The curious ability of polyethylene glycol fusion technologies to restore lost behaviors after nerve severance.

Traumatic injuries to PNS and CNS axons are not uncommon. Restoration of lost behaviors following severance of mammalian peripheral nerve axons (PNAs) relies on regeneration by slow outgrowths and is typically poor or nonexistent when after ablation or injuries close to the soma. Behavioral recovery after severing spinal tract axons (STAs) is poor because STAs do not naturally regenerate. Current techniques to enhance PNA and/or STA regeneration have had limited success and do not prevent the onset of Wallerian degeneration of severed distal segments. This Review describes the use of a recently developed polyethylene glycol (PEG) fusion technology combining concepts from biochemical engineering, cell biology, and clinical microsurgery. Within minutes after microsuturing carefully trimmed cut ends and applying a well-specified sequence of solutions, PEG-fused axons exhibit morphological continuity (assessed by intra-axonal dye diffusion) and electrophysiological continuity (assessed by conduction of action potentials) across the lesion site. Wallerian degeneration of PEG-fused PNAs is greatly reduced as measured by counts of sensory and/or motor axons and maintenance of axonal diameters and neuromuscular synapses. After PEG-fusion repair, cut-severed, crush-severed, or ablated PNAs or crush-severed STAs rapidly (within days to weeks), more completely, and permanently restore PNA- or STA-mediated behaviors compared with nontreated or conventionally treated animals. PEG-fusion success is enhanced or decreased by applying antioxidants or oxidants, trimming cut ends or stretching axons, and exposure to Ca(2+) -free or Ca(2+) -containing solutions, respectively. PEG-fusion technology employs surgical techniques and chemicals already used by clinicians and has the potential to produce a paradigm shift in the treatment of traumatic injuries to PNAs and STAs.

Polyethylene glycol-fused allografts produce rapid behavioral recovery after ablation of sciatic nerve segments.

Restoration of neuronal functions by outgrowths regenerating at ∼1 mm/day from the proximal stumps of severed peripheral nerves takes many weeks or months, if it occurs at all, especially after ablation of nerve segments. Distal segments of severed axons typically degenerate in 1-3 days. This study shows that Wallerian degeneration can be prevented or retarded, and lost behavioral function can be restored, following ablation of 0.5-1-cm segments of rat sciatic nerves in host animals. This is achieved by using 0.8-1.1-cm microsutured donor allografts treated with bioengineered solutions varying in ionic and polyethylene glycol (PEG) concentrations (modified PEG-fusion procedure), being careful not to stretch any portion of donor or host sciatic nerves. The data show that PEG fusion permanently restores axonal continuity within minutes, as initially assessed by action potential conduction and intracellular diffusion of dye. Behavioral functions mediated by the sciatic nerve are largely restored within 2-4 weeks, as measured by the sciatic functional index. Increased restoration of sciatic behavioral functions after ablating 0.5-1-cm segments is associated with greater numbers of viable myelinated axons within and distal to PEG-fused allografts. Many such viable myelinated axons are almost certainly spared from Wallerian degeneration by PEG fusion. PEG fusion of donor allografts may produce a paradigm shift in the treatment of peripheral nerve injuries.

Cellular mechanisms of plasmalemmal sealing and axonal repair by polyethylene glycol and methylene blue.

Mammalian neurons and all other eukaryotic cells endogenously repair traumatic injury within minutes by a Ca²âº-induced accumulation of vesicles that interact and fuse with each other and the plasmalemma to seal any openings. We have used uptake or exclusion of extracellular fluorescent dye to measure the ability of rat hippocampal B104 cells or rat sciatic nerves to repair (seal) transected neurites in vitro or transected axons ex vivo. We report that endogenous sealing in both preparations is enhanced by Ca²âº-containing solutions and is decreased by Ca²âº-free solutions containing antioxidants such as dithiothreitol (DTT), melatonin (MEL), methylene blue (MB), and various toxins that decrease vesicular interactions. In contrast, the fusogen polyethylene glycol (PEG) at 10-50 mM artificially seals the cut ends of B104 cells and rat sciatic axons within seconds and is not affected by Ca²âº or any of the substances that affect endogenous sealing. At higher concentrations, PEG decreases sealing of transected axons and disrupts the plasmalemma of intact cells. These PEG-sealing data are consistent with the hypothesis that lower concentrations of PEG directly seal a damaged plasmalemma. We have considered these and other data to devise a protocol using a well-specified series of solutions that vary in tonicity, Ca²âº, MB, and PEG content. These protocols rapidly and consistently repair (PEG-fuse) rat sciatic axons in completely cut sciatic nerves in vivo rapidly and dramatically to restore long-lasting morphological continuity, action potential conduction, and behavioral functions.

