Designing artificial senses: steps from physiology to clinical implementation.

Our senses are the main information channels through which we perceive and interact with the world. Consequently, the physical and social functioning of patients suffering from severe sensory disabilities is limited on several levels. This has motivated the development of a novel therapeutic alternative: “artificial senses”, more commonly known as sensory neuroprostheses. In order to restore lost function, sensory neuroprostheses attempt to take advantage of the information transfer pathway common to all senses: (i) transduction of the physical stimulus by sensory receptors, (ii) transmission of relevant information to primary sensory areas in the brain by sensory afferents, and (iii) analysis and integration of the information at multiple levels in the central nervous system. Neurosensory deficits might occur upon damage to any of the structures involved in this process. However, damage to the peripheral sensory receptor is often the cause of neurosensory loss. Most sensory neuroprostheses attempt to “replace” the malfunctioning or missing peripheral sensory organ by directly delivering basic sensory information to the brain using electrical currents. If the prosthesis is able to deliver enough consistent information, the brain will be able to correctly interpret it and useful rehabilitation can be achieved. This review presents the main challenges related to the development, implementation and translation to clinical practice of these devices: (i) sensory information needs to be efficiently delivered to specific neural targets (e.g., peripheral afferents or specific central nuclei); (ii) then the expected physiological response must be evoked and quantified; (iii) the restoration of basic sensory abilities can lead to useful rehabilitation in meaningful everyday activities; (iv) optimal prospects require specific rehabilitation therapy and lifelong medico-technical follow-up. To conclude, the current state and future of sensory neuroprostheses will be discussed. This will include current clinical and technical challenges, future prospects, and the potential of these devices to improve our fundamental knowledge of sensory physiology and neurosensory deficits.

The impact of motor and sensory nerve architecture on nerve regeneration.

Sensory nerve autografting is the standard of care for injuries resulting in a nerve gap. Recent work demonstrates superior regeneration with motor nerve grafts. Improved regeneration with motor grafting may be a result of the nerve's Schwann cell basal lamina tube size. Motor nerves have larger SC basal lamina tubes, which may allow more nerve fibers to cross a nerve graft repair. Architecture may partially explain the suboptimal clinical results seen with sensory nerve grafting techniques. To define the role of nerve architecture, we evaluated regeneration through acellular motor and sensory nerve grafts. Thirty-six Lewis rats underwent tibial nerve repairs with 5 mm double-cable motor or triple-cable sensory nerve isografts. Grafts were harvested and acellularized in University of Wisconsin solution. Control animals received fresh motor or sensory cable isografts. Nerves were harvested after 4 weeks and histomorphometry was performed. In 6 animals per group from the fresh motor and sensory cable graft groups, weekly walking tracks and wet muscle mass ratios were performed at 7 weeks. Histomorphometry revealed more robust nerve regeneration in both acellular and cellular motor grafts. Sensory groups showed poor regeneration with significantly decreased percent nerve, fiber count, and density (p<0.05). Walking tracks revealed a trend toward improved functional recovery in the motor group. Gastrocnemius wet muscle mass ratios show a significantly greater muscle mass recovery in the motor group (p<0.05). Nerve architecture (size of SC basal lamina tubes) plays an important role in nerve regeneration in a mixed nerve gap model.

Regeneration of baroafferents after implantation into different vessels.

Regeneration of peripheral nerves involves an essential contribution by surrounding tissues. This study focuses on the role of the target tissue on the regeneration of afferent peripheral nerves. We hypothesized that nerves implanted into the appropriate target tissue regain their function, whereas they degenerate when implanted into a different tissue. Therefore, aortic nerves of rabbits were transected and implanted into arteries or veins, and their function and structure was reevaluated after 1.5, 3, and 10 months. In a subset of animals, the nerves were again severed and implanted into the other vessel. Twelve of 18 nerves implanted into arteries regained typical neurophysiological activity, but none of those implanted into veins. Two times even baroreflexes were elicited through the newly built nerve endings. The structure of the nerve endings implanted into arteries resembled baroreceptors, whereas no fiber growth was detected in veins. Morphometrically, the fiber number and diameter increased over the observed time period after implantation into arteries. Nerves implanted into veins, transected after 3 months, and then implanted into arteries also regained neurophysiological activity. Again, they rebuilt baroreceptors and significantly increased their fiber number and diameter. In conclusion, when severed baroafferents are implanted into arteries, they regenerate new baroreceptors and restore the normal myelination and fiber size of the nerve over time, whereas veins seem to inhibit nerve fiber sprouting and regeneration of severed fibers.

Use of motor nerve material in peripheral nerve repair with conduits.

