VU Research Portal
Gene therapy to repair the injured nerve
Eggers, R.
2020
document version
Publisher's PDF, also known as Version of record
Link to publication in VU Research Portal
citation for published version (APA)
Eggers, R. (2020). Gene therapy to repair the injured nerve.
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal ?
Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
E-mail address:
vuresearchportal.ub@vu.nl
SCOPE AND OUTLINE
7
Scope and outline
Performing a seemingly simple task such as reaching for a cup of coffee requires our nervous system to perform an intricate interplay between sensory input and motor output. To precisely position the hand towards a coffee cup, motoneurons located in the pyramidal tract and in the spinal cord are provided with sensory feedback from the eyes, joints, muscles and skin and subsequently transmit coordinated signals via motor axons towards the arm musculature. When, following a traumatic nerve lesion, this delicate balance becomes disrupted, picking up a cup turns into an impossible task. As a result of the nerve lesion, damaged axons can no longer transmit signals from the spinal cord towards the hand musculature, resulting in a loss of function.
Depending on the location and severity of the lesion, spontaneous function recovery is frequently unsatisfactory. A neurotmetic lesion such as a spinal root avulsion is the most severe proximal nerve lesion possible. Root avulsion lesions are typically part of a brachial plexus traction injury which occurs during traffic accidents and complicated childbirth. Following an avulsion lesion, the rupture of nerve root filaments from the surface of the spinal cord leads to a combined central and peripheral nervous system lesion. This lesion is often not limited to only one nerve root, but consists of the avulsion of multiple roots. Despite neurosurgical repair, axonal regeneration over long distances is limited and the degree of recovery of function in patients suffering a brachial plexus lesion often remains poor. This results in lifelong dysfunction and pain. Thus, in order to regain useful function following neurosurgical repair, supplementary regenerative treatment strategies are required. Glial cell line-derived neurotrophic factor (GDNF) is a compelling treatment candidate due to its role in neuronal differentiation and identification as a potent motoneuron survival and axon outgrowth factor. Furthermore, in motoneurons following axotomy, the GDNF receptors c-RET and GFRα-1 are strongly upregulated. However, GDNF and other neurotrophic factors have a short half-life, exhibit poor tissue penetration and systemic or topical delivery of GDNF results in unwanted side effects in non-targeted tissues. In this thesis, we aim to improve recovery of function following a ventral root avulsion lesion by enhancing motoneuron survival and distal axonal regeneration using GDNF gene therapy. The advantage of gene therapy is the sustained production of GDNF protein by viral vector transduced cells, resulting in the availability of biologically active therapeutic protein restricted to the site of viral vector application.
In Chapter 1, we review clinical repair strategies and recent progress made in experimental ventral root avulsion lesion models. The gold standard in patients with a brachial plexus avulsion lesion is intra- or extra-plexal nerve transfer and nerve grafting. Reimplantation of the avulsed spinal roots into the spinal cord has been performed in a limited number of patients and is considered controversial. In contrast, the neurosurgical repair strategy in experimental models almost exclusively consists
8
of reimplantation of spinal nerve roots. In this chapter, we provide an overview of experimental treatment strategies based upon cellular, pharmacological and neurotrophic factor interventions following experimental reimplantation surgeries. From this overview, it becomes clear that although many treatments demonstrate some degree of neuroprotection, these effects are transient and long distance axonal regeneration is not achieved.
To obtain a more comprehensive understanding of mechanisms underlying the lack of functional recovery following a severe proximal nerve lesion, in Chapter 2 we per-formed a detailed spatio-temporal analysis of motoneuron survival, axonal regenera-tion, and neurotrophic factor expression after a lumbar ventral root avulsion. By com-paring animals with acute reimplantation to those with avulsion only at 1, 2, 4, 8, and 16 weeks post lesion, we were able to show that surgical reimplantation of ventral roots has neuroprotective effects on motoneuron survival. However, this effect was not maintained beyond a period of 4 weeks and the number of axons able to regener-ate long distances remained very low. The failure of axons to regenerregener-ate over long distances is associated with a decline of endogenous peripheral neurotrophic support, including decreasing glial cell line-derived neurotrophic factor (GDNF) protein levels. The observations in this chapter show that following neurosurgical repair without ad-ditional regenerative treatment, the prospect of recovery of function is poor.
To improve distal axonal outgrowth and overcome the loss of endogenous peripheral trophic support, in Chapter 3 lentiviral vectors expressing GDNF (CMV-GDNF) were injected into the peripheral nerve. Previously, we have shown that persistent local high levels of viral vector-mediated GDNF expression inside the reimplanted ventral roots promotes motor neuron survival but also leads to axonal entrapment and formation of axon coils at the site of GDNF expression. The formation of axon coils in the nerve roots prevents distal regeneration beyond the area of GDNF expression. In this chapter, we therefore examined whether creating an increasing proximal to distal GDNF gradient in the peripheral nerve could stimulate axonal growth towards the increasing GDNF concentration and would prevent proximal axonal entrapment. By injecting increasing dosages of lentiviral particles towards the distal nerve, we were able to generate a GDNF gradient in the peripheral nerve over a distance of 4 cm. Although peripheral GDNF expression induced distal axonal sprouting and enhanced axon numbers, we observed the formation of axon coils already at relatively low GDNF concentrations. Surgical reimplantation combined with distal GDNF treatment did not enhance motoneuron survival, underlining the significance of the treatment location.
In Chapter 4, we examined whether regulating the duration of expression would overcome the detrimental effect of persistent high levels of GDNF expression. The most widely used system to control transgene expression is the classical tetracycline-dependent transactivator system. In this system, gene expression can be regulated by
SCOPE AND OUTLINE
9
administration of doxycycline, a tetracycline analog. The immunogenic properties of the transactivator prompted us to employ a previously developed immune-evasive doxycycline inducible GDNF gene switch (dox-i-GDNF). By comparing our previously used CMV-GDNF vector with regulatable expression mediated by the dox-i-GDNF system in the reimplanted ventral root, we investigated whether time restricted GDNF expression improves motoneuron survival and decreases axonal trapping and coil formation. We show that proximal GDNF expression enhanced motoneuron survival and four week timed GDNF expression attenuates axonal coil formation. In addition, we observed increased distal axonal outgrowth and electromyographical recovery of the distal musculature following dox-i-GDNF treatment, suggesting target reinnervation had improved. Timed GDNF expression was however insufficient to stimulate voluntary hind paw function.
A critical factor in the failure of human nerve regeneration is the long distance axons need to regenerate to reach the end organs. Following human brachial plexus lesion, axonal regeneration over a distance of 80 cm is required to reach the hand musculature. With a growth speed of 1 to 3 mm per day, in ideal circumstances this is a protracted process which takes many months or even years. To examine the influence of the regeneration distance, in Chapter 5, we established a ventral root avulsion lesion in the brachial plexus. In this cervical model, the distal target muscles are located at half the distance compared to our previous chapters in which a lumbar avulsion lesion is performed. By using an identical dox-i-GDNF treatment paradigm as described in chapter 4, we were able to replicate the previously observed beneficial effects on motoneuron survival and distal axonal outgrowth. However, in contrast to our long distance regeneration model, here we achieved enhanced recovery of voluntary function. This shows that timed GDNF treatment is an important addition to neurosurgical repair and the detrimental influence of long regeneration distances.