Neurite transection produces cytosolic oxidation, which enhances plasmalemmal repair.

To survive, cells must rapidly repair (seal) plasmalemmal damage. Cytosolic oxidation has been shown to increase cell survival in some cases and produce cell death in other protocols. An antioxidant (melatonin; Mel) has been reported to decrease the probability of sealing plasmalemmal damage. Here we report that plasmalemmal damage produces cytosolic oxidation, as assayed by methylene blue (MB) color change in rat B104 hippocampal cells. Plasmalemmal sealing is affected by duration of Ca²âº deprivation and length of exposure to, and concentration of, oxidizing agents such as H2O2 and thimerosal (TH). Cytosolic oxidation by 10 µM to 50 mM H2O2 or 100 µM to 2 mM TH increases the probability of Ca²âº-dependent plasmalemmal sealing, whereas higher concentrations of H2O2 decrease sealing probability and also damage uninjured cells. We also show that antioxidants (Mel, MB) or reducing agents (dithiothreitol) decrease sealing. Proteins, such as protein kinase A, SNAP-25, synaptobrevin, and N-ethylmaleimide-sensitive factor (previously reported to enhance sealing in other pathways), also enhance sealing in this oxidation pathway. In brief, our data show that plasmalemmal damage produces cytosolic oxidation that increases the probability of plasmalemmal sealing, which is strongly correlated with cell survival in other studies. Our results may provide new insights into the etiology and treatment of oxidation-dependent neurodegenerative disorders, such as Parkinson's, Huntington's, and Alzheimer's diseases.

Rapid, effective, and long-lasting behavioral recovery produced by microsutures, methylene blue, and polyethylene glycol after completely cutting rat sciatic nerves.

Behavioral function lost in mammals (including humans) after peripheral nerve severance is slowly (weeks to years) and often poorly restored by 1-2-mm/day, nonspecifically directed outgrowths from proximal axonal stumps. To survive, proximal stumps must quickly repair (seal) plasmalemmal damage. We report that, after complete cut- or crush-severance of rat sciatic nerves, morphological continuity, action potential conduction, and behavioral functions can be consistently (>98% of trials), rapidly (minutes to days), dramatically (70-85% recovery), and chronically restored and some Wallerian degeneration prevented. We assess axoplasmic and axolemmal continuity by intra-axonal dye diffusion and action potential conduction across the lesion site and amount of behavioral recovery by Sciatic Functional Index and Foot Fault tests. We apply well-specified sequences of solutions containing FDA-approved chemicals. First, severed axonal ends are opened and resealing is prevented by hypotonic Ca²âº-free saline containing antioxidants (especially methylene blue) that inhibit plasmalemmal sealing in sciatic nerves in vivo, ex vivo, and in rat B104 hippocampal cells in vitro. Second, a hypotonic solution of polyethylene glycol (PEG) is applied to open closely apposed (by microsutures, if cut) axonal ends to induce their membranes to flow rapidly into each other (PEG-fusion), consistent with data showing that PEG rapidly seals (PEG-seals) transected neurites of B104 cells, independently of any known endogenous sealing mechanism. Third, Ca²âº-containing isotonic saline is applied to induce sealing of any remaining plasmalemmal holes by Ca²âº-induced accumulation and fusion of vesicles. These and other data suggest that PEG-sealing is neuroprotective, and our PEG-fusion protocols that repair cut- and crush-severed rat nerves might rapidly translate to clinical procedures.

Quantitative aspects of transmitter release.

The opener-stretcher motor neuron in crayfish makes 50 endings upon each of 1200 muscle fibers. We have calculated the quantal content of junctional potentials produced by individual terminals and by the whole cell at various physiological frequencies. The results show that when the motor neuron is active at 20 impulses/second, it releases 50 quanta/impulse per muscle fiber, or a total of 4.5 x 10(9) quanta/hr. These figures are similar to those for vertebrate muscles per fiber, but larger for the entire neuron because the opener motor unit is so large. On the basis that the quanta correspond to synaptic vesicles each containing 10(3)-10(4) molecules of transmitter, the release rate must be around 10(-11) mole/hr. This value is within an order of magnitude of the release figures obtained for mammalian neurons by collecting transmitter in perfusates, but it is far lower than the value reported for a crustacean inhibitory neuron. If the membrane materials surrounding each vesicle were lost in the release process, the replacement synthesis would involve 24 mm(2) of membrane/hr. We conclude that the metabolic load in terms of transmitter synthesis is probably sustainable, but that the release mechanism must operate in such a way that vesicle membrane materials are neither lost nor incorporated into the terminal membrane.

