The following is the first of a two-part report on research presented at the 2017 Society for Neuroscience annual meeting, which took place November 11-15 in Washington, DC. See Part 2 here. Also see a previous report here.
Evidence continues to accumulate showing that interactions between neurons and non-neuronal cell types, including immune cells, keratinocytes, glial cells, cancer cells, and stem cells, are important in touch, itch, and pain. A number of posters and symposia at the 2017 annual meeting of the Society for Neuroscience highlighted this emerging area of research.
Francie Moehring, Medical College of Wisconsin, Milwaukee, US, presented her work on the role of keratinocytes in mechanotransduction, during a nanosymposium on touch, itch, and pain. (This work was recently published after the meeting in eLife; see Moehring et al., 2018. Disclosure: The current writer works in the same lab as Moehring and is a co-author on the published paper.) Keratinocytes are traditionally known for their role as a protective barrier against environmental damage. Since keratinocytes are the first point of contact with external stimuli, Moehring first sought to determine whether these cells play an active role in sensing mechanical stimuli. To do so, she used inhibitory optogenetic tools, selectively expressing the light-sensitive protein archaerhodopsin (Arch) in mouse Keratin14-expressing cells, most of which are keratinocytes (approximately 95 percent). Whole-cell patch clamp recordings indicated that primary mouse keratinocytes depolarized upon mechanical stimulation, and that those expressing Arch became hyperpolarized when exposed to 590 nm light.
Moehring next used optogenetic inhibition to shut down Keratin14-expressing cells in vivo in naive Keratin14-Arch mice. This significantly increased paw withdrawal thresholds in these animals, assessed with von Frey testing, indicating that Keratin14-expressing cells are important for baseline mechanosensation.
To identify the keratinocyte-derived factor(s) that mediate this process, Moehring looked to adenosine triphosphate (ATP). Based on previous findings (Zappia et al., 2016), she said, “ATP is released from the skin, and ATP release deficits [due to TRPA1 ion channel deletion in Keratin14-expressing cells in mice] decrease mechanical responsiveness of the animals in vivo and decrease the action potential firing rate of A- and C-fibers in situ.” Therefore, she went on to confirm that ATP was released from isolated glabrous (hairless) hindpaw skin upon repeated mechanical stimulation of the skin with von Frey filaments.
To determine whether ATP was released specifically from keratinocytes, Moehring used a novel assay called the cell sniffer assay. Human embryonic kidney (HEK) cells transfected with P2X2, an ion channel specific for ATP, were patch clamped while co-cultured primary mouse keratinocytes were mechanically stimulated. Moehring found that only HEK cells expressing P2X2 and not control HEK cells exhibited inward currents in response to mechanical stimulation of an adjacent keratinocyte. These experiments established that ATP release upon mechanical stimulation of the skin came specifically from keratinocytes.
Finally, additional experiments tested whether ATP release was important for baseline mechanosensation. Here, Moehring used apyrase, an enzyme that breaks down ATP, in von Frey behavioral assays and in a tibial skin-nerve preparation. Intraplantar administration of apyrase increased mechanical thresholds of mice during the behavioral assays. Similarly, apyrase treatment significantly reduced the number of action potentials fired in response to mechanical stimulation of the skin-nerve preparation. To test if enzymatic degradation of ATP had any additive effects on Keratin14-expressing cell inhibition, naive Keratin14-Arch mice were injected in the hindpaw with apyrase. Combining apyrase injection with optogenetic inhibition of Keratin14-expressing cells had no additive effect. The results suggest that ATP released from keratinocytes plays a significant role in baseline mechanosensation. Moehring said that the lab is now investigating P2X4 receptors on sensory neurons as a target for keratinocyte-released ATP.
Overall, the data challenge conventional theories that sensory neurons are the sole responders to mechanical stimuli and suggest that they communicate with a range of non-neuronal epidermal cells such as keratinocytes and Merkel cells to convey touch sensation. The current work lays the foundation for future studies on alterations in neuronal and non-neuronal cell communication under skin disease and injury conditions.
During the nanosymposium, Sven-Eric Jordt, Duke University School of Medicine, Durham, US, presented his work on poison ivy contact dermatitis (skin inflammation) and the subsequent itch, noting that more than 10 million Americans suffer from poison ivy each year. Jordt and his colleagues used urushiol, the allergen in poison ivy that causes contact dermatitis, to sensitize mice. The researchers administered an initial high dose followed by a lower dose given every other day for five days. Urushiol treatment resulted in a high dermatitis score as measured by erythema (reddening of the skin), scarring, excoriation (skin picking), and swelling. Skin thickness was significantly increased in urushiol-treated animals compared to vehicle-treated animals.
Having confirmed that urushiol could cause contact dermatitis, Jordt and colleagues performed transcriptome analysis of the affected skin for itch and inflammatory mediators. Many were upregulated but one cytokine, interleukin 33 (IL-33), stood out. IL-33 drives the T helper-2 (Th2) immune response by binding to its receptor, ST2, which produces type 2 cytokines. The Th2 response is associated with allergic reactions and accompanying itch. The investigators confirmed the involvement of cutaneous IL-33 in contact dermatitis by demonstrating that skin-specific overexpression of IL-33 alone caused dermatitis and scratching in mice.