Based upon these encouraging results, in chapter Chapter 6 we applied a combination strategy where in addition to improving motoneuron survival we aimed to enhance distal axonal outgrowth. During prolonged denervation periods, the chronically denervated distal nerve loses its pro-regenerative properties. In the extracellular matrix of this chronically denervated peripheral nerve, regeneration inhibitory chondroitin sulfate proteoglycans (CSPG) accumulate. By injecting lentiviral vectors expressing the enzyme chondroitinase ABC (ChABC) in the distal peripheral nerve, we show that these inhibitory CSPGs were digested. With a follow-up period of almost one year, this study confirms and extends our previous results from chapters 4 and 5, showing that timed GDNF treatment leads to sustained motoneuron survival. Despite successful removal of the inhibitory CSPGs, the distal ChABC treatment effect is modest and occurs during the later stages of the regeneration process. Our combination treatment of dox-i-GDNF with ChABC did not result in a synergistic effect.
In Chapter 7, we first discuss the value of the ventral root avulsion model as a proxy to study clinical neurotmetic nerve lesions. Subsequently, we highlight the results of this thesis in the context of the underlying key factors limiting distal regeneration and recovery of function. Finally, we provide a perspective on how recovery of function could be achieved, bringing treatment closer to the clinic.
Clinical and neurobiological
advances in promoting
regeneration of the ventral root
avulsion lesion
Ruben Eggers1, Martijn R. Tannemaat1,4, Fred de Winter1,2 , Martijn J.A. Malessy1,2, Joost Verhaagen 1,3
1 Laboratory for Neuroregeneration, Netherlands Institute for Neuroscience, an institute of the Royal
Academy of Arts and Sciences, Amsterdam, the Netherlands.
2 Department of Neurosurgery, Leiden University Medical Center, Leiden, the Netherlands. 3 Department of Molecular and Cellular Neurobiology, Center for Neurogenomics and Cognition
research, Vrije Universiteit Amsterdam, Amsterdam, the Netherlands.
4 Department of Neurology, Leiden University Medical Center, Leiden, the Netherlands.
Eur J Neurosci. (2016); 43(3):318-35
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
13
1
Abstract
Root avulsions due to traction to the brachial plexus causes complete and permanent loss of function. Until fairly recent, such lesions were considered impossible to repair. Here we review clinical repair strategies and current progress in experimental ventral root avulsion lesions. The current gold standard in patients with a root avulsion is nerve transfer, whereas reimplantation of the avulsed root into the spinal cord has been performed in a limited number of cases. These neurosurgical repair strategies have significant benefit for the patient but functional recovery remains incomplete. Developing new ways to improve the functional outcome of neurosurgical repair is therefore essential. In the laboratory, the molecular and cellular changes following ventral root avulsion and the efficacy of intervention strategies have been studied at the level of spinal motoneurons, the ventral spinal root and peripheral nerve and the skeletal muscle. We present an overview of cell-based, pharmacological and neurotrophic factor treatment approaches that have been applied in combination with surgical reimplantation. These interventions all demonstrate neuroprotective effects on avulsed motoneurons often accompanied with various degrees of axonal regeneration. However, effects on survival are usually transient and robust axon regeneration over long distances has as yet not been achieved. Key future areas of research include finding ways to further extend the post-lesion survival period of motoneurons, the identification of neuron-intrinsic factors which can promote persistent and long-distance axon regeneration, and finally prolonging the pro-regenerative state of Schwann cells in the distal nerve.
Introduction
Root avulsions of the brachial- or lumbosacral plexus occur as a consequence of traction forces acting at the zone between the spinal cord and root filaments. Avulsion results in permanent loss of motor function and sensation in the affected limb. Root avulsion in adults occurs in high speed traffic accidents involving motorcyclists, but may also occur during birth when the shoulder of the infant gets stuck behind the mother’s symphyses. Traction forces may be required to carry out a safe delivery, but could result in a neonatal brachial plexus injury (NBPI) with root avulsion (Malessy & Pondaag, 2009). The number of roots involved determines the extent of the motor and sensory function loss (Narakas, 1985). In 70% of the adult cases with root avulsion, multiple adjacent roots are involved (Narakas, 1985), resulting in a lesion also described as a “longitudinal spinal cord lesion” (Carlstedt & Havton, 2012). Until fairly recent this type of lesion was considered impossible to repair (Seddon, 1942; Robotti et al., 1995).
Currently, the surgical options for patients with ventral root avulsion are limited and consist mainly of nerve transfers (Malessy & Thomeer, 1998; Malessy et al., 2004).
14
and coaptated to a recipient plexus nerve. Functional restoration after nerve transfer requires adaptation of the central nervous system to execute the new task properly (Addas & Midha, 2009). Avulsed ventral root reimplantation to reconnect neural pathways is technically demanding due to the central location and the involvement of multiple roots (Carlstedt et al., 1995).
Despite the substantial progress in microsurgical techniques, functional recovery after brachial plexus surgery is far from optimal. Time has a major impact on the results of surgical reconstruction: functional recovery is worse in patients with a longer interval between trauma and surgery. Therefore several clinical studies underscore the importance of early intervention as this may be critical for recovery of function (Carlstedt et al., 2000; Kato et al., 2006; Ahmed-Labib et al., 2007). The quality of
functional recovery following nerve repair furthermore depends on additional factors such as age of the patient and the type and quality of the donor nerve (Malessy et al.,
1999; Haninec et al., 2007). Overall, recovery of distal function, particularly of the
hand, is usually very poor. By the time axons have reached the hand, severe atrophy of the muscles has taken place. Therefore, surgery is focused primarily on reconstruction of nerves innervating the proximal musculature (Tung et al., 2005). Functional
recovery is further impeded by the deleterious effect of misrouting, which is more likely to occur in plexus lesions than in more distally located lesions. Misrouting leads to co-contractures and ineffective muscle contraction (Haninec et al., 2007; Htut et al., 2007; Anguelova et al., 2014). Novel additional treatment strategies are needed
that further improve the outcome of neurosurgical repair. To develop new strategies, fundamental research needs to provide a better understanding of the molecular- and cellular mechanisms that underlie motoneuron degeneration and the failure of axons to regenerate over long distances toward their target.
In this review, we first briefly summarize the current neurosurgical procedures that have been developed for patients with spinal root avulsions. Secondly, the pathophysiology of the ventral root avulsion and recent advances in animal models to study this lesion are discussed. Thirdly, we review the state of the art of experimental intervention strategies in these animal models including cell implantation, pharmacological interventions and neurotrophic support. Finally, a perspective is provided on novel future treatment strategies for ventral root avulsion lesions.
Clinical features
A closed traction lesion to the brachial- or lumbosacral nerve plexus can be located proximal at the level of the motor or sensory root filaments, or more distally at the level of the spinal nerve or plexus elements. The location and extent of a plexus lesion depends on the angle and magnitude of traction forces that act during trauma. In the
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
15
1
most severe case, root filaments are forcefully torn out of the spinal cord and in the case of a ventral root, disconnection between the motoneuron pool and the spinal nerve occurs. In this review, we focus on this type of proximal motor lesion, the ventral spinal root avulsion. Various forms of postganglionic brachial plexus lesions are beyond the scope of this review and are discussed elsewhere (Narakas, 1985; Shin et al., 2005;
Kirjavainen et al., 2007; Rezende et al., 2013).
Following neurapraxia or axonotmesis, the nerve remains in continuity and recovery of function is possible. In contrast, neurotmesis and root avulsion will result in permanent loss of function without surgical treatment. Initially, the neurological deficit is characterized by the site of the nerve lesion only. Patients will show a comparable degree of dysfunction following these nerve lesions and in many cases, monitoring of the recovery of function in time is needed to fully assess the potential for future recovery. As a result, a monitoring period of 3 to 6 months is usually observed in patients with severe traumatic nerve injuries such as ventral spinal root avulsion, before surgical intervention is performed (Clarke & Curtis, 1995; Gilbert, 1995; Waters, 1999).
