 2-DE has many limitations that make it less than optimal in modern proteomic investigations in biomarker discovery. As proteins migrate across the gel they separate based on molecular weight and chemical properties. The location, shape, and size of the resulting protein spots are greatly influenced by experimental conditions. This is critically important because protein identification and quantification depend on the ability to locate spots on the gel and to measure their exact size. Standardizing experimental protocols and equipment and controlling for variation among gels, experiments, and laboratories have been major challenges. This is especially true when different experimental samples are compared on different gels. This lack of reproducibility among gels is a major limitation of 2-DE that has been solved with the introduction of 2D-DIGE (see Chapter 4, and Section 15.4.2). A number of computational approaches to deal with this problem have been developed with limited success because the sources of gel-to-gel variation are numerous, complex, and difficult to model (76).

The limitations of 2-DE in its ability to resolve or detect low-abundance proteins and to distinguish proteins clustered near high-abundance spots could, in theory, be overcome by mass spectrometry techniques such as surface-enhanced laser desorption/ionization (SELDI) or matrix-assisted laser desorption/ionization (MALDI) mass spectrometry. But in practice the promise of the superiority of mass spectrometry over electrophoresis has yet to be realized without extensive pre-fractionation of protein samples. However, mass spectrometry does offer an alternative to the gels and liquids required for 2-DE (see Chapters 5, 6, and 7 for detailed discussions of mass spectrometry techniques in neuroproteomics).

Although the technical challenges are far from resolved, proteomic methods have evolved rapidly in recent years, making it feasible to simultaneously evaluate expression differences in hundreds of proteins in tissues under different experimental conditions. Our ongoing research is utilizing two well-established, widely accepted, and complementary proteomic techniques: SELDI mass spectrometry and 2D-DIGE. The goal of our research is to use these techniques to discover and identify the soluble protein mediators expressed by muscle tissue that play a role in axonal regeneration and PMR following peripheral nerve injury.

15.3.3. Pilot Experiments Using SELDI Mass Spectrometry

We hypothesized that denervated muscle is likely to increase production of specific proteins that are attractive to, and help provide support for, motor neurons and their axons. We have performed pilot experiments examining this hypothesis using SELDI to identify biomarkers associated with denervation of skeletal muscle. Our initial experiments were conducted to demonstrate proof of principle for our proteomics approach.

For our pilot work using SELDI we compared the expression profiles from two groups of rat muscles: 10 normal quadriceps and 10 four-day denervated quadriceps. We refined harvesting and protein-processing protocols to be compatible with the SELDI instrument, focusing first on procedures that simultaneously sample wide categories of proteins (e.g., membrane, cytosolic, etc.). The experiment is outlined in Figure 15.5.

FIGURE 15.5. Experimental flow chart for the pilot protein expression data generated by the SELDI instrument.

FIGURE 15.5

Experimental flow chart for the pilot protein expression data generated by the SELDI instrument.

Dilute solutions of the skeletal muscle lysate are applied directly to spots on the protein chip arrays (Ciphergen, Inc., Fremont, CA, USA). The protein chips differ according to the chemical properties of the spot surfaces. The spots are then washed with a series of buffers of different stringencies. Binding of proteins to these spots is a function of the protein affinity to the spot surface, wash stringency/type, and wash parameters (e.g., time and frequency). In this manner, skeletal muscle proteins are fractionated prior to processing in the ProteinChip reader. The presence of chemically active surfaces on the chips distinguishes SELDI from MALDI, which uses passive surfaces. Like MALDI, SELDI also utilizes an energy-absorbing molecule (matrix) that absorbs laser energy, leading to the desorption and ionization of proteins on spot surfaces. Ciphergen provides chips with surfaces that are cationic, anionic, hydrophobic, normal phase, and capable of binding to metals, antibodies, or bait molecules.

The ProteinChip reader contains the SELDI mass spectrometer with a protein chip loader controlled by an external computer. Under computer control, a UV laser is fired at a limited area of each spot. The laser ionizes bound substances and causes them to fly from the surface. These ionized proteins are accelerated through an electrical field in a vacuum, and then are allowed to drift for a distance prior to encountering a multiple-stage electron multiplier ion detector (for more on ion detectors refer to Chapter 5).

The time of flight (TOF) of a substance depends on its charge and mass. Each laser shot generates a “spectrum” of data points that relate TOF to detector signal intensity. Spectra consist of intensity measures for each TOF value. For ease of interpretation, Ciphergen software converts TOF to the ratio mass/charge (m/z). Spectra can be pre-processed using ProteinReader software to adjust for baseline drift and to normalize across samples within a specified m/z range.