Further experiments showed that cutaneous IL-33 released after urushiol exposure was made by keratinocytes. The hypothesis was that this IL-33 could act on sensory neurons since both dorsal root ganglia (DRG) and cutaneous afferents express ST2. To test this idea, Jordt applied IL-33 to isolated DRG neurons and then performed calcium imaging. IL-33 caused a calcium response in DRG neurons, in support of the idea.
Jordt then showed that ST2 staining significantly overlapped with expression of TRPV1, a marker for nociceptive neurons. Next, using calcium imaging to further determine ST2 distribution in DRG neurons, he found that approximately 85 percent of capsaicin-responsive neurons (TRPV1+), 75 percent of histamine-responsive neurons, 68 percent of mustard oil-responsive neurons (TRPA1+), and 52 percent of chloroquine-responsive neurons (Mrgpr+) also responded to IL-33. This indicated that sensory neurons responding to pain and/or itch stimuli also respond to IL-33. Further, a significantly higher percentage of DRG neurons from mice challenged with urushiol responded to IL-33, compared to unchallenged mice.
Next, Jordt blocked IL-33 or ST2 separately using antibodies to each, discovering that this significantly reduced urushiol-induced scratching behavior in mice. Antibody treatment also significantly decreased both skin thickness and dermatitis score. In addition, IL-33 given with an ST2-blocking antibody to urushiol-treated mice resulted in decreased scratching compared to IL-33 alone, suggesting that IL-33 mediates itch through ST2. To determine if this itch behavior depended on histamine, Jordt injected the anti-histamine cetirizine and IL-33 into mice. Cetirizine had no effect on scratching behavior, indicating that IL-33 acts independently of the histamine pathway.
Finally, Jordt used small interfering RNA (siRNA) to knockdown ST2 in DRG neurons. This reduced scratching behavior in urushiol-treated mice, supporting the hypothesis that IL-33 acts on sensory neurons to cause itch associated with urushiol.
In a poster session, Sarah Rosen, McGill University, Montreal, Canada, presented her work on the role of T cells in opioid analgesia and in sex differences in morphine analgesia, where morphine is a more potent analgesic in males than in females. Pain can be caused by an inflammatory immune response, which involves infiltration of leukocytes to an injury site. Leukocytes include neutrophils, eosinophils, basophils, monocytes, and lymphocytes (B and T cells). Previous research had shown that infiltrating leukocytes can release endogenous opioids, that exogenous opioids have various effects on T cells, and that males and females differ in their T cell distribution. “T cells [may] play a large role in opioid analgesia, and could be a driver in the observed sex differences in morphine analgesia,” Rosen said.
First, Rosen studied baseline pain sensitivity in naive control mice, nude mice (which lack T cells), and Rag1 knockout mice (which lack T and B cells), using both male and female animals. Overall, the immunocompromised mice displayed increased pain sensitivity compared to the controls, according to a number of assays including tail withdrawal from a thermal stimulus, intraplantar injection of formalin, paw withdrawal from a hot plate, and von Frey mechanical thresholds.
Next, control and nude mice received various doses of morphine. Controls exhibited a higher amount of morphine-induced analgesia than nude mice in tail withdrawal and formalin assays. Similarly, control mice exhibited a higher amount of morphine-induced analgesia than Rag1 knockouts. Rosen also examined restraint stress-induced analgesia (SIA) using the hot plate test. Stressors such as physical restraint activate the endogenous opioid system and cause opioid-mediated SIA in mice. After restraint, control mice were less sensitive to the hot plate stimulus. However, restraint had no effect on pain sensitivity in nude mice. To confirm that the SIA observed in control mice was opioid mediated, Rosen administered naloxone, an opioid receptor antagonist, to control and nude mice before hot plate testing. Naloxone prevented SIA in control mice and had no effect in nude mice. Together, the data show that mice missing T cells have blunted opioid analgesia.
To pinpoint which T cell subtype mediated opioid analgesia, Rosen performed adoptive transfer of control mouse splenocytes, all T cell subsets, CD4+ T cells, or CD8+ T cells into nude mice and measured tail withdrawal latencies after morphine administration. Nude mice receiving splenocytes, all T cell subsets, or CD4+ T cells displayed morphine analgesia similar to that of control mice, but adoptive transfer of CD8+ T cells had no effect, suggesting that CD4+ T cells, but not CD8+ T cells, play a role in morphine analgesia.
Interestingly, female control mice displayed less morphine analgesia than male control mice, while there were no differences between female and male nude mice. Rosen hypothesized that these sex differences were mediated by CD4+ T cells. Adoptive transfer of CD4+ T cells from female control mice to female nude mice increased morphine analgesia in the nude mice during the tail withdrawal assay. However, adoptive transfer of CD4+ T cells from male control mice to female nude mice resulted in an upward trend in the amount of morphine analgesia compared to CD4+ T cell transfer from female control mice. These data suggest that CD4+ T cells play a role in the sex differences observed with morphine analgesia.
Together, all the studies from SfN described above show that neuronal interactions with non-neuronal cells play a significant role in somatosensation, from baseline touch to itch to pain.
Ashley Cowie is a PhD candidate at the Medical College of Wisconsin, Milwaukee, US.
Image credit: tonobalaguer/123RF Stock Photo.