Peripheral, but not central, axotomy induces neuropilin-1 mRNA expression in adult large diameter primary sensory neurons

Neuropilin-1 (NP-1) is a component of the receptor for semaphorin3a (Sema3a), a member of a large family of molecules with widespread expression and demonstrable influence (via their ability to repel growing axons) on nervous system development. Recent studies have shown that some types of adult mammalian neurons retain the capacity to respond to Sema3a, particularly in relation to neuronal injury and regeneration. Although variations in expression of Sema3a mRNA have been revealed in neurons in both the central and peripheral nervous systems in this context, relatively little is known about NP-1 expression patterns. In this study we investigated the expression of NP-1 mRNA in adult dorsal root ganglion (DRG) neurons in intact and lesioned animals. We compared the effect of unilateral lesioning of the sciatic nerve or unilateral dorsal rhizotomy at lumbar levels L4/5, and bilateral dorsal funiculus lesioning at thoracic levels T10/11 on NP-1 mRNA expression in the cell bodies of lumbar DRGs. A significantly increased level of NP-1 mRNA expression was detected only following sciatic nerve lesioning (P < 0.001), but not after rhizotomy or dorsal funiculus lesioning. Furthermore, this upregulation was mainly confined to large diameter neurons of DRGs at lumbar levels L4/5, which provide the main sensory contribution to the sciatic nerve. These results suggest a role for NP-1 in the axonal response to peripheral nerve injury, which may be specific to a particular subset of primary sensory neurons.     

Calcium entry via TRPV1 but not ASICs induces neuropeptide release from sensory neurons

Inflammatory mediators induce neuropeptide release from nociceptive nerve endings and cell bodies, causing increased local blood flow and vascular leakage resulting in edema. Neuropeptide release from sensory neurons depends on an increase in intracellular Ca^2 + concentration. In this study we investigated the role of two types of pH sensors in acid-induced Ca^2 + entry and neuropeptide release from dorsal root ganglion (DRG) neurons. The transient receptor potential vanilloid 1 channel (TRPV1) and acid-sensing ion channels (ASICs) are both H⁺-activated ion channels present in these neurons, and are therefore potential pH sensors for this process. We demonstrate with in situ hybridization and immunocytochemistry that TRPV1 and several ASIC subunits are co-expressed with neuropeptides in DRG neurons. The activation of ASICs and of TRPV1 led to an increase in intracellular Ca^2 + concentration. While TRPV1 has a high Ca^2 + permeability and allows direct Ca^2 + entry when activated, we show here that ASICs of DRG neurons mediate Ca^2 + entry mostly by depolarization-induced activation of voltage-gated Ca^2 + channels and only to a small extent via the pore of Ca^2 +-permeable ASICs. Extracellular acidification led to the release of the neuropeptide calcitonin gene-related peptide from DRG neurons. The pH dependence and the pharmacological profile indicated that TRPV1, but not ASICs, induced neuropeptide secretion. In conclusion, this study shows that although both TRPV1 and ASICs mediate Ca^2 + influx, TRPV1 is the principal sensor for acid-induced neuropeptide secretion from sensory neurons.     

Interleukin-2 enhances chick and rat sympathetic, but not sensory, neurite outgrowth

Interactions between the immune and nervous systems could be important for normal development and function of both. To determine if a lymphokine, interleukin-2 (IL-2), represents a link between these two systems, sympathetic and sensory neurons from embryonic chick and neonatal rat were cultured in media containing human recombinant IL-2. In chick sympathetic chain and rat superior cervical ganglia cultures, IL-2 enhanced the number of neurons with neurites and the length of those neurites significantly over control cultures. Sensory neurons from chick and rat dorsal root ganglia were not affected by culture in IL-2. Sympathetic neuron response to IL-2 was concentration-dependent, with an optimum around 2-0.2 U/ml (100-10 pM). Immunofluorescence with an anti-IL-2 receptor antibody demonstrated specific staining of sympathetic neurons, but not sensory neurons, implying that sympathetic neurons may have a receptor for IL-2 on their surface.     