Operational parameters modifiable via SELDI ProteinChip Reader software include laser intensity, detector sensitivity, detector voltage, time-lag focusing, spot position, and number of laser shots per position. These are critical parameters that have a major impact on the quality and reproducibility of results. Our experiments include replicates of homogenized tissue samples for multiple animals within each experimental group. In pilot experiments, for example, 60 protein chip spots were required (10 animals × 2 experimental groups run in triplicate). There are 20 laser positions on each spot. Each of these positions is used to generate a spectrum that may be an average of as many as 99 laser shots. Thus, this experiment could generate 1200 spectra containing information from almost 120 laser shots. Obviously this is too much information to be processed manually, and automated analytic techniques were developed and utilized.

We adjusted for systematic variations across samples, spots, and spot positions with respect to TOF and intensity through the use of known protein standards and pooled experimental samples. Protein standard solutions are available from Ciphergen and are supported by software capabilities that adjust m/z estimates determined from TOF by applying correction factors based on proteins of known molecular weight. Experimental samples are pooled and applied to protein chips for analysis under a variety of possible combinations of laser intensity, detector sensitivities, and detector voltages.

We developed a SAS (SAS, Inc., Cary, NC) macro to generate spot and chip protocols that randomize each combination of laser intensity, detector voltage, and detector sensitivity to multiple laser positions, spots, and chips to identify optimum values of these parameters for pooled samples (i.e., all experimental samples combined). The protocols were generated in the XML format used by the ProteinChip reader. These protocols generate four spectra for each of 12 possible combinations of parameters. Data from the resulting 48 spectra are analyzed using SAS software to identify the parameter values that yield the best signal intensity and the least noise. These parameter values are then used in spot protocols to read protein chips containing individual (i.e., non-pooled) experimental samples. This approach substantially increases reproducibility among samples.

Data from the pre-processed spectra (i.e., baseline corrected and normalized) are analyzed by Ciphergen software to identify spectral peaks. Peaks that are common to a pre-specined proportion of spectra within a narrow molecular weight range are grouped by Ciphergen software into clusters (Figure 15.6). Clusters can be thought of as sets of measures of protein expression (i.e., intensity) across spectra for a substance of a given molecular weight.

FIGURE 15.6. Example of reproducibility of SELDI spectra from triplicate spots on protein chip CM-10.

FIGURE 15.6

Example of reproducibility of SELDI spectra from triplicate spots on protein chip CM-10. Ciphergen software identifies common peaks across samples and groups them into “clusters.” These clusters can then be analyzed across experimental (more...)

The objective of cluster analysis is to identify significant differences in protein expression between experimental groups. In this manner, we have identified several potential biomarkers distinguishing denervated muscle from normal muscle, and these proteins would be candidates for purification using Ciphergen columns, combined with on-chip protein digestion to verify protein identity via peptide fingerprinting, as described in Chapter 5.

15.3.4. Lessons Learned from SELDI Pilot Studies

We were early adopters of SELDI technology. There were few established protocols useful for our specific experimental conditions and tissues at the time of our initial experiments. As we learned more about the technology we found that results were highly sensitive to experimental conditions and variations in tissue sample characteristics and processing. We have found it essential to prepare tissue samples consistently with well-defined and uniform protein concentrations across samples. It is also important to ensure an appropriate experimental design with internal controls allowing the calibration of the instrument. We found substantial variability among individuals within experimental groups and among tissue samples within individuals, suggesting the need to increase experimental group size or refine experimental techniques to reduce that variability.

The resource commitments to refine protocols in order to achieve reproducible results were substantial and difficult to meet in a laboratory not specializing in this technique. We had available within our institution an established proteomics laboratory that was able to produce 2D-DIGE results at lower costs and higher quality than we could achieve with SELDI in our own laboratory. For these reasons we chose to complete this line of research using 2D-DIGE-based proteomics as an optimal choice. Nevertheless, our work with SELDI was informative and provided a foundation for our current work.

15.4. DISCOVERY OF TROPHOMORPHIC MEDIATORS WITH 2D-DIGE

15.4.1. Overview

Our pilot biomarker discovery investigations using SELDI suggested that there are protein expression differences comparing normal to denervated quadriceps muscle. This is consistent with experimental evidence such as our laboratory surgical findings showing that muscle tissue exerts strong trophomorphic effects on regenerating motor axons. We hypothesize the presence of soluble protein trophic factors that are released by denervated muscle tissue at the neuromuscular junction and diffuse in a retrograde fashion through the nerve tract to the site of peripheral nerve injury (Figure 15.7B). These factors then stimulate and direct the regeneration of motor axons to their appropriate target organ.

