CONCEPTS AND INNOVATIONS

MICROSCALE SURGERY
David W. Sretavan, M.D., Ph.D.
Departments of Ophthalmology and Physiology, Program in Neuroscience, Bioengineering Graduate Program, University of California, San Francisco, San Francisco, California

ON

SINGLE AXONS

Wesley Chang, Ph.D.

Departments of Ophthalmology and Physiology, Program in Neuroscience, Bioengineering Graduate Program, University of California, San Francisco, San Francisco, California

Elizabeth Hawkes, M.S.

Departments of Ophthalmology and Physiology, University of California, San Francisco, San Francisco, California

Christopher Keller, Ph.D.

MEMS Precision Instruments, Richmond, California

OBJECTIVE: The lack of meaningful axon regeneration after central nervous system damage and poor functional recovery after serious peripheral nervous system nerve injuries have been long-standing problems of substantial interest to both neurosurgeons and neurobiologists. As an alternative to strategies that seek to promote the regeneration of adult axons, our research group has taken advantage of advances in microtechnology to develop a paradigm of direct axon repair involving the substitution of damaged axon regions with healthy segments from donor axons. METHODS: This repair methodology uses a novel combination of microtechnology, electrokinetic axon manipulation, and the well-established biological principle of cell fusion. These three fields of research have been integrated in a multidisciplinary approach to develop a solution for a significant clinical problem that currently has no specific treatment. RESULTS: The findings reported here provide some initial proof of principle for the core technologies we intend to use for axon repair. Functional recovery from nerve damage of course is clinically challenging, and many obstacles would need to be overcome before such axon repair procedures can be contemplated for therapeutic use. We identify some of the clinical issues that must be addressed for microtechnology-assisted axon repair to transition from the realm of research into actual surgical settings. CONCLUSION: It is hoped that each advance in axon repair technology will spur additional research to provide us with a comprehensive understanding on how best to pursue neurosurgical intervention at the microscale.
KEY WORDS: Axon, Microdevice, Microsurgery, Nanotechnology, Nerve injury, Repair
Neurosurgery 57:635-646, 2005
DOI: 10.1227/01.NEU.0000175545.57795.ac

Michel Kliot, M.D.

Department of Neurosurgery, University of Washington, Seattle, Washington Reprint requests: David Sretavan, M.D., Ph.D., K107, Box 0730, University of California, San Francisco 10 Kirkham Street, San Francisco, CA 94143. Email: SretavanD@vision.ucsf.edu Received, February 22, 2005. Accepted, June 14, 2005.

www.neurosurgery-online.com

erve injuries result in significant functional disability for patients and represent a substantial economic cost for society. For spinal cord injuries, in which statistical data is best documented, roughly 11,000 new cases of cord injury occur each year in the United States. Given that more than 60% of spinal cord injuries occur in patients 15 to 30 years of age who have yet to enter their most productive years, these injuries impose an enormous healthcare burden on society. However, a discussion of the true impact of nerve injuries also must take into account the large numbers of peripheral nerve injuries that result in various degrees of temporary or permanent disability. Statistics show that in North America, as many as 3% of all trauma cases have major peripheral nervous system (PNS) nerve injuries resulting in severe sensory and motor loss. The ensuing

rehabilitation can be protracted over a number of years and, in many instances, can result in only minimal recovery.

NERVE REPAIR AND PNS REGENERATION

Conventional teaching states that injured PNS axons can successfully regenerate and reconnect with their targets. PNS axons that are injured but remain confined within their endoneurial sheaths indeed can sprout a new growth cone and grow back to their postsynaptic partners (8). In actual clinical situations, however, peripheral nerve damage often disrupts the connective tissue layers in the nerve, so that axons are not able to track and regenerate within their original endoneurial sheaths. Regenerating axons in the PNS also grow quite slowly, at 1 to 2 mm/d, and often

NEUROSURGERY

VOLUME 57 | NUMBER 4 | OCTOBER 2005 | 635

SRETAVAN

ET AL.

require up to 1 to 2 years to grow back to their target cells after severe injury. During this extended period, neurons can die from the lack of retrograde trophic support, and denervated target tissues such as muscles can atrophy. All of these factors lead to the fact that although recovery occurs, full recovery after significant peripheral nerve injury is the exception rather than the rule.

CURRENT STRATEGIES FOR CENTRAL NERVOUS SYSTEM AXON REGENERATION

Although PNS axon regeneration is possible under favorable circumstances, there is no meaningful axon regeneration in the central nervous system (CNS). The lack of CNS axon regeneration is thought to be in part the result of inhibitory proteins associated with adult CNS myelin (2, 11, 16), inhibitory proteoglycans upregulated at the site of CNS injury (32), and an intrinsic downregulation of the growth ability of adult neurons (12, 22). Current research and therapeutic strategies include the promotion of axon regeneration from the site of damage back to original neuronal targets (25, 47, 49) and the tension-mediated stretching of axons to bridge regions of nerve injury (50). In approaches using regeneration strategies, axons not only must be coaxed to grow over long distances, but also this growth must be directed back to specific neuronal targets and not form abnormal neuronal circuitry. Furthermore, regenerating axons must somehow reestablish the original highly specific pattern of synaptic contacts on individual postsynaptic cells. Last, all of this must occur rapidly enough before denervation atrophy of postsynaptic neurons (15, 24, 51) or apoptosis of the original damaged neurons (1).

FIGURE 1. Diagram showing the proposed scheme for axon repair. A, the gray region in the middle of the axon represents the injured segment. In Step 1, the damaged axon region is excised. In Step 2, a segment of healthy donor axon is brought into place. In Step 3, the host and donor axons are joined together to reestablish functional continuity. B, in instances where a damaged nerve can be mobilized to eliminate a gap between the distal and proximal ends, a similar sequence of axon alignment and joining can be used without resorting to the use of donor axons. In Step 1, the damaged region is excised. In Step 2, the nerve is mobilized and the axons are aligned. In Step 3, the axons are joined.

SURGICAL REPAIR OF AXONS: VENTURING INTO THE MICROSCALE

As an alternative to axon regeneration, we propose a novel nerve repair paradigm in which the damaged region of individual axons is excised and replaced by healthy donor axon segments to reestablish neuronal connectivity and function (Fig 1A). Direct axon repair can be broken down into three essential steps, beginning first with axon cutting to excise damaged axon regions, leaving clean healthy axon ends. Next, a donor axon segment is brought in to fill the gap between the ends of the host axons. As soon as the ends of the host and donor axons are aligned and apposed, the axon segments then are fused to establish functional integrity. A similar sequence of axon alignment and fusion also would apply in the instance of injuries in which nerve mobilization can eliminate gaps resulting from damage, and thus eliminate the need for donor axon segments (Fig. 1B). Compared with strategies based on axon regeneration, we propose that axon repair should be undertaken soon after nerve injury, before Wallerian degeneration and the loss of the distal axon segment with its synaptic connections. If successful, the reestablishment of axonal continuity, together with the presence of an intact synaptic

branching pattern, could lead to recovery of neuronal function. To our knowledge, a method for the systematic and controlled repair of individual axons has not been proposed previously. The manipulation of axons whose diameter may be in the micron to submicron range is not feasible using current microsurgical tools and techniques. However, it is precisely at these small-length scales that microtechnology and nanotechnology excel. The potential impact of microtechnology and nanotechnology on the neurosurgical operative environment has been recognized, and these emerging fields have been predicted to form the scientific foundations of new paradigms (31, 44). With these ideas in mind, our multidisciplinary research group set out 4 years ago to obtain proof of principle for the basic steps of axon repair. In this article, we discuss and demonstrate some of the microscale core technologies we believe are highly promising for axon repair. In addition, we also present our prototype miniature multifunctional axon surgery platform that is approximately 1 mm3 and that contains some of the functionality necessary for axon surgery. By exploiting advantages of microfabrication, noncontact electrokinetic methods for axon manipulation, and fundamental principles of neurobiology, it may be feasible in the future to develop Micro Electro-Mechanical Systems (MEMS) surgical microdevices that will allow neurosurgeons to engage in direct axon repair at a cellular level.