Regeneration in crustacean motoneurons: evidence for axonal fusion.

Crayfish motor axons remain excitable for over 100 days after severance from their central cell bodies, and continue to store and release normal amounts of transmitter substance. Evidence indicates that regeneration occurs by fusion of the central process with its surviving peripheral segment.

Long-term survival of anucleate axons and its implications for nerve regeneration.

Severed distal segments of nerve axons (anucleate axons) have now been reported to survive for weeks to years in representative organisms from most phyla, including the vertebrates. Among invertebrates (especially crustaceans), such long-term survival might involve transfer of proteins from adjacent intact cells to anucleate axons. In lower vertebrates and mammals, long-term survival of anucleate axons is more likely attributed to a slow turnover of axonal proteins and/or a lack of phagocytosis by macrophages or other cell types. Invertebrate anucleate axons that exhibit long-term survival are often reactivated by neurites that have grown from proximal nucleate segments. In mammals, induction of long-term survival in anucleate axons might allow more time to use artificial mechanisms to repair nerve axons by fusing the two severed halves with polyethylene glycol, a technique recently developed to fuse severed halves of myelinated axons in earthworms.

Differentiation of nerve terminals in the crayfish opener muscle and its functional significance.

Junctional potentials (jp's) recorded from superficial distal fibers of the crayfish opener muscle are up to 50 times larger than jp' in superficial central fibers when the single motor axon that innervates the muscle is stimulated at a frequency of 1/sec or less. At 80/sec, in contrast, central jp's are up to four times larger than those observed in distal fibers. The tension produced by single muscle fibers of either type is directly proportional to the integral of the time-voltage curve minus an excitation-contraction coupling threshold of 3 mv. Distal fibers therefore produce almost all the total muscle tension at low frequencies of stimulation and central fibers add an increasingly greater contribution as their nerve endings begin to facilitate in response to increased rate of motor discharge. Differentiation of muscle membrane characteristics (input resistance, space constant, time constant) cannot account for these differences in facilitation ratios. The mechanism of neuronal differentiation is not based upon the size or effectiveness of transmitter quanta, since equal sized jp's have equal variances;: mjp sizes and variances are also equal. No differences were found between fiber types in rates of transmitter mobilization, density of innervation, or the relationship between transmitter release and terminal depolarization. Single terminals on distal fibers were found to release transmitter with a greater probability than central terminals. More effective invasion of distal terminals by the nerve impulse at low frequencies can account for the difference.

Facilitation of transmitter release at squid synapses.

Facilitation is shown to decay as a compound exponential with two time constants (T1, T2) at both giant and non-giant synapses in squid stellate ganglia bathed in solutions having low extracellular calcium concentrations ([Ca++]o). Maximum values of facilitation (F1) were significantly larger, and T1 was significantly smaller in giant than non-giant synapses. Decreases in [Ca++]o or increases in [Mn++]o had variable effects on T1 and F1, whereas decreases in temperature increased T1 but had insignificant effects on F1. The growth of facilitation during short trains of equal interval stimuli was adequately predicted by the linear summation model developed by Mallart and Martin (1967. J. Physiol. (Lond.). 193:676--694) for frog neuromuscular junctions. This result suggests that the underlying mechanisms of facilitation are similar in squid and other synapses which release many transmitter quanta.

Presynaptic potentials and facilitation of transmitter release in the squid giant synapse.

Presynaptic potentials were studied during facilitation of transmitter release in the squid giant synapse. Changes in action potentials were found to cause some, but not all, of the facilitation during twin-pulse stimulation. During trains of action potentials, there were no progressive changes in presynaptic action potentials which could account for the growth of facilitation. Facilitation could still be detected in terminals which had undergone conditioning depolarization or hyperpolarization. Facilitation could be produced by small action potentials in low [Ca++]o and by small depolarizations in the presence of tetrodotoxin. Although the production of facilitation varied somewhat with presynaptic depolarization, nevertheless, approximately equal amounts of facilitation could be produced by depolarizations which caused the release of very different amounts of transmitter.

Presynaptic facilitation at the crayfish neuromuscular junction. Role of calcium-activated potassium conductance.