FIGURE 15.7. 2D-DIGE results of individual terminal nerve branches of the rat femoral nerve, (a) and (b) Diagram of the two terminal nerve branches of the rat femoral nerve, showing the sensory cutaneous branch (continuing as the saphenous nerve and providing local cutaneous innervation) and the muscle branch to the quadriceps muscle.

FIGURE 15.7

2D-DIGE results of individual terminal nerve branches of the rat femoral nerve, (a) and (b) Diagram of the two terminal nerve branches of the rat femoral nerve, showing the sensory cutaneous branch (continuing as the saphenous nerve and providing local (more...)

To identify candidate muscle proteins involved in axonal regeneration and PMR, we are implementing a series of biomarker discovery and protein identification experiments using two-dimensional difference gel electrophoresis (2D-DIGE). We have completed and describe here proof-of-concept pilot studies that demonstrate findings consistent with our expectations. Currently we are developing a series of comprehensive studies to identify candidate trophic factors using a more rigorous study design.

15.4.2. Two-Dimensional Difference Gel Electrophoresis

A major limitation of the 2-DE technique is gel-to-gel variation leading to poor reproducibility among gels. Two-dimensional difference gel electrophoresis (2D-DIGE) was developed to overcome this limitation (76). Rather than comparing samples run on separate gels as in 2-DE, in 2D-DIGE samples are run on the same gel and are distinguished using different fluorescent dyes. The different dyes are virtually identical with respect to characteristics that influence migration in the gel. Therefore, they do not differentially influence migration of their bound proteins. Identical proteins in one sample superimpose on their differentially labeled counterparts for the other sample in the gel (see Chapter 4).

Our biomarker discovery research to identify trophomorphic mediators in nerve regeneration uses 2D-DIGE-based proteomics. This is a powerful approach to evaluate differential protein expression in comparison tissue samples. Protein samples are covalently labeled with one of three fluorescent dyes. All three labeled samples are mixed and then run on a two-dimensional gel. An internal control is used to normalize the concentrations of each sample across gels and biological replicates. This control is usually a pooled mixture of equal aliquots of all samples (77). The relative abundance of each protein from samples is evaluated based on spot intensity for each CyDye.

Proteins with similar expression levels in the comparison samples will have the same relative concentrations and will thus create spots that appear white on the gel. Spots containing proteins that are expressed differently in one sample relative to the others will tend to take on the color of the dye associated with the highest expressed sample. Optical reading of the gel and evaluation with analytic software identify proteins and statistically evaluate expression differences.

15.4.3. Pilot Experiments with 2D-DIGE

Our proof-of-concept pilot experiment with 2D-DIGE first required the development of protein isolation methods that would work with limited samples of peripheral nerve (e.g., single rat terminal muscle branches). Initial attempts at protein isolation from such small samples of nerve (<5 mg wet weight) proved to be problematic. Peripheral nerve is notoriously difficult to homogenize due to the tough myelin and connective tissues, and the smaller the sample the more difficult it is to get consistent results. Quite a bit of pilot work was carried out with current popular protocols, and these were all disappointing for various reasons. An extensive protein extraction procedure was finally tailored after that of the 1979 procedure of Groswald and Luttges (78) and proved to be quite reliable. This procedure has probably fallen out of fashion because it is a multi-day, multi-step protocol; however, it proved to be quite robust and consistent.

A typical experiment involved rats in three experimental groups. Some animals received two nerve ligations, one just distal to the parent femoral nerve bifurcation and one just proximal to the quadriceps muscle (Figure 15.7A). Other animals received a single nerve ligation just distal to the parent femoral nerve bifurcation (Figure 15.7B). We also included a control group of naive animals receiving no surgical procedures.

The rationale for this experimental design was that both sets of samples receiving ligation would have proteins from degenerating axons due to the proximal ligation, but only the samples from the group with a single ligation would contain muscle-derived proteins. Our hypothesis is that muscle-derived proteins are critical for determining the accuracy of regenerating motor neurons. These would migrate in a retrograde fashion from the denervated muscle via the neuromuscular junction to the site of the first ligation by a combination of diffusion and retrograde transport persisting in axons for a short time after injury.