MICROFABRICATED DEVICES FOR AXON CUTTING

We have used MEMS microengineering fabrication techniques to develop small surgical devices with nanoscale fea-

636 | VOLUME 57 | NUMBER 4 | OCTOBER 2005

www.neurosurgery-online.com

MICROSCALE AXON SURGERY

tures capable of repeatable, precise cutting of single axons. MEMS has its roots in the semiconductor industry in which lithography, chemical etching, and silicon doping are used in the manufacturing of integrated circuits. The ability of these processing methods to create precise microscale features in silicon was adopted for the manufacture of micron-sized devices. Rather than the electronic properties of single crystal silicon, MEMS exploits silicons excellent mechanical properties (40). With the addition of sensors, force generating actuation mechanisms, as well as electronic controls, MEMS microdevices are capable of carrying out useful function at unprecedented microscale levels. Given that the mechanical strength of silicon is superior to that of steel (48), MEMS devices are robust and have been used in everyday products such as automobile accelerometers, ink jet printer heads, and video projectors. An active area of MEMS research is the development of biomedical microdevices for diagnostics and healthcare (42, 44). BioMEMS devices are well suited for handling the small tissue or fluid samples often typical of biotechnology research and medical diagnostics. The small size of BioMEMS devices also opens new opportunities for use within the human body for imaging, cellular diagnosis, and tissue engineering.

based compliant knife suspension (Fig. 2, B and C). The knife edge has a radius of curvature of only approximately 20 nm, similar to the diameter of a single microtubule (32) or the width of synaptic clefts (29). The knife is constructed from a thin silicon nitride membrane and is effectively transparent, allowing the optical monitoring of axons during the cutting procedure (Fig. 2D). The mechanical compliance of the suspension can be tuned to any desired stiffness within a range to deliver sufficient force for cutting various biological tissues from single axons to the harvesting of specific cell populations from histological tissue sections.

CUTTING DEVICE PERFORMANCE

These cutting devices are capable of reliable and precise axon cutting (Fig. 3) that leave clean edges and are superior to current cell harvesting methods using razor blades or glass micropipettes that tend to shear and tear axons. Although such imprecise shearing may be acceptable in tissue harvesting for cellular or biochemical studies, these methods cannot

DESIGN OF CUTTING DEVICE

Our current cutting device consists of a silicon nitride knife with an ultrasharp knife edge (Fig. 2A) mounted onto a silicon-

FIGURE 2. Photographs showing the microfabricated axon cutting device. A, axon knife assembled within a compliant suspension frame. The pyramidal structure in the center is an optically transparent axon knife with a 10- m cutting edge. f, silicon suspension flexures; k, axon knife; h, handle to micromanipulator. The footprint of the frame is 1 mm2. Scale 200 m. Devices with substantially smaller footprints can be fabricated. B, detail of the silicon suspension flexure on each side of the axon knife that provides mechanical compliance during cutting. The number of switchbacks in the flexure can be modified to obtain devices with different mechanical compliances that deliver a range of cutting forces. The compliance of the flexures allows it to act as a suspension, maximizing the durability of the knife edge. C, oblique view of the silicon-nitride knife. Knives with edges from 5 to 200 m have been fabricated and tested. A 200- m long knife is shown. D, scanning electron micrograph showing a cross section of the silicon-nitride knife at its very edge. The radius of curvature at the edge is roughly 20 nm. Scale 100 nm.

FIGURE 3. Images demonstrating an axon cutting sequence and examples of cut axons. A, axon to be cut (arrow) in center of field. Knife is at top (not in the plane of focus). B, knife positioned over axon by a micromanipulator. Note that the axon is visible through the knife. C, knife is lowered for cutting. D, the resulting cut axon. E, another axon before cutting. F, same axon as in (E) after cutting. G, examples of myelinated (top) and unmyelinated (bottom) adult mouse sciatic nerve axons that have been cut. DG, scale 10 m.

NEUROSURGERY

VOLUME 57 | NUMBER 4 | OCTOBER 2005 | 637

SRETAVAN

ET AL.

be used for axon repair. Laser thermal ablation can perform precise cutting (6) but typically involves large instrumentation that is difficult to integrate with other functional components required for repair at the micron scale of axons. Moreover, laser cutting typically is achieved by the thermal vaporization of tissue within a minimum spot size of roughly 0.5 m in diameter. Both adult PNS (sciatic nerve) and CNS (optic nerve) axons from mice rapidly reseal after they have been severed using the microcutting devices. Invertebrate axons and mammalian embryonic axons have a self-repair resealing mechanism after injury involving calcium-mediated exocytosis (14, 21, 23, 53, 61). Based on studies designed to detect the leakage of soluble cytoplasmic green fluorescent protein (GFP) from axons after cutting, we have found that this resealing ability also is evident in adult axons. This finding suggests that axons tolerate cutting by a microfabricated ultrasharp microknife. The protein and cell signaling cascades necessary for physiological self-resealing at the cut ends is operational and the intact axon segments can be used in the subsequent steps of axon repair. Future improvements to the microcutting device include sensors as well as force-generating actuation mechanisms that automatically will deliver a controlled cutting stroke. Both piezoelectric and thermal expansion actuation mechanisms can deliver forces in the range needed for axon cutting. A microcutting device with on-board sensing and actuation can function as a semiautonomous instrument, requiring only initiating commands from the human operator. The elimination of any need for manual manipulation maximally uses the precision of MEMS microdevices.

trodes (Fig. 4). In this example, application of a DEP field caused the axons to be deflected upward from the horizontal electrode toward the top half of the semicircular electrode (Fig. 4B). Axons returned to their original location after the stimulus was turned off (Fig. 4C). The direction, magnitude, and speed of axon movement within the DEP field were maintained over several cycles of DEP field initiation. This particular electrode configuration moved axons at roughly 5 m/ sec. In ongoing experiments, we are testing various electrode configurations that reliably will bring together and align the ends of two separate axons as required for the proposed method of axon repair outlined in Figure 1. The viability of cells that have undergone DEP has been studied in some detail. Fibroblasts and erythroleukemia cells show good viability as determined by trypan blue dye exclusion and continued cell proliferation after DEP (18, 56). In recent work, cortical neurons that have been subjected to DEP are capable of axon outgrowth (26, 27), indicating that axon physiology is not affected by DEP and is directly relevant to our proposed use of DEP for axon manipulation.

