(A) Immunostaining for the glutamate transporter GLAST. The stained cytoplasm and unstained nuclei (arrows) gives a characteristic ‘signet ring’ appearance to the SGCs surrounding the neurons. (B,C) A similar staining pattern for the inwardly rectifying Kir4.1 channel in the trigeminal ganglion (B) and the dorsal root ganglion (C). Immunostaining is confined to the SGCs that surround the neuronal cell bodies, and the characteristic appearance given by the unstained SGC nuclei (arrows) is seen. (D–G) Three, merged 1 mm optical slices from a confocal stack of triple-labeled (SK3, NeuN and DAPI) trigeminal ganglion. * indicates the same neuronal nucleus in each image. (D) The SK3 immunostaining has the same appearance as GLAST and Kir4.1 immunostaining (arrows indicate unstained SGC nuclei). From the merged image (G, shown at higher magnification) there is no overlap between the red SK3 and green NeuN immunostaining, which indicates that SK3 is confined to the SGCs. (H) Another example of DAPI-stained nuclei located in SK3 (red) immunostained SGCs. * marks an en-face view of several SK3-immunopositive SGCs and their DAPI-stained nuclei.

(A) SGCs are immunopositive for the purinergic receptor P2Y4 but neurons are unstained. Arrows indicate SGC nuclei (compare with Fig. 2A–C). (B) Immunostaining for P2X3 receptor (green) labels neurons whereas SGCs are unlabelled and apparent by the absence of staining around DAPI-stained SGC nuclei (arrows). (C) NADPH histochemistry shows that not all ganglion neurons contain nNOS. An unstained neuron (arrow) is next to a medium- (arrowhead) and a densely- (double arrowhead) stained neuron. (D) cGMP immunostaining is present in SGCs but is not seen in neurons. (E) A similar staining pattern is observed for the glial form of guanylyl cyclase (alpha1-subunit).

All primary sensory neurons are surrounded by SGCs that are connected by Cx43-containing gap junctions. Neuronal activity leads to the release of K⁺ into the extracellular space and SGCs take it up via the inward rectifying K⁺ channel Kir4.1 and release K⁺ via the small conductance Ca²⁺-activated K⁺ channel SK3. Gap junctions permit the redistribution of K⁺ between SGCs. Neurons modulate the activity of SGCs by releasing ATP, which activates P2Y4 receptors located on the surface of SGCs and leads to the release of intracellular Ca²⁺. Neurons also modulate the activity of SGCs by releasing NO. NO diffuses in the SCGs where it activates the glial form of the enzyme guanylyl cyclase (GC), which catalyses the formation of the second messenger cGMP. An increase in cGMP and Ca²⁺ both activate K⁺ channels.

(A) The classical appearance of a Nissl-stained frozen section of a trigeminal ganglion. Neurons do not appear to be tightly packed and the nuclei (arrows) of SGCs are scattered around the neurons. (B,C) Epon semi-thin sections. The primary sensory neurons are packed closely together and in some places appear to be in contact with each other (arrowheads). SGC nuclei (arrows) and blood vessels (*) are also visible. (D,E) Electron micrographs showing neurons (N) surrounded by SGC cytoplasm (SGC). The extracellular space (*) containing collagen fibers is visible. The cytoplasm of the SGCs can be extremely attenuated, as in E (arrow). (F) The N and SGC cell membranes are closely apposed (~20 nm, space between arrowheads), which allows the SGCs to buffer ions that are extruded from the neuron.

(A,B) Compared to sham surgery (A), Cx43 immunoreactivity increases in the trigeminal ganglion 7 days after CCI of ION (B). Note the punctate labeling pattern restricted to SGCs that encircle sensory neurons. (C) In vivo RNAi suppresses the expression of Cx43 efficiently. Representative immunoblot showing the expression of Cx43 in trigeminal ganglion from naïve and either globin or Cx43 dsRNA-injected rats. The injection of Cx43 dsRNA results in an 80% decrease in Cx43 intensity compared with a non-treated (naïve) trigeminal ganglion. P0 and P1 represent the non-phosphorylated and phosphorylated forms of Cx43, respectively. (D) Spontaneous eye closure increases after injection of Cx43 dsRNA into the trigeminal ganglion. There is a significant interaction between the treatment and the time post-injection only on the ipsilateral side (RMA-NOVA: ipsilateral, F = 11.0, P<0.001 and contralateral, F = 2.0, P = 0.111). The effect of Cx43 dsRNA is seen on the side ipsilateral to the injection only. It consists of an increase in unevoked eye-closure behavior starting at day 1, reaching a maximum on day 3, then declining starting on day 5 to reach baseline on day 7. *P<0.05, ***P<0.001 compared to globin dsRNA; #P<0.05, ##P<0.01 compared to baseline (implanted rats). (E) The magnitude of behavior effect following Cx43 dsRNA injection is similar to that observed after nerve injury. The effect of dsRNA on day 3 (maximum) is compared to that of a 7-day CCI of the ION (sham surgery used as control). The data are expressed for each group because the difference in the number of blinks post-treatment with baseline. There is a significant difference between the four groups (one-way ANOVA, F = 12.3, P = 0.001). The difference in the number of blinks with baseline is increased after Cx43 dsRNA injection or CCI and these two treatments are not significantly different (P = 0.231). *P<0.05, **P<0.01 compared to appropriate control (globin dsRNA or sham surgery).

(A) GFAP staining of a normal trigeminal nerve shows the dramatic decrease in GFAP staining between the central (*) and the peripheral (pn) portion of the nerve. The arrows outline the limits of the nerve (g, ganglion). (B,C) Photomicrographs taken at the same exposure setting from non-injured (B) and injured ganglia (C). These ganglia are from the same animal, contralateral and ipsilateral to the lesion, and were processed (post-fixation and sectioning) and immunostained together on the same slide. Although GFAP is present in the ganglion on the non-injured side (B, arrows), the normal concentration is low compared with the staining on the injured side (C). * represent the location of unstained neurons. (D,E) Following CCI and systemic injection of BrdU, some SGCs (green, SK3 immunostained) have BrdU-positive nuclei (arrows).