The clinical assessment of the anatomical location and estimation of the extent of the plexus lesion is a stepwise process. Patients with a flail arm following a high speed accident with severe pain are likely to have root avulsions. In total, 84% of patients suffer from chronic pain in the affected arm (Kato et al., 2006) and the description
of pain is so typical, that in many cases diagnosis of a preganglionic lesion can be made from the history of the patient alone (Birch, 1998). A correlation between the pain intensity and the number of avulsed roots has been reported (Htut et al., 2006).
The presence of Bernard-Horner’s sign, i.e ipsilateral ptosis, miosis and anhydrosis, is predictive, but not conclusive for avulsion of C8 and T1 ventral spinal nerves (Hentz & Narakas, 1988; Huang et al., 2008). The electromyographic (EMG) measurements
of patients with brachial plexus injury, in particular in infants with NBPI tends to be overly optimistic. EMG signs of reinnervation are present in the majority of patients and do not always predict adequate functional recovery (van Dijk et al., 1998; Malessy et al., 2009). Novel imaging techniques such as Magnetic Resonance Imaging (MRI)
and Computerized Tomography (CT) Myelography are indispensable in the diagnostic work-up of plexus lesions. The absence of root filaments correlates with the occurrence of a root avulsion (Malessy et al., 2009; Steens et al., 2011; Silbermann-Hoffman &
Teboul, 2013; Tse et al., 2014). However, final verification needs to be performed
during surgical exposure with visual inspection of the foramina (Malessy et al., 1999).
Intraoperative spinal somatosensory evoked potentials are used in addition to nerve action potentials registrations and have been reported to be helpful in the diagnostic differentiation between supra- or infraganglionic lesions (Robert et al., 2009; Clarkson et al., 2011).
16
Incidence of spinal root injury
The incidence of NBPI is around 1 to 3 per 1000 births (Pondaag et al., 2004; Kirjavainen et al., 2007; Backe et al., 2008b). Dependent on the type and degree of nerve damage,
44 to 84% spontaneously regain sensory or motor function, however, restoration of function is in general partial (Evans-Jones et al., 2003; Hoeksma et al., 2004; Pondaag et al., 2004; Backe et al., 2008a). Infants with NBPI with severe lesions can be identified
at one month of age (Malessy et al., 2011). CT myelography in infants born in the
cephalic position with severe NBPI who were scheduled for surgical exploration in more than half of the patients showed one or more root avulsions (Steens et al., 2011).
In adult patients with a brachial plexus injury, 72% were caused by traffic accidents of which the majority were motorcycle accidents (Narakas, 1985; Haninec et al., 2007).
Approximately 22% of these patients suffer from a root avulsion lesion and in the majority of patients multiple nerve roots are affected (Narakas, 1985; Schenker & Birch, 2001; Haninec et al., 2007). Overall, when comparing recovery in infants and
adults, the same lesion in adults exhibits poor functional recovery. This is due to the reduced regenerative capacity at older age and to the diverging anatomical dimensions between infants and adult patients: in adults, both the gap between severed nerve ends and the distance that regenerating axons must traverse are substantially larger (Risling
et al., 2011).
Surgical intervention
Current surgical interventions to treat plexus avulsion injuries consist of two approaches. The current gold standard to compensate for the loss of function due to ventral root avulsion is the transfer of donor nerves or fascicles of lesser functional importance, sutured to the denervated distal nerve to provide new regenerating motor axons. When no cervical nerve roots are available for nerve grafting, extraplexal (phrenic, accessory, or intercostal) nerves are used to supply axons to the denervated distal nerves of the plexus. The results of nerve transfer surgery are limited, with only half of the patients experiencing recovery of function following surgical intervention (Haninec et al., 2007).
An alternative technique to compensate for the loss of function due to ventral root avulsion is to implant an avulsed root or donor nerve graft directly into the spinal cord. Following a series of experimental studies, direct implantation was first performed by Carlstedt and colleagues (Carlstedt et al., 1986; Carlstedt et al., 1995; Havton &
Carlstedt, 2009). Although functional recovery of the hand has been described in a nine year old boy with a complete avulsion of all brachial plexus nerve roots (Carlstedt
et al., 2004), results from this procedure are less successful in older patients and
functional innervation was primarily obtained in the proximal musculature (Carlstedt
et al., 1995; Carlstedt et al., 2000; Bertelli & Ghizoni, 2003; Carlstedt, 2008). Surgical
implantation into the spinal cord may result in damage to the spinal cord and superficial implantation is therefore essential. As a result of the traumatic root avulsion, damage
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
17
1
to the spinal cord and the formation of local hematomas and gliosis may occur, and any additional damage due to surgery may have unwanted and irreversible consequences. Whether implantation of spinal roots- or nerve grafts into the spinal cord should be performed has been extensively discussed (Thomeer et al., 2002; Belzberg et al., 2004;
Haninec et al., 2007). To date, only some surgeons consider direct implantation of
the avulsed ventral root or nerve graft into the spinal cord a viable strategy (Fournier
et al., 2001; Bertelli & Ghizoni, 2003; Fournier et al., 2005; Lin et al., 2005; Amr et al., 2009; Wu et al., 2009), primarily because nerve transfer surgery provides similar
neurological outcome without posing a risk of injury to the spinal cord.
Overall, following ventral root avulsion and damage of multiple roots, both nerve transfer and reimplantation lead to a limited degree of functional recovery in patients. In addition, patients suffering from neuropathic pain generally experience amelioration of these symptoms following these surgical treatments (Kato et al., 2006).
Functional recovery is, however, primarily observed in the proximal musculature and misrouting and sprouting leads to co-contractures (Haninec et al., 2007; Htut et al.,
2007; Anguelova et al., 2014). Therefore, novel additional treatment strategies which
promote axonal regeneration and reinnervation of the target organs are needed to further improve the outcome of neurosurgical repair.
Fundamental research
The impact of nerve avulsion lesions and experimental interventions have been studied at the level of spinal motoneurons, the ventral spinal root and peripheral nerve and the skeletal muscle. In tables 1 and 2, an overview of all experimental studies on ventral root avulsion is presented. Table 1 provides an overview of the different lesion paradigms and methodological advantages and disadvantages. Table 2 lists the experimental studies divided in 4 categories (surgical intervention, cell transplantation, pharmacological and macromolecular intervention), including a summary of the results. Although some of these studies only describe the cellular and molecular processes following root avulsion without root reimplantation (avulsion-only), or following avulsion and implantation, the majority included a therapeutic intervention. A small subset of these studies investigated the timing- and method of implantation itself, whereas a large body of literature documented the effects of cell implantation, pharmacological interventions and neurotrophic support (Table 2).
18
Pathophysiology of experimental ventral root injury
Following experimental ventral root avulsion a multitude of cellular and molecular changes occur at distinct anatomical levels, including 1) the ventral horn containing
the spinal motoneurons, 2) the ventral spinal root and peripheral nerve and 3) the
denervated skeletal muscle.