DEP ELECTRODE ARRAYS

The effective DEP fields for axon manipulation are relatively short ranged and extend only approximately 30 m away from the electrodes. This limited range can be very useful for isolating peak DEP forces to specific axons of interest without much effect on neighboring axons in the vicinity. In addition, it should be possible to use closely spaced elec-

DIELECTROPHORESIS FOR AXON MANIPULATION

After cutting, the host and donor axon segments must be aligned with close membrane apposition in preparation for subsequent fusion. We have chosen dielectrophoresis (DEP) as the method of choice for noncontact axon manipulation at the microscale. DEP is the movement of small polarizable objects such as cells in a nonhomogeneous AC electrical field and has been used effectively as a electrokinetic method to move and sort cells within microdevices (19, 55). In essence, objects in an alternating electrical field experience field-induced polarizations on their surfaces. In the presence of a homogeneous field, dipole forces alternate on an objects surface along with alternations in the electrical field, but no net movement of the object occurs. However, in a nonhomogeneous field with gradients of field strengths, the object experiences a net force in one direction, causing it to move toward one side of the electrical field (for review, see (20)). DEP is fundamentally distinct from electrophoresis in which the net charges on objects lead to movement in one direction within a stationary electrical field. We have found that cylindrical structures such as axons can be manipulated in DEP fields generated by micron-sized elec-

FIGURE 4. Photomicrographs demonstrating noncontact DEP manipulation of axons. Adult segments of sciatic nerve were harvested and digested with an enzyme cocktail to remove the ensheathing tissue and to release individual axons. The DEP electrodes consist of a pair of 10- m wide metal traces (black) deposited onto glass. A, arrows point to axons to be subjected to DEP fields. B, after activation of DEP field, axons moved away from the horizontal electrode toward the semicircular electrode. C, axons returned to original position after termination of DEP field. The axons moved 10 m in 1 to 2 seconds. Scale 20 m.

638 | VOLUME 57 | NUMBER 4 | OCTOBER 2005

www.neurosurgery-online.com

MICROSCALE AXON SURGERY

trode arrays (Fig. 5) to move axons at will to desired locations within a surgical field. The sequential activation of a series of addressable electrodes within the array should generate propagating waves of DEP forces capable of directing the movement and alignment of axons. Because DEP fields are determined by electrical parameters and its theoretical basis is relatively well understood, this method of axon manipulation can be conveniently preprogrammed. We envision that given the starting location of an axon of interest and a sufficient understanding of axon behavior in DEP fields, it should be possible to obtain a computer-generated set of DEP parameters and electrode activation sequences that will move the axon to any desired position. Computer-controlled activation sequences then can be executed on human command to trigger optimal stimulus parameters for axon manipulation and alignment.

AXON ELECTROFUSION
After axon cutting and alignment, the third step of axon repair is the joining of axon ends to achieve functional continuity. Cell fusion in biology is most commonly used in the creation of hybridoma cells for monoclonal antibody production. A number of methods can be used to promote cell fusion. These include chemical fusion using polyethylene glycol, laser-induced fusion, and electrofusion. We have selected electrofusion as our initial technology for axon fusion. Electrofusion results in the transient breakdown of cell membranes and the formation of pores that are unstable and eventually reseal (38). If the pores are formed in the region of contact between neighboring cells, fusion between the two cells will occur. The threshold voltage across a membrane required for this electrical breakdown is approximately 1 V. In cell culture, the typical field strengths required to generate 1 V across membranes are in the range of 2 to 8 kV/cm (5, 38), usually delivered using two to four rectangular pulses of 10 to 100 s each.

Electrofusion has been used extensively to create mammalian cell hybrids. Its ease of use and fusion yield of two orders of magnitude better than chemical polyethylene glycol fusion (28, 30) has made electrofusion frequently the technique of choice for hybridoma production. Electrofusion also has been used to fuse embryonic or fetal somatic cells with enucleated oocytes successfully to clone both sheep and mice (54, 57). The ability of fused embryo cells to develop to term, and subsequently to grow into adult animals, argue for the safety and efficacy of electrofusion. We favor the use of electrofusion for axon surgery. With the identification of appropriate parameters, electrofusion can be preprogrammed to automate axon fusion. In addition, because we will use DEP to align axons, electrodes already will be conveniently placed and electrical stimulation parameters can be switched conveniently from DEP to electrofusion. In axon repair, the pairing of DEP with electrofusion permits both steps in the repair sequence to be accomplished using electrodes, simplifying the design and operation of the MEMS axon surgery platform.

DEMONSTRATION OF AXON FUSION

The successful splicing together of two axons should create cytoplasmic continuity and the flow of protein components between the two axon segments. A direct observation of flow is feasible using axons from transgenic mice in which the gene encoding GFP has been introduced into the germ line, and GFP protein is produced in all the cells of the animal (39). GFP exists as a soluble protein in cells, causing cells to exhibit green fluorescence on appropriate excitation (excitation max, 489 nm; emission max, 508 nm). Axons from nontransgenic wildtype mice do not exhibit fluorescence under the same wavelength excitation. We have demonstrated successful fusion between GFPcontaining and nonGFP-containing axons leading to the spread of GFP between the fusion partners. These studies were performed in culture on pairs of axons that cross and lie on top of each other (Fig. 6). Figure 6A shows two bundles of retinal axons running from left to right in the field of study. The corresponding bright field picture (Fig. 6B) indicates the presence of numerous other nonGFP retinal axons in the field, including the axon of interest indicated by the arrow. Electrical field stimulation was applied locally within 20 m of the target axon pair using a pair of 12- m diameter platinum electrodes spaced 12 m apart. The appearance of GFP fluorescence in this retinal axon was observed within 5 minutes after electrofusion, and the entire axon extending between the two axon bundles exhibited GFP fluorescence by 20 minutes after electrofusion (Fig. 6C). The bright field appearance of axons 20 minutes after electrofusion is shown in Figure 6D. A second example of the appearance of GFP after electrofusion in a previously unlabeled axon is shown in Figure 6, EH. Although these data are promising, additional work must be carried out to verify the central tenet of our proposed axon repair scheme that axons that have been repaired are capable

FIGURE 5. Photomicrograph demonstrating an array of microelectrodes fabricated on a glass substrate by isotropic, wet etching and coated with a thin transparent film of conductive indium tin oxide. When energized, the high gradients of electric field at the sharp electrode tips are designed to generate DEP forces to attract axons to that location. Further developments are underway to allow each electrode to be addressed individually so that axons can be directed to specific locations within the array. Scale 100 m.

NEUROSURGERY

VOLUME 57 | NUMBER 4 | OCTOBER 2005 | 639

SRETAVAN

ET AL.