Membrane potential was recorded intracellularly near presynaptic terminals of the excitor axon of the crayfish opener neuromuscular junction (NMJ), while transmitter release was recorded postsynaptically. This study focused on the effects of a presynaptic calcium-activated potassium conductance, gK(Ca), on the transmitter release evoked by single and paired depolarizing current pulses. Blocking gK(Ca) by adding tetraethylammonium ion (TEA; 5-20 mM) to a solution containing tetrodotoxin and aminopyridines caused the relation between presynaptic potential and transmitter release to steepen and shift to less depolarized potentials. When two depolarizing current pulses were applied at 20-ms intervals with gK(Ca) not blocked, the presynaptic voltage change to the second (test) pulse was inversely related to the amplitude of the first (conditioning) pulse. This effect of the conditioning prepulse on the response to the test pulse was eliminated by 20 mM TEA and by solutions containing 0 mM Ca2+/1 mM EGTA, suggesting that the reduction in the amplitude of the test pulse was due to activation of gK(Ca) by calcium remaining from the conditioning pulse. In the absence of TEA, facilitation of transmitter release evoked by a test pulse increased as the conditioning pulse grew from -40 to -20 mV, but then decreased with further increase in the conditioning depolarization. A similar nonmonotonic relationship between facilitation and the amplitude of the conditioning depolarization was reported in previous studies using extracellular recording, and interpreted as supporting an additional voltage-dependent step in the activation of transmitter release. We suggest that this result was due instead to activation of a gK(Ca) by the conditioning depolarization, since facilitation of transmitter release increased monotonically with the amplitude of the conditioning depolarization, and the early time course of the decay of facilitation was prolonged when gK(Ca) was blocked. The different time courses for decay of the presynaptic potential (20 ms) and facilitation (greater than 50 ms) suggest either that residual free calcium does not account for facilitation at the crayfish NMJ or that the transmitter release mechanism has a markedly higher affinity or stoichiometry for internal free calcium than does gK(Ca). Finally, our data suggest that the calcium channels responsible for transmitter release at the crayfish NMJ are not of the L, N, or T type.

Calcium-activated potassium conductance in presynaptic terminals at the crayfish neuromuscular junction.

Membrane potential changes that typically evoke transmitter release were studied by recording intracellularly from the excitor axon near presynaptic terminals of the crayfish opener neuromuscular junction. Depolarization of the presynaptic terminal with intracellular current pulses activated a conductance that caused a decrease in depolarization during the constant current pulse. This conductance was identified as a calcium-activated potassium conductance, gK(Ca), by its disappearance in a zero-calcium/EGTA medium and its block by cadmium, barium, tetraethylammonium ions, and charybdotoxin. In addition to gK(Ca), a delayed rectifier potassium conductance (gK) is present in or near the presynaptic terminal. Both these potassium conductances are involved in the repolarization of the membrane during a presynaptic action potential.

Developmental and other factors affecting regeneration of crayfish CNS axons.

According to histological and ultrastructural criteria, nongiant CNS axons in newly hatched crayfish regenerate more rapidly and with greater frequency than do similar axons in adult crayfish. Regenerative ability is greater in one species (Procambarus clarkii) than in another species (Procambarus simulans), is greater at 20-25 degrees C than at 15-16 degrees C, and is greater in nongiant axons than in giant axons. In contrast to axonal regeneration, nerve cell bodies do not regenerate in newly hatched or adult crayfish of either species. While the ability to regenerate CNS axons differs between newly hatched and adult crayfish, the ultrastructural appearance of the CNS is very similar at any age it is examined.

Morphology and number of neurons in two species of polychaetes.

Neurons of the ventral nerve cord (VNC) in the polychaete species Clymenella torquata and Nereis virens were ultrastructurally distinguished from glial cells by the smaller diameter and elongated shape of glial nuclei in adult organisms. In contrast to neurons, beta-glycogen-like particles and densely packed microfilaments were found in glial cytoplasm. Using these and other criteria, glial cells were distinguished from nerve cells in histologic preparations. All neuronal nuclei were counted in specified regions of the CNS of both polychaetes. In both species, the number of neuronal nuclei in various CNS regions remained constant in animals of very different body size. Since Clymenella has a set number of ganglia in the VNC and a set number of body segments, the total number of CNS neurons remains constant in adult members of this species. Since adult Nereis adds VNC ganglia in newly forming body segments, the total number of CNS neurons continuously increases, but the total number of CNS neurons in a ganglion does not change after it is formed.

Regeneration of axons and nerve cell bodies in the CNS of annelids.