Figure 1: Schematic representation of the experimental ventral root avulsion lesion and an illustration of cellular events in the ventral horn. A) Cross-section of the intact spinal cord depicting
the motoneurons (green) in the ventral horn and their axonal projections into the ventral root (VR). B) Ventral root avulsion directly at the surface of the spinal cord leads to severe motoneuron atrophy and death, and increased gliosis (red). C) Acute implantation of the avulsed ventral root into the medio-lateral aspect of the spinal cord leads to increased motoneuron survival and acts as a conduit for regenerating axons. Representative images of transverse sections of the rat spinal cord of a control animal (D, G), an animal with avulsion-only (E, H) or acute re-implantation of a ventral nerve root (F, I). Sections stained for Choline acetyltransferase staining (ChAT; D- F) display a clear loss of motoneurons 16 weeks post avulsion in comparison to acute implantation. An increased gliotic response is visible following avulsion in which astrocytes (red; GFAP) are found to enwrap the soma of motoneurons (green; NeuN, G-I). Healthy motoneuron profiles with an apparent normal morphology (closed arrowheads) can be found following acute implantation, whereas primarily atrophic motoneurons (open arrowheads) are present following avulsion only. Scale bar in F, I: 100 µm.
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
19
1
1) The ventral horn: motoneuron degeneration and gliosis
Following avulsion, gradual and progressive degeneration of motoneurons and preganglionic parasympathetic neurons occurs (Fig. 1) (Koliatsos et al., 1994; Novikov et al., 1995; Hoang et al., 2003; Eggers et al., 2010). Motoneurons display severe atrophy
and 20 to 60% of motoneurons appear to be lost within the first 2 weeks post-lesion. A substantial loss of motoneurons occurs at 4 weeks (50%), and 20 weeks (80-90 %) post-avulsion (Chai et al., 2000; He et al., 2003; Gu et al., 2004; Eggers et al., 2010).
Between studies however, variation in the magnitude of motoneuron loss is observed, specifically at early post lesion time-points. Difference in visualization- and counting methodology such as the use of Nissl- or Cholinergic marker staining and inclusion parameters such as cell diameter or nucleoli presence, can influence the outcome of the quantification of the loss of motoneurons (Nagano et al., 2003; Blits et al., 2004;
Penas et al., 2009). Due to a significant degree of shrinkage and nuclear condensation,
atrophied motoneurons may be overlooked during quantification. Indeed it has been shown that facial motoneurons appeared after a second axotomy based on re-expression of GAP43 and increased soma size. This indicates that atrophy following ventral root avulsion could have been mistaken for motoneuron death (McPhail et al.,
2004). Motoneuron loss occurs only following a nerve lesion close to the cell body in adult rats or following peripheral nerve axotomy in neonatal (P 3-30) rats (Kemp
et al., 2015). The precise biological mechanism which underlies the degeneration of
motoneurons is still not fully understood.
Following axotomy, the expression of genes involved in necrosis as well as apoptosis is increased (Haninec et al., 2003; Hoang et al., 2003; Zhang et al., 2005; Ohlsson &
Havton, 2006). Microarray studies performed on the denervated ventral motoneuron pool corroborated and extended these observations, as large cohorts of genes coding for proteins involved in apoptosis and energy metabolism including ATP transporters, were increased in expression (Hu et al., 2002; Yang et al., 2006b; Risling et al., 2011).
The exposure of motoneurons to large amounts of glutamate derived from nerve terminals following spinal lesion, results in an activation of NMDA receptors causing excitotoxicity by a subsequent influx of high levels of Ca2+ into the damaged cells (Doble, 1999; Lau & Tymianski, 2010). Intracellular Ca2+ activates neuronal oxide synthase (nNOS) via calmodulin, thereby increasing the production of nitrous oxide (NO) (Dawson et al., 1991). In turn, NO causes a concentration dependent glutamate
release (McNaught & Brown, 1998) potentially amplifying glutamate excitotoxicity. Dependent on the amount of glutamate exposure, necrosis or apoptosis is initiated in neuronal cells (Bonfoco et al., 1995; Syntichaki & Tavernarakis, 2003). AMPA receptors
lacking the GluA2 subunit have been shown to be more permeable to Ca2+ (Brorson
et al., 1992). The selective downregulation of the GluA2 subunit in motoneurons
found after ventral root avulsion consequently aggravates the influx of Ca2+ (Nagano
et al., 2003). Specific upregulation of nNOS in non-regenerating motoneurons that
20
Novikov et al., 1995; Gu et al., 1997; Yuan et al., 2010a). This increased expression
of nNOS has been suggested to mediate spinal motoneuron apoptosis (Martin et al.,
2005). In parallel to the induction of genes involved in degenerative processes, genes involved in regenerative processes are also induced, i. e. the transcription factors c-Jun and ATF-3 (Wu, 1996; Yuan et al., 2010b; Linda et al., 2011; Cheng et al., 2013) and
receptors for specific growth factors such as the glial cell line-derived neurotrophic factor (GDNF) receptors GFRa and cRET are upregulated. However, receptors for other growth factors, e.g. the ciliary neurotrophic factor (CNTF), brain-derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3) receptors gp130, trkB and trkC are down regulated (Hammarberg et al., 2000). This indicates that the responsiveness to certain
neurotrophic factors is enhanced, whereas the sensitivity for other neurotrophic factors may be diminished. These observations show that avulsed motoneurons initially shift from a stable transmitting to a regenerative state (Piehl et al., 1998; Risling et al., 2011),
but eventually irreversible degeneration of motoneurons occurs.
Avulsion injury induces an activation of astrocytes and microglia and an influx of
Figure 2: The regenerative response of motoneurons following acute implantation.
A) Schematic representation of the different trajectories of regenerating motor axons. Boxed areas
correspond to the representative images of sections stained for Choline acetyltransferase (ChAT) in
B-D. A high magnification image of the reimplanted ventral root shows longitudinal ChAT+ axons that
grow towards the periphery (B). Some axons grow from the ventral horn (top) towards the ventral root exit zone (VREZ) and subsequently grow along the spinal cord surface towards the reimplanted root or to ectopic sites along the surface of the spinal cord (C). At the site of implantation, ChAT+ axons can be found traversing from the motoneuron pool towards the implanted ventral root (D, VR). The spinal cord- ventral root border is indicated by arrowheads. Grey matter (GM), white matter (WM).
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
21
1
macrophages in the ventral horn (Fig. 1) (Koliatsos et al., 1994; Novikov et al., 2000;
Ohlsson et al., 2006). Interestingly, this gliotic response is not increased following direct
ventral or ventrolateral reimplantation of the avulsed root when compared to avulsion alone (Ohlsson et al., 2006; Barbizan et al., 2013). Directly following the avulsion,
activated microglia and astrocytes are found to enwrap the soma of axotomized motoneurons and parasympathic preganglionic neurons (PPN) (Emirandetti et al.,
2006; Ohlsson & Havton, 2006; Penas et al., 2009; Scorisa et al., 2009). Ultrastructural
analysis shows a loss of synapses on the cell bodies of motoneurons following axotomy, a process also referred to as synaptic stripping (Linda et al., 2000; Barbizan et al., 2014b).
Astroglial activation following axotomy appears to be causal to synaptic retraction from motoneuron cell bodies (Aldskogius et al., 1999; Emirandetti et al., 2006; Penas et al.,
2009). In a recent ventral root avulsion study, the application of the immunomodulator Glatiramer resulted in both reduced gliosis and reduced synaptic stripping (Scorisa et al., 2009). The elimination of synaptic input on spinal motoneurons, leads to a 48%
loss of synaptic surface area with a preferential elimination of glutamatergic terminals (Linda et al., 2000). These observations suggest an active reduction of glutamatergic
synaptic input to prevent motoneuron loss due to glutamate excitoxicity (Linda et al.,
2000; Scorisa et al., 2009; De Freria et al., 2012).