Neuronal IL-17 receptor upregulates TRPV4 but not TRPV1 receptors in DRG neurons and mediates mechanical but not thermal hyperalgesia

In addition to the proinflammatory cytokines tumor necrosis factor-α, interleukin-6 and interleukin-1ß, the cytokine interleukin-17 (IL-17) is considered an important mediator of autoimmune diseases such as rheumatoid arthritis. Because tumor necrosis factor-α and interleukin-1ß have the potential to influence the expression of transduction molecules such as transient receptor potential vanilloid 1 (TRPV1) in dorsal root ganglion (DRG) neurons and thus to contribute to pain we explored in the present study whether IL-17A activates DRG neurons and influences the expression of TRPV1. The IL-17A receptor was visualized in most neurons in dorsal root ganglion (DRG) sections as well as in cultured DRG neurons. Upon long-term exposure to IL-17A, isolated and cultured rat DRG neurons showed a significant upregulation of extracellular-regulated kinase (ERK) and nuclear factor κB (NFκB). Long-term exposure of neurons to IL-17A did not upregulate the expression of TRPV1. However, we found a pronounced upregulation of transient receptor potential vanilloid 4 (TRPV4) which is considered a candidate transduction molecule for mechanical hyperalgesia. Upon the injection of zymosan into the paw, IL-17A-deficient mice showed less mechanical hyperalgesia than wild type mice but thermal hyperalgesia was not attenuated in IL-17A-deficient mice. These data show, therefore, a particular role of IL-17 in mechanical hyperalgesia, and they suggest that this effect is linked to an activation and upregulation of TRPV4.     

Corticospinal neurons up-regulate a range of growth-associated genes following intracortical,but not spinal,axotomy

The failure of some CNS neurons to up-regulate growth-associated genes following axotomy may contribute to their failure to regenerate axons. We have studied gene expression in rat corticospinal neurons following either proximal (intracortical) or distal (spinal) axotomy. Corticospinal neurons were retrogradely labelled with cholera toxin subunit B prior to intracortical lesions or concomitantly with spinal lesions. Alternate sections of forebrain were immunoreacted for cholera toxin subunit B or processed for mRNA in situ hybridization for ATF3, c-jun, GAP-43, CAP-23, SCG10, L1, CHL1 or krox-24, each of which has been associated with axotomy or axon regeneration in other neurons. Seven days after intracortical axotomy, ATF3, c-jun, GAP-43, SCG10, L1 and CHL1, but not CAP-23 or krox-24, were up-regulated by layer V pyramidal neurons, including identified corticospinal neurons. The maximum distance between the lesion and the neuronal cell bodies that up-regulated genes varied between 300 and 500 microm. However, distal axotomy failed to elicit changes in gene expression in corticospinal neurons. No change in expression of any molecule was seen in the neocortex 1 or 7 days after corticospinal axotomy in the cervical spinal cord. The expression of GAP-43, CAP-23, L1, CHL1 and SCG10 was confirmed to be unaltered after this type of injury in identified retrogradely labelled corticospinal neurons. Thus, while corticospinal neuronal cell bodies fail to respond to spinal axotomy, these cells behave like regeneration-competent neurons, up-regulating a wide range of growth-associated molecules if axotomized within the cerebral cortex.     

Differences in horseradish peroxidase labeling of sensory, motor and sympathetic neurons following chronic axotomy of the rat sural nerve

In an attempt to clarify the ultimate fate of permanently axotomized adult primary neurons, horseradish peroxidase (HRP) was used as a cell marker to label the motor, sensory and postganglionic sympathetic neurons of rat sural nerves which had been sectioned at the ankle and prevented from regenerating for periods of up to 80 weeks. Axotomy did not affect sympathetic neurons, but resulted 4 weeks later in a sudden reduction in the number of labeled sensory and motor cells which persisted to the end of the study. The missing neuronal population amounted to 44.4% and 45.9% respectively of the normal sensory and motor contingent and included most of the large afferent and efferent neurons. However, examination of sural nerves at the thigh, 30 mm proximal to the neuroma, revealed marked axonal atrophy but no change in the number of myelinated and unmyelinated fibers up to 52 weeks after axotomy. Such prolonged survival of the peripheral processes is indirect evidence that axotomized neurons can endure long-term detachment from their end organs and suggests that the lack of HRP labeling in certain sensory and motor neurons does not imply their degeneration, but expresses one of many retrograde dysfunctions triggered by axotomy.     