We have recently shown in experimental nerve injury models that nerve regeneration is enhanced across a motor nerve graft as compared with a sensory nerve graft. To test the hypothesis that nerve architecture may mediate the beneficial effect of motor nerve grafting, we developed a model of disrupted nerve architecture in which motor and sensory nerve fragments were introduced into silicone conduits. Lewis rats were randomized to 5 experimental groups: nerve repair with motor nerve fragments, sensory nerve fragments, mixed nerve fragments, saline-filled conduit (negative control), or nerve isograft (positive control). At 6, 9, or 12 weeks, animals were sacrificed and nerve tissues were analyzed by quantitative histomorphometry. No significant differences were observed between the motor, sensory, and mixed nerve fragment groups. These findings suggest that intact nerve architecture, regardless of neurotrophic or biochemical factors, is a prerequisite for the beneficial effect of motor nerve grafting.

Repair of motor nerve gaps with sensory nerve inhibits regeneration in rats.

OBJECTIVE: Sensory nerve grafts are often used to reconstruct injured motor nerves, but the consequences of such motor/sensory mismatches are not well studied. Sensory nerves have more diverse fiber distributions than motor nerves and may possess phenotypically distinct Schwann cells. Putative differences in Schwann cell characteristics and pathway architecture may negatively affect the regeneration of motor neurons down sensory pathways. We hypothesized that sensory grafts impair motor target reinnervation, thereby contributing to suboptimal outcomes. This study investigated the effect of motor versus sensory grafts on nerve regeneration and functional recovery. STUDY DESIGN: The authors conducted a prospective, randomized, controlled animal study. METHODS: Fifty-six Lewis rats were randomized to seven groups of eight animals each. Five-millimeter tibial nerve defects were reconstructed with motor or sensory nerve grafts comprised of single, double, triple, or quadruple cables. Tibial nerve autografts served as positive controls. Three weeks after reconstruction, nerves were harvested for histologic examination and quantitative histomorphometric analysis. Wet muscle masses provided an index of functional recovery. RESULTS: Nerve regeneration was significantly greater across motor versus sensory nerve grafts independent of graft cross-sectional area or cable number. Motor grafts demonstrated increased nerve density, percent nerve, and total fiber number (P < .05). Normalized wet muscle masses trended toward improved recovery in motor versus sensory groups. CONCLUSIONS: Reconstruction of tibial nerve defects with nerve grafts of motor versus sensory origin enhanced nerve regeneration independent of cable number in a rodent model. Preferential nerve regeneration through motor nerve grafts may also promote functional recovery with potential implications for clinical nerve reconstruction.

The astrocytic barrier to axonal regeneration at the dorsal root entry zone is induced by rhizotomy.

After dorsal rhizotomy, sensory axons fail to regenerate beyond the astrocytic glia limitans at the dorsal root entry zone (DREZ) but this inhibition can be overcome with the delivery of exogenous neurotrophin-3. We investigated whether axonal inhibition at the DREZ is constitutive or induced after dorsal rhizotomy. Primary afferent neurones from enhanced green fluorescent protein-expressing mice were transplanted into adult rat dorsal root ganglia in the presence or absence of dorsal rhizotomy. In the absence of dorsal rhizotomy mouse axons freely extended into the rat central nervous system. After host dorsal rhizotomy, mouse axons were unable to cross the DREZ. However, in rats that received a dorsal rhizotomy concomitant with intrathecal neurotrophin-3, the mouse axons were able to cross the DREZ. These results indicate that, under normal circumstances, the adult DREZ is permissive to the regeneration of adult sensory axons and that it only becomes inhibitory once dorsal root axons have been injured and astrocytes at the DREZ have become reactive.

Effects of motor versus sensory nerve grafts on peripheral nerve regeneration.

Autologous nerve grafting is the current standard of care for nerve injuries resulting in a nerve gap. This treatment requires the use of sensory grafts to reconstruct motor defects, but the consequences of mismatches between graft and native nerve are unknown. Motor pathways have been shown to preferentially support motoneuron regeneration. Functional outcome of motor nerve reconstruction depends on the magnitude, rate, and precision of end organ reinnervation. This study examined the role of pathway type on regeneration across a mixed nerve defect. Thirty-six Lewis rats underwent tibial nerve transection and received isogeneic motor, sensory or mixed nerve grafts. Histomorphometry of the regenerating nerves at 3 weeks demonstrated robust nerve regeneration through both motor and mixed nerve grafts. In contrast, poor nerve regeneration was seen through sensory nerve grafts, with significantly decreased nerve fiber count, percent nerve, and nerve density when compared with mixed and motor groups (P < 0.05). These data suggest that use of motor or mixed nerve grafts, rather than sensory nerve grafts, will optimize regeneration across mixed nerve defects.

