2. Administration of NTF by Direct Infusion in the CNS
Various techniques have been used to get NTFs into the brain. The best known technique is direct intracerebroventricular (ICV) infusion. In particular, recombinant human (rh) GDNF and
125Iodine-labelled GDNF (
125I-GDNF) have been shown to diffuse into the deep brain structures of rats
[76][77], not only to significantly increase striatal and nigral dopamine (DA) levels, but also to increase hypothalamic DA levels, which could explain the decreased food and water consumption and body weight observed in in vivo experiments
[78][79]. ICV injection of GDNF into 6-hydroxydopamine (6-OHDA)-treated rats, an animal model of PD, also appears to result in improved locomotor performance
[78][79]. Furthermore, the ICV delivery route seems suitable for therapies that need to reach the BFCNs. Early and progressive degeneration of BFCNs contributes substantially to cognitive impairments of AD. Since BFCNs extend their axons through the hippocampus and neocortex, NGF administered in the lateral ventricle can act on the TrkA receptor to transmit trophic support signals to BFCNs. This approach has been shown to be particularly effective in preventing loss of BFCNs in rodents associated with injury and ageing
[80][81][82]. However, the small volume of the rodent brain compared to the human brain raises important questions about the applicability of this technique in clinical studies. Therefore, ICV injections were also performed in non-human primates
[83][84][85]. GDNF has been shown to produce significant improvements in motor activity in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-treated rhesus monkeys, a model of PD
[83][86], and improvements in motor impairment and reductions in l-dopa-induced dyskinesia in marmosets
[84]. In an autoradiographic study of the distribution of
125I-GDNF administered in the lateral ventricles of rhesus monkeys with a MPTP lesion, GDNF was not found to diffuse readily into the putamen. This finding contrasts with similar studies in rodents
[76][77], suggesting that the success of ICV infusion in rodents might be a product of the smaller diffusion distance within their brain
[78][79]. Moreover, the ICV delivery route was associated with serious side effects
[80], such as hyperinnervation of cerebral blood vessels
[82], hypophagia
[80][81], Schwann cell hyperplasia with sprouting of sensory and sympathetic neurons
[87], neuropathic pain
[80], and dyskinesia
[83][84][86], providing profound contra-indications for the applicability in clinical trials.
Because of these ICV-related side effects, the study by Tuszynski et al.
[88] investigated whether intra-parenchymal infusion would be a well-tolerated way to administer NTFs to degenerating cholinergic neurons. In particular, intraparenchymal NGF infusion prevented degeneration of BFCNs, whereas glial responses were minimal in adult rats that underwent complete unilateral fornix transections, followed by intraparenchymal infusions of recombinant human NGF for a 2-week period. In addition, no apparent toxic effects of the infusions were observed, according to the researchers
[88]. Other studies aimed to administer NTFs by a less invasive method. The group of Braschi et al.
[20] tested whether intranasal (IN) administration of different concentrations of BDNF in AD11 transgenic mice, a model of AD, was able to rescue neuropathological and memory deficits. They found that IN administration of BDNF, but not with PBS, was adequate to completely rescue the performance of AD11 mice in both the object recognition test and the object context test. The strong improvement in memory performance in BDNF-treated mice was not accompanied by an improvement in AD-like pathology, amyloid-β (Aβ) load, tau hyperphosphorylation, and cholinergic deficiency
[20]. Similarly, IN administration of NGF to Aβ peptide-expressing traumatic brain injury (TBI) rats, which are at risk of AD in later life, caused a marked reduction in Aβ42 deposits and restored motor and behavioural function
[19]. Features such as non-invasive manipulations, rapid absorption rate, easy repetitive dosing, and reduction of non-target biodistribution make IN administration superior to the systemic and ICV routes of administration
[18][19].
Finally, studies examined the effects of continuous intraputamenal administration of GDNF in both aged and MPTP-lesioned non-human primates
[89][90][91]. Histological and biochemical analysis showed an increase in cell size and the number of dopaminergic neurons within the substantia nigra, as well as increased fibre density in the caudate nucleus, putamen, and globus pallidus. Primates with MPTP lesions showed improvements in the primate PD rating scale, whilst aged monkeys demonstrated improvement in general motor performance at high doses and increases in hand speed
[89][90][91]. To assess the possible side effects of continuous administration of GDNF, a six-month toxicity study was conducted in rhesus monkeys. The results cast considerable doubt about the neuro-restorative potential of GDNF for the treatment of PD, given that they identified a number of pathological markers of toxicity, including reduced food intake and weight loss, meningeal thickening, and most concerning, multifocal cerebellar Purkinje cell loss
[30]. Apart from the above-mentioned side effects, direct administration of NTFs into the brain also had some practical problems, such as invasiveness, BBB permeability
[7][8][92], poor half-life, and rapid degradation
[93]. This led to studies using cell therapy, where cells were modified to produce a specific protein.