To demonstrate this concept, we designed and fabricated a prototype of an integrated axon surgery platform that is 1 mm3 in size (Fig. 7). This prototype uses a space frame (Fig. 7A) in which interlocking and latching features are used for self-fixation and to facilitate assembly. Individual components needed for axon repair such as the microcutting device and DEP/fusion electrode arrays are inserted and attached to the space frame. With the base of the various components fixed within the space frame, their relevant functional elements such as the axon knife and electrodes all extend toward the bottom surface of the cubic device closest to FIGURE 6. Examples of axon electrofusion resulting in spread of green fluorescent protein between retinal axons in culture. A, two bundles of GFP axons in the field of study. B, bright field picture of the field shown in the target axons, where they are (A) indicating an unlabeled axon (arrow). C, appearance of GFP in the same axon (arrow) 20 minutes after positioned in an optimal spatial electrofusion. D, bright field picture 20 minutes after electrofusion. Arrow indicates same axon as in C. E, GFP- relationship with one another to labeled axons. F, bright field image of the same field. Arrow points to unlabeled axons. G, within 1 minute of execute axon repair (Fig. 7, BD). electrofusion, GFP appeared in a previously unlabeled axon (arrow). H, arrow indicates same axon as in G 15 This modular design also allows minutes after fusion. AH, scale 10 m. components to be switched conveniently in and out during exof normal function and survival. Aspects of axonal function of perimentation and with design improvements. In addition, the interest include the ability for electrical conduction and the open structure allows for optical monitoring of the repair promaintenance of axonal transport. cess. Features such as microfluidics to control the fluid environment or to deliver reagents to assist axon repair can be readily incorporated into the platform. Last, a cubic framework can ASSEMBLING CORE TECHNOLOGIES INTO incorporate on-board actuation mechanisms to move the knife in A MULTIFUNCTIONAL AXON REPAIR an up-and-down motion, allowing the autonomous execution of cutting on human command, and maximally can benefit from the MICRODEVICE precision offered by microscale devices. The results presented thus far were obtained using early An example of an axon surgery platform prototype consistprototypes of MEMS axon knives and electrodes, thus undering of an assembled space frame with a microcutting device is scoring the robustness and ease of use of the underlying shown in Figure 7, E and F. This 1-mm3 prototype platform principles. However, to develop a practical methodology for was assembled using customized microfabricated microgripaxon repair, the individual axon repair steps have to be coorpers and microtools specifically designed for micron-scale dinated into an efficient sequence that is applicable at the assembly. With existing microfabrication technologies, it microscale of axons. Our strategy is to provide all of the should be quite possible to develop improved versions of axon functions necessary for axon repair within a multifunctional surgery platforms that are substantially smaller in size. AlMEMS axon surgery platform. Although it may be technically though the current cubic design offers many advantages, this feasible to perform each step involved in axon repair by opdesign is by no means fixed, and eventual multifunctional erating multiple microscale tools, each independently devices capable of the full range of axon repair steps may take mounted on its own positioner (similar to probe stations for on quite different shapes. testing integrated circuits), such an arrangement would be extremely bulky, inefficient, and time consuming to use clinINTEGRATING MICROSCALE AXON ically. Our approach solves these problems by assembling all repair functions onto a single MEMS surgical platform that SURGERY PLATFORMS INTO THE incorporates as much autonomy as possible into each repair SURGICAL FIELD step. Such a semiautonomous microscale surgical platform will achieve the degree of miniaturization and performance To use any microscale device for surgery effectively, methefficiency required for device use at a cellular level. ods must be developed to integrate such small devices for use

640 | VOLUME 57 | NUMBER 4 | OCTOBER 2005

www.neurosurgery-online.com

MICROSCALE AXON SURGERY

FIGURE 8. Conceptual drawing of a surgery chamber for axon repair. The surgery chamber provides mechanical support and environmental control for the host and donor axons as well as the axon surgery platform. (Components are not drawn to scale). Such platforms also can be used in cases where injured nerves are mobilized and proximal and distal ends of host axons are joined without the insertion of donor axons (see Fig. 1B).

FIGURE 7. Diagrams showing the design and assembly of prototype multifunctional axon surgery platform. A, schematic representation of the space frame. Carrier holes are used for positioning the platform. B, modular axon repair components such as cutting devices and electrode arrays are inserted into the space frame to preposition their functional elements for efficient axon repair. (Generic components are shown here.) In current concept, the space frame also contains a force generation (actuation) mechanism to execute the updown motion of the axon knife during cutting. C, oblique view of platform assembled with a microcutting device and electrode arrays. D, view from bottom. E, 16 individual microfabricated parts on display on a penny. A microgripper and a microprobe are shown on the left. F, scanning electron microscope image of the assembled axon surgery device prototype containing a functioning axon microcutting device and supporting pieces locked in place to form a 1 mm3 superstructure. Scale 200 m.

within a surgical field. One method is through the use of a specialized surgical chamber such as the one shown in Figure 8. Some features that are likely important for such chambers include a means of holding the host and donor nerves in place, maintenance of an appropriate fluid environment, and a mechanical anchor and micromanipulator that can position the small axon surgery platform for repair. In Figure 8, the components within the chamber, including nerves, axons, and the axon surgery platform, are designed to be mechanically isolated from the rest of the body (components are not drawn to scale). Such chambers could be positioned adjacent to peripheral nerves relatively easily, where large open fields of surgical exposure are used. A greater challenge would be to position such a device within the spinal cord or brain. Possible solutions include the use of access ports and minimally invasive techniques, coupled with the enhanced dexterity made possible by robotic devices. Together, these technologies conceivably could allow the placement and operation of a MEMS surgical platform anywhere in the nervous system.

severe nerve injuries often are viewed as requiring chronic rehabilitative care, with those with CNS injuries as having a particularly poor outlook. We think that with the advent of microscale axon repair, the perception of nerve and axon injuries may shift and may begin to be viewed as an acute surgical situation. The ability to intervene and perform axon repair before irreversible changes take place, such as Wallerian degeneration, loss of synaptic contacts, or neuronal cell death, will shift the emphasis from support and rehabilitation to earlier intervention and treatment. Given that our proposed method of axon repair is significantly different from current practice and takes place at an unprecedented small-length scale, a host of clinical issues must be considered and evaluated. Below, we lay out a number of these issues with the goal of stimulating discussion and further research into areas prompted by the possibility of neurosurgery at the microscale level.

When Should Axon Repair Take Place?

An advantage of axon repair is that, if performed soon after injury, it may be possible to reuse the synaptic connections of the distal axon segment and to achieve rapid functional recovery. This consideration sets the onset of Wallerian degeneration as the upper limit for the interval from damage to repair. In the PNS, the interval between injury and onset of axonal changes is 24 to 48 hours (8). In the CNS, Wallerian degeneration occurs more slowly or to a lesser extent after axon damage (3, 34), and the interval after injury available for intervention to conduct axonal repair may be longer. The existence of a critical interval for axon repair will necessitate adjustments in clinical management in the immediate period after injury. In addition to recognition of nerve damage as potentially requiring acute care, improved and faster methods for determining the severity of axon injury would be critical. It also should be noted that basic neurobiological research is beginning to identify some of the cell signaling pathways in Wallerian degeneration (60). It is possible that in the future, one may intervene to delay the onset of Wallerian degeneration and increase the window of opportunity for axon repair.

CLINICAL ISSUES
Currently, there is no specific treatment for axon damage from trauma either to the PNS or the CNS. Patients with

NEUROSURGERY

VOLUME 57 | NUMBER 4 | OCTOBER 2005 | 641

SRETAVAN

ET AL.

Sources of Donor Axons

In current clinical management, procedures involving peripheral nerve grafts are all performed at the nerve connective tissue level, with the sole purpose of providing a connective tissue conduit to encourage possible axon regeneration. Nerve repair therefore is not aimed directly at repairing individual axons. However, donor axons to be used as fusion partners also can be obtained from this type of autologous donor nerve segment (e.g., often the sural nerve in the lower leg). For axon repair, the harvested nerve segments can provide a source of healthy donor axon segments to be used in axon fusion. Axons will be separated from the surrounding connective tissues by enzymatic digestion of collagen and other means to eliminate the Schwann cells in peripheral nerve (Fig. 9). The exposed axons are then ready for cutting, alignment, and fusion.

Which Axons to Repair?

