Intraganglionic interactions between satellite cells and adult sensory neurons

Molecular and Cellular Neuroscience 67 (2015) 1–12

Contents lists available at ScienceDirect

Molecular and Cellular Neuroscience journal homepage: www.elsevier.com/locate/ymcne

Intraganglionic interactions between satellite cells and adult sensory neurons Kimberly Christie 1, Dilip Koshy 1, Chu Cheng 1, GuiFang Guo, Jose A. Martinez, Arul Duraikannu, Douglas W. Zochodne ⁎ Division of Neurology, University of Alberta

a r t i c l e

i n f o

Article history: Received 5 January 2015 Revised 12 April 2015 Accepted 11 May 2015 Available online 12 May 2015 Keywords: Dorsal root ganglia Perineuronal satellite cells Peripheral nerve Nerve injury Neuron-glia communication

a b s t r a c t Perineuronal satellite cells have an intimate anatomical relationship with sensory neurons that suggests close functional collaboration and mutual support. We examined several facets of this relationship in adult sensory dorsal root ganglia (DRG). Collaboration included the support of process outgrowth by clustering of satellite cells, induction of distal branching behavior by soma signaling, the capacity of satellite cells to respond to distal axon injury of its neighboring neurons, and evidence of direct neuron-satellite cell exchange. In vitro, closely adherent coharvested satellite cells routinely clustered around new outgrowing processes and groups of satellite cells attracted neurite processes. Similar clustering was encountered in the pseudounipolar processes of intact sensory neurons within intact DRG in vivo. While short term exposure of distal growth cones of unselected adult sensory neurons to transient gradients of a PTEN inhibitor had negligible impacts on their behavior, exposure of the soma induced early and substantial growth of their distant neurites and branches, an example of local soma signaling. In turn, satellite cells sensed when distal neuronal axons were injured by enlarging and proliferating. We also observed that satellite cells were capable of internalizing and expressing a neuron fluorochrome label, diamidino yellow, applied remotely to distal injured axons of the neuron and retrogradely transported to dorsal root ganglia sensory neurons. The findings illustrate a robust interaction between intranganglionic neurons and glial cells that involve two way signals, features that may be critical for both regenerative responses and ongoing maintenance. © 2015 Published by Elsevier Inc.

1. Introduction The response of sensory neurons to injury involves local changes at the level of the injured axon but also a series of retrograde alterations in parent neuronal perikarya. It has long been held that local satellite cells surrounding neurons within parent ganglia, contribute to the survival and plasticity of the sensory neuron, especially after injury. Perineuronal satellite cells surround, support and possibly protect neurons in sensory ganglia. Features include an intimate association with neurons, scant cytoplasm, a high surface-volume laminar structure and a basement membrane (Pannese, 1981; Pannese et al., 1999). They are thought to secrete growth factors, and may scavenge free radicals (Calcutt et al., 1992; Pannese, 1981; Powell et al., 1991; Zhou et al., 1999). Satellite cells also express p75, the low affinity NGF receptor (Zhou et al., 1996). There is literature to suggest that perineuronal satellite cells have dynamic properties and alter their behavior in

⁎ Corresponding author at: Division of Neurology, 7-132A Clinical Sciences Building, 11350-83 Ave, Edmonton, Alberta T6G 2B7, Canada. E-mail address: [email protected] (D.W. Zochodne). 1 Previous affiliation: Hotchkiss Brain Institute and Department of Clinical Neurosciences, University of Calgary, Calgary, Alberta, Canada.

http://dx.doi.org/10.1016/j.mcn.2015.05.001 1044-7431/© 2015 Published by Elsevier Inc.

response to neuronal changes after injury (Fenzi et al., 2001; Hanani et al., 2002). Satellite cells somehow sense neuron injuries involving axon processes far remote from the ganglia. Thus, a two way interchange between this unique cell population with neurons must exist, but its mechanism and importance are not well understood. No evidence that molecules can be interchanged and internalized between perineuronal satellite cells and primary sensory neurons has been reported. While the analysis of the behavior of adult sensory neurons in culture is crucial in understanding their responses to injury, most investigations have supposed that harvested neurons are studied in isolation. While many laboratories add steps to remove glial cells from harvested sensory neurons, few routinely stain or analyze what complement of cells remain. Their presence offers an opportunity to examine peripheral neuron-glial interaction. In this work, we show that adult sensory neurons in vitro cultures capture a population of adherent satellite cells. We identify relationships between clusters of these retained cells and neurite outgrowth. In vivo we show that perineuronal satellite cells sense injury to distal axons and respond by enlarging and by proliferating. Finally we show that sensory neurons can transfer small molecules to their surrounding glial partners. Overall these investigations provide evidence that there is ongoing and bilateral interaction between sensory neurons and perineuronal satellite cells.