3. Cells Modified to Express Neurotrophic Factors
During the last years, different cell types have been utilized to deliver NTFs to the injured sites. Mesenchymal stromal cells (MSCs) are described as adherent, fibroblast-like cells with prominent proliferation capacity
[41][49][50][94]. Because of their low immunogenicity (low expression levels of major histocompatibility complex (MHC) class II), MSCs can survive after administration
[95]. The existence of such capabilities makes MSCs a safe, tolerable, and efficient biological vector for the generation and delivery of therapeutic agents, such as NTFs, to the target sites
[41][49][50]. Furthermore, different routes of administration were used to administer the modified MSCs, resulting in different outcomes. In a study by Suzuki et al.
[50], human MSCs (hMSCs), derived from neonatal bone marrow aspirates which were modified to express GDNF, were administered intramuscularly as a "Trojan horse" to superoxide dismutase (SOD1)
G93A rats, a rat model of familial ALS, to deliver GDNF to the terminals of motor neurons and to skeletal muscle. hMSC-GDNF survived in the muscle, secreted GDNF, and significantly increased the number of neuromuscular connections and motor neuron cell bodies in the spinal cord in the mid-stage of the disease. Moreover, hMSC-GDNF significantly slowed down disease progression
[50]. In addition, several improvements have been reported when CNTF-
[51], NT-3-
[52], and BDNF-modified
[53] MSCs were administered directly into the spinal cord of SCI rats, such as improvement in behavioural scores, motor function, axonal regeneration, and neuronal survival
[51][52], and restoration of diaphragm muscle function
[53]. Positive results with MSCs expressing NTFs were also observed after intravenous (iv) administration. A remarkable recovery of neuronal function was observed and demyelination was significantly reduced in EAE mice: the cumulative clinical scores were significantly decreased, and the disease onset was statistically delayed, after iv MSC-CNTF
[54] and MSC-BDNF administration
[96]. Moreover, BDNF-expressing MSCs can also reduce striatum atrophy and increase neurogenesis in HD mouse models
[21]. In summary, MSCs represent a promising tool for cell therapy. There is currently much interest in the use of MSCs for the treatment of neurodegenerative diseases. There are several studies using the innate trophic support of MSCs or increased support by NTFs, such as the administration of BDNF, CNTF NTF-3, or GDNF to the CNS to support damaged neurons, using genetically engineered MSCs as delivery tools. Biosafety could be a potential difficulty in cell therapies when using genetically engineered MSCs. The random integration of vectors with genes for neurotrophic or other factors may pose the risk of insertional integration. However, homologous recombination and targeted gene transfer are advancing rapidly.
Neural stem cells (NSCs) are also used as a NTF vector, resulting in several positive effects. NSCs are characterised as multipotent and self-renewing cells with the capacity to differentiate into mature neurons and neuroglia cells
[22][23][24][25]. In a rodent model of cervical SCI, it was shown that GDNF-expressing human induced pluripotent stem cell-derived NSCs (hiPSC-NSCs) showed greater differentiation into a neuronal phenotype than unmodified hiPSC-NSCs
[26]. Furthermore, several improvements were seen with NSCs expressing GDNF in SOD1
G93A ALS rats, when administered in the motor cortex
[23] and in the spinal cord
[24]. The results show improved survival, as well as enhanced proliferative and neuroprotective properties
[23][24]. Moreover, human GDNF-expressing NSCs duly migrated to the disease site and integrated into the CNS after administration into the spinal cord of SOD1
G93A ALS rats
[24]. In addition, it has been shown that GDNF-expressing NSCs administration in the lateral ventricle promotes axonal regeneration and remyelination in chronic EAE rats
[25].
A number of studies have indicated that immune cells are also useful as therapeutic biosystems to deliver various molecules into target areas
[27][28]. Among the subsets of immune cells, macrophages are the most suitable target cells, as they are activated soon after the onset of the inflammatory response, can cross the BBB, and move to sites of neuronal degeneration
[27][28]. In this regard, the monocyte-macrophage lineage could represent an efficient cellular system to deliver NTFs at the site of injury within the CNS. To support this hypothesis, Biju et al. used ex vivo transduced bone marrow-derived macrophages to deliver GDNF
[27]. Axonal regeneration and retention of tyrosine hydroxylase (TH+) neurons were observed in both the striatum and substantia nigra regions
[27]. Moreover, GDNF-expressing macrophages could successfully cross the BBB and deliver GDNF into the neuro-generated DA neurons after systemic administration
[28].