In clinical applications, the careful selection of which axons to repair will contribute significantly to surgical success. Because the surgical repair must be completed before the onset of distal Wallerian degeneration of axons, it will be possible to electrophysiologically map and identify cut distal axons by electrically stimulating them and recording from their target structures. One strategy is to focus on damaged axons that are physically most accessible for MEMS-assisted repair. In addition, it may be useful to target large-diameter axons for repair, because these axons mediate motor and somatosensory function. Axon repair also may be accomplished more easily in larger axons than in small axons that are less than 1 m in diameter. A further consideration is to identify situations where the reconnection of a subset of axons can make a large difference in clinical management and in the patients quality of life. An example is the restoration of partial movement in the hand and digits, which would greatly impact how tetraplegic patients interact with their environment. A second example is the repair of axon pathways mediating bladder control. After spinal cord injury, loss of voluntary control of voiding leads to urinary retention and eventually to incontinence (58). Management of these conditions by catheterization leads to urinary tract infections, which is a major source of morbidity in these patients (7). Recovery of some degree of bladder control would contribute significantly to their quality of life. The development of microscale axon repair, matched with a well thought out surgical plan and realistic expectations, will be important ingredients for clinical success.

How Many Axons Need to be Repaired?

An important point to consider is how many axons need to be repaired to obtain functional recovery. Results from studies aiming to promote CNS axon regeneration can serve as a guide. After spinal cord lesions in rodents, the regeneration of several hundred axons has been reported to correlate with some recovery of function (10, 13, 37), and it is generally agreed that the survival of a fraction of axons within a pathway is sufficient to maintain some degree of useful function (4, 33, 43, 46, 59). This finding is thought to reflect the potential for substantial plasticity in the injured nervous system. If so, this indicates that it may not be necessary to repair all axons within a PNS nerve or a CNS axon tract. The immediate repair of a subset of axons may provide a sufficient degree of functional recovery to alleviate the disability normally associated with nerve injuries. It should be recognized that only very crude estimates of the required number of repaired axons could be given at the present time. It seems likely that the number of axons to be repaired will depend on the site of injury, and that rational guidelines may be developed in the future based on direct clinical experience.

Selecting Axon Fusion Partners

A key element in axon repair is the identification of appropriate axon partners for reconnection. A characteristic feature of both CNS axon tracts and PNS nerves is the topographic organization of axons. Throughout the nervous system, axons that originate from adjacent neurons tend to run together and to maintain their neighbor relationship within a nerve or axon tract. For example, motor axons innervating a group of neighboring muscles typically are found next to each other in the same subregion within an axon tract or nerve. Similarly, axons carrying sensory information from adjacent areas of the skin surface are found next to one another both in a peripheral nerve and in the spinal cord. During repair, it thus may not be necessary to use donor axon segments specifically to reconnect the proximal axon segment with its original distal segment. Reconnection with an adjacent axon, as long as it is not an extreme mismatch, may restore acceptable function. The plasticity and remodeling of neuronal circuitry that is known to occur in the adult nervous system also may work in favor of axon repair. The extent to which this can occur is illustrated by the cortical plasticity that occurs in patients who have received intercostal nerve to musculocutaneous nerve transfers and in whom biceps movement is gradually dissociated from respiratory effort and eventually comes under volitional control (35, 36). Exactly how much of a mismatch will

FIGURE 9. Photomicrograph showing individual axons at the end of a segment of adult mouse sciatic nerve that was subjected to connective tissue digestion. A single myelinated axon is seen in the center of the field. Scale 20 m.

642 | VOLUME 57 | NUMBER 4 | OCTOBER 2005

www.neurosurgery-online.com

MICROSCALE AXON SURGERY

be tolerated may depend on the specific part of the nervous system involved and the function of the axon population under consideration. Advances in new methodologies for direct axon repair we hope will encourage development of preoperative or intraoperative methods of rapid functional mapping of axon groups in both the PNS and CNS.

Remyelination of Repaired Axons

Schwann cells or oligodendrocyte glia cells extend cytoplasmic protrusions to form myelin around axons to enhance electrical conduction. The importance of myelination is illustrated by demyelinating diseases such as multiple sclerosis, in which myelin is lost, resulting in unreliable signal conduction by CNS axons and a breakdown in neuronal communication. Thus, remyelination of axons after repair is likely to be important for enhancing functional recovery. Research indicates that transplanted cells can help remyelinate axons. Olfactory ensheathing cells (which resemble Schwann cells), oligodendrocytes, and oligodendrocyte precursor cells all have been demonstrated to be effective in providing myelin for axons in vivo after experimentally induced demyelination (9, 17). A recent study, in fact, reported that adult stem cells injected intravenously can remyelinate CNS axons (41), potentially providing a simple and convenient cell delivery method without requiring intracranial access. Cell transplantation currently is being tested in clinical trials as a treatment for demyelinating diseases (52). These advances suggest that axon remyelination after MEMS-assisted repair also may be achieved using a cell replacement strategy.

How Long Will It Take?

Axon repair will be a practical approach only if it can be carried out as a surgical procedure of reasonable duration. Our investigations of the core technologies and how quickly each repair step currently can be accomplished provide guidance for a minimum time required. As soon as positioned, actuated microdevices with reasonable force generation and gearing design should be able to deliver an axon cutting stroke consisting of a 50- m length of travel in under 5 seconds. We have observed that axons move at approximately 5 m/s in DEP fields generated by very simple electrode configurations. If we estimate a need to move axons 25 to 50 m for proper alignment with their fusion partner, this second step of repair should take 5 to 10 seconds. Because the same electrodes will be used for the final fusion step, electrofusion can be initiated immediately after alignment. Typical electrofusion parameters are applied over 2 to 3 seconds. From these estimates, we arrive at roughly 20 seconds for each axon. If the protocol calls for each of these steps to be completed on one axon at a time, 180 cycles of axon repair can be performed in 1 hour. A much more efficient way to use microtechnology, however, is to exploit its potential for batch processing of function. In particular, the overall time required for axon repair can be shortened significantly by cutting and trimming multiple axons simultaneously. In addition, DEP electrode arrays with individually addressable electrodes will permit the simultaneous manipulation of multiple axons within the surgical field. After alignment, the entire set of axons can then be subjected to electrofusion. Such a parallel processing approach will increase significantly the number of axons that can be repaired per unit time and will make the repair of several thousand axons a reasonable surgical goal.

Assessment of Successful Repair

During surgery, an immediate assessment of whether the repair on a particular pair of axons is successful provides an ongoing tabulation of the number of repairs accomplished. One convenient test of axonal function is the demonstration of electrical conduction from the host axon into the donor axon through the fusion site. Because electrode arrays are already in place for axon manipulation and axon electrofusion, it should be possible to use different elements within the array close to the axon of interest to act as stimulating and recording electrodes. This assessment of electrical conduction can be preprogrammed to occur after each electrofusion, and the information will allow an objective intraoperative scoring of repair success.

Is Axon Repair Feasible in the CNS?

From a basic neurobiological perspective, axon repair methodologies for PNS axons also should be applicable to CNS axons because of their fundamental similarities in biological structure. Beyond basic neurobiology, however, access to the CNS may be more demanding and likely will require surgical chambers of different design. A second issue is whether PNS axons can be used as grafts to replace a segment of CNS axon. A definitive answer is not at hand because such an operation has yet to be performed. It is well known, however, that CNS axons will grow into a piece of PNS sciatic nerve graft for significant distances (1, 2). Although it is unclear whether and how a CNS/PNS hybrid axon will function, the highly similar cell biological make-up of PNS and CNS axons suggests that it may be worthwhile exploring this approach.

