Corneal Sensory Nerve Activity in an Experimental Model of UV Keratitis

Abstract

Purpose.: To produce in guinea pigs a UV-induced keratitis, to analyze the effects of this pathology on corneal nerve activity.

Methods.: In anesthetized animals, one eye was exposed to 254 nm UV-C radiation (500–1000 mJ/cm²), excised 24 to 48 hours later and superfused in vitro. Nerve impulse activity was recorded in ciliary nerve filaments or in corneal sensory terminals of intact and UV-irradiated eyes. Impulse activity in response to mechanical (von Frey hairs), chemical (98.5% CO₂ gas jets), and thermal stimulation (cooling from 34°C to 20°C; heating to 50°C) was analyzed. Duration of eyelid closure and blinking and tearing rates were evaluated in control and in UV-irradiated eyes, before and after application of TRPV1, TRPA1, and TRPM8 agonists (100 μM capsaicin; 10 mM AITC, and 200 μM menthol, respectively).

Results.: After irradiation, mechanical threshold of mechano-nociceptor corneo-scleral fibers was reduced (0.59 ± 0.4 vs. 0.27 ± 0.07 mN; P < 0.05) while polymodal nociceptors increased their response to chemical stimulation (1.7 ± 0.2 vs. 3.4 ± 0.5 imps/s; P < 0.05). In contrast, cold thermoreceptors showed a significantly lower ongoing activity at 34°C (8.6 ± 0.5 vs. 6.1 ± 0.9 imp/s; P < 0.05) and a reduced responsiveness to cooling pulses (peak frequency = 29.8 ± 1.3 vs. 18.9 ± 1.8 imp/s; P < 0.001). Blinking but not tearing rate was significantly higher; behavioral responses to topical capsaicin and AITC, but not to menthol were enhanced in UV-irradiated animals.

Conclusions.: Sensitization of nociceptor and depression of cold thermoreceptor activity following UV radiation appear to result from an action of inflammatory mediators on TRP channels selectively expressed by sensory nerve terminals. Changes in nerve activity possibly underlie discomfort sensations associated with corneo-conjunctival inflammation induced by UV exposure.

Introduction

Inflammation accompanies multiple pathological conditions of the cornea and conjunctiva, such as mechanical or chemical injury, exposure to excessive temperatures or radiation, desiccation, bacterial and viral infections, or allergic reactions. ^1–10 A common symptom of ocular surface inflammation is the presence of conscious feelings of unpleasantness, ranging from moderate discomfort to intense pain. The subjective quality of these unpleasant ocular sensations varies widely and has been described as itch, dryness, grittiness, burning, pricking, or stinging pain, depending on the cause, intensity, location, and the extension of the eye inflammation. ^11–13  

Ocular surface sensations normally result from the stimulation of peripheral axon terminals of trigeminal ganglion sensory neurons innervating the cornea and conjunctiva. These sensory nerve afferents are functionally different, corresponding to mechanonociceptor, polymodal nociceptor, and cold thermoreceptor sensory neurons. ^12–19 Their experimental selective stimulation in healthy humans evokes qualitatively distinct sensations of irritation and pain that include a variable thermal component. ^20,21  

After corneal damage, a large number of endogenous chemical mediators are locally released by injured corneal cells and infiltrating immune cells. These include kinins, amines, prostanoids, purines, nerve growth factor (NGF), and protons, depending on the triggering process. ¹² It is well established that inflammatory mediators act on polymodal nociceptor fibers, exciting them directly and also producing a change in nociceptor responsiveness (sensitization) characterized by an enhanced responsiveness and lowered threshold to natural stimulation, as well as by the development of spontaneous, ongoing activity. ^22–24 Sensitization of nociceptor fibers innervating inflamed ocular tissues is the cause of spontaneous pain and altered sensitivity to innocuous and noxious stimuli (primary allodynia and hyperalgesia, respectively) observed during eye surface inflammation. ^8,12–14,17 In contrast, recent experimental evidence has shown that inflammatory agents attenuate the response of corneal cold thermoreceptors to low temperatures through an inhibition of the cold-activated channel TRPM8 through a Gαq protein. ²⁵ Accordingly, during allergic keratoconjunctivitis, cold thermoreceptor activity and responsiveness are reduced. ³  

Ultraviolet (UV) radiation is naturally generated by sunlight and it is classified into three wavebands (UV-A, UV-B, and UV-C) with differences in their intensity and biological effects. UV radiation in solar or artificial form is responsible for a substantial burden of human skin and eye diseases, including erythema, skin cancer, photokeratitis, pterygium, or cataract. ^26–28 It causes cellular DNA damage, with UV-C appearing as the strongest genotoxin of the three broad-spectrum components. ^27,28 Solar UV-C is normally blocked by the ozone layer, but accidental exposure can occur from artificial sources, such as germicidal lamps, producing injury to the skin and the eye. ^29,30 The cornea of the eye absorbs UV-C and a substantial amount of UV-B. Hence, an acute exposure of the eyes to UV-B radiation and nonsolar UV-C produces an inflammation of the cornea called photokeratitis. ^30,31 This is a painful transient inflammatory condition that typically appears 6 to 12 hours after UV exposure and resolves within 48 hours, usually without long-term consequences. Symptoms include tearing, ocular redness and pain, swollen eyelids, headache, gritty feeling in the eyes, halos around lights, hazy vision, and temporary loss of vision. ^31,32  

Local inflammation evoked by experimental UV irradiation of rat eyes induces in second-order neurons at the Vc/C1 region of the brainstem, strong expression of c-FOS, a well-known marker of neuronal activity, after peripheral stimulation, ³³ suggesting that this area of the brain is the target of peripheral ocular sensory fibers processing ocular pain evoked by UV radiation. ³⁴ However, the effect of photokeratitis on the activity of the different functional classes of corneal sensory nerves during photokeratitis, leading ultimately to the symptoms of discomfort and pain, is still undefined. The purpose of this study was to analyze in guinea pigs the effects of eye exposure to UV-C radiation on behavioral nocifensive responses and tearing and blinking rates, and on the spontaneous and stimulus-evoked nerve impulse activity of the different functional types of corneal sensory neurons. Preliminary results of this work have been reported in abstract form (Luna C, et al. IOVS 2010;51:ARVO E-Abstract 1962). 