The neuron addition hypothesis predicts that species that add CNS neurons during a particular ontogenetic stage should regenerate ablated somata better than species that add few, if any, neurons to the CNS during that same stage. We report that CNS nerve cells do not regenerate in three species of adult (reproductively competent) leeches (Hirudo medicinalis, Haemopus grande, and Macrobdella decora), which do not increase the number of neurons in any portion of the CNS. Nereis virens, a polychaete that adds CNS neurons to newly forming ganglia in the adult stage, also does not regenerate CNS neurons. Conversely, CNS neurons, including a pair of uniquely identifiable somata, do regenerate in Clymenella torquata, a polychaete that has a constant number of neurons in the adult stage. Hence, the results of our study suggest that several versions of the neuron addition hypothesis cannot predict CNS regenerative abilities in adult annelids. Finally, we report that severed stumps of CNS axons do not degenerate rapidly in Nereis or Clymenella, and that both species can regenerate severed CNS axons presumably by morphologic fusion or physiologic activation of surviving stumps.

Analysis of neuritic outgrowth from severed giant axons in Lumbricus terrestris.

This study analyzes the detailed morphometric pattern at various postoperative times of neuritic outgrowths from the proximal and distal stumps of two uniquely identifiable axons. Morphological patterns of neuritic outgrowths from stumps of severed axons were compared for medial and lateral giant axons in the central nervous system of the earthworm Lumbricus terrestris. Outgrowths from proximal and distal stumps were labeled by injection of fluorescent dye into axonal stumps and assessed according to morphometric parameters. Outgrowths from axonal stumps of severed giant axons were statistically indistinguishable for most morphometric measures of neuritic quantity, shape, direction, and location. There were two exceptions to this general rule: 1) proximal stumps of medial giant axons produced significantly more neurites than distal stumps of medial giant axons, and 2) proximal stumps of lateral giant axons produced significantly longer neurites than proximal stumps of medial giant axons. No measure of neuritic outgrowth showed a significant change from the second through seventh postoperative week, suggesting that most outgrowth occurred in the first two postoperative weeks and that neuritic morphology remained stable through the seventh postoperative week. Neurites grew across the lesion site in relatively straight trajectories parallel to the longitudinal axis of the ventral nerve cord and often grew alongside the appropriate axonal stump across the lesion site. The length of neurites growing in close apposition to appropriate axonal stumps or giant axons was much greater than expected, had outgrowth been randomly directed. These data provide a basis for future investigations of the mechanisms that regulate neuritic outgrowth.

Effects of fibinolysis on neurite growth from dorsal root ganglia cultured in two- and three-dimensional fibrin gels.

The mechanism of neurite penetration of three-dimensional fibrin matrices was investigated by culturing embryonic chick dorsal root ganglia (DRGs) within fibrin gels, upon fibrin gels, and upon laminin. The length of neurites within three-dimensional matrices of fibrin was decreased in a concentration-dependent manner by agents that inhibited plasmin, e.g. aprotinin, or that inhibited plasminogen activation, e.g., epsilon-aminocaproic acid (EACA), or plasminogen antiserum. In contrast, such agents increased the length of neurites growing out from DRGs cultured upon two-dimensional substrates of fibrin and had no effect on the length of neurites growing out from DRGs cultured upon laminin. Visualization of neurites within three-dimensional fibrin matrices demonstrated that the distance between fibrin strands was much smaller than the diameter of neurites. All these data were consistent with the hypothesis that fibrinolysis localized to the region of the neurite tip is an important mechanism for neurite penetration of a physical barrier of fibrin strands arranged in a three-dimensional matrix.

Effect of temperature on long-term survival of anucleate giant axons in crayfish and goldfish.

The effect of temperature on the electrophysiology and morphology of anucleate axons was examined following severance of crayfish medial giant axons and goldfish Mauthner axons from their respective cell bodies. Although anucleate segments of each giant axon exhibited long-term survival for weeks to months at 5-25 degrees C in crayfish and 10-30 degrees C in goldfish, the two axons differed in their survival characteristics. All measures of long-term survival in crayfish medial giant axons were independent of animal holding temperature, whereas all measures in Mauthner axons were dependent on holding temperature. Medial giant axons survived for at least 90 days in crayfish maintained at 5-25 degrees C in this and previous studies. Mauthner axons survived for over 5 months in goldfish maintained at 10 degrees C but survived for 1 month at 30 degrees C. Postoperative time had different effects on many single measures of long-term survival (axonal diameter, amplitude of action or resting potentials) in medial giant axons compared to Mauthner axons. For example, resting and action potentials in crayfish medial giant axons remained remarkably constant at all holding temperatures for 0-90 postoperative days. In contrast, resting and action potentials in goldfish Mauthner axons declined abruptly in the first 10-20 postoperative days followed by a slower decline at each holding temperature. We suggest that the mechanism of long-term survival is not necessarily the same in all anucleate axons.