Finally, other cells, such as fibroblasts, were also used as vectors to deliver NTFs. Specifically, fibroblasts modified to express BDNF were inoculated into SCI sites in rats, and these caused regenerative and sprouting responses at the sites of injury
[97][98][99]. Similarly, genetically modified baby hamster kidney (BHK) cells and primary cells expressing NGF showed that they were able to rescue cholinergic function in damaged neurons in ageing models of both rodents and non-human primates
[100][101][102]. More interestingly, the implanted cells maintained NGF secretion for at least 8 months in primate brains and did not cause the adverse side effects observed in studies with direct administration
[103][104][105].
To date, research advances in cell-based therapies offer promising methods for treating neurodegenerative diseases. Although much work remains to be done, the increasing focus on preclinical studies and the recent translation of some of these therapies into clinical trials have paved the way for further progress. The use of modified cells expressing NTFs is likely to play a key role in future clinical strategies to treat neurodegenerative diseases by replacing dysfunctional neurons and providing neuroprotective functions. As mentioned earlier, a potential drawback that remains today is the biosafety.
4. Viral Delivery of Neurotrophic Factors
Viral vector-mediated gene delivery might be a more optimal approach instead of the techniques that have been previously described. Virus administration would permanently alter the cells’ ability to make its own NTF, requiring a single injection at the site of administration, rather than multiple injections
[106][107][108][109], and eliminating the cumbersome cell preparation associated with the cell transfer technique
[29][31][32][33][35][106][107][108][109][110].
Nakajima et al.
[29] reported that injection of adenovirus (AV)-BDNF into bilateral sternomastoid muscles transferred vectors to the damaged sites, via retrograde transport using spinal accessory motor neurons, in SCI rats. The AV-BDNF was able to reach the spinal cord and reduce apoptotic signalling in neurons and oligodendrocytes
[29]. Likewise, the application of retrograde AV-BDNF in bilateral sternomastoid muscles of chronically compressed SCI mice led to the recovery of oligodendrocyte progenitors and neurofilament expression via the axons of spinal accessory nerves
[31]. However, there are some drawbacks using AV vectors, including immunogenicity, replicability, and the small insertion size of the vectors
[29][31].
To date, adeno-associated virus (AAV)-mediated gene transfer of GDNF has been used and evaluated in a number of studies in rodents and primates, particularly for PD
[109], HD
[106][107], and SCI
[32]. Eslamboli et al.
[109] showed that unilateral intrastriatal injection of AAV-GDNF, resulting in the expression of high levels of GDNF in the striatum, induced a significant bilateral increase in tyrosine hydroxylase protein levels and DA turnover in a 6-OHDA lesion in marmosets. In addition, AAV-GDNF-treated rats scored better on a blinded semi-quantitative neurological scale compared to rats receiving the control AAV- Green Fluorescent Protein (GFP), which was supported by histological analyses
[107]. Interestingly, Fouad et al.
[32] reported that rats, with complete thoracic SCI, that received combined treatment, including self-complementary AAV-BDNF and NT-3 administration in the spinal cord, showed not only improved axonal regeneration, but also improved motor function of the hind limbs
[32]. AAV vectors offer many of the same advantages as AV vectors, including a wide host-cell range and a relatively high transduction efficiency. In addition, AAV vectors do not express their own proteins and, therefore, would not elicit an immune response, making the technique even more attractive. However, the major drawback is the limited cloning capacity of the vector, which restricts its use in the gene delivery of large genes
[32][106][107][109].
Next to AV- and AAV- mediated NTF delivery, viral delivery of GDNF by lentivirus (LV) reversed motor deficits and prevented nigrostriatal degeneration in MPTP-treated monkeys
[110]. The delivery of LV expressing GDNF to AD mice models enhanced learning and memory function, while simultaneously improving the cognition capacity
[33]. In addition, the group of Pereira de Almeida et al.
[111][112] conducted two studies using tetracycline-regulated LV-mediated delivery of CNTF in a quinolinic acid (QA) rat model of HD. The 2001 study
[111] showed that the extent of striatal damage was significantly reduced in the CNTF-treated rats, and the volume of the lesion was significantly reduced
[111]. In 2002, they reported CNTF′s dose-dependent effects
[112]. Remarkably, LV-based administration has numerous advantages, such as long-term transgene expression, low inflammation rate, and large-size gene insertion
[34][35][113]. Despite these advantages, in some cases, oncogenic mutation may occur after integration of the LV gene into the host cell genome. This is cited as the main concern of safety in in vivo conditions.