2

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

2. Methods 2.1. Animals, surgery Adult male Sprague–Dawley rats weighing 200–300 g were used for this protocol. To address cell turnover in adult rats after injury we studied ipsilateral L4 and L5 DRGs following sciatic nerve transection (left midsciatic transection and resection of the distal nerve and its branches to prevent regeneration) or sciatic nerve crush (compression of the nerve for 30 s between the jaws of a plastic coated forceps). All interventions were carried out with pentobarbital anesthesia (65 mg/kg ip). Rats were sacrificed at endpoint using an overdose of pentobarbital. The protocol was reviewed and approved by the University of Calgary Animal Care Committee in conjunction with the guidelines of the Canadian Council of Animal Care (CCAC). 2.2. In vitro analysis of adult sensory neurons Prior to tissue harvesting, rats were anesthetized using isoflurane (Halocarbon Laboratories, River Edge, NJ) then sacrificed. DRG neurons were dissociated and maintained in vitro using a modification from the method of Lindsay (Andersen et al., 2000; Lindsay, 1988). Briefly, L4 and L5 DRGs were dissected from rats and placed into L15 (Invitrogen, Burlington, Ontario) medium where the axon roots and dural tissue were manually removed. The DRGs were rinsed three times in L15 medium and then transferred to a tube containing 2 ml 0.1% collagenase (Invitrogen,Burlington, Ontario)/L15. Following a 90 min incubation at 37 °C, the DRGs were placed into single-cell suspension by triturating 10–15 times every 5 min through three 1 ml pipette tips and then three 200 μl pipette tips. The single-cell suspension was spun for 5 min at 800 rpm at 4–8 °C and the cell pellet was washed three times in 2 ml L15. After the final 5 min 800 rpm spin, the cells were resuspended in L15 and passed through a 70 μm mesh (VWR International Co., Mississauga, Ontario). In an additional procedure to partially remove glial cells in some of the experiments, the cell suspension was loaded onto 15% BSA (Sigma, St. Louis, MO) in L15 and spun at 900 rpm for 10 min and the pellet collected (passaged). The cells were then washed once more with L15 and spun again then placed into a culture medium of Dulbecco's Modified Eagle Medium/F12 (DMEM/F12; Invitrogen, Burlington, Ontario) + 1:100 dilution N2 (Invitrogen, Burlington, Ontario), 0.5–0.8% BSA (Sigma Aldrich, Oakville, Ontario) and 0.2 ng/ml NGF (Cedarlane Laboratories Ltd, Horby, Ontario) plus 50U Penicillin (1 ml), 50 μg Streptomycin (1 ml) (Invitrogen, Burlington, Ontario) and plated onto poly-L-lysine (Sigma Aldrich, Oakville, Ontario) and 10 μg/ml mouse laminin (Invitrogen, Burlington, Ontario) coated plates. Half of the cell medium was changed every 2 days. The cultures involved 6 rats with axotomy and 6 sham injured rats (2 each per culture day) for nonpassaged analysis and similar animal and culture numbers for passaged neurons. 2.3. Soma signaling experiments The procedures were modifications of a growth cone turning assay modified for adult mammalian DRG neurons from previous embryonic neuron cultures in Xenopus laevis (Lohof et al., 1992; Nishiyama et al., 2003; Zheng et al., 1994) as reported in our laboratory and others (Murray et al., 2012; Webber et al., 2008a) (Guo et al., 2014). Two DRG samples were harvested from 1 rat to generate 6 plates. Briefly, gradients were created and maintained using a Picospritzer II (Parker Hannafin, Fairfield, NJ) ejecting at a pressure of 3 psi, frequency of 2Hz, for 20 milliseconds using a pulse generator (SD9; Grass Instrument Co., Quincy, MA) from a micropipette with a 0.5- to 1 μm opening. The micropipette was positioned 100 μm from soma of the neuron or the growth cone center at an angle of 45° with respect to the last 10 μm segment of the axon shaft. The turning assay was performed on preconditionally lesioned DRG cultures at 1 to 3 DIV. For analysis, the

trajectories of the outgrowing neurites were traced onto a graph and we calculated the percentage of total growth cones examined that exhibited a branch (% of growth cones with branches), and the ratio of primary or secondary branch numbers to total growth cones examined (ratio of branches to growth cones [as a %]). Primary branch were classified as having emerged from the main advancing growth cone process and secondary branches as arising from a primary branch. All analyses were carried out 60 min following onset of the continuous picospritzer application. 2.4. Immunocytochemistry and immunohistochemistry For in vitro studies, cell cultures were fixed in 2% paraformaldehyde for 10 min, washed with PBS, blocked for 30 min in PBS with 5% fetal bovine serum and 0.1% BSA then incubated with primary and secondary antibodies (Table 1). Tissue samples were fixed in modified Zamboni's fixative (2% paraformaldehyde, 0.5% picric acid and 0.1% phosphate buffer) overnight at 4 °C. Tissues were then washed in PBS 5 times, cryoprotected in 20% sucrose/PBS and left at 4 °C overnight. After embedding in optimum cutting temperature (OCT) compound (Miles), 20 μm thick sections were placed onto poly-L-lysine coated slides. For indirect immunofluorescence, slides were incubated for 48 h at 4 °C with primary antibodies. Slides were then washed with PBS and incubated with secondary antibodies for 1 h at room temperature. After further PBS washing, cover slips were mounted with bicarbonate-buffered glycerol (PH 8.6) or Vectashield Mounting Medium with DAPI (Vector Laboratories,Burlington, Canada) and slides were viewed with a fluorescent microscope (Zeiss, Axioskope, Zeiss Canada, Toronto, Canada). Negative controls included omission of primary antibodies or secondary antibodies on parallel sections. To examine retrograde neuron labeling, we administered DY (diamidino yellow, Sigma), a fluorochrome retrograde tracer, by transecting the sural nerve distal to a crush zone through a separate incision and immersing the stump into a capsule filled with 1.5 μl of 2.5% DY for 30 min, then gently rinsing with saline. Seven days after crush the L5 DRGs were fixed in Zamboni's fixative and sections made at 14 μm for additional immunohistochemistry. In separate experiments, we administered DY to the site of a sciatic transection, then Table 1 Antibodies and Primers. (i) Primary antibodies used anti-β-Tubulin-Ш, (1:100), mouse monoclonal, Sigma, St. Louis, Mo anti-NF200 (neurofilament 200, 1:800), mouse monoclonal, Sigma, St. Louis, Mo anti-S100 (1:200), mouse monoclonal S100 beta chain, Santa Cruz, Santa Cruz, Ca anti-BrdU (1:50), monoclonal, Cedarlane, Hornby, Ont anti-BrdU (1: 200), monoclonal, Sigma, St. Louis, Mo, anti-NeuN (1: 100), monoclonal, Chemicon International, Temecula, Ca anti-GFAP (1:250), polyclonal rabbit DakoCytomation, Carpinteria, CA (ii) Secondary antibodies used anti-mouse IgG CY3 conjugate (1:100) sheep, Sigma, St. Louis, Mo anti-goat (1:200) Alexa Fluor 488 donkey, Invitrogen Canada, Burlington, Ont anti-rabbit IgG (H + L) conjugate (1:400) Alexa Fluor 488 goat Cedarlane, Burlington, Ont anti-mouse (1:400) Cy3 sheep, Alexis Biochemicals, San Diego, Ca (iii) Primers used CNTF R F 5′- TTCTGCCTTTGCCTACCAGCT-3′ CNTF R R 5′-AGACCACCATCTCCAACTGTGG-3′ HGF RM F 5′-AACACAGCTTTTTGCCTTCGAG-3′ HGF RM R 5′-CTGGATTGCTTGTGAAACACCA-3′ Protrudin R F 5′- TGATGAGGCGGAAGTACCACA-3′ Protrudin R R 5′- GGCTTCCAGTCGGATCAAGAA-3′ ErbB2 R F 5′- ACATCTCAGCATGGCCAGACA-3′ ErbB2 R R 5′- TGTCAATGAGTACGCGCCATC-3′ Connexin 43 (GJA1 R) F 5′- AGCACGGCAAGGTGAAAATG-3′ Connexin 43 (GJA1 R) R 5′- TACCACTGGATGAGCAGGAAGG-3′ Connexin 32 (GJB1 R) F 5′-TTTTTCCCCATCTCCCATGTG-3′ Connexin 32 (GJB1 R) R 5′-ATGTGTTGTTGGTGAGCCACG-3′ RPLP0 F 5′- TACCTGCTCAGAACACCGGTCT-3′ RPLP0 R 5′- GCACATCGCTCAGGATTTCAA-3′