The regenerative response of motoneurons includes the accumulation of motor axons at the avulsion site and at the remaining nerve tufts (Fig. 2C). Furthermore, axons exit the avulsion site and grow along the spinal cord surface towards ectopic locations (Risling et al., 1992; Hallin et al., 1999; Gu et al., 2005; Lang et al., 2005; Eggers et al., 2010). Interestingly, a ventral root avulsion lesion also induced degeneration of
intramedullary sensory afferents in the dorsal funiculus (Bigbee et al., 2008). In
addition, a significant decrease of calcitonin gene-related peptide positive innervation was found in the dorsal horn at thoracolumbar - but not at lumbosacral - levels after sacral ventral root avulsion (Wu et al., 2012). These studies show that even remote
motor injuries can have an effect on intact sensory systems that anatomically share a peripheral nerve or ganglia. In addition, following root avulsion, animals develop at-level neuropathic pain at the adjacent dermatome (Li et al., 2003; Rodrigues-Filho et al., 2003; Bigbee et al., 2007; Chew et al., 2014; Wang et al., 2015)
2) The ventral spinal nerve root and peripheral nerve
Directly following axotomy, denervated Schwann cells in the ventral nerve roots and in the more distal peripheral nerves proliferate, align and convert to a neurite outgrowth-promoting state by enhancing their expression of a myriad of pro-regenerative proteins including the neurotrophic factors BDNF and GDNF. The beneficial effect on axonal regeneration is transient as neurotrophic factor expression in Schwann cells declines in time (Meyer et al., 1992; Funakoshi et al., 1993; Hoke et al., 2002; Boyd & Gordon,
2003; Omura et al., 2005; Michalski et al., 2008; Eggers et al., 2010). The outgrowth
promoting environment in the peripheral nerve is only maintained for 4 to 8 weeks after which endoneurial tube fibrosis and fragmentation of Schwann cell basal lamina
22
occurs (Giannini & Dyck, 1990; Fu & Gordon, 1995; Sulaiman & Gordon, 2000; Hoke, 2006).
The simultaneous loss of neurotrophic factor expression in the peripheral nerve and the reduced responsiveness of motoneurons to certain neurotrophic factors (discussed above), highlights the need for early intervention and the development of strategies which prolong the post-axotomy pro-regenerative period, in order to promote motoneuron survival, long-distance axonal regeneration and target reinnervation.
3) End organ atrophy and articular contractures
Prolonged periods of denervation will result in severe muscle wasting, myofiber death and deposition of interstitial collagen (Fu & Gordon, 1995; Lu et al., 1997).
When measuring recovery of muscle weight and tetanic force following delayed surgical repair, results are satisfactory when reinnervation occurs within 1 month, but increasingly poor and with permanent deficits if reinnervation is delayed by 3 months (Kobayashi et al., 1997; Bain et al., 2001; Aydin et al., 2004). Innervation of the muscle
at late time points does occur, but the number of available motor-units decreases (Fu & Gordon, 1995) and the denervated neuromuscular junction is less receptive to regenerating axons (Aydin et al., 2004). Neurosurgical intervention or repetitive
mechanical stimulation such as stretching of the denervated hind limb does improve cross sectional muscle fiber area (Bain et al., 2001; Sakakima & Yoshida, 2003; Agata et al., 2009).
In summary, following ventral root avulsion, time is an essential factor. The intrinsic ability of motoneurons to regenerate declines. The Schwann cells in the peripheral nerve initially support regeneration but gradually lose their regenerative potential and the end organs deteriorate over the course of months. In addition, regenerating axons are misdirected and potentially reinnervate inappropriate targets. The pathophysiological response after ventral root avulsion shares many of the characteristics also observed after chronic denervation of an injured peripheral nerve (Sulaiman & Gordon, 2000; Gordon et al., 2003; Furey et al., 2007; Sulaiman & Gordon, 2009).
Models of ventral root avulsion
Fundamental research on ventral root avulsions is performed in four animal models as summarized in Table 1. This said, there are significant differences in surgical approaches between laboratories which may have an impact on the regeneration process and the effect of a therapeutic intervention. These differences include the methodology of root avulsion and the location and number of nerve roots that are avulsed. An overview of the advantages and disadvantages of the animal models currently in use are presented
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
23
1
The methodology of root avulsion
The most common method of ventral root avulsion is performed via unilateral laminectomy, opening of the dura and identification of individual spinal nerve roots. This is followed by application of lateral traction on the roots directly at the spinal cord exit site resulting in an avulsion (Fig. 1). This procedure results in a reproducible injury of anatomically well-defined ventral roots at the spinal cord surface without injury to the sensory roots. The second approach, an extravertebral avulsion, is performed by applying lateral traction on the peripheral nerve outside of the spinal column. This results in a rupture of the nerve roots out of the foramen and spinal column (He et al., 2000; Noguchi et al., 2013). In this type of lesion no laminectomy or manipulation
of the dura mater is performed and reimplantation of the avulsed nerve roots is not possible. Extra vertebral avulsion results in both sensory and motor dysfunction (Table 1). This method is similar to the clinical mechanism of the preganglionic intradural avulsion in human patients, but this similarity is at the expense of reproducibility. The completeness of the avulsion can only be assessed after sacrifice of the animal and there is variation between the length of the proximal nerve stumps attached to the cord (Livesey & Fraher, 1992). The importance of the length of the remaining nerve stump and distance between axotomy and cell body was shown by (Gu et al., 1997)
in which only an axotomy closer than 4 mm to the cell body results in reproducible motoneuron death. When comparing gene expression profiles in motoneurons after axotomy 10 mm distal from the cell body or following ventral root avulsion at the surface of the spinal cord, a two-fold downregulation of 147 regeneration- and survival related genes was found following ventral root avulsion (Yang et al., 2006b). These
results are consistent with earlier observations on the differential expression of c-JUN and ATF-3 following proximal or distal axotomy (Herdegen et al., 1997; Tsujino et al.,
2000; Zhou et al., 2008; Linda et al., 2011). In contrast, peripheral axotomy in neonatal
(P3-30) rats leads to comparable motoneuron loss and poor function regeneration, emphasizing the importance of the age of the experimental animals and location of the axotomy (Kemp et al., 2015).
