activation of fibres in rat sciatic nerve alters phosphorylation state of connexin-43 at astrocytic...

ACTIVATION OF FIBRES IN RAT SCIATIC NERVE ALTERS PHOSPHORYLATIONSTATE OF CONNEXIN-43 AT ASTROCYTIC GAP JUNCTIONS IN SPINAL CORD:

EVIDENCE FOR JUNCTION REGULATION BY NEURONAL–GLIALINTERACTIONS

W. E. I. LI and J. I. NAGY*Department of Physiology, Faculty of Medicine, University of Manitoba, 730 William Avenue, Winnipeg, Manitoba, Canada R3E 3J7

Abstract—Intercellular communication via gap junction channels composed of connexin-43 is known to be regulated byphosphorylation of this protein. We investigated whether connexin-43 at astrocytic gap junctions is similarly regulated in responseto neural activation. The effect of peripheral nerve stimulation on connexin-43 phosphorylation state in the spinal cord of rats wasexamined with a monoclonal antibody (designated 13-8300) shown previously to recognize selectively a dephosphorylated form ofconnexin-43. Immunolabelling with 13-8300 was absent in the lumbar spinal cord in control animals, but was induced in the dorsalhorn ipsilateral to sciatic nerve electrical stimulation for 15 min or 1 h at a frequency of 1 or 100 Hz. Immunorecognition ofconnexin-43 by a polyclonal anti-connexin-43 antibody, shown previously to undergo epitope masking under various conditions,was reduced in the dorsal horn on the stimulated side. These responses were abolished by local anaesthetic or tetrodotoxinapplication proximal to the site of nerve stimulation. Selective electrical stimulation of A-fibres or activation of cutaneous C-fibresby capsaicin evoked labelling with 13-8300 in deep and superficial laminae of the dorsal horn, respectively. Nerve stimulationincreased the number of 13-8300-positive astrocytic gap junctions, as well as the levels of dephosphorylated connexin-43 in thedorsal horn on the stimulated side. Sciatic nerve transection produced results similar to those seen after C-fibre activation withcapsaicin.

Thus, peripheral nerve stimulation evokes astrocytic connexin-43 dephosphorylation in the spinal cord dorsal horn, suggestingthat gap junctional coupling between astrocytesin vivo is subject to regulation by neuronal–glial interactions following neuralactivation.q 2000 IBRO. Published by Elsevier Science Ltd.

Key words:astrocytes, connexin-43, phosphorylation, gap junctions, nerve stimulation, neuron–astrocyte interaction.

Gap junctions occur at apposing plasma membranes of manycell types and are thought to contribute to local metabolichomeostasis and synchronization of cellular activities byallowing direct intercellular movement of ions, metabolitesand second messengers. These junctions are composed ofclustered hemichannels, termed connexons, which formaqueous conduits linking the intracellular compartments ofjunctionally coupled cells. Each connexon consists of ahexameric arrangement of gap junction proteins termedconnexins.3,15,25 Gap junctional intercellular communication(GJIC) or channel conductance is regulated by a variety offactors, including pH, Ca21 concentration, transjunctionalvoltage and signalling mechanisms.3,10,15 Numerous studiesof cells coupled via gap junctions composed of connexin-43(Cx43), one member of the multigene connexin family,3

indicate that channel gating is also regulated by connexinphosphorylation.3,28 A number of serine phosphorylationsites has been identified within the Cx43 sequence62 andinduction of Cx43 dephosphorylation by various methodshas been correlated with either a reduction in GJIC6,13,48 oran increase in unitary channel conductance.35

In the CNS, it is well established that astrocytesin vivoexpress Cx43 and connexin-30 (Cx30),39–42,46,70,71and that

these cells are extensively coupled10,36,43,50with as many as30,000 gap junctions per astrocyte.67 Although the functionalrequirement for this high level of coupling is not yet clear, ithas been suggested that astrocytic GJIC serves a role in K1

spatial buffering66 and intercellular exchange of Na1.53 Thereis now reason to consider that the conductance state of astro-cytic junctional channels may be dynamically regulated byneuronal–glial interactions. In isolated optic nerve, forexample, dye coupling between astrocytes is increased byelectrical stimulation of the nerve.32 Furthermore, GJICbetween astrocytesin vitro can be rapidly altered by elevatedK1, as well as by a variety of neuronally released substances,including ATP, glutamate, noradrenaline, endothelin andanandamide.10 As in peripheral tissue, Cx43 in the CNS existsas a phosphoprotein,18,30 and control of its phosphorylationstate by kinase- and phosphatase-associated signalling path-ways may be one mechanism whereby astrocytic GJIC isregulated by these substances. Indeed, agents known to altercoupling between astrocytes can also influence the activity of,for example, protein kinase C (PKC) and mitogen-activatedprotein kinases in these cells,5,20,29 and it is known thatactivation of PKC reduces astrocytic GJIC.9,23

To date, most studies of astrocytic GJIC and Cx43phosphorylation state have been conducted using culturedastrocytes or CNS tissues subjected to variouslesions.4,6,10,16,19,30,33,39,43Thus, it is not clear whether Cx43phosphorylation mechanisms potentially contribute to astro-cytic GJIC regulation under normal physiological conditions.To address this, and to examine the possible role of neuronal–glial interactions in GJIC regulation between astrocytesin

Connexin-43 dephosphorylation after nerve stimulation 113

113

NeuroscienceVol. 97, No. 1, pp. 113–123, 2000Copyrightq 2000 IBRO. Published by Elsevier Science Ltd

Printed in Great Britain. All rights reserved0306-4522/00 $20.00+0.00PII: S0306-4522(00)00032-4

Pergamon

www.elsevier.com/locate/neuroscience

*To whom correspondence should be addressed. Tel.:11-204-789-3767;fax:11-204-789-3934.

E-mail address:[email protected] (J. I. Nagy).Abbreviations:Cx, connexin; GJIC, gap junctional intercellular communi-

cation; IR, immunoreactivity; PAP, peroxidase–antiperoxidase; PB,phosphate buffer; PBS, phosphate-buffered saline; PBST, phosphate-buffered saline with Triton X-100; PKC, protein kinase C.

vivo, we investigated the phosphorylation status of astrocyticCx43 in the spinal cord after electrical stimulation of sciaticnerve or topical capsaicin application to the skin, whichselectively activates unmyelinated C-fibres.21,31,57 Spinalcord tissues were analysed immunohistochemically and bywestern blotting with various anti-Cx43 antibodies, includingone that we characterized previously as selectively recogniz-ing a dephosphorylated form of Cx43 in at least neural tissuesand some cell lines.30,40