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

harvested L4/5 sensory neurons 7 days later in vitro. Neurons were cultured for approximately 18 h. 2.5. BrdU administration, analysis 5-Bromo-2′-deoxyuridine (BrdU) (Sigma, St. Louis, MO) was dissolved in 1 mM Tris, 0.8% NaCl, 0.25 mM EDTA solution. For peripheral nerve injury studies and their control group, an initial dose of 100 mg BrdU/kg at the day of sciatic nerve injury, followed by a second, third and fourth BrdU injection each day (50 mgBrdU/kg) after injury were administered (ip). For detection of BrdU, sections were rinsed in PBS for 15 min in three changes, and then hydrolyzed in 2 N HCl in PBS for 30 min. After further PBS washing, the sections were digested with 0.01% Trypsin for 3 min at 37 °C, rinsed with distilled water for 1 min, washed with PBS for 10 min in 2 changes and then incubated with 1% BSA, 0.3% Triton X-100 in PBS for 30 min. Sections were then treated with the primary monoclonal mouse anti-BrdU antibody or monoclonal rat BrdU as described above. DAPI (4′, 6-diamidino-2-phenylindole), a marker of nuclei, was used to counterstain DNA. After immunohistochemical labeling for BrdU, Vectashield Mounting Medium with DAPI (Vector Laboratories,Burlington, Canada) was used for mounting slides. To assess numbers of BrdU labeled nuclear profiles in DRG, six transverse sections representing each rat DRG (n = 3 rats for each group) were used to count profiles by fluorescence microscopy and a mean calculated for each rat. The area of each DRG section was measured to yield BrdU profiles per mm2. 2.6. Satellite cells and neurite outgrowth For in vitro analysis of the association of neurite outgrowth with satellite cells we labeled neurons with antibodies to either Nf200, the heavy subunit of neurofilament or βIII tubulin and colabelled with DAPI to stain nuclei. Satellite cell nuclei were easily distinguished from those of neurons by their location, smaller size and more intense DAPI staining. In contrast, neuronal nuclei, identified in the center of clusters of satellite cell nuclei were larger, oval and less intensely stained. This imaging distinction was confirmed by colabelling. Merged images of cells with a neuronal label and DAPI stain were created. Images were overlaid with an overlying grid of squares (each square side measured approximately 13 μm). For each satellite cell nucleus we counted the number (and percentage) that were found within a one grid square radius of a neuron process (neurite), the number (and percentage) of neurites that had a satellite cell nucleus within a one grid square distance, the number of neurites and the number of satellite cells associated with each neuron. Neurons were studied from intact rats, rats with a pre-existing sciatic nerve section (axotomy) 3 days earlier to precondition them, and neurons with sham exposure of the sciatic nerve but without frank injury. 2.7. Western immunoblots DRG cells were separated by centrifuging through a density gradient. Cells were added to a tube containing 15% BSA in L15 without mixing the phases then spun at 900 rpm for 10 min to provide glial-enriched (top fraction) and neuronal-enriched (bottom fraction; not used in this work) fractions. Protein samples were separated from glial and neuronal fractions, loaded for SDS-PAGE then incubated with primary antibodies diluted as per the instructions from the manufacturer and incubated overnight at 4 °C. The blots were then washed thrice with high salt followed by low salt buffer, incubated with 1:1000 dilution of horseradish peroxidase conjugated secondary antibodies for 1 h at room temperature. Proteins were visualized using enhanced chemiluminescence (ECL) detection reagents. Quantification of bands was through Adobe Photoshop and the band densities were normalized with those of the loading control. Antibodies used were Nf200

3

(1:2000; Mouse Monoclonal; Sigma-Aldrich, St. Louis, Mo), S100 (1:1000; Rabbit Monoclonal; Cell Signaling Tech,Beverly, Ma), alpha tubulin (1:4000; Mouse Monoclonal; Sigma-Aldrich, St. Louis, Mo), Actin (1:2000; Rabbit Polyclonal; Santa Cruz, Dallas,Tx). 2.8. qRT-PCR Total RNA was extracted using Trizol reagent according to the manufacturer's protocol (Invitrogen, Inc., Burlington, Ontario, Canada). Approximately 1 μg of DNAse 1-treated RNA was used to synthesize first strand DNA utilizing SuperScript II First-strand Synthesis Kit (Invitrogen, Inc). Random hexamers (50 mg) were utilized as per the manufacturer's protocol. First strand DNA was then used for PCR reactions. Real time quantitative PCR was performed on the ABI prism 7000 sequence detection system (Applied Biosystems, Foster City, CA). Each PCR mixture contained 1U Taq DNA polymerase mixture (Invitrogen, Carlsbad, CA), 3 mM MgCl2, 2.5 μM concentrations of primers of interest (synthesized by the University of Calgary DNA Lab), and 4 μL of cDNA in a total volume of 30 μL. The PCR was run according to the following conditions: 95 °C for 10 min, followed by 40 cycles of 95 °C for 30 s and 60 °C for 1 min. The purity of each amplicon was determined using melting curve analysis. All samples and standards were run in duplicate. The primer sequences used were designed with Primer Express 2.0 (Applied Biosystems, Foster City, CA) and are listed in Table 1. Quantification of amplified products was done on a cycle-by-cycle basis using the fluorescence signal generated by binding of SybrGreen I fluorophore (Invitrogen, Carlsbad, CA) to double-stranded DNA. The cycle number at which the fluorescence signal crossed a fixed threshold (threshold cycle, CT) with an exponential growth of PCR product during the linear phase was recorded. Relative expression values were generated using the comparative CT method (2−ΔΔCT) where all genes of interest were standardized to expression of RPLO (ribosomal protein, large, P0). To prepare neuron depleted ganglion samples, harvested material including sensory neurons were placed on top of a 15% BSA in L15 gradient and spun for 8 min at 600 rpm. The neurons pellet and the neuron depleted material that remained on top of the gradient were placed in Trizol and its RNA extracted. 2.9. Analysis Results were calculated as means ± sem and comparisons made among groups using unpaired Student's t-tests. 3. Results 3.1. Perineuronal satellite cells are intimate and retained partners of harvested sensory neurons Routine harvesting of cultured primary sensory neurons usually involves steps to eliminate or reduce the number of associated glial cells. These steps include the passage through mesh filters, as routinely done in this work. Direct staining for residual satellite cells is usually not done. We observed that most harvested adult sensory neurons carried with them a coterie of adherent satellite cells not removed during the harvesting procedure [Fig. 1]. Neurons without satellite cells were uncommon. Satellite cell nuclei were easily distinguished from those of primary neurons by their perineuronal location, smaller nuclei and more intense DAPI labeling. Satellite cells frequently indented the plasmalemmae of the neurons. We also labeled cells with a glial marker, S100 to confirm their identity (Levy et al., 2007). This cytoplasmic labeling also indicated that independent glial processes emerged from the neuron-satellite cell partnership. Additional findings were that satellite cells were adherent to their original neuronal partners in the ganglia, confirmed by harvesting primary sensory neurons administered DY (below). Passage of neurons through with a BSA gradient

