The following is Part 2 of a three-part series of selected talks from the 34th Annual Scientific Meeting of the American Pain Society (APS) held May 13-16, 2015, in Palm Springs, California, US. Also see Part 1 and Part 3.
Keratinocytes, the cells that make up the epidermis, form the barrier between the body and the outside world, and they regularly interact with immune cells. But detecting painful stimuli is not in their wheelhouse … or is it? Three researchers argued that keratinocytes actively contribute to pain sensation emanating from the skin.
First, Jose Moron-Concepcion, Columbia University, New York, US, provided evidence that keratinocytes express an ion channel classically found only in neurons: AMPA-type glutamate receptor channels. Moron-Concepcion previously found that AMPA receptor expression was elevated in spinal cord neurons from mice with hyperalgesia (Cabañero et al., 2013). “Then we thought, maybe AMPA receptors are not just in the spinal cord—maybe they are also in nerve terminals in the periphery,” he said.
Searching for AMPA channels in the peripheral terminals of nociceptive afferents, Moron-Concepcion and colleagues stained mouse epidermis using CD34, an antibody specific to the GluA4 AMPA subunit. To his surprise, he found robust staining in the keratinocytes of the skin. “At first we thought it was an artifact,” he said, but the finding held up to further scrutiny. Keratinocytes from human skin also stained positive for GluA4. In cultured human cells, the staining disappeared after genetic knockdown of GluA4 by small-hairpin RNA (shRNA) interference, confirming the subunit’s presence.
The team next performed whole-cell electrophysiology on the cultured human cells and found AMPA-evoked inward currents. “They were really slow compared to AMPA currents in neurons,” Moron-Concepcion said. The currents were blocked by a drug that only blocks currents through AMPA channels lacking the GluA2 subunit—a channel component that prohibits calcium from entering. That finding suggests that the keratinocyte AMPA channels do pass calcium, an important intracellular signal.
In collaboration with Frank Rice (see comment below), Integrated Tissue Dynamics, Rensselaer, New York, US, the researchers then compared skin samples from healthy people with samples from patients with the chronic itch condition atopic dermatitis (AD) or painful post-herpetic neuropathy (PN). “The results were quite striking,” said Moron-Concepcion. Skin from AD patients displayed intense immunolabeling for GluA4 and for the inflammatory cytokine thymic stromal lymphopoietin (TSLP), which is known to contribute to itch in AD. Skin from patients with PN, in contrast, contained less GluA4 than healthy skin samples. What that differential expression of GluA4 in the two conditions means is still a mystery, and the role of keratinocytes in contributing to pain and itch remains to be further investigated. But Moron-Concepcion concluded, “We believe there is a reciprocal interaction between keratinocytes and peripheral nerve termini of C- and Aδ-fibers, and that GluA4-containing AMPA receptors may play a role in such interactions.”
Evidence for skin-nerve interaction in vivo came from data presented by Michael Caterina, Johns Hopkins University, Baltimore, US, and published in April in Pain (Pang et al., 2015). Keratinocytes are poised to participate in sensing noxious stimuli: they express thermally and mechanically sensitive ion channels, and they are closely juxtaposed with nerve endings, Caterina said in his presentation, and “there is a long list of what can be released by keratinocytes”—substances to which nociceptors respond. Caterina set out to determine whether stimulation of skin cells could activate sensory neurons in vivo and whether that could evoke the sensation of pain.
To do that, the team needed a way to activate keratinocytes specifically, so they adopted a chemogenetic approach. They began with mice lacking the transient receptor potential vanilloid type 1 receptor (TRPV1), which were insensitive to the TRPV1 activator capsaicin. Caterina’s team then crossed those mice with two other transgenic mouse lines to produce mice that conditionally expressed TRPV1 following treatment with tamoxifen only in keratinocytes. “We can control both in space and time when and where TRPV1 gets expressed in the mice,” Caterina told PRF.
Would activation of TRPV1—found solely in keratinocytes in the mice—also activate sensory neurons? After swabbing the paws of the transgenic mice with capsaicin, the researchers used immunocytochemistry to show upregulation of c-fos, a marker of neuronal activation, in laminae I and II of the dorsal horn of the spinal cord, where nociceptive neurons terminate. “The only way that could have happened was if there was activation of sensory neurons,” Caterina said.