EXPERIMENTAL PROCEDURES

Antibodies

Two rabbit polyclonal anti-Cx43 antibodies designated 18A and 71-0700, and one monoclonal Cx43 antibody designated 13-8300, wereused in this study. The latter two antibodies were obtained from ZymedLaboratories (South San Francisco, CA, U.S.A.). Anti-Cx43 18A and13-8300 were produced against peptides corresponding to amino acids346–363 and 360–376, respectively, in the carboxy-terminal sequenceof Cx43.1 The Cx43 sequence recognized by antibody 71-0700 isproprietary, but does not correspond to sequences recognized by either18A or 13-8300.30 All three antibodies have been extensively charac-terized for specific detection of Cx43 by immunohistochemical,western blotting and immunoprecipitation methods, as well as by anti-body preadsorption with synthetic peptide, in a variety of tissuesincluding normal and lesioned CNS, cardiac myocytes and culturedcells.18,19,30,40,44,46,47,59,70,71Selective detection of a dephosphorylatedform of Cx43 by antibody 13-8300 in some tissues has also beendescribed previously by us.30,40 Synthetic peptide preadsorption of13-8300 in control experiments performed in the present studies wasconducted as described previously.30 Selective immunohistochemicalrecognition of dephosphorylated Cx43 by 13-8300 is supported by theconsistent presence or absence of immunostaining with this antibody intissues that contain or lack, respectively, dephosphorylated Cx43 asseen by western blotting.30,40 Specificity of 13-8300 for Cx43 in thespinal cord is further indicated by the similar appearance of 18A label-ling before and 13-8300 labelling after sciatic nerve stimulation, asseen in the present study, the localization of 13-8300 immunoreactivityat astrocytic gap junctions and the elimination of 13-8300 immuno-labelling by peptide preadsorption.

Animal preparation

Sixty-five male Sprague–Dawley rats weighing 280–330 g wereused in this study. A minimum of three animals was used for eachtreatment condition and six animals served as sham-operated controls.All animals were obtained from the local Central Animal Care Services(CACS) at the University of Manitoba and utilized according toapproved protocols by the CACS Care Committee with minimizationof stress to, and the numbers of, animals used. Anaesthesia wasinduced with 5% halothane and maintained with 1–2% halothane in70% O2/30% N2O using a tracheal catheter. Animals were immobilizedwith gallamine triethiodide (40 mg/kg, i.v.) in order to avoid muscleactivation and limb movement during sciatic nerve stimulation. Toenable control of anaesthesia during surgery and stimulation, the tailartery was cannulated for measurement of blood gases (pH,pO2, pCO2)and for monitoring blood pressure, which was maintained between 70and 100 mmHg in all experiments. Body temperature was monitoredand maintained at 37.5oC.

The left and right sciatic nerves were exposed and carefully freed ofconnective tissue along a length of 2–3 cm. The sciatic nerve of oneside was subjected to electrical stimulation, while the other side servedas a control for nerve exposure and manipulation. With the sciaticnerve resting in warm mineral oil, nerve stimulation was achievedwith bipolar silver electrodes linked to a Grass SD9 electrical stimu-lator, which was set to generate 1-ms electrical pulses of either 1 or100 Hz. Stimulation intensity was varied from 20mA to 2 mA to allowselection of current that produced excitation of both large- and small-diameter fibres or only A-fibres. Additional electrodes, placed proxi-mal to the site of stimulation, were used to monitor compound actionpotentials, allowing selection of appropriate current strength for nerveactivation. Nerves were stimulated for durations of either 15 min or1 h. In some experiments, cotton soaked with either 0.25% bupivacainehydrochloride or 1mM tetrodotoxin was applied to the sciatic nerve

proximal to the stimulating electrodes in order to block impulseconduction to the spinal cord, with verification of blockade by moreproximally placed electrodes. In separate animals, selective C-fibreactivation in animals anaesthetized with equithesin was attained bytopical application of a 1% capsaicin solution to the left hindlimb.Capsaicin was dissolved in 48% ethanol, 25% propylene glycol,25% distilled water and 2% methyl laurate to facilitate maximum C-fibre excitation.31 The solvent alone was applied to the right hindlimb.In another group of animals, the sciatic nerve was unilaterally ligatedwith surgical thread and cut distal to the ligation.

Immunohistochemistry

Immediately after nerve stimulation, or 1 h after either capsaicinapplication or nerve transection, animals were given an overdose ofnembutol and perfused transcardially with cold 4% formaldehydedissolved in 0.1 M sodium phosphate buffer at pH 7.4 (PB). The L4to S1 segment of the spinal cord was removed, postfixed for 2 h at 4oCin the same fixative and cryoprotected in PB containing 25% sucroseand 10% glycerol. Sections were cut transversely at a thickness of20mm on a sliding microtome and collected free floating in PBcontaining 0.9% saline (PBS). For electron microscopy, spinal cordswere stored in PBS and cut at 30mm with a Vibratome. Sections forlight microscopy were incubated for 48 h with either antibody 18Adiluted 1:2000, antibody 71-0700 diluted 1:500 or antibody 13-8300diluted 1:500 in PBS containing 0.3% Triton X-100 (PBST). Sectionswere then processed for labelling with peroxidase–antiperoxidase(PAP) as described previously.30,46,47In brief, goat anti-rabbit or goatanti-mouse immunoglobulin G were used as secondary antibodiesat 1:100 dilution in PBST and rabbit PAP or mouse clonoPAP(Sternberger Monoclonals) were used at 1:500 dilution in PBST.Secondary and tertiary antibody incubations were conducted for1.5 h. Sections were washed with PBST for at least 60 min betweenantibody incubations. Final enzyme reaction was conducted for 10–15 min in 50 mM Tris–HCl buffer (pH 7.4) containing 0.00005% H2O2and 0.02% diaminobenzidine. The immunostained sections weremounted on to slides, dehydrated in alcohol, cleared in Histoclearand coverslipped with Lipshaw mounting medium. Sections wereprocessed similarly for electron microscopy, except that 0.2% Photo-Flo 200 (Kodak) instead of Triton X-100 was included in the PBS forall of the incubation and wash steps. Sections were plastic embeddedand counterstained as described previously.30,46,47

Western blotting

A minimum of three animals was subjected to electrical stimulationof the sciatic nerve and capsaicin stimulation of unmyelinated primaryafferent fibres. After 1 h of nerve stimulation and laminectomy at alumbar level, spinal cord segments containing L4 to S1 were rapidlyremoved, and dorsal horns from control and stimulated sides weredissected, stored at270oC and analysed separately by western blot-ting. Tissue was homogenized in 1 mM bicarbonate buffer (pH 7.4)containing 1 mM phenylmethylsulphonyl fluoride and 1 mg/ml each ofleupeptin and pepstatin. The phosphatase inhibitors sodium fluorideand sodium orthovanadate, at concentrations of 10 and 1 mM, respec-tively, were also included in the homogenization buffer to inhibit Cx43dephosphorylation during tissue processing. The total protein in homo-genates was assayed with a Bio-Rad protein assay kit. Sodium dodecylsulphate–polyacrylamide gel electrophoresis and western blottingwere performed as described previously18,30,40 using antibodies 18Aand 13-8300 at 1:200,000 and 1:750 dilutions, respectively.