Difference between cervical, lumbar and sacral lesions
Implantation of the avulsed root or peripheral nerve graft leads to substantial regeneration of axons into the implant irrespective of the location of surgery. Following cervical lesions in the rat, axons have to regenerate approximately 4 to 6 cm to reach the most distal targets, whereas after lumbar lesions this distance is 10 to 12 cm. In agreement with clinical observations, following ventral root implantation, functional innervation of the proximal musculature occurs almost exclusively in the cervical avulsion model (Chuang et al., 2002; Gu et al., 2004; Wu et al., 2004; Jivan et al.,
2006). Thus, in the cervical model regenerating axons are able to successfully bridge the shorter distance and reach the end organs in time which prevents the formation of severe articular contractures (Bontioti et al., 2003; Jivan et al., 2006). In the lumbar
24
is 1 - 3 mm/day but this velocity declines to 0.7 mm/day in the later phases of the regeneration process (Eggers et al., 2010). In addition, the number of regenerating
axons in the peripheral nerve reduces significantly in the distal nerve portion and recovery of function is much less common (Eggers et al., 2010; Torres-Espin et al.,
2013). The regenerative capacity of axons that do not form functional connections for prolonged periods of time, declines significantly (Fu & Gordon, 1995; Sulaiman & Gordon, 2000; Furey et al., 2007). In the lumbar model, the larger distance between
avulsion site and end organs leads to long term denervation of the distal nerve, severe muscle atrophy and articular contractures (Table 2)
From a neurosurgical point of view the lumbar and cervical lesion models differ in complexity. At the cervical level, the location of the ventral root exit zone and corresponding intervertebral foramen are closely aligned. This allows for easy identification of the ventral nerve roots. This close alignment however, partially obstructs surgical access to the ventral root due to the anatomical positioning of the dorsal root and dorsal root ganglion (DRG) directly above the ventral root exit zone. Therefore most studies perform a DRG excision prior to avulsion and (graft-) implantation (Wu et al., 1994b; Chai et al., 2000; Gu et al., 2005) resulting in sensory
denervation. In a recent study, sparing of the DRG is described, however, the short length and fragility of the ventral root make manipulation and implantation surgically challenging (Fu et al., 2014). The lumbar ventral roots exit the spinal cord 4 to 5
segments proximal to their corresponding invertebral foramen, resulting in a longer segment of ventral root available for reimplantation and no obstruction due to the DRG. However, due to the lack of clear intraspinal anatomical landmarks and a certain degree of anatomical variability between animals in the proximo-distal positioning of the ventral root exit site, a substantial amount of experience is required for the positive identification of the L4, L5 and L6 ventral roots.
Total number of roots avulsed and anatomical variation
The motoneuron pool in the rat that contributes axons to the cervical and lumbar plexi are located at the C5-T1 and L3-L6 spinal cord segments. Anatomical variation in the contribution of specific spinal nerve roots to the peripheral nerves in the fore and hind paw exists between animals and specific strains. An anatomical connection between the L5 and L6 ventral nerve root is absent in almost 50% of the Wistar or Sprague Dawley rats (Asato et al., 2000; Rigaud et al., 2008). This natural anatomical variation
between individual rats accounts for at least some of the incidental variability in post-operative electophysiological measurements and the detection of variable numbers of uninjured neurons following retrograde tracing (Eggers et al., 2013; Torres-Espin et al., 2013; Hoeber et al., 2015).
Between studies and laboratories, differences in the number of avulsed- and/or reimplanted roots have been reported. Some laboratories have employed the avulsion
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION 25
1
Model Adv antages Limitations Pr imar y e xper imental impr ov ements Ex tra ver tebr al Avulsion Surgicall y less in vasi ve Clinical r ele vanceNo implantation possible Axotom
y of aff
er
ent and eff
er ent ax ons Impr ov ed MN sur vi val and r educed gliosis Cer vical Avulsion Shor t distance and r egener ation time to w ar ds target Limited ar tic ular contr ac tur e f or mation Easy v entr al ner ve r oot identification PN is surgicall y a vailable up to musc le DR G e xcision r equir ed f or implantation Mul tiple r oot a vulsions needed to pr ev ent spr outing Inabilit y to st ud y pr
olonged beneficial eff
ec
ts
of inter
vention due to fast inner
vation Motoneur on sur vi val (++) Regener
ating fiber ingr
owth (++) Regener ating fiber in ‘ distal ner ve’ (+) Elec troph ysiolog y i.e. CMAP , T etanic f or ce (+) For epa w gr ip and T erzis scor e (+/-) Lumbar Avulsion
Long distance and r
egener ation time to w ar ds target Ventr al ner ve r oot leng th a vailable f or manipulation Affer ent ner ve intac t Clinical r ele
vance due to distance
Ventr
al ner
ve r
oot identification comple
x
Long r
egener
ation time needed
Access to PN obstr
uc
ted b
y spinal column and
iliac cr est Func tion impair ed due to ar tic ular contr ac tur es and musc le atr oph y Motoneur on sur vi val (++) Regener
ating fiber ingr
owth (++) Regener ating fiber s pr oximal (1-3 cm) (+) Regener ating fiber s distal (8- 10 cm) (+/-) Elec troph ysiolog y; CMAP , T etanic f or ce (-) Pa w f unc
tion, Open field, toe spr
ead (-) Sacr al A vulsion Shor t distance/ r egener ation time to w ar ds target Bilater al a vulsion of v entr al r oot possible Clinical r ele vance f or Conus medullar is injur y
Both motor and autonomic MN degener
ation
Ventr
al ner
ve r
oot identification comple
x Dail y bladder v oiding needed Motoneur on sur vi val (++) Regener
ating fiber ingr
owth (++) Bladder f unc tion (Eph ys , V oiding) (+) Table 1: O ver vie w of the e xp erimen tal v en tr al r oot a vulsion par
adigms with their adv
an
tages and limita
tions . T he pr imar y e xper imen tal impr ov emen ts descr
ibes the most c
ommonly r
epor
ted r
esults in these models and fr
equenc y of obser ved eff ec t (bet w een br ackets): ++ fr equen t, + pr esen t, +/- oc casional , - r ar e. (DR G: dorsal r
oot ganglion; PN per
ipher
al ner
ve; MN: mot
oneur
26
of a single cervical or lumbar root (Haninec et al., 2004; Nogradi et al., 2007; Noguchi et al., 2013). More extensive lesions include avulsion of multiple roots and reimplantation
of a single root (Blits et al., 2004; Ding et al., 2014) or avulsion and reimplantation of
multiple roots (Huang et al., 2009; Eggers et al., 2013). Alternatively, an avulsion and
implantation of a single root is followed by peripheral ligation of the adjacent nerves
(Gu et al., 2004; Gu et al., 2005). The avulsion of a small number of ventral roots
and leaving adjacent nerve roots intact influences the outcome of functional recovery. Following avulsion of spinal roots L4 – 5, electrophysiological (CMAP) function was absent in the gastrocnemius and tibial muscle and although reduced, was still present in plantar muscles (Penas et al., 2009). The contribution of an intact spinal root following
cervical lesion was found to result in improved recovery in human neonates but not in adults (Vredeveld et al., 2000). This observation was confirmed in rat neonates where
following a spinal root lesion, the adjacent intact C7 motoneuron contribution to the biceps muscle increased four-fold (Korak et al., 2004). Using functional-MRI imaging
of the sensory motor cortex or spinal electrophysiological measurements, it was shown that after nerve injury, adjacent spinal cord segments and corresponding cortical areas may become significantly over-activated, compensating for the loss of function (Havton & Kellerth, 2004; Li et al., 2013). These results suggest that intact axons from adjacent
segments could potentially take over the innervation as a result of sprouting. This is primarily of importance when quantifying the effects of the applied intervention on the number of distal axons or recovery of function. Therefore, in studies where the effect of an intervention on functional recovery is examined a more extensive lesion or re-lesioning should be considered (Lin et al., 2014).
Intervention strategies
Experimental strategies to promote repair following ventral root avulsion can be subdivided in four approaches: (i) Surgical implantation, (ii) cell transplantation, (iii)
pharmacological intervention and (iv) macromolecular intervention. An overview of
the current literature applying such intervention strategies following experimental ventral root avulsion is provided in Table 2.