4

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

Fig. 1. Satellite cells are associated with neuronal processes. Primary sensory neurons studied in vitro (A–I) are associated with asymmetric distributions of glial cells. Neurons are labeled with an antibody directed against neurofilament (red) and glial cell nuclei are identified with a DAPI (blue) label. Neuronal nuclei are distinguished from those of satellite cells by their location within neurons, their larger size and less intense DAPI staining. Note that clusters of satellite cell nuclei are associated with the side of the neuron with neurite outgrowth either as an early process (A–C) or as a later, more complex process (G–I). In D–F, note that the process from the neuron is directed upward toward a cluster of satellite cells. In J–L, a primary sensory neuron from an intact in vivo L4/5 ganglion labeled with neurofilament is illustrated showing a very similar pattern of satellite cells clustered around a pseudounipolar neuron process (arrows). [Bar = 50 μm (A–I) and 20 μm (J–L)]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

was associated with smaller numbers of satellite cells in association each neuron but did not eliminate adherent satellite cells (3.4 ± 0.2/ neuron in BSA passaged compared to 10.6 ± 0.8/neuron without BSA passage; uninjured neurons). Taken together, these results indicate that cultures of primary sensory neurons from adult rodents routinely include a complement of intimate satellite cells that retain their relationship previously established in vivo. 3.2. Clusters of perineuronal satellite cells are associated with neurite outgrowth from adult sensory neurons Over the first 2 d of culture, adult sensory neurons send out one or several neuritic processes, the precursors of mature axons. We analyzed

whether satellite cells were centered (using a nuclear stain) about areas of outgrowth, or were randomly distributed around the circumference of the primary neuron. A very high proportion of satellite cell nuclei (75–95%) were found within a one square grid distance of an adjacent neurite process. Alternatively, a very high proportion of neurites (80– 95%) had at least one satellite cell nucleus within a one square grid distance. These findings were consistent and essentially identical whether or not the neurons had been passaged through BSA and had fewer overall satellite cells as discussed above (data not shown). The results from nonpassaged neurons, in culture for 3 days, with or without axotomy are shown in Fig. 2. There was also a relationship between the number of satellite cells found in association with a sensory neuron and the number of processes it possessed. Preconditioning involves the transection of the nerve at mid-thigh level 3 days before culturing the

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

DRG perikarya, a procedure that accelerates both neurite outgrowth and the number of emerging neurites in harvested neurons. These neurons had larger numbers of neurite processes for each neuron together with larger numbers of associated satellite cells. Taken together, the results indicated that initial neurite outgrowth from adult sensory neurons was closely aligned with clusters of satellite cells. Since neurites can grow from any position around the

5

circumference of the neuron, satellite cell clustering might either facilitate or follow neurite outgrowth. However, in labeling studies described below, there was little or no evidence that dissociated satellite cells freely associated with neurons or their processes. Instead they appeared adherent, largely fixed in position with their original neuron partners (see below). In some instances unassociated clusters of satellite cells appeared to attract long neurites toward them. Finally,

Fig. 2. A large majority of satellite cells are associated with adjacent neurite processes. Data illustrating numbers and percentages of satellite cells found adjacent to primary sensory neurites in vitro and numbers and percentages of neurites found to be associated with satellite cells. A substantially larger percentage of satellite cells were adjacent to a neurite than not adjacent (A) [*p b 0.001; n = 41; Student's t-test]. Similarly a larger percentage of neuritis were associated with satellite cells than without (B) [*p b 0.001; n = 41; Student's ttest]. This association was present in neurons with or without a pre-existing sciatic axotomy injury lesion. On average, neurons (unpassaged) had approximately 10–12 satellite cells routinely associated with them irrespective of whether they had been injured before (C). Axotomy itself resulted in an expected rise in the number of outgrowing neurons through preconditioning (D; [*p b 0.001; n = 41; Student's t-test]) and in the numbers of neurites/neuron associated with a satellite cell (E) [*p b 0.001; n = 41; Student's t-test].