RESULTS

Light microscopy of connexin-43 immunolabelling in thedorsal horn

In sham-operated animals and in the control unstimulatedside of spinal cords from animals given unilateral electricalstimulation of the sciatic nerve, immunoreactivity (IR) withantibody 18A (Fig. 1A, C) was similar to patterns that wedescribed previously in normal unoperated rats.46 This wasalso true of animals treated unilaterally with topical capsaicinor nerve transection, indicating that none of these treatmentshad an effect on the appearance of labelling contralateral to

W. E. I. Li and J. I. Nagy114

the side of treatment. Labelling consisted primarily ofpunctate profiles in gray matter and was localized to punctaor fibrous astrocytic elements in white matter. Immuno-staining was distributed throughout the dorsal and ventralgray matter, but was most intense in the superficial layersof the dorsal horn and around the central canal. Antibody71-0700 produced qualitatively similar, but a lower intensityof, immunostaining in control spinal cord (not shown).Monoclonal anti-Cx43 antibody 13-8300 gave a near totalabsence of labelling in spinal cords of normal or sham-operated animals (not shown), and 13-8300-IR was alsovirtually absent in the spinal cord contralateral to the varioustreatments (Fig. 1B, D).

The effect of sciatic nerve electrical stimulation on 18A-IRand 13-8300-IR in the spinal cord is shown in Fig. 1. Stimula-tion for 1 h at 1 Hz with a current strength of 2 mA, leading tothe activation of both myelinated and unmyelinated fibres asjudged from observation of proximal compound action poten-tials, resulted in a consistent reduction of 18A-IR in the L4 toS1 segment of dorsal horn ipsilateral to the stimulation (Fig.1A). In adjacent sections processed with antibody 13-8300,immunoreactivity was induced on the stimulated side in anarea of the dorsal horn (Fig. 1B) corresponding closely to theregion of 18A reduction, which was demarcated ventrally atthe level of laminae VI/VII. Nerve stimulation also induced13-8300-IR in the dorsal portion of the dorsolateral funiculus(Fig. 1B). Cx43 remained undetectable by antibody 13-8300in all other areas of the spinal cord, and no alterations in either18A-IR or 13-8300-IR were evident in dorsal column nucleior thalamus of animals receiving nerve stimulation (notshown). In addition, nerve activation had no discernible effect

on Cx43 detection by antibody 71-0700 in the spinal cord orthe more rostral regions examined (not shown). Nerve stimu-lation for 1 h at a frequency of 100 Hz gave results similar tothose seen after stimulation at 1 Hz for 1 h (not shown).Reduction of the stimulation time to 15 min at 1 Hz dimin-ished the loss of 18A-IR (Fig. 1C) and reduced the expanse ofdorsal horn exhibiting intense 13-8300-IR on the stimulatedside (Fig. 1D). Although it was already evident from sham-operated controls that these changes in Cx43 detection werenot due to surgery alone, a relationship to nerve activationwas further tested by blockade of impulse conduction withbupivacaine or tetrodotoxin application to the sciatic nerveproximal to the site of stimulation. Both agents abolished thealterations in 18A-IR and 13-8300-IR seen after nervestimulation for 1 h at 1 Hz (not shown).

We next determined whether alterations in Cx43 detectionby antibodies occur only after activation of all fibres in thesciatic nerve or whether this could be elicited by activation ofA-fibres or C-fibres alone. Selective activation of A-fibreswas achieved by sciatic nerve stimulation at a reduced inten-sity of about 100mA, at which C-fibre potentials recordedfrom a proximal electrode were not evident. After 1 h stimu-lation at 1 Hz, 18A-IR was only minimally reduced in deepand medial superficial dorsal horn laminae on the stimulated(Fig. 2B) compared with the contralateral control side(Fig. 2A). Immunoreactivity with antibody 13-8300 wasinduced in most areas of the dorsal horn, although not asprominently as seen after stimulation of both A- and C-fibres,and lateral regions of laminae I, II and III remained unlabelled(Fig. 2C). After selective C-fibre activation by topical appli-cation of capsaicin unilaterally to the hind leg for 1 h, 18A-IR


Fig. 1. Photomicrographs showing Cx43 detection with antibody 18A and 13-8300 in rat spinal cord dorsal horn after unilateral sciatic nerve stimulation on theright side compared with control unstimulated left side. In all sections, labelling on the control side is dense with antibody 18A and virtually absent with 13-8300. (A, B) After 1 h of nerve stimulation, 18A-IR is reduced in the dorsal horn (A, arrows), while 13-8300-IR is induced in the corresponding area of anadjacent section (B, arrows). (C, D) After 15 min of sciatic nerve stimulation, 18A-IR is slightly reduced in deeper laminae (C, arrows) and 13-8300-IR is

induced in the superficial half of the dorsal horn (D, arrows). Magnifications:× 50 (A–D).

was unaltered (not shown) and 13-8300-IR was increasedrobustly in the lateral half of dorsal horn regions encompass-ing laminae I, II and III (Fig. 2E). As shown in the case ofcapsaicin-induced labelling with 13-8300 (Fig. 2F), 13-8300-IR was punctate and, whenever present after nerve activation,had an appearance similar to that seen in normal tissue stainedwith antibody 18A. Furthermore, as shown in the case ofA-fibre stimulation, the specificity of Cx43 detection by anti-body 13-8300 is indicated by abolition of immunolabellingafter preadsorption of 13-8300 with synthetic peptide antigen(Fig. 2D).

In the dorsal horn of animals receiving sciatic nerve

transection followed by a survival time of 1 h, 18A-IR wasreduced slightly in the superficial half of lamina II on thetransected (Fig. 3B) compared with the control side (Fig.3A) and 13-8300-IR was consistently induced in lateralregions of laminae I, II and III (Fig. 3D), which correspondedclosely to areas of induced 13-8300-IR seen after capsaicintreatment. Nerve transection did not elicit Cx43 detectionwith antibody 13-8300 in the ventral horn on the transectedside (not shown) or in the dorsal (Fig. 3C) or ventral horn onthe non-transected side. Sciatic nerve transection produced noalterations in immunolabelling for Cx43 with antibody71-0700 (not shown).


Fig. 2. Photomicrographs showing Cx43 detection with antibody 18A and 13-8300 in spinal cord dorsal horn after 1 h electrical stimulation of A-fibres in thesciatic nerve or C-fibre activation by topical capsaicin applied to the left hind skin. (A–C) A-fibre stimulation slightly reduced 18A-IR in the dorsalhorn on thestimulated (B, arrows) compared with control unstimulated (A) side, and induced 13-8300-IR in dorsal horn laminae subjacent to the substantia gelatinosa (C,arrows). (D) The 13-8300-IR seen in C is eliminated in an adjacent section after preadsorption of 13-8300 with peptide antigen. (E) After cutaneous C-fibreactivation, dense 13-8300-IR is seen in superficial laminae on the side (left, arrows) corresponding to that of capsaicin application, while no labelling is seen onthe control side or elsewhere in the spinal cord. (F) Higher magnification of 13-8300-IR in E shows the punctate appearance of labelling. Magnifications: × 110

(A, B); × 50 (C–E); × 330 (F).