Next, the team wanted to see whether keratinocyte activation alone would elicit pain behaviors in the transgenic mice. The mice did lick their paws after application of capsaicin, but that might just be a reflexive behavior, Caterina said, rather than a pain response. So the team employed a conditioned place avoidance test to see whether the mice experienced the capsaicin as unpleasant. The animals were allowed to explore two interconnected chambers with different interior designs: one had vertically striped walls and a smooth floor, whereas the other had walls decorated with triangles and a floor with a grooved surface. Mice received a paw injection of saline and were confined to one chamber for 20 minutes. Later that day, they received an injection of capsaicin and were restricted to the other chamber. The next day, Caterina said, when allowed to roam between the cages, “the mice spent more time in the saline-conditioned chamber and less in the capsaicin chamber,” suggesting that the capsaicin was painful or at least aversive to the mice.
Kathryn Albers, University of Pittsburgh, US, reported that changing keratinocyte activity directly influenced sensory neuron firing. Albers and her team used optogenetics to selectively activate keratinocytes and found that they, in turn, provoked sensory nerve firing and pain behavioral responses in mice, whereas inhibiting keratinocytes dampened nerve firing.
Albers first made transgenic mice that expressed channel rhodopsin 2 (ChR2) tagged with a fluorescent protein under control of the promoter for peripherin, a protein expressed solely in peripheral nerves, which put the light-activated ion channel in myelinated and unmyelinated sensory neurons. Similar to previous published studies, blue light directed at the skin evoked instantaneous paw withdrawal in the mice, indicating that the light had produced a noxious sensation (see PRF related news). “This was a very potent signal,” Albers said in her presentation. Using an ex vivo skin-nerve preparation, Albers’ team recorded activity from dorsal root ganglia (DRG) neurons, which fired action potentials in response to the light stimulus, but “the recordings did not look the same as if you had applied a mechanical or heat stimulus” in wild-type animals, Albers said. Because keratinocytes can also respond to those stimuli, she hypothesized that activation of nerves and keratinocytes together might more closely replicate those recordings.
Consequently, the team decided to add ChR2 to keratinocytes as well as to neurons. To begin, they expressed ChR2 under control of the promoter for K14, a protein expressed solely in keratinocytes, without expressing ChR2 in neurons. Unexpectedly, when they shined light on the skin of mice to activate keratinocytes alone, the mice again showed nocifensive behaviors, although more slowly than with neuron activation—some within two seconds and others over the course of 20 seconds. “We interpret that lag to mean that keratinocytes are releasing something, and it takes that long to reach the neurons,” Albers said. The researchers are working to determine what those compounds might be, but they have determined that “ATP is one of the players underlying these effects,” she added.
The investigators again employed the skin-nerve preparation, recording from DRG neurons while stimulating keratinocytes with light. After characterizing the neurons that responded to keratinocyte activation (based on their responses to other stimuli and their firing properties), the team found that small, unmyelinated C-fibers were preferentially activated by keratinocytes. Slowly adapting type I (SAI) Aβ afferents also fired action potentials with light stimulation; those nerve endings form functional complexes with specialized, mechanically sensitive keratinocytes called Merkel cells (Woo et al., 2015).
The investigators then explored how ChR2-expressing keratinocytes might work in concert with mechanical and thermal stimuli to help activate neurons. For example, an Aβ high-threshold mechanoreceptor fiber that responded to a 10-millinewton (mN) force but not to 5 mN or to light activation of keratinocytes alone did fire action potentials in response to a 5 mN force delivered together with light. The team saw similar summation of light- and heat-evoked action potentials in C-fibers.
Albers next made transgenic mice that expressed halorhodopsin, a protein that responds to yellow light and alters chloride balance to inhibit excitable cells, in keratinocytes. Intriguingly, when keratinocyte activity was shut down with yellow light, “we saw complete inhibition in some nerve fibers and decreased firing in others,” in response to mechanical or thermal stimuli. “That means that whatever signal comes from keratinocytes can be modulated by changing chloride balance,” Albers said. “We don’t yet understand how, but it decreases afferent responses.”
Caterina called that experiment provocative. “That suggests there are ongoing roles for keratinocytes in thermal and mechanical pain sensation, even without engineering the cells,” he said. “It opens a lot of exciting questions for the field to address.”
Some of those future questions on the list for these researchers include finding the endogenous signals to which keratinocytes respond and what they release to activate neurons. The researchers also wonder whether keratinocyte-neuron communication is layer specific within the epidermis and how skin cells contribute to pain in disease states.
Image credit: American Pain Society
Stephani Sutherland, PhD, is a neuroscientist, yogi, and freelance writer in Southern California.