Acute and delayed implantation
An implanted avulsed ventral root or a peripheral nerve graft acts as a conduit for regenerating axons and results in a significant increase in motoneuron survival when compared to avulsion without implantation (Fig. 1) (Wu et al., 1994a; Chai et al., 2000;
Gu et al., 2004; Hoang & Havton, 2006; Ohlsson et al., 2006; Eggers et al., 2010). The
reimplanted ventral root contains Schwann cells secreting high levels of neurotrophic factors. These neurotrophic factors diffuse from the root into the spinal cord and exert positive effects on motoneuron survival, but can also act as neurotropic cues, attracting motor axons from ventral horn towards the implantation site (Havton & Kellerth, 1987;
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
27
1
Hoang & Havton, 2006). In addition to motor axons, a reimplanted intercostal nerve graft will allow regeneration of both afferent and efferent axons simultaneously when distally connected to both avulsed ventral and dorsal root (Lin et al., 2014). Substantial
axonal outgrowth into the nerve implant is accompanied by electrophysiological improvement and innervation of the proximal musculature is frequently observed (Chuang et al., 2002; Gu et al., 2004; Wu et al., 2004; Hoang et al., 2006; Jivan et al.,
2006). Additional beneficial effects following acute implantation include a reduction of the inflammatory response (Risling et al., 2011), reduced nNOS expression (Wu et al., 1994a; Wu, 1996; Chai et al., 2000), a significant reduction of pain behavior and
amelioration of the degeneration of sensory projections in the dorsal column (Bigbee
et al., 2007; 2008). If the implantation is performed superficially into the ventrolateral
spinal cord or by direct ventral reattachment, no additional intramedullary gliosis has been observed in comparison to ventral root avulsion without implantation (Ohlsson
et al., 2006; Barbizan et al., 2013).
However, as discussed above, the expression of neurotrophic factors by denervated Schwann cells in the reimplanted root is transient and the levels of neurotrophic factors that reach the motoneurons declines, eventually leading to motoneuron loss (Chai et al., 2000; Bergerot et al., 2004; Gu et al., 2004; Eggers et al., 2010). A delayed
implantation of nerve grafts at 1, 2 or 3 weeks after the avulsion strongly reduces the positive effects on motoneuron survival as seen following acute implantation, whereas the percentage of the surviving motoneurons projecting an axon into the implant remains constant at around 80% (Wu et al., 2004; Gu et al., 2005; Jivan et al., 2006;
Su et al., 2013). These observations show that acute implantation only delays, but not
prevents, motoneuron degeneration and a relatively short post-lesion time-window of 2 to 4 weeks is available to promote motoneuron survival and regeneration, after which irreversible motoneuron loss occurs.
The type of nerve graft and the Schwann cell phenotype plays a pivotal role in the degree of regeneration and experimental outcome. Following implantation of a sensory (saphenous) nerve graft, significantly lower levels of BDNF and GDNF mRNA were found in the nerve and suboptimal ingrowth of motor axons occurs in comparison to ventral root implantation (Chu et al., 2008; Su et al., 2013). The expression profile
of neurotrophic factors differs substantially between denervated motor and sensory Schwann cells. GDNF and Pleiotrophin (PTN) are predominantly expressed in the ventral nerve root (Hoke et al., 2006; Brushart et al., 2013). In addition, Osteopontin
is upregulated in motor but not sensory Schwann cells and preferentially induces motor, but not sensory axonal outgrowth (Wright et al., 2014). Preconditioning of
cultured sensory Schwann cells with GDNF has been shown to overcome Schwann cell phenotypic memory and improve motor axon outgrowth in vitro (Marquardt & Sakiyama-Elbert, 2015). In addition to their phenotype, the age of Schwann cells is an important factor. In a sciatic nerve transection model, grafts containing Schwann cells of aged animals lead to impaired axonal regeneration in comparison to young Schwann
28
cells. Aged Schwann cells failed to rapidly activate the transcriptional repair program that is induced in young Schwann cells (Painter et al., 2014). Although implantation of
various nerve grafts, including an autologous saphenous nerve, acellular nerve graft, intercostal nerve or artificial collagen guide, support axonal outgrowth up to 2 cm (Kassar-Duchossoy et al., 2001; Chuang et al., 2002; Huang et al., 2007; Chu et al.,
2012; Su et al., 2013; Lin et al., 2014), axonal regeneration in these conduits remains
suboptimal when compared to implantation of the ventral root itself.
In conclusion, acute implantation of nerve conduits has a beneficial effect on motoneuron survival and axonal regeneration occurs. The results are most optimal after ventral root reimplantation and results are less encouraging when a sensory nerve graft is used or when implantation is delayed. This underscores the need for additional treatments which further promote motoneuron survival and enhance the pro-regenerative properties of the spinal root, nerve graft and more distal peripheral nerve.
Cell implantation
The experimental application of cells, including mesenchymal stem cells (MSC), induced pluripotent stem cells (iPSC), neuroectodermal stem cells (ESC), neural progenitor cells (NPC) and bone marrow stem cells (BMC) in the ventral root avulsion paradigm are summarized in Table 2. The implantation of MSC’s, ESC’s or iPSC’s in the ventral horn following a ventral root avulsion promotes the survival of motoneurons, reduced synaptic stripping and gliosis (Rodrigues Hell et al., 2009; Spejo et al., 2013), improved
axon regeneration into the reimplanted root and some studies report a certain degree of recovery of function (Pajer et al., 2014; Pajer et al., 2015). Dependent on the cell type
used, cell transplantation results in enhanced expression of the neurotrophic factors NGF, BDNF, GDNF (Rodrigues Hell et al., 2009; Su et al., 2009), modulatory cytokines
and anti-inflammatory IL-10 (Pajer et al., 2014) or cytokines and neurotrophic factors
simultaneously (Pajer et al., 2015).
Transplantation of MSC’s following avulsion of ventral roots L4, 5 and 6 delays motoneuron loss when compared to reimplantation only and initially enhanced axonal outgrowth. EMG of hind limb muscles showed evidence for functional muscle re-innervation (Torres-Espin et al., 2013). Despite this electrophysiological improvement,
articular contractures occurred and there was no improvement in recovery of hind paw motion. The number of transplanted cells gradually decreased in time which may explain the transient beneficial effects on motoneuron survival and the limited effects on distal axon regeneration (Torres-Espin et al., 2013). A comparison between the
number of surviving transplanted ESC’s and the number of regenerating motoneurons, shows a positive correlation between the number of transplanted cells and the survival of motoneurons (Pajer et al., 2014). The transplanted cells show little migration from
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
29
1
astrocytes after 7 days post-implantation (Pajenda et al., 2013; Torres-Espin et al., 2013;
Pajer et al., 2014; Pajer et al., 2015). Transplantation of ESC’s into either the injured
spinal cord segment or reimplanted nerve root reduces motoneuron loss following transplantation itself (Pajenda et al., 2013). In contrast, following topical application
of ESC’s or BMC’s around the reimplanted root using fibrin glue no improvement on motoneuron survival occurs, suggesting that the beneficial factors produced by these cells are unable to reach the lesioned motoneuron pool (Pajenda et al., 2013; Barbizan et al., 2014a; Barbizan et al., 2014b).
Olfactory ensheathing glial cells (OECs) are the resident glia cells in the primary olfactory system. This specialized type of glia cell forms cellular pathways that actively supports axon regeneration following transplantation in the injured spinal cord and dorsal root lesions (Ramon-Cueto & Nieto-Sampedro, 1994; Gomez et al., 2003; Ramer et al., 2004; Ibrahim et al., 2009). Following application of an OEC matrix at the ventral
root implantation site, OECs migrate towards the periphery and a four-fold increased ingrowth of motor axons in the implanted root occurred at two weeks (Li et al., 2007).