Electron microscopy of connexin-43 immunolabelling in thedorsal horn

Ultrastructural immunolabelling patterns obtained withantibody 18A in the dorsal horn of unstimulated spinal cordwere similar to those reported previously.46 This consisted ofan exclusively astrocytic localization where label was eitherdiffuse and dispersed throughout the cytoplasm of fine astro-cyte processes, particularly those ensheathing glomeruli,axons and dendrites, or concentrated at gap junctions betweenthese processes (Fig. 4A). After sciatic nerve stimulation at1 Hz for 1 h at a strength activating both A- and C-fibres,18A-IR (Fig. 4B, C) was largely comparable qualitativelywith that in normal dorsal horn. Some gap junctions, however,tended to exhibit lighter labelling of their inner membranes(Fig. 4C) than seen in control tissue, and unlabelled astrocyticgap junctions were observed in the vicinity of labelledprocesses (Fig. 4D), indicating that lack of labelling wasnot simply due to failure of antibody penetration into sections.In the dorsal horn of the non-stimulated side, such unlabelledastrocytic gap junctions were never observed in tissueprocessed with antibody 18A.

In representative dorsal horn areas of sections processedwith antibody 13-8300, very little labelling was seen on theside contralateral to nerve stimulation (Fig. 5). Astrocyteprocesses were almost entirely devoid of 13-8300-IR (Fig.5A) and the majority of astrocytic gap junctions examinedwas unlabelled (Fig. 5B). However, lightly or partiallylabelled gap junctions were seen (Fig. 5C) and the occurrenceof these in the vicinity of unlabelled junctions (Fig. 5D) again

indicated that lack of labelling, in this case with antibody 13-8300, was not due to failure of antibody penetration to thedepth of section analysed by electron microscopy.

After stimulation of both A- and C-fibres in sciatic nerve at1 Hz for 1 h, 13-8300-IR of gap junctions in the dorsal horn onthe stimulated side was similar to the labelling of junctionswith 18A in normal dorsal horn and in dorsal horn contra-lateral to the side of stimulation. Labelling of astrocytic gapjunctions was intense (Fig. 6A, B) and junctions characteris-tically displayed dense, uniform immunoreaction productdeposited along each side of their inner membranes (Fig.6C), resulting in a symmetric appearance of labelling at themajority of junctions. Occasionally, however, labelling wasasymmetric, such that one side of a junction formed by animmunoreactive process was labeled, while the other sidealong with the contributing process was unstained (Fig. 6D).Labelling within astrocytic processes with 13-8300 was lesswidely distributed than seen with 18A in control animals, andwas mostly seen near labelled gap junctions, where it mayhave arisen by diffusion of immunoreaction product.

A survey of labelled and unlabelled gap junctions in thedorsal horns of animals receiving sciatic nerve stimulation asdescribed above in tissues examined by electron microscopyis given in Table 1. In control dorsal horn, the number ofjunctions unlabelled with 13-8300 was about twice that oflabelled junctions. After nerve stimulation, the number of13-8300-immunoreactive gap junctions on the stimulatedside was significantly increased to 98% of junctions counted,compared with 35% on the control side. A similar analysiswas attempted in material stained with antibody 18A, but the


Fig. 3. Photomicrographs showing antibody detection of Cx43 in the spinal cord dorsal horn 1 h after unilateral sciatic nerve transection. (A, B) A slightreduction in 18A-IR is seen in dorsal horn laminae I and II of the transected (B) compared with the control (A) side. (C, D) 13-8300-IR is induced in superficiallaminae ipsilateral to nerve transection (D, arrows), but remains undetected on the contralateral control side (C). Magnifications:× 110 (A, B); × 90 (C, D).

presence of numerous very lightly stained junctions oftenmade it difficult to designate these unequivocally as stainedor unstained.

Western blotting of connexin-43 after nerve stimulation

Western blotting of Cx43 was conducted to determine therelative levels of the phosphorylated and non-phosphorylatedforms of this protein in the spinal cord dorsal horn after nerve

stimulation. Phosphorylation status is typically deduced fromthe progressively slower mobility of Cx43 in gels, as itbecomes phosphorylated at multiple sites. Comparable toresults in the rat brain probed with antibody 18A,18,30 whichdetects all phosphorylated forms of Cx43, a higher proportionof astrocytic Cx43 in control dorsal horn was found tomigrate at a slower mol. wt of 43,000 compared with thenon-phosphorylated 41,000 mol. wt form (Fig. 7A, lane 1).Following activation of C-fibres by topical capsaicin application,


Fig. 4. Electron micrographs of immunolabelling for Cx43 with antibody 18A in spinal cord dorsal horn contralateral and ipsilateral to unilateral sciatic nerveelectrical stimulation. (A) In contralateral control areas, diffuse labelling is seen in astrocyte processes (arrows) and at gap junctions betweenthese processes(arrowheads, magnified in inset). (B–D) After stimulation of the sciatic nerve for 1 h, labelling is seen in astrocyte processes (B, arrows), and astrocytic gapjunctions are either densely stained (C, arrows), lightly stained (C, arrowhead) or unlabelled (D, arrowhead), despite the presence of cytoplasmic18A-IR in

nearby processes (D, arrows). Magnifications:× 13,000 (A); × 82,000 (inset);× 20,000 (B); × 95,000 (C);× 46,000 (D).

Table 1. Comparison of the proportion of astrocytic gap junctions that are unlabelled and immunolabelled with antibody 13-8300 in control spinal cord dorsalhorn and in dorsal horn after sciatic nerve electrical stimulation†

Control SNS

Unlabelled Labelled Unlabelled Labelled

Average 118̂¯10 (65%) 66̂

¯9 (35%) 4̂

¯1* (2%) 186^

¯7* (98%)

Total counted 353 197 11 558

†Control, control side; SNS, sciatic nerve-stimulated side. Values are means^¯S.E.M. of three animals. Values in parentheses indicate percentage of

unlabelled and labelled junctions in control or stimulated animals. Total counted indicates the cumulative number of gap junctions counted in threeanimals.*P, 0.01.

as in the anatomical studies above, a small increase in thenon-phosphorylated 41,000 mol. wt form was evident (Fig.7A, lane 2), and this was reproducible in the three animalsexamined. However, resolution of bands was poor due tonecessary inclusion of phosphatase inhibitors, which inexplic-ably deteriorate band separation, as noted previously.34,40

Also consistent with results obtained using brain tissue,30,40

antibody 13-8300 was found to detect only a dephosphory-lated 41,000 mol. wt form of Cx43 in control dorsal horn (Fig.7B, lanes 1 and 2). After activation of C-fibres with topicalcapsaicin, blots probed with antibody 13-8300 showed anincreased level of the 41,000 mol. wt form of Cx43 in thedorsal horn on the side of capsaicin application (Fig. 7B, lane4) compared with the control side (Fig. 7B, lane 3). Thisincrease was not likely to be due to surgical or dissectionartifact, since no such differences were seen between theleft and right sides of animals receiving topical applicationof capsaicin vehicle (Fig. 7B, lanes 1 and 2). Similar resultswere seen in spinal cord dorsal horns from animals givenelectrical stimulation of the sciatic nerve (not shown).

DISCUSSION

We show that Cx43 in the spinal cord is largely phosphory-lated and that the response to sciatic nerve input is

dephosphorylation of a specific pool of Cx43 located at astro-cytic gap junctions. These results demonstrate that remoteactivation of primary fibres in the sciatic nerve can influencethe phosphorylation state of astrocytic Cx43 in the dorsal hornof the spinal cord and provide indirect evidence for theinvolvement of neuronal–glial interactions in the regulationof astrocytic GJIC under physiological conditionsin vivo.