Individual muscle branch samples (<5 mg wet weight) were harvested 7 days after the ligations were placed on the nerve. Nerve samples were harvested from naive animals at that time as well. Nerve samples were placed in liquid N2, and stored at −80°C until they were transported to the proteomics laboratory where they were immediately processed. Total time from sacrifice to dissection is less than 10 minutes per rat.

For proteomics analysis nerves were rinsed in PBS, centrifuged, the supernatant discarded, and lysis buffer was added. The tissue was crushed and homogenized on ice with a sonicator, vortexed, centrifuged, and the protein supernatant was saved. The protein concentration was measured with a 2D-Quant kit (GE Healthcare, Piscataway, NJ, USA). We obtained a yield on average of 100 ng of protein per nerve, an amount sufficient for 2D-DIGE analyses. We labeled 40 µg each of the three samples (i.e., normal, single-ligation, double-ligation) with one of the CyDyes (Cy2, Cy3, Cy5, respectively). Two independent experiments were carried out with different nerve samples from different animals. Labeled samples from single animals in each of the three experimental groups were combined and run on the same gel.

The first dimension of the 2D-DIGE was run on an IPG strip 13 cm, pH 3–10 (GE Healthcare). We used a separation pH range from 3 to 10, and the conditions were as follows: active rehydration at 30 V, step voltage up to 500 V for 1 hour, step voltage up to 1000 V for 1 hour, finally a step voltage up to 8000 V until it reaches a total of 26,000 KVh. The second dimension (molecular weight) used a standard 12% Polyacrylamide gel with sodium dodecyl sulfate (SDS-PAGE) stacking gel.

The fluorescently labeled spots (proteins) generated from the 2D-DIGE resolved gels were imaged with a Typhoon 9410 (GE Healthcare) Variable Mode Imager (Figure 15.7). Because the excitation and emission energy for each dye is specific, the imager can digitally separate them, visualized as three colors (blue Cy2, green Cy3, and red Cy5). With these images, we performed automated analysis of individual spots as peaks characterized by their intensity and shape with the DeCyder (GE Healthcare) Differential Analysis software for proteomics (see Chapter 4). A ratio threshold is set by determining variation of comparable spots among the samples. We used a two-fold increase or decrease in spot intensity as a cutoff for pair-wise comparisons. Selected spots were cut from the gels with an Ettan spot picker, and then transferred into 96-well plates for further digestion and spotting for protein identification using liquid chromatography tandem mass spectrometry (LC-MS/MS) to identify protein fingerprints and peptide sequences (as described in Chapter 5).

The combined image of a representative gel from one experiment is shown in Figure 15.7C. One can quickly appreciate protein spots that are more highly expressed in either normal nerve (appears as blue), double ligated nerve (green), or single ligated nerve (red), as well as subsets of proteins from two samples in relation to the third, e.g., yellow equals more highly expressed in single- and double-ligated nerves compared to normal. Spots that appear white are equally expressed in all three samples.

Since our hypothesis centers on proteins originating from muscle, we focused our attention on spots that are more highly expressed in the single-ligated nerve with an intact connection to muscle (i.e., red spots in this gel). DeCyder 6.5 Differential Analysis software was used for automated analysis of individual relative spot intensities and shape. This system characterizes spots based on parameters such as area, maximum peak, maximum volume, and relative abundance. It produces quantitative (intensities) and qualitative (graphical) representations for protein expression comparisons.

Two independent sets of experimental groups in two 2D-DIGE gels were run and analyzed as described. About 2700 individual protein spots were resolved on each gel. Relative abundance measures based on spot volume ratios were used to identify proteins of interest. Expression was judged to be similar where the ratio of the highest spot volume to the lowest spot volume was 2.5 or less. More than 80% of the resolved spots were similar based on this criterion. Expression was elevated in double-ligated samples compared to single-ligated samples in about 6% of resolved spots, and the maximum spot volume ratio was less than seven for these. In about 10% of the spots, expression was higher in the single-ligated sample compared to the double-ligated sample. The magnitudes of the spot volume ratios in this case were much higher (maximum = 238). The spot volume ratio exceeded 100 for more than 10 of these single-ligated high expression spots.

The relatively high abundance of multiple proteins in single-ligated versus double-ligated nerve samples is consistent with our hypothesis that denervated muscle produces substances that migrate in a retrograde fashion from the muscle to the site of nerve injury. Any of these proteins can be considered a candidate mediator of axonal regeneration and PMR. The determination of which, if any, actually are mediators of trophomorphism is an active area of investigation in our laboratory.