Towing of sensory axons by their migrating target cells in vivo.

Many pathfinding axons must locate target fields that are themselves positioned by active migration. A hypothetical method for ensuring that these migrations are coordinated is towing, whereby the extension of axons is entirely dependent on the migration of their target cells. Here we combine genetics and time-lapse imaging in the zebrafish to show that towing by migrating cells is a bona fide mechanism for guiding pathfinding axons in vivo.

A model for implanting neuronal tissue into the cochlea.

Immature dorsal root ganglion (DRG) neurons have previously been shown to survive implantation to the cavity of extirpated adult native DRG, send axons via the dorsal root into the host spinal cord and make functional sypnatic connections. Regeneration or replacement of the auditory nerve would provide a major intervention in the clinical treatment of severe hearing impairment. In this study we have exploited the potential of fetal DRG neurons to survive allografting into the cochlea of adult guinea pigs. In some animals implantation of fetal DRGs was combined with infusion of neurotropic substances into the cochlea. Survival of the implanted DRG neurons was found in the majority of grafted animals. Treatment with neurotrophic factors significantly increased the number of surviving implanted DRG neurons. However, even in the absence of neurotrophic substances survival of DRG neurons was found in a majority of the animals, indicating the presence of endogenous growth promoting factors within the cochlea and/or an intrinsic capacity of fetal DRG neurons themselves to survive in this heterotropic location.

Olfactory ensheathing cells induce less host astrocyte response and chondroitin sulphate proteoglycan expression than Schwann cells following transplantation into adult CNS white matter.

Both Schwann cells and olfactory ensheathing cells (OECs) create an environment favorable to axon regeneration when transplanted into the damaged CNS. However, transplanted cells can also exert an effect on the host tissue that will influence the extent to which regenerating axons can grow beyond the transplanted area and reenter the host environment. In this study equivalent numbers of Lac-Z-labeled Schwann cells and OECs have been separately transplanted into normal white matter of adult rat spinal cord and the host astrocyte response to each compared. Schwann cell transplantation resulted in a greater area of increased glial fibrillary acidic protein (GFAP) expression compared to that associated with OEC transplantation. This was accompanied by a greater increase in the expression of axon growth inhibitory chrondroitin sulfate proteoglycans (CSPGs) following Schwann cell transplantation compared to OEC transplantation. However, no differences were detected in the increased expression of the specific CSPG neurocan following transplantation of the two cell types. These results mirror differences in the interactions between astrocytes and either Schwann cells or OECs observed in tissue culture models and reveal one aspect of the complex biology of creating regeneration-promoting environments by cell transplantation where transplanted OECs have favorable properties compared to transplanted Schwann cells.

Allografted fetal dorsal root ganglion neuronal survival in the guinea pig cochlea.

Neural grafting is a potential strategy to help restore auditory function following loss of spiral ganglion cells. As a first step towards the reconstruction of a neural pathway from the cochlea to the brainstem, we have examined the survival of fetal dorsal root ganglion (DRG) neurons allografted into the cochlea of adult guinea pigs. In some animals implantation of DRGs was combined with a local infusion of neurotrophic substances whereas in others auditory sensory receptors were chemically destroyed prior to DRG implantation by injection of the ototoxin neomycin into the middle ear. The results show that many transplanted DRG neurons attached close to the cochlear spiral ganglion neurons. The survival of the implant was significantly increased by treatment with neurotrophic factors, but not reduced by the absence of auditory sensory structures. This study shows that implanted sensory neurons can survive heterotopic grafting immediately adjacent to the eighth cranial nerve, thereby providing a basis for further studies of the anatomical and functional influence of neural grafts in the inner ear.

Grafted swine neuroepithelial stem cells can form myelinated axons and both efferent and afferent synapses with xenogeneic rat neurons.