Immunorecognition of connexin-43 by 13-8300

The above conclusions rely on the proposition that anti-body 13-8300 selectively detects a dephosphorylated formof Cx43 in our spinal cord preparations. As demonstrated inbrain tissue, cardiac tissue and cultured cells,30,40 this anti-body also fails to recognize the slower migrating, phosphory-lated 43,000 mol. wt forms of Cx43 in the spinal cord, butdoes react with a faster migrating 41,000 mol. wt form, whichcorresponds to dephosphorylated Cx43.7,26,37,55 We havesuggested that the lack of 13-8300 reaction with slowermobility forms of Cx43 is likely due to epitope blockade byphosphate groups.40 This is supported by the presence of PKCphosphorylation sites (Ser368 and Ser372) within the Cx43sequence (amino acids 360–376) against which 13-8300 wasgenerated,54,55 and by results showing that phosphorylatedforms of Cx43 ordinarily undetectable by 13-8300 are


Fig. 5. Electron micrographs of immunolabelling for Cx43 with antibody 13-8300 in normal spinal cord dorsal horn. (A–C) Cytoplasmic 13-8300-IR is verysparse and typical gap junctions distributed in neuropil are generally unlabelled (A, B, arrowheads) or occasionally lightly labelled (C, arrowheads). (D) Apartly labelled gap junction (arrow; magnified in lower inset) is seen in the same field as two unlabelled junctions (arrowheads; magnified in upper inset).

Magnifications:× 20,000 (A); × 140,000 (B);× 95,000 (C);× 32,000 (D); × 87,000 (upper inset);× 76,000 (lower inset).

detected following Cx43 dephosphorylation with alkalinephosphatase.40

It should be noted that Cx43 phosphorylation within the13-8300 epitope remains to be demonstrated specifically in

astrocytes. Furthermore, although Cx43 is known to bemultiply phosphorylated,14,37 it is not clear which sitescontaining phosphate groups cause a slower mobility onwestern blots. There is evidence, however, that the formwith the fastest mobility is entirely non-phosphory-lated,7,26,37,55and it is this form that appears to be detectedby 13-8300. Thus, for tissues we have examined so far, nowincluding the spinal cord, we suggested that dephosphory-lation at the 13-8300 epitope may be accompanied bydephosphorylation at other, if not all other, sites in themolecule. However, we did not exclude the possibility that13-8300 detection of phosphorylated forms of Cx43 havingslower mobility on western blots could occur in other systemswhere phosphate removal at the 13-8300 epitope may not beaccompanied by dephosphorylation at other sites. Evidencefor this has, in fact, been reported recently in culturedfibroblasts.8

Immunorecognition of astrocytic connexin-43 by 18A

In several CNS lesion preparations, antibody 18A consis-tently exhibits loss of Cx43 immunorecognition in tissuesections, which cannot be accounted for by reduced levelsor degradation of Cx43.19,30,47,56,59,63We now show that thisepitope masking also occurs in normal spinal cord followingnerve stimulation and that it often coincides spatially andtemporally with Cx43 dephosphorylation, suggesting arelationship between the two processes. The molecular basis


Fig. 6. Electron micrographs of immunolabelling for Cx43 with antibody 13-8300 in spinal cord dorsal horn after 1 h of sciatic nerve electrical stimulation. (A)Labelling is seen dispersed in astrocyte processes (arrows) as well as at gap junctions between labelled processes (arrowheads). (B) Two fibrous astrocyteprocesses containing intermediate filaments (if) are seen forming a densely labelled gap junction (arrowhead). (C, D) Gap junctions either display an equaldensity of labelling on each side of the junctional membranes (C, arrowheads) or occasionally exhibit labelling on only one side (D, arrow) and not theother (D,

arrowhead). Magnifications:× 22,000 (A); × 35,000 (B); × 90,000 (C);× 124,000 (D).

Fig. 7. Western blots showing Cx43 phosphorylation state in spinal corddorsal horn after cutaneous C-fibre activation by capsaicin. (A) Blot probedwith antibody 18A shows detection of both phosphorylated and non-phos-phorylated Cx43 migrating at mol. wts of 43,000 and 41,000, respectively,with the former predominating on both control (lane 1) and stimulated (lane2) sides of the dorsal horn. (B) Blot probed with antibody 13-8300 showsdetection of only the non-phosphorylated form migrating at mol. wt 41,000.Control animals display an equal level of this form in the left (lane 1) andright (lane 2) dorsal horn. Topical capsaicin application to the right hindflank for 1 h induces an increase in the non-phosphorylated form in theipsilateral (lane 4) compared with the contralateral control dorsal horn

(lane 3).

for this masking is as yet unclear, but is not likely due tophosphorylation of the 18A epitope (amino acids 346–363of Cx43), which does not appear to contain any knownphosphorylation sites.62 The nearby carboxy-terminus ofCx43, however, has been reported to contain a PDZ inter-action domain that is able to bind to a PDZ domain in thetight junction protein zona occludens-1.12,60 In addition,myotonic dystrophy protein kinase was reported to be co-localized with Cx43 at gap junctions in cardiac tissue.38

Thus, we speculate that the 18A epitope of Cx43 in astrocytesmay undergo blockade of antibody recognition followingassociation of an as yet unidentified protein with possiblythe PDZ interaction domain of Cx43. Since epitope maskingand Cx43 dephosphorylation almost always occur concomi-tantly, we further speculate that Cx43 association with aprotein that blocks the 18A epitope may be promoted bydephosphorylation at the 13-8300 epitope.

Connexin-43 immunolabelling in the spinal cord

We have demonstrated previously that Cx43 in both brainand spinal cord is localized exclusively to astrocytes.46,70,71Astriking difference, however, was that labelling with 18A inthe brain was localized largely at or near astrocytic gap junc-tions, while in the spinal cord it was present at junctions,along non-junctional membranes and intracellularly withinastrocytic processes.46 This is noteworthy for two reasons.First, our electron microscopy counts indicate some lightand partial labelling of gap junctions with 13-8300 in normaldorsal horn, consistent with western blot detection of somedephosphorylated Cx43 in this tissue. This may represent anormal complement of dephosphorylated gap junctionalCx43, or it may be due to rapidpost mortemCx43dephosphorylation, which we have characterized in thebrain.18 This low level of labelling appears to be below thesensitivity of light-microscopic detection, since no 13-8300-IR was seen by light microscopy in normal tissue. It isperhaps curious that little cytoplasmic labelling is foundwith 13-8300 in control tissue, suggesting that a cytoplasmicpool of Cx43 detectable with 18A in the cord remains largelyphosphorylated.