A graphical example of a marked expression difference favoring single-ligated nerves is shown in two images of Figure 15.7C on the far right. A second independent gel with nerve samples from different animals verified the increased expression of this same spot in the single-ligated nerves. This spot was subsequently picked by a robot from both gels and subjected to protein sequencing with LC-MS/MS and identified as vimentin both times, a marker of reactive Schwann cells, which are known to have a large influence on nerve regeneration.

15.5. SUMMARY AND FUTURE DIRECTIONS

Our proof-of-concept pilot experiments using two prominent techniques in proteomics, SELDI and 2D-DIGE, have demonstrated that patterns of protein expression in skeletal muscle are influenced by denervation and that protein products from denervated muscle are conveyed in a retrograde fashion from muscle to the site of nerve injury. These findings are consistent with prior research in our laboratory and from other groups that muscles produce soluble protein mediators that control and direct axonal regeneration following nerve injury.

Ongoing and future research will utilize these and other proteomic techniques to identify and quantify all candidate biomarkers potentially related to axonal regeneration and PMR. Once candidates are identified, their individual effects will be tested in vitro and in experimental animals to evaluate the therapeutic potential of these factors in nerve repair. One can envision that this work might identify substances that could be introduced in muscle distal to recently injured nerve. These substances would enhance and guide axonal regeneration to appropriate end organs, thus reducing the substantial deficits experienced by persons with peripheral nerve injury. We are likely far from the realization of that dream, but neuroproteomics offers a promising pathway toward that goal.

ACKNOWLEDGMENTS

Supported by the NIH (to RDM, NS061106) and the Office of Research and Development, Biological Laboratory Research and Development (BLRD) Service, Department of Veterans Affairs (to RDM). RDM is a research career scientist for the BLRD service.