Mechanical Support after Axon Repair

A number of considerations suggest that an engineered microstructure acting as a supporting scaffold may be useful in the postrepair period. Recent studies have examined the suitability of various polymers and biogels for use as scaffolds in promoting axon regeneration (45). After axon repair, the seeding of cells for axon myelination will benefit from a threedimensional structure providing a means of cell localization and a substrate for cell anchoring. Furthermore, the polymeric or biogel scaffolds can be infiltrated with chemicals and substances that enhance axon viability and recovery, encourage vascularization, and decrease inflammation as well as scarring. Last, the engineered scaffold can be used to provide mechanical support.

NEUROSURGERY

VOLUME 57 | NUMBER 4 | OCTOBER 2005 | 643

SRETAVAN

ET AL.

SUMMARY
In this article, we propose a new paradigm for the treatment of nerve injuries based on the direct repair of damaged axons. This repair methodology uses a novel combination of microtechnology, electrokinetic axon manipulation, and the wellestablished biological principle of cell fusion. These three fields of research have been integrated in a multidisciplinary approach to develop a solution for a significant clinical problem that currently has no specific treatment. The findings reported here provide some initial proof of principle for the core technologies we intend to use for axon repair. This new treatment paradigm for nerve injuries raises a number of clinical issues such as defining the optimal time, number, and type of axons for repair. Although existing knowledge suggests possible answers to some of these questions, solutions to other problems will have to come from additional research before microtechnology-based axon repair can become a possibility in patients. It is hoped that each advance in axon repair technology will spur additional research to provide us with a comprehensive understanding on how best to pursue neurosurgical intervention at the microscale level.

REFERENCES
1. Aguayo AJ, Clarke DB, Jelsma TN, Kittlerova P, Friedman HC, Bray GM: Effects of neurotrophins on the survival and regrowth of injured retinal neurons. Ciba Found Symp 196:135144; discussion 144138, 1996. 2. Aguayo AJ, Rasminsky M, Bray GM, Carbonetto S, McKerracher L, VillegasPerez MP, Vidal-Sanz M, Carter DA: Degenerative and regenerative responses of injured neurons in the central nervous system of adult mammals. Philos Trans R Soc Lond B Biol Sci 331:337343, 1991. 3. Aldskogius H, Kozlova EN: Central neuron-glial and glial-glial interactions following axon injury. Prog Neurobiol 55:126, 1998. 4. Basso DM, Beattie MS, Bresnahan JC: Descending systems contributing to locomotor recovery after mild or moderate spinal cord injury in rats: Experimental evidence and a review of literature. Restor Neurol Neurosci 20:189218, 2002. 5. Bates G, Saunders J, Sowers A: Electrofusion: Principles and applications, in Sowers A (ed): Cell Fusion. New York, Plenum Press, 1987, pp 376395. 6. Berns MW, Tadir Y, Liang H, Tromberg B: Laser scissors and tweezers. Methods Cell Biol 55:7198, 1998. 7. Biering-Sorensen F: Urinary tract infection in individuals with spinal cord lesion. Curr Opin Urol 12:4549, 2002. 8. Bisby M: Regeneration of peripheral nervous system axons, in Waxman S, Kocis J, Stys P (eds): The Axon: Structure, Function, and Pathophysiology. New York, Oxford University Press, 1995, pp 553578. 9. Blakemore WF, Franklin RJ: Transplantation options for therapeutic central nervous system remyelination. Cell Transplant 9:289294, 2000. 10. Bradbury EJ, Moon LD, Popat RJ, King VR, Bennett GS, Patel PN, Fawcett SB McMahon JW: Chondroitinase ABC promotes functional recovery after spinal cord injury. Nature 416:636640, 2002. 11. Caroni P, Schwab ME: Antibody against myelin-associated inhibitor of neurite growth neutralizes nonpermissive substrate properties of CNS white matter. Neuron 1:8596, 1988. 12. Chen DF, Schneider GE, Martinou JC, Tonegawa S: Bcl-2 promotes regeneration of severed axons in mammalian CNS. Nature 385:434439, 1997. 13. Coumans JV, Lin TT, Dai HN, MacArthur L, McAtee M, Nash C, Bregman BS: Axonal regeneration and functional recovery after complete spinal cord transection in rats by delayed treatment with transplants and neurotrophins. J Neurosci 21:93349344, 2001.

14. Detrait ER, Yoo S, Eddleman CS, Fukuda M, Bittner GD, Fishman HM: Plasmalemmal repair of severed neurites of PC12 cells requires Ca(2 ) and synaptotagmin. J Neurosci Res 62:566573, 2000. 15. Eysel UT, Wolfhard U: The effects of partial retinal lesions on activity and size of cells in the dorsal lateral geniculate nucleus. J Comp Neurol 229: 301309, 1984. 16. Fouad K, Dietz V, Schwab ME: Improving axonal growth and functional recovery after experimental spinal cord injury by neutralizing myelin associated inhibitors. Brain Res Brain Res Rev 36:204212, 2001. 17. Franklin RJ: Remyelination by transplanted olfactory ensheathing cells. Anat Rec 271B:7176, 2003. 18. Fuhr G, Glasser H, Muller T, Schnelle T: Cell manipulation and cultivation under a.c. electric field influence in highly conductive culture media. Biochim Biophys Acta 1201:353360, 1994. 19. Gascoyne P, Pethig R, Satayavivad J, Becker FF, Ruchirawat M: Dielectrophoretic detection of changes in erythrocyte membranes following malarial infection. Biochim Biophys Acta 1323:240252, 1997. 20. Gascoyne PR, Vykoukal J: Particle separation by dielectrophoresis. Electrophoresis 23:19731983, 2002. 21. Godell CM, Smyers ME, Eddleman CS, Ballinger ML, Fishman HM, Bittner GD: Calpain activity promotes the sealing of severed giant axons. Proc Natl Acad Sci U S A 94:47514756, 1997. 22. Goldberg JL, Klassen MP, Hua Y, Barres BA: Amacrine-signaled loss of intrinsic axon growth ability by retinal ganglion cells. Science 296:1860 1864, 2002. 23. Goslin K, Banker G: Experimental observations on the development of polarity by hippocampal neurons in culture. J Cell Biol 108:15071516, 1989. 24. Guillery RW: Quantitative studies of transneuronal atrophy in the dorsal lateral geniculate nucleus of cats and kittens. J Comp Neurol 149:423438, 1973. 25. He Z, Koprivica V: The Nogo signaling pathway for regeneration block. Annu Rev Neurosci 27:341368, 2004. 26. Heida T, Rutten WL, Marani E: Dielectrophoretic trapping of dissociated fetal cortical rat neurons. IEEE Trans Biomed Eng 48:921930, 2001. 27. Heida T, Rutten WL, Marani E: Experimental investigation on neural cell survival after dielectrophoretic trapping. Arch Physiol Biochem 110:373 382, 2002. 28. Hui SW, Stenger DA: Electrofusion of cells: Hybridoma production by electrofusion and polyethylene glycol. Methods Enzymol 220:212227, 1993. 29. Kandel E, Schwartz J, Jessell T (eds): Principles of Neural Science. New York, McGraw-Hill, 2000. 30. Karsten U, Stolley P, Seidel B: Polyethylene glycol and electric fieldmediated cell fusion for formation of hybridomas. Methods Enzymol 220: 228238, 1993. 31. Liu CY, Spicer M, Apuzzo MLJ: The genesis of neurosurgery and the evolution of the neurosurgical operative environment: Part IIConcepts for future development, 2003 and beyond. Neurosurgery 52:2035, 2003. 32. Lodish H, Berk A, Matsudaira P, Kaiser CA, Krieger M, Scott MP, Zipursky L, Darnell J: Molecular Cell Biology, 5th ed, W.H. Freeman, 2004. 33. Loy DN, Talbott JF, Onifer SM, Mills MD, Burke DA, Dennison JB, Fajardo LC, Magnuson DS, Whittemore SR: Both dorsal and ventral spinal cord pathways contribute to overground locomotion in the adult rat. Exp Neurol 177:575580, 2002. 34. Ludwin S: Pathology of the myelin sheath, in Waxman S, Kocis J, Stys P (eds): The Axon: Structure, Function, and Pathophysiology. New York, Oxford University Press, 1995, pp 412437. 35. Malessy MJ, Bakker D, Dekker AJ, Van Duk JG, Thomeer RT: Functional magnetic resonance imaging and control over the biceps muscle after intercostal-musculocutaneous nerve transfer. J Neurosurg 98:261268, 2003. 36. Mano Y, Nakamuro T, Tamura R, Takayanagi T, Kawanishi K, Tamai S, Mayer RF: Central motor reorganization after anastomosis of the musculocutaneous and intercostal nerves following cervical root avulsion. Ann Neurol 38:1520, 1995. 37. Merkler D, Metz GA, Raineteau O, Dietz V, Schwab ME, Fouad K: Locomotor recovery in spinal cord-injured rats treated with an antibody neutralizing the myelin-associated neurite growth inhibitor Nogo-A. J Neurosci 21:36653673, 2001.