Secondly, nerve stimulation increased the number andimmunoreaction density of gap junctions labelled with13-8300, but produced a less striking enhancement ofcytoplasmic Cx43-IR. Given the rapid appearance of dephos-phorylated Cx43 after nerve stimulation, together with itspresence almost exclusively at gap junctions, it is unlikelythat induction of labelling with 13-8300 was due tode novosynthesis of Cx43. This is further supported by the lack of anychanges in Cx43 labelling with antibody 71-0700, which wasgenerated against an epitope in the molecule different fromthose recognized by either 18A or 13-8300. Thus, conditionsin the spinal cord which cause Cx43 dephosphorylation mayactivate a phosphatase that appears to act selectively on Cx43localized at gap junctions. Lack of action on the cytoplasmicpool may explain the large proportion of Cx43 that remainsphosphorylated after nerve stimulation as seen by westernblotting with antibody 18A.

Neuronal activity and connexin-43 dephosphorylation

Neurons in the dorsal horn are mostly quiescent in restinganimals and display burst discharges upon peripheral nerve

stimulation. Activation of either myelinated fibres by low-intensity electrical stimulation or unmyelinated C-fibresincreases neuronal discharge in widespread areas of the dorsalhorn, despite the largely superficial and deep dorsal horntermination areas of unmyelinated and myelinated primaryafferent fibres, respectively.27,68,69Thus, our observation thatCx43 dephosphorylation occurs in superficial or deep laminaecorresponding to the termination area of the fibre types acti-vated suggests that this response is perhaps more related to theactivity of primary afferents than that of dorsal horn neurons.In any event, the nature of neuronal–glial interactions leadingto astrocytic Cx43 dephosphorylation remains to be eluci-dated. Transmitter substances released from primary afferentsor dorsal horn neurons may mediate such interactions via adirect action on astrocytes, which are endowed with a host ofneurotransmitter receptors linked to intracellular signallingcascades.17,22 Alternatively, such interactions may be medi-ated by changes in ionic milieu, particularly elevation ofextracellular K1 in the dorsal horn, even by low levels ofsciatic nerve stimulation.24,66 This is less likely, however,since extracellular K1 concentration in the ventral horn canbe increased to levels similar to those observed in the dorsalhorn following antidromic activation of motor axons atfrequencies similar to those used here,24,64 yet no Cx43dephosphorylation was observed in the ventral horn.

Astrocytic connexin-43 after sciatic nerve transection

Facial nerve transection was reported by Rohlmannetal.51,52 to cause a rapid increase (within 45 min) in Cx43immunostaining in the facial nucleus. Using 13-8300, whichwas generated against a similar, although not exactly thesame, Cx43 sequence as the antibody employed by Rohlmanet al.,51,52 we observed no changes in Cx43 labelling in theventral horn after sciatic nerve transection. These differingresults may be due to differences in the properties of astro-cytes in the facial nucleus compared with those in spinal cordmotor nuclei. Alternatively, the Cx43 epitope recognized bythe antibody in earlier reports51,52 may be subtly differentfrom that of 13-8300, and may indicate additional complexityin molecular modifications that occur in Cx43 followingvarious manipulations. Sciatic nerve transection did, how-ever, elicit a pattern of superficial dorsal horn 13-8300 label-ling similar to that seen after skin application of capsaicin,suggesting that Cx43 dephosphorylation following nervetransection is due to activation of C-fibres. This is consistentwith findings that sciatic nerve transection induces prolongedunmyelinated fibre discharge comparable in duration andfrequency to that evoked in C-fibres by capsaicin,2,65 andsuggests that the Cx43 dephosphorylation response is unlikelyto be due to Wallerian degenerative changes occurring in thedorsal horn shortly after nerve transection.

Regulation of astrocyte coupling

Relationships between Cx43 phosphorylation state andGJIC have been examined in a few studies, but have yet tobe fully clarified. In cell lines and peripheral tissues, bothincreases and decreases in GJIC have been seen followingCx43 dephosphorylation.13,35,48 In cultured astrocytessubjected to hypoxia, Cx43 dephosphorylation occurredafter reduced dye coupling.6 We have found that hypoxia-induced Cx43 dephosphorylation is accompanied by reduced


GJIC in cultured astrocytes and that both of these events canbe partially reversed by inhibitors of protein phosphatase (LiW. E. I. and Nagy, J. I., unpublished observations). Thus, itmay be tentatively concluded that Cx43 dephosphorylation inastrocytesin vitro and perhapsin vivo results in reducedGJIC. Astrocytes are thought to be essential for maintainingneuronal excitability by accumulating extracellular K145,49,66

and possibly by providing neurons with an energy substrate inthe form of lactate.61 These processes are probably facilitatedby gap junction-mediated intercellular flow of ions and meta-bolites between astrocytes distributed among active andinactive neurons. What purpose could then be served by thecounterintuitive process of channel closure at junctionscomposed of Cx43 within the astrocytic syncytium if suchclosure indeed occurs under physiological conditions? Asdiscussed in detail elsewhere,39 this may be related to thedynamic creation of open and closed junctional pathwaysfor routing substances between astrocytes according toneuronal needs. In addition, transiently reduced coupling

may be important for maintenance of normal metabolism,since it has been observed11,58 that glucose uptake and glyco-lysis within astrocytes are linked to their state of gap junc-tional coupling such that the former is increased whenastrocytic GJIC is reduced. Further studies of these issueswill need to take into consideration the additional presenseof Cx30 at astrocytic junctions41,42 and the regulatoryresponses of this connexin to neuronal activity. Although noimmunohistochemical alterations in the distribution of Cx30were noted in the dorsal horn after sciatic nerve stimulation(not shown), this connexin does contain several kinaseconsensus sequences that, if confirmed, may be subject tomodification following nerve stimulation.

Acknowledgements—We thank A. Ochalski for excellent technicalhelp, and Drs D. McCrea and R. Brownstone for allowing the use ofinstruments for nerve stimulation. This study was supported by grantsfrom the Medical Research Council of Canada (MRC) and theManitoba Medical Service Foundation (MMSF) to J.I.N.

REFERENCES

1. Beyer E. C., Paul D. L. and Goodenough D. A. (1987) Connexin43: a protein from rat heart homologous to a gap junction protein from liver.J. Cell Biol.105,2621–2629.

2. Blenk K. H., Janig W., Michaelis M. and Vogel C. (1996) Prolonged injury discharge in unmyelinated nerve fibres following transection of the suralnerve in rats.Neurosci. Lett.215,185–188.

3. Bruzzone B., White T. W. and Paul D. L. (1996) Connections with connexins: the molecular basis of direct intercellular signalling.Eur. J. Biochem.238,1–27.

4. Campos de Carvalho A. C., Roy C., Hertzberg E. L., Tanowitz H. B., Kessler J. A., Weiss L. M., Wittner M., Dermietzel R., Gao Y. and Spray D. C.(1998) Gap junction disappearance in astrocytes and leptomeningeal cells as a consequence of protozoan infection.Brain Res.790,304–314.