The biological mechanism underlying the beneficial effect of cell grafting possibly consists of an interplay between prolonged neurotrophic support, reduction of gliosis, enhanced axon guidance and myelination. The improved motoneuron survival and enhanced regeneration into the reimplanted root without the observation of unwanted side effects such as aberrant sprouting, axon coil formation (Eggers et al., 2008) or
mechanical allodynia (Hofstetter et al., 2005) suggests that transplanted cells are
capable of providing significant therapeutic support after ventral root avulsion (Torres-Espin et al., 2013; Pajer et al., 2015). It is currently not known if the beneficial effects of
cell grafting following ventral root implantation could be further improved, as gradual loss of grafted stem cells due to differentiation or cell death occurs.
Pharmacological interventions
Pharmacotherapeutic interventions following ventral root avulsion have primarily targeted the motoneurons in the ventral horn using neuroprotective or immune-modulatory agents (Table 2). Neuroprotective approaches aimed at reducing glutamate excitotoxicity (Nogradi & Vrbova, 2001; Nagano et al., 2003; Bergerot et al., 2004;
Nogradi et al., 2007; Pinter et al., 2010; Penas et al., 2011; Wu et al., 2013; Chew et al.,
2014; Fu et al., 2014), preventing apoptosis by inhibition of the caspase pathway (Chan et al., 2001; Chan et al., 2003; Zhang et al., 2005) or the inhibition of nitric-oxide
synthase (Wu & Li, 1993; Zhou & Wu, 2006a).
Motoneurons in the cervical and lumbar enlargement undergo glutamate induced excitotoxicity primarily via AMPA dependent mechanisms, whereas thoracic motoneuron excitotoxicity occurred via NMDA receptors (Gerardo-Nava et al., 2013).
30
osmotic minipumps significantly reduced motoneuron loss (Nagano et al., 2003).
Riluzole inhibits presynaptic glutamate release and blocks voltage-gated Na+ and Ca2+ channels and NMDA receptors (Doble, 1996). Riluzole promoted motoneuron survival up to 70% when applied within 2 weeks post lesion (Nogradi & Vrbova, 2001; Nogradi et al., 2007; Pinter et al., 2010). The application of riluzole or GDNF protein,
leads to equal levels of motoneuron survival at short time points. However, prolonged motoneuron survival was only obtained following combined treatment with riluzole and GDNF (Bergerot et al., 2004). This indicates that simultaneously increasing
neurotrophic support and inhibition of glutamate excitotoxicity is required to exert long-term effective neuroprotection. In addition, riluzole decreases thermal and tactile hypersensitivity following both afferent and efferent spinal root avulsion by reducing microglial cell activation (Chew et al., 2014) .Other pharmacological approaches that
have been implicated in reduction of glutamate-induced excitotoxicity include the systemic application of Pre-084, a Sigma-1 receptor agonist, and lithium. Lithium increased motoneuron survival but not regeneration (Fu et al., 2014), whereas daily
injections of Pre-084, leads to increased motoneuron survival, reduced astrogliosis and increased GDNF expression (Penas et al., 2011).
Caspases are proteases thought to be instrumental in the induction of apoptosis and cell death. N-acetyl-cysteine blocks mitochondrial events that activate the caspase cascade. N-acetyl-cysteine promotes motoneuron survival for a period of 4 weeks (Zhang et al.,
2005). In contrast, a general caspase inhibitor (Boc-D-FMK) and a specific caspase-3 inhibitor (Ac-DEVD-CHO) improved motoneuron survival in neonates, but not adult rats up to 2 weeks (Chan et al., 2001; Chan et al., 2003).
In addition to apoptotic pathways, other mechanisms of neurodegeneration may also be involved in motoneuron death after avulsion. Nitric oxide-synthase (NOS) is induced in motoneurons after ventral root avulsion and its enzymatic product nitric oxide has been linked to neurodegeneration. Nitro-arginine, an inhibitor of nitric oxide-synthase, rescues motoneurons (Wu & Li, 1993), however, in vivo knock-down of NOS with an antisense-oligonucleotide induced motoneuron degeneration (Zhou & Wu, 2006a). These contradictory observations indicate that further studies are necessary to shed more light on the involvement of the NOS-pathway in motoneuron survival. Moreover, mechanisms of necrosis come into play, as suggested by increased complement activation and deposition of membrane attack complexes on dying motoneurons (Ohlsson & Havton, 2006).
Reduction of local inflammation and gliosis using the immune-modulatory compounds glatiramer and minocycline has beneficial effects after a ventral root avulsion (Table 2; Hoang et al., 2008; Scorisa et al., 2009; Chew et al., 2014).Daily injections with
Glatiramer, a 4 aminoacid polymer of myelin basic protein thought to act as an immune-modulator, leads to reduced synaptic stripping and gliosis (Scorisa et al., 2009). The
ADVANCES PROMOTING REGENERATION AFTER VENTRAL ROOT AVULSION
31
1
mitochondrial factor release and caspase activation and inhibition of inflammatory responses (Hua et al., 2005; Ledeboer et al., 2005). Minocycline has been shown to
be neuroprotective following ventral root avulsion and leads to reduced microglial activation (Hoang et al., 2008). However, in a second study the use of minocycline
prevents the onset of evoked pain hypersensitivity, but did not lead to neuroprotective effects on avulsed motoneurons (Chew et al., 2014).
Several of these pharmacological compounds have been clinically applied in the treatment for other conditions, making these compounds accessible for future therapeutic intervention following spinal root lesion. The application of such interventions directly following a lesion could prevent early motoneuron death leading to a prolonged time-window in which surgical intervention would be beneficial. The beneficial effects of riluzole on motoneuron survival, and the reduction of microgliosis and pain, its approval as a drug for ALS patients (Bensimon et al., 1994) and the current
phase III trial in spinal cord lesion patients (Fehlings et al., 2015), may render riluzole
one of the first agents to be applied in patients suffering a spinal root lesion.
Macromolecular intervention: neurotrophic factors and interference
with neurite outgrowth inhibitors
Neurotrophic factors. Neurotrophic factors modulate neuronal survival, axonal growth
and synaptic plasticity and are considered to be the most potent pro-regenerative proteins to date (Koliatsos et al., 1993; Henderson et al., 1994; Vejsada et al., 1995;
Kishino et al., 1997; Jones et al., 2001). The pro-regenerative effects of neurotrophic
factors of the neurotrophin (BDNF), transforming growth factor-β (GDNF), cytokine (CNTF and PTN), insulin growth factor (IGF1) and FGF families (acidic fibroblast growth factor, aFGF) that have been studied following ventral root avulsion are summarized
in Table 2. Neurotrophic factors have been delivered to injured ventral roots locally by
means of gelfoam, fibrin glue, slow releasing polymers, osmotic minipumps, direct protein injection, or viral vector-mediated gene transfer. The main findings of these studies indicate that 1) GDNF consistently promotes the survival of spinal motoneurons and stimulates axon regeneration into the reimplanted roots or nerve grafts; 2) The effects of BDNF are generally less robust and more variable than GDNF, however, this neurotrophin shows considerable promise as a motoneuron survival factor in several studies; 3) Variable and small effects are obtained using IGF1 on motoneuron survival, whereas aFGF leads to improved electrophysiological responses; 4) CNTF and PTN show no effects; 5) Irrespective of cervical or lumbar ventral root implantation, similar motoneuron survival and regeneration into the implant occurs following neurotrophic application. Long-term positive effects on functional recovery following cervical lesions have been reported, however, after lumbar lesions much room for improvement remains as prolonged motoneuron survival, distal outgrowth and functional recovery remain poor.