644 | VOLUME 57 | NUMBER 4 | OCTOBER 2005

www.neurosurgery-online.com

MICROSCALE AXON SURGERY

38. Neil GA, Zimmermann U: Electrofusion, in Duzgunes N (ed): Methods in Enzymology. New York, Academic Press, 1993, pp 174196. 39. Okabe M, Ikawa M, Kominami K, Nakanishi T, Nishimune Y: Green mice as a source of ubiquitous green cells. FEBS Lett 407:313319, 1997. 40. Petersen KE: Silicon as a mechanical material. Proc IEEE 70:420457, 1982. 41. Pluchino S, Quattrini A, Brambilla E, Gritti A, Salani G, Dina G, Galli R, Del Carro U, Amadio S, Bergami A, Furlan R, Comi G, Vescovi AL, Martino G: Injection of adult neurospheres induces recovery in a chronic model of multiple sclerosis. Nature 422:688694, 2003. 42. Polla DL, Erdman AG, Robbins WP, Markus DT, Diaz-Diaz J, Rizq R, Nam Y, Brickner HT, Wang A, Krulevitch P: Microdevices in medicine. Annu Rev Biomed Eng 2:551576, 2000. 43. Rafuse VF, Gordon T: Self-reinnervated cat medial gastrocnemius muscles. I. Comparisons of the capacity for regenerating nerves to form enlarged motor units after extensive peripheral nerve injuries. J Neurophysiol 75: 268281, 1996. 44. Roy S, Ferrara LA, Fleischman AJ, Benzel EC: Microelectromechanical systems and neurosurgery: A new era in a new millennium. Neurosurgery 49:779799, 2001. 45. Schmidt CE, Leach JB: Neural tissue engineering: strategies for repair and regeneration. Annu Rev Biomed Eng 5:293347, 2003. 46. Schucht P, Raineteau O, Schwab ME, Fouad K: Anatomical correlates of locomotor recovery following dorsal and ventral lesions of the rat spinal cord. Exp Neurol 176:143153, 2002. 47. Schwab ME: Repairing the injured spinal cord. Science 295:10291031, 2002. 48. Shigley J, Mischke C, Budynas R: Mechanical Engineering Design. New York, McGraw-Hill, 2003. 49. Silver J, Miller JH: Regeneration beyond the glial scar. Nat Rev Neurosci 5:146156, 2004. 50. Smith DH, Wolf JA, Meaney DF: A new strategy to produce sustained growth of central nervous system axons: continuous mechanical tension. Tissue Eng 7:131139, 2001. 51. Somogyi J, Eysel U, Hamori J: A quantitative study of morphological reorganization following chronic optic deafferentation in the adult cat dorsal lateral geniculate nucleus. J Comp Neurol 255:341350, 1987. 52. Stangel M, Hartung HP: Remyelinating strategies for the treatment of multiple sclerosis. Prog Neurobiol 68:361376, 2002. 53. Stoppini L, Buchs PA, Muller D: Lesion-induced neurite sprouting and synapse formation in hippocampal organotypic cultures. Neuroscience 57: 985994, 1993. 54. Wakayama T, Perry AC, Zuccotti M, Johnson KR, Yanagimachi R: Full-term development of mice from enucleated oocytes injected with cumulus cell nuclei. Nature 394:369374, 1998. 55. Wang X, Becker FF, Gascoyne PR: Membrane dielectric changes indicate induced apoptosis in HL-60 cells more sensitively than surface phosphatidylserine expression or DNA fragmentation. Biochim Biophys Acta 1564:412420, 2002. 56. Wang X, Yang J, Gascoyne PR: Role of peroxide in AC electrical field exposure effects on friend murine erythroleukemia cells during dielectrophoretic manipulations. Biochim Biophys Acta 1426:5368, 1999. 57. Wilmut I, Schnieke AE, McWhir J, Kind AJ, Campbell KH: Viable offspring derived from fetal and adult mammalian cells. Nature 385:810813, 1997. 58. Yoshimura N, Smith CP, Chancellor MB, de Groat WC: Pharmacologic and potential biologic interventions to restore bladder function after spinal cord injury. Curr Opin Neurol 13:677681, 2000. 59. You SW, Chen BY, Liu HL, Lang B, Xia JL, Jiao XY, Ju G: Spontaneous recovery of locomotion induced by remaining fibers after spinal cord transection in adult rats. Restor Neurol Neurosci 21:3945, 2003. 60. Zhai Q, Wang J, Kim A, Liu Q, Watts R, Hoopfer E, Mitchison T, Luo L, He Z: Involvement of the ubiquitin-proteasome system in the early stages of wallerian degeneration. Neuron 39:217225, 2003. 61. Ziv NE, Spira ME: Induction of growth cone formation by transient and localized increases of intracellular proteolytic activity. J Cell Biol 140:223 232, 1998.

was received in conjunction with the generation of this manuscript submission and none have any personal or institutional financial interest in drugs, materials, or devices described in their submissions. Patent Pending to the Regents of the University of California for microdevices and methods of axon repair.

COMMENTS
his is a well-written paper on an innovative new nerve repair paradigm that employs microscale neurosurgery to repair single axons using micro/nanotechnology. The authors report a multidisciplinary strategy to manipulate and operate at the single axon level and discuss the design of microfabricated devices for axon cutting, dielectrophoresis (DEP), for axon manipulation, and electrofusion of axons. They also present interesting preliminary data to support their novel strategy. As predicted by a brilliant article in Neurosurgery in 2003 that foresaw that microtechnology and nanotechnology surely would make important contributions to clinical neurosurgery in the future (1), this report is just one excellent example of research advances in that direction. This type of creative thinking is exactly what is needed to tackle the difficult issues involved in the repair of neural injuries. There currently is no technique available for determining acutely which closed peripheral nervous system injuries will recover spontaneously and which will require surgical intervention. We will need viable answers to many such basic questions before this type of technology can reach clinical testing. In our lab, we have successfully elongated axons of both human and rat dorsal root ganglia in vitro with a computer-controlled mechanical device and started transplantation experiments using in vitro grown GFP labeled axons into host animals with acute sciatic nerve injuries (2, 3). It is conceivable that the technology described by Dr. Sretavan et al. could take advantage of our in vitro elongated axons as donor axons in the repair of acutely injured axons. Obviously, there are numerous technical hurdles that need to be overcome before this strategy can be applied to patients, but we concur with the authors that this report will stimulate additional research in this direction to provide neurosurgeons and neuroscientists with a better understanding of how to advance neurosurgical intervention at the microscale and nanoscale. Jason H. Huang Eric L. Zager Philadelphia, Pennsylvania