5. Chisamore B., Solc M. and Dow K. (1996) Excitatory amino acid regulation of astrocyte proteoglycans.Devl Brain Res.97, 22–28.6. Cotrina M. L., Kang J., Lin J. H. C., Bueno E., Hansen T. W., He L., Liu Y. and Nedergaard M. (1998) Astrocytic gap junctions remain open during

ischemic conditions.J. Neurosci.18, 2520–2537.7. Crow D. S., Beyer E. C., Paul D. L., Kobe S. S. and Lau A. F. (1990) Phosphorylation of connexin43 gap junction protein in uninfected and Rous sarcoma

virus-transformed mammalian fibroblasts.Molec. cell. Biol.10, 1754–1763.8. Cruciani V. and Mikalsen S. O. (1999) Stimulated phosphorylation of intracellular connexin43.Expl Cell Res.251,285–298.9. Enkvist M. O. and McCarthy K. D. (1992) Activation of protein kinase C blocks astroglial gap junction communication and inhibits the spread of calcium

waves.J. Neurochem.59, 519–526.10. Giaume C. and McCarthy K. D. (1996) Control of gap-junctional communication in astrocytic networks.Trends Neurosci.19, 319–325.11. Giaume C., Tabernero A. and Medina J. M. (1997) Metabolic trafficking through astrocytic gap junctions.Glia 21, 114–123.12. Giepmans B. N. and Moolenaar W. H. (1998) The gap junction protein connexin43 interacts with the second PDZ domain of the zona occludens-1

protein.Curr. Biol. 8, 931–934.13. Godwin A. J., Green L. M., Walsh M. P., McDonald J. R., Walsh D. A. and Fletcher W. H. (1993)In situ regulation of cell–cell communication by the

cAMP-dependent protein kinase and protein kinase C.Molec. cell. Biochem.127/128,293–307.14. Goldberg G. S. and Lau A. F. (1993) Dynamics of connexin43 phosphorylation in pp60v-src-transformed cells.Biochem. J.295,735–742.15. Goodenough D. A., Goliger J. A. and Paul D. L. (1996) Connexins, connexons, and intercellular communication.A. Rev. Biochem.65, 475–502.16. Hofer A., Saez J. C., Chang C. C., Trosko J. E., Spray D. C. and Dermietzel R. (1996) C-erbB2/neu transfection induces gap junctional communication

incompetence in glial cells.J. Neurosci.16, 4311–4321.17. Hosli E. and Hosli L. (1993) Receptors for neurotransmitters on astrocytes in the mammalian central nervous system.Prog. Neurobiol.40, 477–506.18. Hossain M. Z., Murphy L. J. and Nagy J. I. (1994) Phosphorylated forms of connexin43 predominate in rat brain: demonstration by rapid inactivationof

brain metabolism.J. Neurochem.62, 2394–2403.19. Hossain M. Z., Peeling J., Sutherland G. R., Hertzberg E. L. and Nagy J. I. (1994) Ischemia-induced cellular redistribution of the astrocytic gap junctional

protein connexin43 in rat brain.Brain Res.652,311–322.20. Kasuya Y., Abe Y., Hama H., Sakurai T., Asada S., Masaki T. and Goto K. (1994) Endothelin-1 activates mitogen-activated protein kinases through two

independent signalling pathways in rat astrocytes.Biochem. biophys. Res. Commun.204,1325–1333.21. Kenins P. (1982) Responses of single nerve fibres to capsaicin applied to the skin.Neurosci. Lett.29, 83–88.22. Kimelberg H. K. (1995) Receptors on astrocytes—what possible functions?Neurochem. Int.26, 27–40.23. Konietzko U. and Muller C. M. (1994) Astrocytic dye coupling in rat hippocampus: topography, developmental onset, and modulation by protein kinase

C. Hippocampus4, 297–306.24. Kriz N., Sykova E., Ujec E. and Vyklicky L. (1974) Changes of extracellular potassium concentration induced by neuronal activity in the spinal cord of

the cat.J. Physiol.238,1–15.25. Kumar N. M. and Gilula N. B. (1996) The gap junction communication channel.Cell 84, 381–388.26. Laird D. W., Puranam K. L. and Revel J.-P. (1991) Turnover and phosphorylation dynamics of connexin43 gap junction protein in cultured cardiac

myocytes.Biochem. J.273,67–72.27. Laird J. M. A. and Bennett G. J. (1993) An electrophysiological study of dorsal horn neurons in the spinal cord of rats with an experimental peripheral

neuropathy.J. Neurophysiol.69, 2072–2085.28. Lau A. F., Kurata W. E., Kanemitsu M. Y., Loo L. W. M., Warn-Cramer B. J., Eckhart W. and Lampe P. D. (1996) Regulation of connexin43 function by

activated tyrosine protein kinases.J. Bioenerg. Biomembr.28, 359–368.29. Lazarini F., Strosberg A. D., Couraud P. O. and Cazaubon S. M. (1996) Coupling of ETB endothelin receptor to mitogen-activated protein kinase

stimulation and DNA synthesis in primary cultures of rat astrocytes.J. Neurochem.66, 459–465.30. Li W. E. I., Ochalski P. A. Y., Hertzberg E. L. and Nagy J. I. (1998) Immunorecognition, ultrastructure and phorphorylation status of astrocytic gap

junctions and connexin43 in rat brain after cerebral focal ischaemia.Eur. J. Neurosci.10, 2444–2463.



31. Lynn B., Ye W. and Cotsell B. (1992) The actions of capsaicin applied topically to the skin of the rat on C-fibre afferents, antidromic vasodilatation andsubstance P levels.Br. J. Pharmac.107,400–406.

32. Marrero H. and Orkand R. K. (1996) Nerve impulses increase glial intercellular permeability.Glia 16, 285–289.33. Martinez A. D. and Saez J. C. (1999) Arachidonic acid-induced dye uncoupling in rat cortical astrocytes is mediated by arachidonic acid byproducts.

Brain Res.816,411–423.34. Mikalsen S.-O., Husoy T., Vikhamar G. and Sanner T. (1997) Induction of phosphotyrosine in the gap junction protein, connexin43.Fedn Eur. Biochem.

Socs Lett.401,271–275.35. Moreno A. P., Saez J. C., Fishman G. I. and Spray D. C. (1994) Human connexin43 gap junction channels: regulation of unitary conductances by

phosphorylation.Circulation Res.74, 1050–1057.36. Mugnaini E. (1986) Cell junctions of astrocytes, ependyma, and related cells in the mammalian central nervous system with emphasis on the hypothesis

of a generalized functional syncytium of supporting cells. In:Astrocytes, (eds Fedoroff S. and Vernadakis A.), Vol. 1. Academic, New York37. Musil L. S., Cunningham B. A., Edelman G. M. and Goodenough D. A. (1990) Differential phosphorylation of the gap junction protein connexin43 in

junctional communication-competent and -deficient cell lines.J. Cell Biol.111,2077–2088.38. Mussini J., Biral D., Marin O., Furlan S. and Salvatori S. (1999) Myotonic dystrophy protein kinase expressed in rat cardiac muscle is associated with

sarcoplasmic reticulum and gap junctions.J. Histochem. Cytochem.47, 383–392.39. Nagy J. I. and Dermietzel R. (2000) Gap junctions and connexins in the mammalian central nervous system. InAdvances in Molecular and Cell Biology

