This Week in The Journal

Fibroblast Protein Limits Pain Sensitivity

Shuo Zhang, Bing Cai, Zhen Li, Kaikai Wang, Lan Bao, et al.

(see pages 4069–4086)

It is now well established that glia contribute to neural transmission. For example, oligodendrocytes form myelin to speed action potential conduction; microglia prune synapses; and astrocytes enwrap synapses to limit the spatial and temporal spread of neurotransmitter action. In peripheral ganglia, a unique type of glia, satellite glia cells (SGCs), encapsulate individual or small clusters of neuronal somata, limiting direct communication between neurons. But SGCs also form gap junctions with each other and with neurons, forming regulated communication channels. Intriguingly, gap-channel coupling between SGCs and neurons increases in dorsal root ganglia (DRGs) after injury, and this may contribute to pain hypersensitivity.

Zhang et al. have discovered that interactions between SGCs and sensory neurons are unexpectedly regulated by fibroblasts. They found that fibroblasts surround the SGC envelope around neurons and release a protein called secreted modular calcium-binding protein 2 (SMOC2), which colocalizes with a component of the basement membrane. Remarkably, knocking out SMOC2 reduced mechanical pain thresholds in mice. In addition, knocking out SMOC2 reduced the proportion of DRG neurons that were separated from others by basement membrane, increased the proportion of neurons that showed coupled responses to mechanical stimulation, and increased the total number of neurons activated by mechanical stimulation. Notably, injection of an inflammatory agent that increases pain sensitivity reduced fibroblast and SMOC2 levels and increased neuronal coupling in the DRGs. Furthermore, injection of exogenous SMOC2 reduced neuronal coupling and the number of neurons responding to mechanical stimulation in both SMOC2-deficient mice and mice treated with the inflammatory agent.

Because SMOC2 did not physically interact with the gap junction protein Cx43, the authors conclude that SMOC2 does not regulate electrical coupling. Instead, SMOC2 interacted with P2X7, a purine receptor expressed exclusively in SGCs in the DRGs, and reduced ATP-induced calcium currents through P2X7 channels in HEK cells. Importantly, a P2X7 antagonist replicated and occluded the effects of exogenous SMOC2 on DRG neuron activation in SMOC2-deficient mice.

These results suggest that SMOC2, secreted by DRG fibroblasts, helps maintain the basement membrane and limits the activation of P2X7 in SGCs. Inflammation disrupts this regulation and increases coupled activation of nociceptive neurons, which likely reduces thresholds for mechanical pain. Future work should determine whether other molecules secreted by fibroblast regulate the activity of other types of neurons in DRGs.

• Download figure
• Open in new tab
• Download powerpoint

SMOC2 (green) surrounds SGCs (red), which encapsulate neurons (blue) in the DRGs. See Zhang et al. for details.

M1 Retinal Ganglion Cells Underlie Photoaversion

Franklin S. Caval-Holme, Marcos L. Aranda, Andy Q. Chen, Alexandre Tiriac, Yizhen Zhang, et al.

(see pages 4101–4115)

Before wiring of the retina is complete, acetylcholine released from amacrine cells produces propagating waves of activity in retinal ganglion cells (RGCs). This activity is thought to instruct the wiring of downstream visual areas. During the same postnatal period—before RGCs receive input from photoreceptors—newborn mice respond to bright light by turning away from the light and emitting ultrasonic vocalizations. These responses are initiated by intrinsically photosensitive RGCs (ipRGCs). Which of the six subtypes of ipRGC are responsible for photoaversion and how mice can distinguish light-induced activity from retinal wave-induced activity have been unclear. Caval-Holme et al. now answer these questions.

The authors first determined the threshold light intensity for photoaversion. Comparing this threshold to previously published activation thresholds for different subtypes of ipRGCs suggested that light just bright enough to induce photoaversion will activate M1 and M3 ipRGCs, but not M2, M4, or M5 ipRGCs. Consistent with a role for M1 cells in photoaversion, knocking out TRPC6 and TRPC7 channels, which mediate phototransduction predominantly in M1 ipRGCs, eliminated photoaversion responses at threshold light intensities. In contrast, photoaversion was unaffected by killing all ipRGCs except M1 cells. Moreover, preventing communication via gap junctions, which contributes to the photosensitivity of all ipRGCs other than M1 cells, had no effect on photoaversion. These results indicate that M1 ipRGCs are the principal cells responsible for producing photoaversion in neonatal mice.

Unexpectedly, M1 cells were less likely than other ipRGCs to participate in retinal waves. Moreover, whereas photostimulation evoked greater depolarization than retinal waves in M1 ipRGCs, retinal waves evoked greater depolarization than photostimulation in other ipRGCs. This suggests that M1 ipRGCs provide a separate channel of information flow to the brain, allowing light-induced activity to be distinguished from spontaneous retinal activity. The authors suggest that the minimal participation of M1 ipRGCs in retinal waves may also help explain why brain regions involved in photoaversion lack the topographic organization seen in image-forming areas.