1. Liu CY, Spicer M, Apuzzo ML: The genesis of neurosurgery and the evolution of the neurosurgical operative environment: Part IIconcepts for future development, 2003 and beyond. Neurosurgery 52:2033, 2003. 2. Huang JH, Zager EL, Pfister BJ, Smith DH: Human Dorsal Root Ganglia Neurons as Live Nerve Constructs: Implications for Transplantation. Presented at the AANS Annual Meeting, Orlando, Florida, May 2004. 3. Huang JH, Groff RF, Zhang J, Pfister BJ, Zager EL, Smith DH: Living Nerve Constructs Bridging Peripheral Nerve Lesions Demonstrate Long-term Survival and Restoration of Function. Platform Presentation (373.7) at the 34th Annual Meeting of the Society for Neuroscience, San Diego, California, October 2004.

Acknowledgments
This study was supported by the That Man May See Foundation and the Sandler Family Supporting Foundation. No financial support

r. Stretavan et al. have presented a wonderful glimpse into the future of high tech Neurosurgery. In this regard, they have exploited microtechnology to its fullest current potential. Although, what they have presented seems like rocket science bordering on science fiction to most, their observations, accomplishments, and technological

NEUROSURGERY

VOLUME 57 | NUMBER 4 | OCTOBER 2005 | 645

SRETAVAN

ET AL.

developments are truly insightful. At the outset, most would suggest that what the authors ultimately propose, clinically meaningful and useful surgery on axons, is impossible or even ludicrous. However, the strategy and methodologies they outline should lead us to believe that some day they (or those who they intrigue) will actually pull it off. Dr. Stretavan et al. have introduced several unique strategies and have suggested a global strategy for their implementation. These include 1) a technique for axon cutting with a suspension mounted silicon nitride knife; 2) a technique for axon manipulation, dielectrophoresis (DEP), that causes the movement of small polarizable objects, such as neurons in a nonhomogeneous alternating current electrical field; 3) axon electrofusion techniques; 4) the confirmation of electrophysiologically successful electrofusion; 5) a surgical platform for accomplishing the aforementioned; and 6) a mechanism to complete such tasks in bulk (parallel processing) so that hundreds, thousands, or millions of axons could be repaired in a reasonable amount of time. What Dr. Stretavan et al. have proposed is not science fiction. They are sitting at the forefront of, and are hopefully going to guide us into, this most exciting of neurosurgical frontiers. For this, they are to be heartily congratulated. Edward C. Benzel Cleveland, Ohio his report illustrates and discusses the concept of repairing damaged axons by combining three separate microtechniques into a single surgical platform. Sequentially, these include an axon micro-cutting device made from silicon with a 20-nanometer tip, electrodes capable of noncontact manipulation of individual axons using DEP, and the fusion of aligned axon segments by delivering a pulse of electricity through the same electrodes used for DEP. They provide illustrations depicting the cutting of axons, as well as the electro-fusion of overlapping, adjacent axon segments. However, the precise (two-dimensional) alignment of cut axons to be fused using DEP is not shown. Perhaps this is because it is not feasible with the technology currently available. Although the authors do not provide evidence of a successful axon spliced in vitro using their proposed surgical platform, we believe it should be readily possible. However, the important question remains as to whether it can be applied to the central or peripheral nervous system in vivo. As with other technologies, if adequate resources are available and commercial interest is strong enough, a clinically applicable prototype may eventually appear. As the authors allude, axon repair could be computer automated, with even multiple surgical platforms being simultaneously applied to an injury site. For efficiency, the process would utilize automated, real-time recognition of those axon terminals that are in the need of repair. This could be accomplished with visual pattern recognition software. One of the more obvious barriers to its eventual use would be the preparation and alignment of axons in vivo prior to repair. Chemical skeletonization may be one option, as the authors point out. But do they necessarily need to be skeletonized to such a degree? Considering that we (us and the computers) can probably identify cut axon terminals on end, even in their connective tissue environment, then one may extrapolate that with proper alignment and electro-fusion, axon repair could occur without extensive, and possible damaging, skeletonization. Once prepared, axon alignment in three-dimensional space would likely be required, not only in a two-dimensional plane as described in this report. Even if axon repair becomes possible, the question remains whether the distal axon segment will remain viable after repair. To our knowledge, there have been no studies addressing this. Furthermore, what happens to their terminal connections? It is plausible that with a quick enough repair these distal axon segments will not undergo degener-

ation. More information is required. An experiment using the larger, and thus more readily manipulated, squid axon may be a good starting point. For example, the squid axons functional outcome can be investigated following transection and electro-fusion at various times. We believe that the window for this type of axon repair may be quite short, and therefore, it require pharmacological adjuvants to prolong this potential state in which axon salvage is possible. We commend the authors on their innovative and thoughtprovoking application, which is a great example of how crossdisciplinary collaboration allows us to achieve new heights in the field of neurosurgery. Stephen M. Russell Patrick J. Kelly New York, New York

r. Sretavan et al. publish what can be considered the next era in nerve repair: nanotechnology. The authors applied microscale knives to resect an axon, and, with the aid of an electrical field, were able to interpose a healthy piece of axon, or graft, between the proximal and distal stumps of a divided axon. This technique is promising for peripheral nerve and spinal cord regeneration. Although the authors postulate multiple ways of achieving these goals, there are many hurdles that still need to be overcome. They feel that with these techniques, 180 axons could be repaired within an hour. Performing this on an injured patient at a nano-level to a major peripheral nerve with about 20,000 fibers would require many hours and also would be difficult in deeply seated nerves or spinal cord in the depths of the spinal cord with surrounding spinal fluid. These considerations provide only challenges for future work and do not negate the importance of this contribution. Restoring electrical impulses in these axons is what is going to make the concept work. Whether this can be accomplished along with remyeliniation is topic for future investigation. Although this now may seem to be a far-fetched concept, the future may see a more practical application of this new technology, which will translate into better outcomes for our patients. We wholeheartedly congratulate the authors for their results and look forward to further developments. Rashid M. Janjua David G. Kline New Orleans, Lousiana

n this paper, the authors describe an intriguing approach to axonal repair. They have taken advantage of advances in microtechnology to consider the direct repair of damaged single axons by substituting segments of healthy donor axons. The basic steps to this mode of axon repair are described, along with prototypes of surgical devices that may bring this concept to reality in the future. Without a doubt, the development of neurosurgery has been marked by a progressive minimalism. The surgical paradigm described here may represent yet another step in this direction. At the present time, considerable development in technologies may eventually allow manipulation and surgery at even smaller scales. Although the concept of single axon repair may seem fantastic, future realities enabled by the emerging field of nanotechnology may be even more astonishing. Irrespective of the eventual clinical utility of the devices proposed in this work, I applaud the authors for their creativity and hope that work such as this serve to stimulate all neurosurgeons. I read this paper with great enthusiasm. Charles Y. Liu Los Angeles, California

646 | VOLUME 57 | NUMBER 4 | OCTOBER 2005

www.neurosurgery-online.com