(ed. Hertzberg E. L.), Vol. 18. JAI, New York (in press).40. Nagy J. I., Li W. E. I., Roy C., Doble B. W., Gilchrist J. S., Kardami E. and Hertzberg E. L. (1997) Selective monoclonal antibody recognition and cellular

localization of an unphosphorylated form of connexin43.Expl Cell Res.236,127–136.41. Nagy J. I., Patel D., Ochalski P. A. Y., Li J. and Hertzberg E. L. (1997) Evidence for the co-localization of another connexin with connexin-43 at

astrocytic gap junctions in rat brain.Neuroscience78, 533–548.42. Nagy J. I., Patel D., Ochalski P. A. Y. and Stelmack G. L. (1999) Connexin30 in rodent, cat and human brain: selective expression in gray matter

astrocytes, co-localization with connexin43 at gap junctions and late developmental appearance.Neuroscience88, 447–468.43. Nagy J. I. and Rash J. E. (2000) Connexins and gap junctions of astrocytes and oligodendrocytes in the CNS.Brain Res. Rev.(in press).44. Nagy J. I., Yamamoto T., Sawchuk M. A., Nance D. M. and Hertzberg E. L. (1992) Quantitative immunohistochemical and biochemical correlates of

connexin43 localization in rat brain.Glia 5, 1–9.45. Newman E. A. (1985) Regulation of potassium levels by glial cells in the retina.Trends Neurosci.8, 156–159.46. Ochalski P. A. Y., Frankenstein U. N., Hertzberg E. L. and Nagy J. I. (1997) Connexin-43 in rat spinal cord: localization in astrocytes and identification of

heterotypic astro-oligodendrocytic gap junctions.Neuroscience76, 931–945.47. Ochalski P. A. Y., Sawchuk M. A., Hertzberg E. L. and Nagy J. I. (1995) Astrocytic gap junction removal, connexin43 redistribution, and epitope

masking at excitatory amino acid lesion sites in rat brain.Glia 14, 279–294.48. Oelze I., Kartenbeck J., Crusius K. and Alonso A. (1995) Human papillomavirus type 16 E5 protein affects cell–cell communication in an epithelialcell

line. J. Virol. 69, 4489–4494.49. Orkand R. K., Nicholls J. G. and Kuffler S. W. (1966) Effect of nerve impulses on the membrane potential of glial cells in the central nervous system of

amphibia.J. Neurophysiol.29, 788–806.50. Rash J. E., Duffy H. S., Dudek F. E., Bilhartz B. L., Whalen L. R. and Yasumura T. (1997) Grid-mapped freeze-fracture analysis of gap junctions in gray

and white matter of adult rat central nervous sytem, with evidence for a “panglial syncytium” that is not coupled to neurons.J. comp. Neurol.388,265–292.

51. Rohlmann A., Laskawi R., Hofer A., Dobo E., Dermietzel R. and Wolff J. R. (1993) Facial nerve lesions lead to increased immunostaining of theastrocytic gap junction protein (connexin 43) in the corresponding facial nucleus of rats.Neurosci. Lett.154,206–208.

52. Rohlmann A., Laskawi R., Hofer A., Dermietzel R. and Wolff J. R. (1994) Astrocytes as rapid sensors of peripheral axotomy in the facial nucleus of rats.NeuroReport5, 409–412.

53. Rose C. R. and Ransom B. R. (1997) Gap junctions equalize intracellular Na1 concentration in astrocytes.Glia 20, 299–307.54. Saez J. C., Berthoud V. M., Moreno A. P. and Spray D. C. (1993) Gap junctions: multiplicity of controls in differentiated and undifferentiated cells and

possible functional implications. In: (eds Shenolikar S. and Nairn A. C.), Vol. 27.Advances in Second Messenger and Phosphoprotein Research, Raven,New York.

55. Saez J. C., Nairn A. C., Czernik A. J., Fishman G. I., Spray D. C. and Hertzberg E. L. (1997) Phosphorylation of connexin43 and the regulation ofneonatal rat cardiac myocyte gap junctions.J. molec. cell. Cardiol.29, 2131–2145.

56. Sawchuk M. A., Hossain M. Z., Hertzberg E. L. and Nagy J. I. (1995)In situ transblot and immunocytochemical comparisons of astrocytic connexin43responses to NMDA and kainic acid in rat brain.Brain Res.683,153–157.

57. Szolcsanyi J. (1987) Selective responsiveness of polymodal nociceptors of the rabbit ear to capsaicin, bradykinin and ultraviolet radiation.J. Physiol.388,9–23.

58. Tabernero A., Giaume C. and Medino J. M. (1996) Endothelin-1 regulates glucose utilization in cultured rat astrocytes by controlling intercellularcommunication through gap junctions.Glia 16, 187–195.

59. Theriault E., Frankenstein U. N., Hertzberg E. L. and Nagy J. I. (1997) Connexin43 and astrocytic gap junctions in the rat spinal cord after acutecompression injury.J. comp. Neurol.382,199–214.

60. Toyofuku T., Yabuki M., Otsu K., Kuzuya T., Hori M. and Tada M. (1998) Direct association of the gap junction protein connexin-43 with ZO-1 incardiac myocytes.J. biol. Chem.273,12,725–12,731.

61. Tsacopoulos M. and Magistretti P. J. (1996) Metabolic coupling between glia and neurons.J. Neurosci.16, 877–885.62. Unger V. M., Kumar N. M., Gilula N. B. and Yeager M (1999) Electron cryo-crystallography of a recombinant cardiac gap junction channel.Novartis

Foundation Symp.219,22–37.63. Vukelic J. I., Yamamoto T., Hertzberg E. L. and Nagy J. I. (1991) Depletion of connexin43-immunoreactivity in astrocytes after kainic acid-induced

lesions in rat brain.Neurosci. Lett.130,120–124.64. Vyklicky L., Sykova E., Kriz N. and Ujec E. (1972) Post-stimulation changes of extracellular potassium concentration in the spinal cord of the rat. Brain

Res.45, 608–611.65. Wall P. D., Waxman S. and Basbaum A. (1974) Ongoing activity in peripheral nerve: injury discharge.Expl. Neurol.45, 576–589.66. Walz W. (1989) Role of glial cells in the regulation of the brain ion microenvironment.Prog. Neurobiol.33, 309–333.67. Wolburg H. and Rohlmann A. (1995) Structure–function relationships in gap junctions.Int. Rev. Cytol.157,315–373.68. Woolf C. J. and Fitzgerald M. (1983) The properties of neurones recorded in the superficial dorsal horwn of the rat spinal cord.J. comp. Neurol.221,

313–328.69. Woolf C. J. and King A. E. (1987) Physiology and morphology of multireceptive neurons with C-afferent fiber inputs in the deep dorsal horn of the rat

lumbar spinal cord.J. Neurophysiol.58, 460–479.70. Yamamoto T., Ochalski A., Hertzberg E. L. and Nagy J. I. (1990) LM and EM immunolocalization of the gap junctional protein connexin 43 in rat brain.

Brain Res.508,313–319.71. Yamamoto T., Ochalski A., Hertzberg E. L. and Nagy J. I. (1990) On the organization of astrocytic gap junctions in rat brain as suggested by LM and EM

immunohistochemistry of connexin43 expression.J. comp. Neurol.302,853–883.

(Accepted20 January2000)

activation of fibres in rat sciatic nerve alters phosphorylation state of connexin-43 at astrocytic...

Documents