5. Biomaterials to Deliver Neurotrophic Factors
Several of the above-mentioned strategies to deliver NTF to the site of injury in the spinal cord or brain, such as direct delivery, genetically engineered cells, and viral vectors, have a number of drawbacks, including viral vector spread beyond the target area, uncontrolled transgene expression, and immune rejection of transplanted cells. Therefore, there is a growing interest in using biomaterials as vehicles to deliver NTFs. Natural biomaterials are biocompatible, biodegradable, have remodelling advantages and a lower toxicity rate
[114], while synthetic biomaterials have a more favourable mechanical and thermal resistance, no immune response capacity, and can be produced on large scales
[36][37].
A recent study by Zhijiang et al.
[114] used the natural biomaterial methylcellulose (MC), combined with hyaluronic acid (HAMC) hydrogel modified with the peptide KAFAK-LAARLYRKALARQLGVAA (KAFAK) and BDNF. They injected these into a lesion area of SCI rats and showed that locomotor function and axonal regeneration improved 8 weeks after SCI
[114]. A similar study with NT-3 also showed that HAMC could release NT-3 for 28 days. The persistence of NT-3 in the target areas confirmed the regeneration and expansion of axons, without induction of the astroglial response, which can cause an inflammatory reaction
[38]. Furthermore studies have used other natural bio-materials, such as bioactive scaffolds, to create a microenvironment conducive to endogenous regeneration of neuronal tissue in the SCI site. In particular, gelatin sponge scaffold, silk fibroin, chitosan, or a more developed multichannel nanofibrous gelatin scaffold have been used. These scaffolds were integrated into NT-3, with or without NSCs
[43], adipose-derived stem cells
[42], or MSCs
[44][115]. The in vivo experiments have significantly improved neuronal differentiation, synaptic connection, and axonal remyelination, with reduced local inflammation at the SCI sites following bioactive scaffold implantation with NT-3. In addition the treatment has shown significant improvement in locomotor functionality
[39][42][43][44][115].
Poly-lactide-co-glycolide (PLG) is one of the most frequently used synthetic biomaterials for drug delivery, due to its controlled and sustained release properties, low toxicity, and biocompatibility with tissue and cells
[45][46]. PLG has been widely used as a material for spinal cord repair or peripheral nerve conduits
[46]. Khalin et al. found that iv injection of poloxamer 188 (PX)-coated PLG nanoparticles with BDNF (PLG-BDNF) in TBI mice restored cognition and showed that this system is eligible to cross the BBB and deliver BDNF into the brain of the TBI model
[37]. Furthermore, several studies with PLG-BDNF in animal models of SCI observed robust axon growth and remyelination 6 months after initial injury
[38][46][47]. These positive findings of PLG-BDNF were not confirmed with CNTF. The latter would not be sufficient in vivo to promote oligodendrocyte remyelination in the glial-depleted environment of unilateral ethidium bromide lesions
[48]. Similar to the PLG-BDNF results in SCI rats, poly N-isopropylacrylamide (PNIPAAm) with BDNF improved the axonal regeneration in SCI rats
[36]. Finally, intrathecal infusion of N-terminal pegylated (PEG) BDNF (PEG-BDNF) was also used in an attempt to increase NTF release
[116]. The authors showed that the PEG-BDNF was able to reach the spinal cord and that its expression was induced in that area. However, they could not observe an improved axonal response or recovery of motor function, which suggests that the amount of BDNF was insufficient
[116].
As mentioned earlier, most NTFs have difficulties passing through the BBB and are, therefore, delivered directly into the brain in animal models and some clinical trials with patients using expensive and risky intracranial surgery
[69][70][71]. The efficiency of delivery and the poor distribution of some NTFs in the brain are considered the main problems behind their modest effects in clinical trials. There is a great need for NTFs that can be administered systemically to avoid intracranial surgery. Nanoparticles (NPs) can be used to stabilise NTFs and facilitate their transport through the BBB
[117]. For example, one study used plasmid DNA NPs encoding human GDNF (pGDNF) that were administered IN to a rat model of PD
[118]. The amphetamine-induced rotational behaviour was reduced, and dopaminergic fibre density and cell counts in the lesioned substantia nigra and nerve terminal density in the lesioned striatum were significantly preserved in rats given IN pGDNF
[118].