REFERENCES

1.
Campbell W. W. Evaluation and management of peripheral nerve injury. Clin Neurophysiol. 2008;119:1951–65. [PubMed: 18482862]
2.
Robinson L. R. Traumatic injury to peripheral nerves. Muscle Nerve. 2000;23:863–73. [PubMed: 10842261]
3.
Noble J., Munro C. A., Prasad V. S., Midha R. Analysis of upper and lower extremity peripheral nerve injuries in a population of patients with multiple injuries. J Trauma. 1998;45:116–22. [PubMed: 9680023]
4.
Sunderland S. The anatomy and physiology of nerve injury. Muscle Nerve. 1990;13:771–84. [PubMed: 2233864]
5.
Ross M. H., Reith E. J., Romrell L. J. Histology. Baltimore: Williams and Wilkins; 1989. Nervous tissue; pp. 241–82.
6.
Farel P. B., Meeker M. L. Developmental regulation of regenerative specificity in the bullfrog. Brain Res Bull. 1993;30:483–90. [PubMed: 8457898]
7.
Holmes W., Young J. Z. Nerve regeneration after immediate and delayed suture. J Anat. 1942;77(10):63–96. [PMC free article: PMC1252769] [PubMed: 17104917]
8.
Ide C., Tohyama K., Yokota R., Nitatori T., Onodera S. Schwann cell basal lamina and nerve regeneration. Brain Res. 1983;288:61–75. [PubMed: 6661636]
9.
Lee M. T., Farel P. B. Guidance of regenerating motor axons in larval and juvenile bullfrogs. J Neurosci. 1988;8:2430–7. [PubMed: 3266877]
10.
Scherer S. S., Easter S. S. Jr. Degenerative and regenerative changes in the trochlear nerve of goldfish. J Neurocytol. 1984;13:519–65. [PubMed: 6481411]
11.
Brown M. C., Hardman V. J. A reassessment of the accuracy of reinnervation by motoneurons following crushing or freezing of the sciatic or lumbar spinal nerves of rats. Brain. 1987;110:695–705. 3. [PubMed: 3580830]
12.
Brown M. C., Hopkins W. G. Role of degenerating axon pathways in regeneration of mouse soleus motor axons. J Physiol. 1981;318:365–73. [PMC free article: PMC1245496] [PubMed: 7320895]
13.
Nguyen Q. T., Sanes J. R., Lichtman J. W. Pre-existing pathways promote precise projection patterns. Nat Neurosci. 2002;5:861–7. [PubMed: 12172551]
14.
Witzel C., Rohde C., Brushart T. M. Pathway sampling by regenerating peripheral axons. J Comp Neurol. 2005;485:183–90. [PubMed: 15791642]
15.
Grinnell A. D. Dynamics of nerve-muscle interaction in developing and mature neuromuscular junctions. Physiol Rev. 1995;75:789–834. [PubMed: 7480163]
16.
Fox M. A., Sanes J. R., Borza D. B., et al. Distinct target-derived signals organize formation, maturation, and maintenance of motor nerve terminals. Cell. 2007;129:179–93. [PubMed: 17418794]
17.
Chadaram S. R., Laskowski M. B., Madison R. D. Topographic specificity within membranes of a single muscle detected in vitro. J Neurosci. 2007;27:13938–48. [PubMed: 18094231]
18.
McCann C. M., Nguyen Q. T., Santo Neto H., Lichtman J. W. Rapid synapse elimination after postsynaptic protein synthesis inhibition in vivo. J Neurosci. 2007;27:6064–7. [PubMed: 17537978]
19.
Seddon H. Three types of nerve injury. Brain. 1943;66:237–88.
20.
Sunderland S. A classification of peripheral nerve injuries producing loss of function. Brain. 1951;74:491–516. [PubMed: 14895767]
21.
Brushart T. M. Nerve repair and grafting. In: Green D., editor. Green’s operative hand surgery. New York: Churchill Livingston; 1998. pp. 1381–1403.
22.
Madison R. D., Archibald S. J., Karup C. Peripheral nerve injury. In: Cohen I. K., Diegelman F., Lindblad W. J., editors. Wound healing: Biochemical and clinical aspects. Philadelphia: W. B. Saunders Co; 1992. pp. 450–80.
23.
Sunderland S. Nerve injuries and their repair: A critical appraisal. New York: Churchill Livingstone; 1991.
24.
Weiss P., Edds M. V. Sensory-motor nerve crosses in the rat. J Neurophysiology. 1945;8:173–93.
25.
Brushart T. M. Preferential reinnervation of motor nerves by regenerating motor axons. J Neurosci. 1988;8:1026–31. [PubMed: 3346713]
26.
Brushart T. M., Gerber J., Kessens P., Chen Y. G., Royall R. M. Contributions of pathway and neuron to preferential motor reinnervation. J Neurosci. 1998;18:8674–81. [PubMed: 9786974]
27.
Brushart T. M. Preferential motor reinnervation: A sequential double-labeling study. Restor Neurol Neurosci. 1990;1:281–87. [PubMed: 21551568]
28.
Brushart T. M. Motor axons preferentially reinnervate motor pathways. J Neurosci. 1993;13:2730–8. [PubMed: 8501535]
29.
Franz C. K., Rutishauser U., Rafuse V. F. Polysialylated neural cell adhesion molecule is necessary for selective targeting of regenerating motor neurons. J Neurosci. 2005;25:2081–91. [PubMed: 15728848]
30.