Cell death of corticospinal neurons is induced by axotomy before but not after innervation of spinal targets

The response of corticospinal neurons to axotomy at postnatal ages from 5 days to adulthood was studied in the golden hamster (Mesocricetus auratus). Corticospinal neurons were retrogradely labeled with fluorescent rhodamine latex beads injected into the cervical or lumbar spinal cord. A unilateral lesion of the medullary pyramidal tract was made 1-2 days later and the brains fixed 1-30 days after axotomy. Comparisons of labeled axotomized corticospinal neurons with labeled normal corticospinal neurons in the contralateral cortex showed that axotomy at 14 days or later caused cell shrinkage but not cell death. Axotomy prior to 14 days caused cell death of corticospinal neurons. More neurons died the earlier the lesion was made, culminating in virtual complete cell death of corticospinal neurons following axotomy at 5 days. Axotomy at a given age did not affect all corticospinal neurons uniformly. Lumbar projection neurons underwent cell death ranging from slight to complete following axotomy at 13 and 9 days, respectively. Cervical projection neurons, in contrast, survived axotomy after a lesion at 9 days but underwent complete cell death if the lesion occurred at 5 days. Since corticospinal axons innervate the cervical cord from postnatal days 4-8 and the lumbar cord from 10-14 days (Reh and Kalil, '81; J. Comp. Neurol. 200:55-67), the ability of corticospinal neurons to survive axotomy appears to be temporally well correlated with their innervation of spinal targets. These neurons die if their axons are cut prior to target innervation but are able to survive if axotomy occurs after their axons innervate spinal targets. The results show that plasticity in the corticospinal pathway documented in previous reports cannot take the form of regrowth of severed axons, since early lesions cause extensive corticospinal cell death. Aberrant corticospinal pathways resulting from early lesions must therefore arise from undamaged axons. Additional retrograde labeling experiments showed that the opposite cortex responded to contralateral pyramidotomy by sprouting into denervated areas of the spinal cord. Thus another source of plasticity after early pyramidal tract lesions is sprouting from corticospinal axons arising from the intact cortex.     

Hypertonicity sensing in organum vasculosum lamina terminalis neurons: a mechanical process involving TRPV1 but not TRPV4

Primary osmosensory neurons in the mouse organum vasculosum lamina terminalis (OVLT) transduce hypertonicity via the activation of nonselective cation channels that cause membrane depolarization and increased action potential discharge, and this effect is absent in mice lacking expression of the transient receptor potential vanilloid 1 (Trpv1) gene (Ciura and Bourque, 2006). However other experiments have indicated that channels encoded by Trpv4 also contribute to central osmosensation in mice (Liedtke and Friedman, 2003; Mizuno et al., 2003). At present, the mechanism by which hypertonicity modulates cation channels in OVLT neurons is unknown, and it remains unclear whether Trpv1 and Trpv4 both contribute to this process. Here, we show that physical shrinking is necessary and sufficient to mediate hypertonicity sensing in OVLT neurons isolated from adult mice. Steps coupling progressive decreases in cell volume to increased neuronal activity were quantitatively equivalent whether shrinking was evoked by osmotic pressure or mechanical aspiration. Finally, modulation of OVLT neurons by tonicity or mechanical stimulation was unaffected by deletion of trpv4 but was abolished in cells lacking Trpv1 or wild-type neurons treated with the TRPV1 antagonist SB366791. Thus, hypertonicity sensing is a mechanical process requiring Trpv1, but not Trpv4.     

Regulation of calcium homeostasis in sensory neurons by bradykinin

The nonapeptide bradykinin (BK) activates sensory neurons and stimulates the transmission of nociceptive information into the CNS. We investigated the effect of this peptide on rat dorsal root ganglion neurons (DRG) grown in vitro. BK stimulated the synthesis of inositol trisphosphate (IP3) and the breakdown of phosphatidylinositol bisphosphate, the synthesis of diacylglycerol, and the release of arachidonic acid from DRG cells. The release of IP3 and arachidonic acid was not inhibited by pretreatment of the cells with pertussis toxin. BK also mobilized intracellular Ca2+ stores in DRG cells as assessed by fura-2-based microfluorimetry. Two types of Ca2+ stores appeared to exist in DRG neurons. One type could be mobilized by caffeine (10(-2) M), and this effect could be blocked by ryanodine in a use-dependent manner. These stores occurred primarily in the cell soma and were virtually absent from cell processes. A second type of store could be mobilized by BK, presumably through the mediation of IP3. These latter stores were distributed equally between the cell soma and processes. Experiments with combinations of caffeine and BK suggested that the stores mobilized by these 2 agents may be separate entities. Both the caffeine and BK sensitive Ca2+ storage sites appeared to participate in buffering a Ca2+ load induced in DRG neurons by cell depolarization. The relevance of these observations to the mechanism of action of BK on sensory neurons is discussed.     