6

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

in vivo, we identified similar clustering of satellite cell nuclei around the pseudounipolar initial axon process that emerged from sensory neurons. This suggested that clustering persists in stable adult ganglia and may be a requirement for maintenance. 3.3. Perineuronal satellite cells sense distal axonal injuries and respond by proliferating Given the close anatomical relationship between satellite cells and neurons, we tested whether satellite cells might be altered by a distant axon injury of neurons, providing evidence of satellite cell-neuron communication. If identified, it would imply a functional connection subserved by their close relationship and exchange of an injury signal. To determine if satellite cells proliferate following axonal nerve injury, we counted the number of mitotically active cells in the ipilateral ganglion following either sciatic nerve transection or crush injury. Following either injury, there was a striking increase of BrdU uptake in a circumferential pattern around primary ganglion sensory neurons at day 5 after injury. Uninjured ganglia had baseline levels of BrdU uptake, indicating ongoing satellite cell turnover. The rises in BrdU uptake after transection were larger than those following crush. Satellite cells also exhibited hypertrophy, as demonstrated by rises in the perineuronal expression of cytoplasmic GFAP, not normally observed in circumferential processes of the satellite cell [Fig. 3]. 3.4. Soma directed signals alter the behavior of distal neurites We have previously shown that immunoselected GFRα1 neurons with high PTEN (phosphatase and tensin homolog deleted on chromosome 10) expression exhibit distal neurite branching in response to inhibition of PTEN at the neuronal soma (Guo et al., 2014). In an identical setup, we examined whether this strategy might apply to unselected sensory neurons generally irrespective of their phenotype. For this experiment we directed a pipette to picospritz bpV(pic) toward the neuron soma and compared applied doses of a PTEN inhibitor with carrier. Outgrowing neurites were exposed to the gradient for 60 min prior to analysis. As observed in GFRα1 neurons after bpV(pic) application there was enhanced and dose dependent outgrowth of distal primary and secondary branches of neurites [Fig. 4]. This was not observed when bpV(pic) was applied directly to growth cones (data not shown). In some instances there was neurite branching polar opposite to the side of the soma to which picospritzer application was provided. While not observed after exposure to PBS carrier at the cell body, some neurons exposed to bpV(pic) developed new tertiary branches as well. A neuron exposed to 50 nM of bpV(pic) developed 4 new neurites from the cell body, not seen after carrier or higher doses of bp(V)pic treatment. 3.5. Molecular transfer between neurons and satellite cells The apparent encouragement of neurite outgrowth by satellite cells and the capability of satellite cells to sense a remote axon injury of the neuron implied that a molecular interaction between the two cells must take place. To address this possibility we examined retrograde uptake and transfer of the fluorochrome DY. In L4/5 ganglia ipsilateral to axotomy, DY was taken up by a proportion of sensory neurons and labeled both the nucleus and cytoplasm of neurons. Since the fluorochrome was applied at a transection site distal to the crush zone, outgrowing axons were required to grow beyond the crush to access DY. Thus uptake was by outgrowing axons. The uptake was discrete, not diffuse and did not nonspecifically permeate the tissue. Around each neuron however, there was also distinct uptake of DY in satellite cells surrounding labeled neurons. The uptake pattern was that of rings of labeled satellite nuclei, within the milieu of the otherwise intact ganglion. Other closely adjacent neurons and these satellite cells were not labeled despite their close proximity indicating that the labeling of

satellite cells was only by interaction with the closely apposed (and DY labeled) DRG cell body. Harvested neurons examined in vitro that had DY within their cytoplasm and nucleus were similarly surrounded by clusters of satellite cells with DY nuclear labeling. DY did not label satellite cells from other unlabelled neurons. Results are illustrated in Fig. 5. Taken together these findings indicate that DY was taken up by axons severed peripherally, retrogradely transported to ganglia then transferred from neurons to their surrounding satellite cells. In vitro, the finding that labeled satellite cells retained their relationship to labeled neurons indicated that these cells otherwise routinely adhere and retain their in vivo relationship after harvesting. 3.6. Ganglionic extracts express selected glial-axon signals Ganglia are a complex cellular milieu and it is challenging to determine whether specific signals might be generated by satellite glial cells alone. To partly isolate non-neuronal ganglion samples we separated neuronal and noneuronal samples by a BSA gradient. mRNAs were compared among uninjured DRGs, or DRGs with a sciatic nerve transection 3d or 7d earlier. Our hypothesis was that selected glial-neuron signals or connexin channels might be expressed in this milieu and potentially be upregulated after injury. We confirmed that the glial cell fraction was enriched in S100, a marker of satellite and Schwann cells and depleted in Nf200, a marker of neurons. The glial enriched samples expressed several potential mediators including neuregulin receptor ErbB2, connexin 43 and 32, protrudin, CNTF, and HGF [Fig. 6]. However, after injury, levels of ErbB2, connexin 43 and 32 all trended toward lower levels. CNTF and protrudin levels were unchanged whereas HGF trended toward higher levels after injury. Overall, the findings did identify several potential signals that might be elaborated by satellite glial cells to support facilitated outgrowth after injury. 4. Discussion The results presented highlight, confirm and extend longer standing suppositions about how ganglia neurons and their glial partners interact. This relationship is very intimate, is retained during harvesting and culture in vitro and appears to involve two way interactions. Since most neurons identified had satellite cells associated with them, it is possible that neurons stripped of all their satellite cells during harvesting are less likely to survive. The distribution of satellite cells indicated that they clustered around sites of neurite outgrowth in vitro and around the pseudounipolar process in vivo. Since the timepoints studied were early after harvesting it is unlikely that satellite cells divided, migrated and secondarily moved to these sites, but instead suggest that they facilitate neurite outgrowth from existing sites of adherence. Along these lines, long neurites were also attracted to ‘free’ clusters of satellite cells not attached to neurons. We also did not observe mitotic processes surrounding neurons at the early timepoints we analyzed in these in vitro preparations. We cannot exclude the possibility that sites of the neuron where adherence is more robust have independent properties that also facilitate subsequent outgrowth. Our findings extend interesting ultrastructural observations by Pannese and colleagues (Pannese et al., 1999) who showed that neuronal projections identified in vivo from rabbit ganglionic neurons were associated with the portion of the neuron adjacent to the satellite cell, and not to the extracellular matrix. The authors postulated that perineuronal satellite cells have the capability to promote or facilitate these microvilli (Pannese, 2002). Such structures may be sources of molecular exchange. Since neurons are fully surrounded by basement membrane that express laminin and fibronectin, neither constituent is likely to account for microvilli formation. We believe it highly likely that such projections are analogous to the early outgrowing neurites that we observed in vitro.

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

7

Fig. 3. Satellite cells form clusters, are labeled with S100 and proliferate and hypertrophy ipsilateral to an injury. In A–F, clusters of satellite cells ipsilateral to injury are labeled with S100. In A–C, a cluster is clearly associated with a neuron, identified with its larger less intense central nucleus (arrow, B). Note that the satellite cells form in a crescent shape around the neuron. In D–F, it is more difficult to identify a neuron associated with this dense cluster of satellite cells. One satellite cell (F, arrow) has elongated processes. In G, BrdU labeled proliferating perineuronal satellite cells are observed (green) surrounding NeuN labeled neurons (red). In H, faint staining of GFAP is identified diffusely within an intact L4 ganglion whereas GFAP labeling around neurons becomes intense and widespread ipsilateral to a sciatic nerve injury, indicating satellite cell hypertrophy (arrow, I). [Bar = 20 μm for A–F, 100 μm for G and 50 μm for H,I]. Quantitation of BrdU labeling is given in J. Note that rises in BrdU labeled cells in ipsilateral L4 ganglia are observed after both crush and transection, but is greater after more severe transection injuries (p = 0.002 ANOVA; *p b 0.05 vs. control). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