Neuroepithelial stem cells derived from the swine mesencephalic neural tube were examined regarding their eligibility for neural xenografting as a donor material, with the aim of evaluating myelinated axon formation and both types of synaptic formation with xenogeneic host neurons as part of possible neural circuit reconstruction. The mesencephalic neural tube tissues were dissected out from swine embryos at embryonic days 17 and 18 and were implanted immediately into the striatum of the Parkinsonian model rat. The swine-derived grafts had many nestin-positive rosette-forming, neurofilament-positive, and tyrosine hydroxylase-positive cells in the rat striatum. Electron microscopic study revealed both efferent and afferent synaptic formations in the donor-derived immature neurons or tyrosine hydroxylase-positive donor cells in the grafts. Myelinated axons, both positive and negative for swine-specific neurofilament antibody, were mingled together in the graft. These results indicated that implanted neuroepithelial stem cells could survive well and divide asymmetrically into both nestin-expressing precursors and differentiated neurochemical marker-expressing neurons in the xenogeneic rat striatum, with the help of an immunosuppressant. Donor-derived immature neurons formed both efferent and afferent synapses with xenogeneic host neurons, and donor-derived axons were myelinated, which suggests that implanted swine neuroepithelial stem cells could possibly restore damaged neuronal circuitry in the diseased brain.

Sensory restoration of the skin graft on a free muscle flap: experimental rabbit study.

Transplantation of a muscle flap with free skin graft for wound coverage is a common procedure in reconstructive microsurgery. However, the grafted skin has little or no sensation. Restoration of the sensibility of the grafted skin on the transferred muscle is critically important, especially in palmar hand, plantar foot, heel, and oral cavity reconstruction. The purpose of this study was to investigate the possibility of sensory restoration of the grafted skin on a trimmed muscle surface that has been sensory neurotized after sensory nerve-to-motor nerve transfer, using the rabbit gracilis muscle as an animal model. The ipsilateral saphenous nerve (sensory) was transferred to the motor nerve of the gracilis muscle for sensory neurotization. A 4 x 4-cm2 area of skin island over the midportion of the gracilis muscle was harvested as a full-thickness skin graft. The upper half of the gracilis muscle was then excised, becoming a rough surface. The harvested skin was reapplied on the trimmed rough surface of the muscle. After 6 months, retrograde and antegrade horseradish peroxidase labeling studies were performed through skin and muscle injection. The group with a free skin graft was compared with the group with an intact surface of the gracilis muscle. This study clearly shows that sensory nerves can regenerate and penetrate into the trimmed muscle surface and grow into the overlying grafted skin. However, if the muscle surface is intact as with the compared group, sensory reinnervation of the grafted skin is not possible.

Transplantation of nasal olfactory tissue promotes partial recovery in paraplegic adult rats.

Recent reports have highlighted the potential therapeutic role of olfactory ensheathing cells for repair of spinal cord injuries. Previously ensheathing cells collected from the olfactory bulbs within the skull were used. In humans a source of these cells for autologous therapy lies in the nasal mucosa where they accompany the axons of the olfactory neurons. The aim of the present study was to test the therapeutic potential of nasal olfactory ensheathing cells for spinal cord repair. Olfactory ensheathing cells cultured from the olfactory lamina propria or pieces of lamina propria from the olfactory mucosa were transplanted into the transected spinal cord. Three to ten weeks later these animals partially recovered movement of their hind limbs and joints which was abolished by a second spinal cord transection. Control rats, receiving collagen matrix, respiratory lamina propria or culture medium, did not recover hind limb movement. Recovery of movement was associated with recovery of spinal reflex circuitry, assessed using the rate-sensitive depression of the H-reflex from an interosseous muscle. Histological analysis of spinal cords grafted with olfactory tissue demonstrated nerve fibres passing through the transection site, serotonin-positive fibres in the spinal cord distal to the transection site, and retrograde labelling of brainstem raphe and gigantocellularis neurons from injections into the distal cord, indicating regeneration of descending pathways. Thus, olfactory lamina propria transplantation promoted partial restoration of function after relatively short recovery periods. This study is particularly significance because it suggests an accessible source of tissue for autologous grafting in human paraplegia.

Functional connections are established in the deafferented rat spinal cord by peripherally transplanted human embryonic sensory neurons.

Functionally useful repair of the mature spinal cord following injury requires axon growth and the re-establishment of specific synaptic connections. We have shown previously that axons from peripherally grafted human embryonic dorsal root ganglion cells grow for long distances in adult host rat dorsal roots, traverse the interface between the peripheral and central nervous system, and enter the spinal cord to arborize in the dorsal horn. Here we show that these transplants mediate synaptic activity in the host spinal cord. Dorsal root ganglia from human embryonic donors were transplanted in place of native adult rat ganglia. Two to three months after transplantation the recipient rats were examined anatomically and physiologically. Human fibres labelled with a human-specific axon marker were distributed in superficial as well as deep laminae of the recipient rat spinal cord. About 36% of the grafted neurons were double labelled following injections of the fluorescent tracers MiniRuby into the sciatic and Fluoro-Gold into the lower lumbar spinal cord, indicating that some of the grafted neurons had grown processes into the spinal cord as well as towards the denervated peripheral targets. Electrophysiological recordings demonstrated that the transplanted human dorsal roots conducted impulses that evoked postsynaptic activity in dorsal horn neurons and polysynaptic reflexes in ipsilateral ventral roots. The time course of the synaptic activation indicated that the human fibres were non-myelinated or thinly myelinated. Our findings show that growing human sensory nerve fibres which enter the adult deafferentated rat spinal cord become anatomically and physiologically integrated into functional spinal circuits.