MAP kinase phosphatase 1 is expressed and enhanced by FK506 in surviving mamillary, but not degenerating nigral neurons following axotomy

The MAP kinase phosphatase 1 (MKP-1), a dual serine–threonine phosphatase, inactivates the MAP kinases ERK and JNK/SAPK which are involved in neuronal survival and neuronal cell death following injury and degenerative stimuli. We have studied by immunocytochemistry whether regulation of MKP-1 is part of the cell-body response following nerve fiber transection. The expression of MKP-1 was investigated in axotomized neurons of the corpus mamillaris (CMm) and substantia nigra pars compacta (SNC) following transection of the mamillo-thalamic tract (MT) and the medial forebrain bundle (MFB), respectively. In contrast to the surviving CMm neurons, the vast majority of SNC neurons undergoes cell death following axotomy. MKP-1 immunoreactivity which is absent in untreated adult rats, appeared in CMm neurons 24 h following MT transection, reached a maximum after 2 days and persisted in a substantial proportion of CMm neurons until 20 days, the end of observation period. In contradistinction, MKP-1 could not be detected in the SNC neurons. MKP-1 immunoreactivity was virtually restricted to the nuclei of neurons. Subcutaneous injection of the immunosuppressant FK506 that protects axotomized SNC neurons against neuronal cell death, enhanced the expression of MKP-1 in CMm, but failed to do so in SNC neurons. The selective expression of MKP-1 in CMm is the first finding on a different regulation of components in the stress kinase signal pathway in surviving vs. degenerating axotomized neurons.     

Hydroxy-alpha-sanshool activates TRPV1 and TRPA1 in sensory neurons

Sanshools are major active ingredients of Zanthoxylum piperitum and are used as food additives in East Asia. Sanshools cause irritant, tingling and sometimes paresthetic sensations on the tongue. However, the molecular mechanism underlying the pungent or tingling sensation induced by sanshools is not known. Because many transient receptor potential (TRP) channels are responsible for the sensations induced by various spices and food additives, we expressed 17 TRP channels in human embryonic kidney (HEK) cells and investigated their activation by hydroxy-alpha-sanshool (HalphaSS) or hydroxy-beta-sanshool (HbetaSS) isolated from Zanthoxylum piperitum. It was found that HalphaSS, but not HbetaSS, depolarized sensory neurons with concomitant firing of action potentials and evoked inward currents. Among 17 TRP channels expressed in HEK cells, HalphaSS caused Ca(2+) influx in cells transfected with TRPV1 or TRPA1, and evoked robust inward currents in cells transfected with TRPV1 or TRPA1. In primary cultured sensory neurons, HalphaSS induced inward currents and Ca(2+) influx in a capsazepine-dependent manner. Moreover, HalphaSS-induced currents and Ca(2+) influx were greatly diminished in TRPV1(-/-) mice. HalphaSS evoked licking behavior when injected into a single hind paw of wild-type mice, but this was much reduced in TRPV1-deficient mice. These results indicate that TRPV1 and TRPA1 are molecular targets of HalphaSS in sensory neurons. We conclude that the activations of TRPV1 and TRPA1 by HalphaSS explain its unique pungent, tingling sensation.     

Lesioning of TRPV1 expressing primary afferent neurons prevents PAR-2 induced motility, but not mechanical hypersensitivity in the rat colon