8

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

Fig. 4. Local soma directed PTEN inhibition generates distal neurite outgrowth. Bp(v)pic (BPV), a pharmacological inhibitor of PTEN was applied by picospritzer to the soma of adult sensory neurons as described in separate work (Guo et al., 2014). The gradient was continuously applied then analyzed after 60 min. Existing and new neurite branches are shown after PBS carrier (A), BPV at 50 nM (B), and BPV at 200 nM (C). The original configuration of the neurons is in blue. New primary branches are illustrated in pink, new secondary branches in black and tertiary branches in green. There was a rise in the number primary neurites/neuron (D) [p = 0.028 paired before vs. after treatment, n = 5 PBS, 6 50 nM, 6 200 nM; one-tailed Student's t-test] and secondary neurites/neuron (E) [p = 0.05 paired before vs. after treatment, n = 5 PBS, 6 50 nM, 6 200 nM; one-tailed Student's t-test] after administration of 200 nM of BPV. An individual neuron is shown under brightfield before and after 60 min of growth in response to 200 nM BPV directed to the soma (F). Neurites emerge distant from the side of application of the PTEN inhibitor. Bar = 50 μm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Here we also confirm our recent observation in selected GFRα1 neurons that a gradient of PTEN inhibition applied to the soma has a more striking impact on distal growth cone behavior than that applied directly to a growth cone (Guo et al., 2014). Here we show very similar

behavior to a wider repertoire of unselected neurons. This is a response that we recorded over 60 min of neurite outgrowth, indicative of a short term but potent signaling event. We judged its significance important enough in the current context to again analyze the pattern of response,

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

9

Fig. 5. Neurons retrogradely labeled with DY transfer the fluorochrome to adjacent, adherent satellite cells. Panels A–H (G and H are magnified insets) show primary sensory neurons ipsilateral to a backlabelled sural axon labeled with DY. Triple labeling shows two neurons (labeled with antibody to the heavy subunit of neurofilament, red, and given as neurons 1 and 2), one of which is backlabelled with DY (green). Nuclei of neurons and satellite cells are labeled with DAPI (blue; neuronal nuclei are larger, less intensely labeled, and found within the neurofilament labeled neurons). Neuron #1 (A–F) is labeled with DY within its cytoplasm and nucleus and its adjacent satellite cell nuclei are similarly labeled with DY. Neuron#2 (A–F) is labeled with neurofilament and DAPI but it and its satellite cells do not label with DY. Other adjacent satellite cells (C, arrowhead) are also not labeled with DY. Images C (neurofilament and DAPI),E (neurofilament and DY), and F (neurofilament, DAPI and DY) are merged images. Higher magnification merged images G (neurofilament and DAPI) and H (neurofilament and DY) illustrate the differences in nuclear labeling between neurons and satellite cells (G) and to show that DY had neuronal nuclear, cytoplasmic and satellite cell nuclear localization of the fluorochrome. [Bar = 50 μm for A–F and 25 μm for G,H]. Images I–K are from L4/5 DRGs labeled with an antibody directed toward the heavy neurofilament subunit (red;I). In these images (J,K) DY is highlighted through a different filter and appears blue. Note that labeled satellite cell nuclei in blue are found in interesting rosette patterns circumferentially around neuron profiles. [Bar = 50 μm for I–K]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

similar to previously reported, but without preselecting the neuron population. The response appears to be a more generalized neuronal response than we had explored in the prior work. The signals that endogenously regulate local PTEN levels to facilitate growth are uncertain but the findings provide evidence that a local soma signal is capable of accomplishing it. This local soma signal may originate or be influenced by surrounding satellite cells due to their position and close contact with the neuronal soma. Potential signals between satellite cells and early neurites are unknown. Our search for mRNAs that might be expressed or upregulated

in ganglionic isolates depleted of neurons identified several important candidates to facilitate interactions. Of particular interest was the neuregulin receptor, an important signaling interaction that can signal Schwann cell proliferation and migration (Buonanno and Fischbach, 2001; Esper et al., 2006). The neuregulin co-receptor ErbB2 was expressed in our samples, and trended toward lower levels after injury. Protrudin is a molecule that promoted directional neurite outgrowth in cortical neurons, hippocampal neurons and PC12 cells (Shirane and Nakayama, 2006). Since NGF can induce adult sensory neuron outgrowth in vivo and steer injured growth cones, it might

10

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

Fig. 6. Relative mRNA expression in DRG lysates enriched in glial cells (A–F). Control samples are from intact, uninjured ganglia samples. Samples are ipsilateral to a sciatic nerve transection carried out 3d and 7d earlier. Note that connexins, protrudin and ErbB2 mRNAs trend toward declines after injury. HGF but not CNTF was increased [changes in HGF and Connexin 43 are borderline significant at p = 0.075 and p = 0.054 respectively; one way ANOVA; n = 3 for control and 3d glia, n = 3 for HGF, protrudin and n = 4 for the remaining markers in 7d glia]. Western immunoblots (G,H) illustrate that the glial fraction is depleted in NF200 as a marker for neurons, and enriched in S100, as a marker for satellite cells and Schwann glial cells [*p b 0.05; Student's t test; n = 3/group].

account for early neurite protrusion (Webber et al., 2008b). However NGF is known to arise from other targets, not ganglia, and distal injury that promotes satellite cell proliferation instead blocks the retrograde uptake of NGF. These findings suggest that NGF may be less important in changing satellite cell behavior following injury. Similarly BDNF is thought to arise in neurons within ganglia. CNTF, a potential messenger from glial cells to neurons, was also present but unchanged after injury. Levy et al. (Levy et al., 2007) identified FGF-2 in both satellite cells and neurons of DRG with rises in the satellite cells after injury. However, our investigation was limited. One or more of several other growth factors, not explored here, could also account for the changes we observed. Given the caveats above, in this work we focused on HGF, a growth factor we have recently linked to growth cone turning (Guo et al., 2014) and CNTF, reported to be expressed in Schwann cells following axonal contact (Lee et al., 1995).