Striatal dopaminergic afferents concentrate in GDNF-positive patches during development and in developing intrastriatal striatal grafts.

Glial cell line-derived neurotrophic factor (GDNF) has potent trophic action on fetal dopaminergic neurons. We have used a double immunocytochemical approach with antibodies that recognize GDNF and tyroxine hydroxylase (TH) or the phosphoprotein DARPP-32, to study the developmental pattern of their interactions in the rat striatum and in intrastriatal striatal transplants. Postnatally, at one day and also at 1 week, GDNF showed a patchy distribution in the striatum, together with a high level of expression in the lateral striatal border, similar to that observed for the striatal marker DARPP-32 and also for TH. In the adult striatum, there was diffuse, weak immunopositivity for GDNF, together with widespread expression of DARPP-32-positive neurons and TH-immunoreactive (TH-ir) fibers. In 1-week-old intrastriatal striatal transplants, there were some GDNF immunopositive patches within the grafts and although there was not an abundance of TH-positive fibers, the ones that were seen were located in GDNF-positive areas. This was clearly evident in 2-week-old transplants, where TH-ir fibers appeared selectively concentrated in GDNF-positive patches. This pattern was repeated in 3-week-old grafts. In co-transplants of mesencephalic and striatal fetal tissue (in a proportion of 1:4), TH-ir somata were located mainly at the borders of areas that were more strongly immunostained for GDNF, and TH-ir fibers were also abundant in these areas and were found in smaller numbers in regions that were weakly positive for GDNF. These results demonstrate that GDNF-ir is coincident with that for TH and DARPP-32, and suggest that GDNF release by fetal striatal neurons both in normal development and in developing striatal grafts may have not only a trophic but also a tropic influence on TH-ir fibers and may be one of the factors that regulate dopaminergic innervation of the striatum.

Target neuron specification of short-term synaptic facilitation and depression in the cricket CNS.

We investigated the role of retrograde signals in the regulation of short-term synaptic depression and facilitation by characterizing the form of plasticity expressed at novel synapses on four giant interneurons in the cricket cercal sensory system. We induced the formation of novel synapses by transplanting a mesothoracic leg and its associated sensory neurons to the cricket terminal abdominal segment. Axons of ectopic leg sensory neurons regenerated and innervated the host terminal abdominal ganglion forming monosynaptic connections with the medial giant interneuron (MGI), lateral giant interneuron (LGI), and interneurons 7-1a and 9-2a. The plasticity expressed by these synapses was characterized by stimulating a sensory neuron with pairs of stimuli at various frequencies or with trains of 10 stimuli delivered at 100 Hz and measuring the change in excitatory postsynaptic potential amplitude recorded in the postsynaptic neuron. Novel synapses of a leg tactile hair on 7-1a depressed, as did control synapses of cercal sensory neurons on this interneuron. Novel synapses of leg campaniform sensilla (CS) sensory neurons on MGI, like MGI's control synapses, always facilitated. The form of plasticity expressed by novel synapses is thus consistent with that observed at control synapses. Leg CS synapses with 9-2a also facilitated; however, the plasticity expressed by these sensory neurons is dependent on the identity of the postsynaptic cell since the synapses these same sensory neurons formed with LGI always depressed. We conclude that the form of plasticity expressed at these synaptic connections is determined retrogradely by the postsynaptic cell.

Olfactory ensheathing cells: potential for glial cell transplantation into areas of CNS injury.

Ensheathing cells are the glial cells that ensheath olfactory axons within both the PNS and CNS portions of the primary olfactory pathway. These glial cells express a mixture of astrocyte-specific and Schwann cell-specific phenotypic features, support axonal growth by olfactory as well as by non-olfactory neurons, and survive transplantation into injured areas of the CNS. This review article focuses on those phenotypic features that are expressed by ensheathing cells that make them ideal candidates for transplantation into wound cavities in the damaged spinal cord of humans. Although much work remains to be done before such a therapeutic approach can be tried, the likelihood that ensheathing cells could simultaneously perform the roles of both astrocytes and Schwann cells following transplantation is the justification for developing such a therapeutic approach using animal models of spinal cord injury.