Background Proteinase activated receptor 2 (PAR‐2) is expressed by many neurons in the colon, including primary afferent neurons that co‐express transient receptor potential vanilloid 1 (TRPV1). Activation of PAR‐2 receptors was previously found to enhance colonic motility, increase secretion and produce hypersensitivity to mechanical stimuli. This study examined the functional role of TRPV1/PAR‐2 expressing neurons that innervate the colon by lesioning TRPV1 bearing neurons with the highly selective and potent TRPV1 agonist resiniferatoxin. Methods Colonic motility in response to PAR‐2 activation was evaluated in vitro using isolated segments of descending colon and in vivo using manometry. Colonic mechanical nociceptive thresholds were measured using colorectal distension. Transient receptor potential vanilloid 1 expressing neurons were selectively lesioned with resiniferatoxin. Key Results In vitro, the PAR‐2 agonists, trypsin and SLIGRL did not alter contractions of colon segments when applied alone, however, the agents enhanced acetylcholine stimulated contraction. In vivo, PAR‐2 agonists administered intraluminally induced contractions of the colon and produced hypersensitivity to colorectal distention. The PAR‐2 agonist enhancement of colonic contraction was eliminated when TRPV1 expressing neurons were lesioned with resiniferatoxin, but the PAR‐2 agonist induced hypersensitivity remained in the lesioned animals. Conclusions & Inferences Our findings indicate that TRPV1/PAR‐2 expressing primary afferent neurons mediate an extrinsic motor reflex pathway in the colon. These data, coupled with our previous studies, also indicate that the recently described colospinal afferent neurons are nociceptive, suggesting that these neurons may be useful targets for the pharmacological control of pain in diseases such as irritable bowel syndrome.     

Dorsal root ganglion neurons up-regulate the expression of laminin-associated integrins after peripheral but not central axotomy

The favorable prognosis of regeneration in the peripheral nervous system after axonal lesions is generally regarded as dependent on the Schwann cell basal lamina. Laminins, a heterotrimeric group of basal lamina molecules, have been suggested to be among the factors playing this supportive role. For neurons to utilize laminin as a substrate for growth, an expression of laminin binding receptors, integrins, is necessary. In this study, we have examined the expression of laminin binding integrin subunits in dorsal root ganglion (DRG) neurons after transection to either their peripherally projecting axons, as in the sciatic nerve, followed by regeneration, or the centrally projecting axons in dorsal roots, followed by no or weak regenerative activity. In uninjured DRG, immunohistochemical staining revealed a few neurons expressing integrin subunit alpha6, whereas integrin subunits alpha7 and foremost beta1 were expressed in a majority of neurons. After an injury to the sciatic nerve, mRNAs encoding all three integrins were up-regulated in DRG neurons. By anterograde tracing, immunoreactivity for all studied integrins was also found in association with growing axons after a sciatic nerve crush lesion in vivo. In contrast, mRNA levels remained constant in DRG neurons after a dorsal root injury. Together with previous findings, this suggests that integrin subunits alpha6, alpha7, and beta1 have an important role in the regenerative response following nerve injury and that the lack of regenerative capacity following dorsal root injury could in part be explained by the absence of response in integrin regulation.     

Lithium chloride alters cytoskeletal organization in growing, but not mature, cultured chick sensory neurons.

Biochemical analysis indicates that lithium ion (Li+) has deleterious effects on the metabolism of at least two elements of the cytoskeleton in cultured chick dorsal root ganglia (DRG) neurons. Phosphorylation of newly synthesized middle molecular mass neurofilament polypeptide (NF-M) is inhibited by 10-25 mM LiCl, and tubulin (Tb) synthesized in the presence of Li+ is subject to rapid degradation. These Li(+)-induced metabolic abnormalities are accompanied by alterations in cellular and cytoskeletal morphology. Treatment of cultures having vigorously growing neurites with 25 mM LiCl results in the cessation of net neurite growth, without causing neurite retraction. Indirect immunofluorescence reveals that in these cultures Li+ provokes an aggregation of NF protein into a dense knot in the cell body/proximal neurite region. The knots contain accumulations of all three NF polypeptides and electron microscopic observation demonstrates that the knots contain intact, but disorganized, filaments. Both the inhibition of neurite outgrowth and NF collapse are reversible. Tubulin and intact microtubules are redistributed in immature cultures treated with Li+ insofar as they are excluded from the NF knots. Neurons in established cultures (e.g., 7 days and beyond) fail to show any difference between Li+ treatment and control conditions in the morphology of the cytoskeletal elements examined.     

Regulation of locus coeruleus neurons and splanchnic, sympathetic nerves by cardiovascular afferents

The brain norepinephrine (NE) neurons in the nucleus locus coeruleus (LC) have been claimed to be involved both in the regulation of behavioral functions, e.g. vigilance and arousal reactions, and in cardiovascular control. Recent studies from this laboratory have also shown that cardiovascular, vagal afferents can participate in the regulation of the LC neurons in the rat.

Utilizing electrophysiological techniques, we have now studied the effects of activation of blood volume receptors or arterial baroreceptors on the firing rate of single cells in the LC and, parallelly, on splanchnic, sympathetic discharge in the chloral hydrate anesthetized rat.