Madison R. D., Archibald S. J., Brushart T. M. Reinnervation accuracy of the rat femoral nerve by motor and sensory neurons. J Neurosci. 1996;16:5698–5703. [PubMed: 8795625]
31.
Mears S., Schachner M., Brushart T. M. Antibodies to myelin-associated glycoprotein accelerate preferential motor reinnervation. J Peripher Nerv Syst. 2003;8:91–9. [PubMed: 12795713]
32.
Robinson G. A., Madison R. D. Preferential motor reinnervation in the mouse: Comparison of femoral nerve repair using a fibrin sealant or suture. Muscle Nerve. 2003;28:227–31. [PubMed: 12872328]
33.
Robinson G. A., Madison R. D. Manipulations of the mouse femoral nerve influence the accuracy of pathway reinnervation by motor neurons. Exp Neurol. 2005;192:39–45. [PubMed: 15698617]
34.
Robinson G. A., Madison R. D. Motor neurons can preferentially reinnervate cutaneous pathways. Exp Neurol. 2004;190:407–13. [PubMed: 15530879]
35.
Abercrombie M. Estimation of nuclear population from microtome sections. Anat Rec. 1946;4:239–46. [PubMed: 21015608]
36.
Uschold T., Robinson G. A., Madison R. D. Motor neuron regeneration accuracy: balancing trophic influences between pathways and end-organs. Exp Neurol. 2007;205:250–6. [PubMed: 17368445]
37.
Hassig R., Tavitian B., Pappalardo F., Di Giamberardino L. Axonal transport reversal of acetylcholinesterase molecular forms in transected nerve. J Neurochem. 1991;57:1913–20. [PubMed: 1940908]
38.
Miledi R., Slater C. R. Electrophysiology and electron-microscopy of rat neuromuscular junctions after nerve degeneration. Proc R Soc Lond B Biol Sci. 1968;169:289–306. [PubMed: 4384567]
39.
Miledi R., Slater C. R. On the degeneration of rat neuromuscular junctions after nerve section. J Physiol. 1970;207:507–28. [PMC free article: PMC1348721] [PubMed: 5499034]
40.
Smith R. S., Bisby M. A. Persistence of axonal transport in isolated axons of the mouse. Eur J Neurosci. 1993;5:1127–35. [PubMed: 8281318]
41.
Stanley E. F., Drachman D. B. Denervation and the time course of resting membrane potential changes in skeletal muscle in vivo. Exp Neurol. 1980;69:253–9. [PubMed: 7409044]
42.
Watson D. F., Glass J. D., Griffin J. W. Redistribution of cytoskeletal proteins in mammalian axons disconnected from their cell bodies. J Neurosci. 1993;13:4354–60. [PubMed: 8410191]
43.
Curtis R., Scherer S. S., Somogyi R., et al. Retrograde axonal transport of LIF is increased by peripheral nerve injury: Correlation with increased LIF expression in distal nerve. Neuron. 1994;12:191–204. [PubMed: 7507340]
44.
McGraw T. S., Mickle J. P., Shaw G., Streit W. J. Axonally transported peripheral signals regulate alpha-internexin expression in regenerating motoneurons. J Neurosci. 2002;22:4955–63. [PubMed: 12077192]
45.
McPhail L. T., Oschipok L. W., Liu J., Tetzlaff W. Both positive and negative factors regulate gene expression following chronic facial nerve resection. Exp Neurol. 2005;195:199–207. [PubMed: 15935349]
46.
Murphy P. G., Borthwick L. S., Johnston R. S., Kuchel G., Richardson P. M. Nature of the retrograde signal from injured nerves that induces interleukin-6 mRNA in neurons. J Neurosci. 1999;19:3791–3800. [PubMed: 10234011]
47.
Richardson P. M., Verge V. M. The induction of a regenerative propensity in sensory neurons following peripheral axonal injury. J Neurocytol. 1986;15:585–94. [PubMed: 3772404]
48.
Easter S. S. Jr., Purves D., Rakic P., Spitzer N. C. The changing view of neural specificity. Science. 1985;230:507–11. [PubMed: 4048944]
49.
Levi-Montalcini R. The nerve growth factor 35 years later. Science. 1987;237:1154–62. [PubMed: 3306916]
50.
Purves D. Body and brain. Cambridge, MA: Harvard University Press; 1988.
51.
Crutcher K. A., Saffran B. N. Developmental remodeling of neuronal projections: Evidence for trophomorphism? Comments Developmental Neurobiology. 1990;1:119–41.
52.
Saffran B. N., Crutcher K. A. NGF-induced remodeling of mature uninjured axon collaterals. Brain Res. 1990;525:11–20. [PubMed: 2245317]
53.
Campenot R. B. Development of sympathetic neurons in compartmentalized cultures. II. Local control of neurite survival by nerve growth factor. Dev Biol. 1982;93:13–21. [PubMed: 7128928]
54.
Campenot R. B. NGF and the local control of nerve terminal growth. J Neurobiol. 1994;25:599–611. [PubMed: 8071664]
55.
Takahashi Y., Maki Y., Yoshizu T., Tajima T. Both stump area and volume of distal sensory nerve segments influence the regeneration of sensory axons in rats. Scand J Plast Reconstr Surg Hand Surg. 1999;33:177–80. [PubMed: 10450574]
56.
Libelius R., Lundquist I., Templeton W., Thesleff S. Intracellular uptake and degradation of extracellular tracers in mouse skeletal muscle in vitro: The effect of denervation. Neuroscience. 1978;3:641–7. [PubMed: 724113]