The second major theme of this paper was that satellite cells in turn alter their behavior in response to remote axonal injury of the axons. Satellite cells develop hypertrophy and proliferate, strictly around their associated neurons, not spreading out to adjacent parts of the ganglia. We confirmed this finding by identifying glial cell hypertrophy, as described previously (Fenzi et al., 2001) and show that this is accompanied by rises in proliferation of the cells, as also suggested previously by Friede and Johnstone using thymidine labeling after sciatic transection (Friede and Johnstone, 1967) and by others (Humbertson et al., 1969; Shinder et al., 1999). Shinder et al. (1999) demonstrated a timeline for satellite cell proliferation that matched the findings in the present work. A similar impact on heightened GFAP expression by satellite cells surrounding neurons, and spreading to adjacent cells, in trigeminal ganglia innervating an area of tooth injury has been described (Stephenson and Byers, 1995). This relationship suggests

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

that either a local diffusible or directly transferred factor(s) released from reprogrammed, injured neurons ramps up baseline proliferative plasticity of the satellite cell population. These results are analogous to “nests of Nageotte”, clumps of proliferating satellite cells within ganglia that occupy “ghost” profiles left when neurons die. It is possible that intact neurons normally suppress satellite plasticity, and that derepressed cells proliferate when neurons are injured or disappear. Neuregulin, for example, may suppress SC proliferation in favor of myelination (Esper et al., 2006; Michailov et al., 2004). The trend toward lower ErbB2 levels could suggest that loss of neuregulin suppression of proliferation accounts for nesting. The findings from our DY experiments provide direct evidence that a molecular exchange from neuron to satellite cells can exist. We think it highly unlikely that the selective DY uptake into satellite nuclei, strictly around labeled neurons represents an artifact of DY diffusibility or leakage, processes that would prevent much retrograde signal from appearing in the first place. DY did not spread nonspecifically and no diffusion to other cell types, to adjacent neurons or their satellite cells was observed. Several possible routes of transfer of DY to satellite cells might exist. DY could be transferred directly from neurons to satellite cells through connexin channels, known to exist in DRGs and thought to allow direct connections among satellite cells. We did confirm presence of mRNAs for connexin 43 and 32, but no change after injury was detected. Connexins are known to be expressed by satellite cells (Procacci et al., 2008) but do not, to our knowledge, connect satellite cells and neurons. The possibility that there may be dye coupling, or a direct connection between neurons and satellite cells was explored by Hanani and colleagues (Hanani et al., 2002; Huang et al., 2005) using direct injections of satellite cells or neurons with Lucifer yellow. While there were direct connections among satellite cells, no coupling between satellite cells and neurons was detected. Satellite cell coupling through gap junctions increased after axotomy. These findings were also confirmed by Dublin and Hanani after axon injury; neurons were occasionally coupled to eachother, but not to satellite cells (Dublin and Hanani, 2007). Given these findings, a more plausible transfer mechanism might involve release by neurons through a process known as retroendocytosis, then uptake by satellite cells. This process involves the rapid exocytosis of intact ligands from neurons, a process that may also be utilized by other cell proteins, peptides or small molecules. Retroendocytosis of insulin has been described in kidney epithelial cells, fibroblasts transfected with insulin receptors, adipocytes and other cells (Dahl et al., 1989; Levy et al., 1988; Marshall, 1985). While largely unexplored in neurons, this pathway may be important in bidirectional signaling. That neurons and satellite cells communicate has been confirmed in other ways as well. Electrical stimulation of DRG sensory neurons generates calcium influx through L-type channels, followed by exocytotic release of ATP that in turn activates P2X7 receptors on satellite cells (Zhang et al., 2007). This form of activation, in turn, releases intraganglionic TNFα that can further sensitize P2X3 responses in neurons and increase sensory neuron excitability. Moreover, perineuronal signaling does not simply involve neuron perikarya and satellite cells. Chronic axon injury can also promote the formation of axonal rings, for example IB4 GDNF sensitive axons, around neurons that intermingle with satellite cells, (Li and Zhou, 2001). This form of plasticity developed later (5–12 weeks) than the injury paradigms studied here. Once activated after injury however, perineuronal glial cells may generate a variety of signals that alter axon plasticity. Overall, our findings provide evidence of a unique bidirectional interaction between sensory neurons and satellite cells. Evidence of this interaction is prominent after injury and includes a physical association of satellite cells with newly emergent neurite processes, as well as a proliferative and hypertrophic response of satellite cells in response to remote axon injuries. These responses imply an important form of signaling in which satellite cell plasticity is kept in check by uninjured