Blood volume load (0.5–5 ml heparinized blood, intravenously administered) induced a reduction in both LC neuronal firing rate and splanchnic nerve activity (SNA), effects which were readily and completely reversed by withdrawal of the corresponding amount of blood. In comparison, the central LC neurons were more sensitive to blood volume expansion than the peripheral splanchnic nerves. The effects of blood volume load on LC and SNA remained unaffected after deafferention of arterial baroreceptors.

Blood pressure elevation, induced by slow intravenous infusion of NE or angiotensin (AII) (total dose 2 μg/kg), caused an immediate reduction in both the firing rate of most of the LC cells tested as well as in SNA. While the effect on SNA was abolished by deafferentation of arterial baroreceptors, the effect on central LC activity remained largely unaffected.

Consequently, these data strengthen the concept that brain NE neurons in the LC are subject to control by peripheral blood volume receptors, analogously to peripheral sympathetic nerves. Arterial baroreceptors may still participate in the control of central noradrenergic nerve activity, but in contrast to their function for SNA they are not critical for the inhibition of LC neurons by blood pressure elevation. Rather, these two cardiovascular afferent systems may participate in the physiological regulation of the LC activity in a complimentary and convergent fashion.     

Phorbol esters stimulate 2-deoxyglucose uptake in glia, but not neurons

Activation of protein kinase C by phorbol esters caused a time- and dose-dependent stimulation (270% of control) of glucose uptake in cultured glia, but not in neurons from rat brain. The phorbol ester stimulation of Vmax of glial glucose uptake occurred only in glia despite nearly 2.5-fold greater phorbol ester binding in neurons. These differences in cellular responses to protein kinase C activation may be the key to understanding brain glucose regulation.     

Patterns of innervation of sympathetic vascular neurons by peptide-containing primary sensory fibers

The purpose of this study was to determine whether there is a specific organization of the primary sensory innervation on to identified vascular neurons in the inferior mesenteric ganglion (IMG) in guinea-pig. Retrograde tracers were placed intraluminally in inferior mesenteric artery (IMA) or inferior mesenteric vein (IMV) in vitro to identify ganglionic neurons as arterial, venous or unlabeled neurons. The distribution of primary sensory nerve fibers containing calcitonin gene-related peptide (CGRP), neuronal nitric oxide synthase (NOS) and substance P immunoreactivity (SP-IR) was compared before and after treatment with capsaicin. In control animals the density of immunoreactivity varied both with the transmitter and the type of neuron innervated. The density of immunoreactivity for all the three substances was reduced by capsaicin treatment. The degree of reduction of immunoreactivity in the fibers varied with the transmitter and the type of neuron. The density of CGRP and SP immunoreactive fibers was greatest around unlabeled neurons; 78% of the CGRP fibers were of primary sensory origin and all of the SP fibers were primary sensory. Around arterial neurons 44% of the CGRP fibers were of primary sensory origin and around venous 68% were primary sensory. NOS positive innervation around venous neurons was denser than around arterial neurons and all of it was completely (97%) eliminated by capsaicin, indicating that it was solely of primary sensory origin. We conclude that the primary sensory fibers innervating the IMG are differentially distributed to arterial and venous neurons and that the pattern of distribution is characteristic for each sensory neurotransmitter.     

Influence of the extracellular potassium environment on neurite growth in sensory neurons, spinal cord neurons and sympathetic neurons

The influence of the extracellular potassium concentration ([K+]o) on neurite growth in rat sensory neurons, spinal cord neurons and sympathetic neurons was investigated. Experiments carried out in 3-compartment culture dishes showed that although neurites from sensory and spinal cord neurons were capable of growing in both 5 mM [K+]o and 20 mM [K+]o, they were virtually unable to grow from a region of 5 mM [K+]o into a region of 20 mM [K+]o. Neurites from sympathetic neurons behaved similarly although [K+]o exceeding 20 mM was required to exclude sympathetic neurites. We suggest the possibility of a negative chemotaxis to [K+]o by growth cones in these neurons. Neurite regeneration following axotomy in sensory neurons was partially inhibited distal to a proximo-distal increase in [K+]o. The nature of this inhibition was somewhat different from that described previously in sympathetic neurons. The possibility is raised that [K+]o plays a role in the development of the nervous system.