57.
Vult von Steyern F., Kanje M., Tagerud S. Protein secretion from mouse skeletal muscle: Coupling of increased exocytotic and endocytotic activities in denervated muscle. Cell Tissue Res. 1993;274:49–56. [PubMed: 7694802]
58.
deLapeyriere O., Henderson C. E. Motoneuron differentiation, survival and synaptogenesis. Curr Opin Genet Dev. 1997;7:642–50. [PubMed: 9388781]
59.
Henderson C. E., Bloch-Gallego E., Camu W., et al. Motoneuron survival factors: Biological roles and therapeutic potential. Neuromuscul Disord. 1993;3:455–8. [PubMed: 8186693]
60.
Libelius R. Lysosomes in skeletal muscle. In: Libelius R., editor. Neuromuscular junction. Elsevier; 1989. pp. 481–85.
61.
Magnusson C., Svensson A., Christerson U., Tagerud S. Denervationinduced alterations in gene expression in mouse skeletal muscle. Eur J Neurosci. 2005;21:577–80. [PubMed: 15673457]
62.
Thesleff S., Libelius R. Neuromuscular stimulation: Basic concepts and clinical applications. New York: Demos Publications; 1989. Some aspects of long term regulations of nervemuscle relations.
63.
Jennische E., Ekberg S., Matejka G. L. Expression of hepatocyte growth factor in growing and regenerating rat skeletal muscle. Am J Physiol. 1993;265:C122–8. [PubMed: 8338120]
64.
Lie D. C., Weis J. GDNF expression is increased in denervated human skeletal muscle. Neurosci Lett. 1998;250:87–90. [PubMed: 9697925]
65.
Tonra J. R., Curtis R., Wong V., et al. Axotomy upregulates the anterograde transport and expression of brain-derived neurotrophic factor by sensory neurons. J Neurosci. 1998;18:4374–83. [PubMed: 9592114]
66.
Wallenius V., Hisaoka M., Helou K., et al. Overexpression of the hepatocyte growth factor (HGF) receptor (Met) and presence of a truncated and activated intracellular HGF receptor fragment in locally aggressive/malignant human musculoskeletal tumors. Am J Pathol. 2000;156:821–9. [PMC free article: PMC1876854] [PubMed: 10702398]
67.
Wehrwein E. A., Roskelley E. M., Spitsbergen J. M. GDNF is regulated in an activity-dependent manner in rat skeletal muscle. Muscle Nerve. 2002;26:206–11. [PubMed: 12210384]
68.
Yamaguchi A., Ishii H., Morita I., Oota I., Takeda H. mRNA expression of fibroblast growth factors and hepatocyte growth factor in rat plantaris muscle following denervation and compensatory overload. Pflugers Arch. 2004;448:539–46. [PubMed: 15118860]
69.
Zhao C., Veltri K., Li S., Bain J. R., Fahnestock M. NGF, BDNF, NT-3, and GDNF mRNA expression in rat skeletal muscle following denervation and sensory protection. J Neurotrauma. 2004;21:1468–78. [PubMed: 15672636]
70.
Lawoko G., Tagerud S. High endocytotic activity occurs periodically in the endplate region of denervated mouse striated muscle fibers. Exp Cell Res. 1995;219:598–603. [PubMed: 7543856]
71.
Lefkovits I., Kettman J. R., Frey J. R. Global analysis of gene expression in cells of the immune system I. Analytical limitations in obtaining sequence information on polypeptides in two-dimensional gel spots. Electrophoresis. 2000;21:2688–93. [PubMed: 10949147]
72.
Harrison P. M., Kumar A., Lang N., Snyder M., Gerstein M. A question of size: The eukaryotic proteome and the problems in defining it. Nucleic Acids Res. 2002;30:1083–90. [PMC free article: PMC101239] [PubMed: 11861898]
73.
Anderson N. G., Matheson A., Anderson N. L. Back to the future: The human protein index (HPI) and the agenda for post-proteomic biology. Proteomics. 2001;1:3–12. [PubMed: 11680895]
74.
Weinberg A. M. The birth of Big Biology. Nature. 1999;401:738. [PubMed: 10548091]
75.
O’Farrell P. H. High resolution two-dimensional electrophoresis of proteins. J Biol Chem. 1975;250:4007–21. [PMC free article: PMC2874754] [PubMed: 236308]
76.
Viswanathan S., Unlu M., Minden J. S. Two-dimensional difference gel electrophoresis. Nat Protoc. 2006;1:1351–8. [PubMed: 17406422]
77.
Friedman D. B., Lilley K. S. Optimizing the difference gel electrophoresis (DIGE) technology. Methods Mol Biol. 2008;428:93–124. [PubMed: 18287770]
78.
Groswald D. E. Changes in sciatic nerve protein composition during postnatal development of mice. Dev Neurosci. 1979;2:51–64.
Copyright © 2010 by Taylor and Francis Group, LLC.
Bookshelf ID: NBK56006PMID: 21882439

Views

  • PubReader
  • Print View
  • Cite this Page

Other titles in this collection

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Similar articles in PubMed

See reviews...See all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...