11

neurons whereas satellite cells remain in close contact with neuron processes. The full repertoire of molecular signals involved in this communication could potentially be wide. Acknowledgments The work was supported by an operating grant from the Canadian Institutes of Health Research (FRN15686). DK was supported by a summer studentship from the Alberta Heritage Foundation for Medical Research (AHFMR). Dr. Xia-Qing Li helped to collect data for this manuscript. AD was an ‘Eyes High’ scholar of the University of Calgary. References Andersen, P.L., Webber, C.A., Kimura, K.A., Schreyer, D.J., 2000. Cyclic AMP prevents an increase in GAP-43 but promotes neurite growth in cultured adult rat dorsal root ganglion neurons. Exp. Neurol. 166, 153–165. Buonanno, A., Fischbach, G.D., 2001. Neuregulin and ErbB receptor signaling pathways in the nervous system. Curr. Opin. Neurobiol. 11, 287–296. Calcutt, N.A., Muir, D., Powell, H.C., Mizisin, A.P., 1992. Reduced ciliary neurontrophic factor-like activity in nerves from diabetic or galactose-fed rats. Brain Res. 575, 320–324. Dahl, D.C., Tsao, T., Duckworth, W.C., Mahoney, M.J., Rabkin, R., 1989. Retroendocytosis of insulin in a cultured kidney epithelial cell line. Am. J. Physiol. 257, C190–C196. Dublin, P., Hanani, M., 2007. Satellite glial cells in sensory ganglia: their possible contribution to inflammatory pain. Brain Behav. Immun. 21, 592–598. Esper, R.M., Pankonin, M.S., Loeb, J.A., 2006. Neuregulins: versatile growth and differentiation factors in nervous system development and human disease. Brain Res. Rev. 51, 161–175. Fenzi, F., Benedetti, M.D., Moretto, G., Rizzuto, N., 2001. Glial cell and macrophage reactions in rat spinal ganglion after peripheral nerve lesions: an immunocytochemical and morphometric study. Arch. Ital. Biol. 139, 357–365. Friede, R.L., Johnstone, M.A., 1967. Responses of thymidine labeling of nuclei in gray matter and nerve following sciatic transection. Acta Neuropathol. 7, 218–231. Guo, G., Singh, V., Zochodne, D.W., 2014. Growth and turning properties of adult glial cellderived neurotrophic factor coreceptor alpha1 nonpeptidergic sensory neurons. J. Neuropathol. Exp. Neurol. 73, 820–836. Hanani, M., Huang, T.Y., Cherkas, P.S., Ledda, M., Pannese, E., 2002. Glial cell plasticity in sensory ganglia induced by nerve damage. Neuroscience 114, 279–283. Huang, T.Y., Cherkas, P.S., Rosenthal, D.W., Hanani, M., 2005. Dye coupling among satellite glial cells in mammalian dorsal root ganglia. Brain Res. 1036, 42–49. Humbertson Jr., A., Zimmermann, E., Leedy, M., 1969. A chronological study of mitotic activity in satellite cell hyperplasia associated with chromatolytic neurons. Z. Zellforsch. Mikrosk. Anat. 100, 507–515. Lee, D.A., Zurawel, R.H., Windebank, A.J., 1995. Ciliary neurotrophic factor expression in Schwann cells is induced by axonal contact. J. Neurochem. 65, 564–568. Levy, J.R., Ullrich, A., Olefsky, J.M., 1988. Endocytotic uptake, processing, and retroendocytosis of human biosynthetic proinsulin by rat fibroblasts transfected with the human insulin receptor gene. J. Clin. Invest. 81, 1370–1377. Levy, B.F., Cunha, J.C., Chadi, G., 2007. Cellular analysis of S100Beta and fibroblast growth factor-2 in the dorsal root ganglia and sciatic nerve of rodents. focus on paracrine actions of activated satellite cells after axotomy. Int. J. Neurosci. 117, 1481–1503. Li, L., Zhou, X.F., 2001. Pericellular Griffonia simplicifolia I isolectin B4-binding ring structures in the dorsal root ganglia following peripheral nerve injury in rats. J. Comp. Neurol. 439, 259–274. Lindsay, R.M., 1988. Nerve growth factors (NGF, BDNF) enhance axonal regeneration but are not required for survival of adult sensory neurons. J. Neurosci. 8, 2394–2405. Lohof, A.M., Quillan, M., Dan, Y., Poo, M.-M., 1992. Asymmetric modulation of cytosolic cAMP activity induces growth cone turning. J. Neurosci. 12, 1253–1261. Marshall, S., 1985. Dual pathways for the intracellular processing of insulin. Relationship between retroendocytosis of intact hormone and the recycling of insulin receptors. J. Biol. Chem. 260, 13524–13531. Michailov, G.V., Sereda, M.W., Brinkmann, B.G., Fischer, T.M., Haug, B., Birchmeier, C., Role, L., Lai, C., Schwab, M.H., Nave, K.A., 2004. Axonal neuregulin-1 regulates myelin sheath thickness. Science 304, 700–703. Murray, A.J., Peace, A.G., Tucker, S.J., Shewan, D.A., 2012. Mammalian growth cone turning assays identify distinct cell signalling mechanisms that underlie axon growth, guidance and regeneration. Methods Mol. Biol. 846, 167–178. Nishiyama, M., Hoshino, A., Tsai, L., Henley, J.R., Goshima, Y., Tessier-Lavigne, M., Poo, M.-M., Hong, K., 2003. Cyclic AMP/GMP-dependent modulation of Ca2 + channels sets the polarity of nerve growth-cone turning. Nature 423, 990–995. Pannese, E., 1981. The satellite cells of the sensory ganglia. Adv. Anat. Embryol. Cell Biol. 65, 1–111. Pannese, E., 2002. Perikaryal surface specializations of neurons in sensory ganglia. Int. Rev. Cytol. 220, 1–34. Pannese, E., Procacci, P., Berti, E., Ledda, M., 1999. The perikaryal surface of spinal ganglion neurons: differences between domains in contact with satellite cells and in contact with the extracellular matrix. Anat. Embryol. 199, 199–206. Powell, H.C., Garrett, R.S., Kador, P.F., Mizisin, A.P., 1991. Fine-structural localization of aldose reductase and ouabain-sensitive, K(+)-dependent p-nitro-phenylphosphatase in rat peripheral nerve. Acta Neuropathol. 81, 529–539.

12

K. Christie et al. / Molecular and Cellular Neuroscience 67 (2015) 1–12

Procacci, P., Magnaghi, V., Pannese, E., 2008. Perineuronal satellite cells in mouse spinal ganglia express the gap junction protein connexin43 throughout life with decline in old age. Brain Res. Bull. 75, 562–569. Shinder, V., Govrin-Lippmann, R., Cohen, S., Belenky, M., Ilin, P., Fried, K., Wilkinson, H.A., Devor, M., 1999. Structural basis of sympathetic-sensory coupling in rat and human dorsal root ganglia following peripheral nerve injury. J. Neurocytol. 28, 743–761. Shirane, M., Nakayama, K.I., 2006. Protrudin induces neurite formation by directional membrane trafficking. Science 314, 818–821. Stephenson, J.L., Byers, M.R., 1995. GFAP immunoreactivity in trigeminal ganglion satellite cells after tooth injury in rats. Exp. Neurol. 131, 11–22. Webber, C.A., Xu, Y., Vanneste, K.J., Martinez, J.A., Verge, V.M., Zochodne, D.W., 2008a. Guiding adult Mammalian sensory axons during regeneration. J. Neuropathol. Exp. Neurol. 67, 212–222. Webber, C.A., Xu, Y., Vanneste, K.J., Martinez, J.A., Verge, V.M.K., Zochodne, D.W., 2008b. Guiding adult mammalian sensory axons during regeneration. J. Neuropathol. Exp. Neurol. 67, 212–222.

Zhang, X., Chen, Y., Wang, C., Huang, L.Y., 2007. Neuronal somatic ATP release triggers neuron-satellite glial cell communication in dorsal root ganglia. Proc. Natl. Acad. Sci. U. S. A. 104, 9864–9869. Zheng, J.Q., Felder, M., Connor, J.A., Poo, M.-M., 1994. Turning of nerve growth cones induced by neurotransmitters. Nature 368, 140–144. Zhou, X.F., Rush, R.A., McLachlan, E.M., 1996. Differential expression of the p75 nerve growth factor receptor in glia and neurons of the rat dorsal root ganglia after peripheral nerve transection. J. Neurosci. 16, 2901–2911. Zhou, X.F., Deng, Y.S., Chie, E., Xue, Q., Zhong, J.H., McLachlan, E.M., Rush, R.A., Xian, C.J., 1999. Satellite-cell-derived nerve growth factor and neurotrophin-3 are involved in noradrenergic sprouting in the dorsal root ganglia following peripheral nerve injury in the rat. Eur. J. Neurosci. 11, 1711–1722.