Methods

The experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and also adhered to the European Union Directive (2010/63/EU) on the protection of animals used for scientific purposes. The procedures were authorized by the Ethics Committee of the University Miguel Hernández. 

Animals and UV Radiation

Albino and pigmented guinea pigs of both sexes weighing 300 to 900 g (1–3 months of age) were anesthetized (80 mg/kg ketamine plus 4 mg/kg xylazine, intraperitoneally [IP]); immediately afterward, 254 nm UV-C radiation was delivered to one eye of the animal with a UV lamp (VL-4.C 230V 50/60 Hz; Vilber Lourmat, Marne-la-Vallée, France) placed at a distance of 17 cm from the eye for different times (5, 7.5, 24.5, and 49 minutes) to deliver different intensities (100, 150, 500, and 1000 mJ/cm², respectively). Animals not exposed to UV radiation served as controls. Data from contralateral (unexposed) eyes were not included. 

In preliminary experiments, we determined that UV intensities of stimulation below 500 mJ/cm² evoked only mild clinical signs of photokeratitis and did not change the spontaneous or stimulus-evoked response of mechano-nociceptors (n = 53) and polymodal nociceptors (n = 50; Luna C, et al. IOVS 2010;51:ARVO E-Abstract 1962). We also confirmed that no clinical signs of inflammation were observed 72 hours after UV exposure. Thus, only intensities of 500 and 1000 mJ/cm² at 24 and 48 hours were analyzed and included in the study. 

Electrophysiological Recordings

Two types of preparations were used for electrophysiological recordings from the excised, superfused eye of the guinea pig. ^3,35 The “whole eye” preparation was particularly suitable for recording polymodal and mechanosensory nociceptive units in the ciliary nerves, whereas in the “isolated cornea” preparation, activity of cold-sensitive units could be more easily identified. ³ Animals were killed with an IP injection of 100 mg/kg sodium pentobarbitone, and the eyes, together with the bulbar and tarsal conjunctiva, were enucleated. The whole eye or the excised cornea was placed in a recording chamber and superfused with physiological solution of the following composition (in mM): NaCl, 133.4; KCl, 4.7; CaCl₂, 2.0; MgCl₂, 1.2; NaHCO₃, 16.3; NaH₂PO₄, 1.3; glucose, 7.8, gassed with 95% O₂–5% CO₂ to pH = 7.4, maintained at 34°C with a homemade feedback-controlled Peltier device. The characteristics of the spontaneous and stimulus-evoked impulse activity were analyzed in control and UV-irradiated eyes 24 or 48 hours after UV exposure. 

Recording of Corneal Ciliary Nerve Axons.

In the whole eye preparation, connective tissue and extraocular muscles in the back of the eye were carefully removed to expose and isolate the ciliary nerves around the optic nerve. The eye was then placed in a chamber divided in two compartments by an elastomer-coated plastic wall (Sylgard 184; Dow Corning, Midland, MI, USA). The front of the eye was introduced into a round perforation made in the center of the dividing wall to which the bulbar conjunctiva was pinned, thereby isolating the front from the back of the eye. This prevented direct exposure of the ciliary nerves located in the back compartment to the chemical substances applied onto the corneal surface. The anterior compartment was continuously bathed with warmed (34°C) physiological saline solution and the rear compartment was filled with warm mineral oil. Thin nerve filaments were teased apart from the ciliary nerve trunks and placed on an Ag-AgCl electrode for monopolar recording of unitary impulse activity in single nerve axons, using conventional electrophysiological equipment (DAM50 amplifier; WPI, Sarasota, FL, USA). Electrical signals were fed into a PC through an acquisition system (CED Micro-1401; Cambridge Electronic Design, Cambridge, UK), and analyzed with Spike 2 software (v6.0; Cambridge Electronic Design). Receptive fields of afferent fibers innervating the corneoconjunctival surface were located by using mechanical stimulation with a fine paint brush and mapped thereafter using a von Frey hair (5.88 mN). Mechanical threshold was determined assessing the first impulse response evoked by calibrated von Frey hairs of increasing force (range, 0.078–4 mN; Bioseb, Vitrolles, France). For chemical stimulation, a jet of gas containing 98.5% CO₂ was applied onto the corneal receptive field for 30 seconds. 

The following parameters of the stimulus-evoked firing discharge were measured: Frequency of the spontaneous activity in impulses per second (imp/s) measured during 1 minute in recordings performed before any stimulation was applied. Firing frequency evoked by the stimulus, mean discharge in imp/s during the 30-second CO₂ pulse. Latency of the impulse discharge, time in s lapsed between the onset of the CO₂ pulse and the beginning of the impulse response. Postdischarge, mean discharge rate in imp/s during 30 seconds after the end of the CO₂ pulse (Fig. 1). 

Figure 1

 

Response to a 30-second CO₂ pulse of polymodal nociceptors recorded from a control eye (left) and an UV-irradiated eye (right,) 48 hours after 1000 mJ/cm². Two different units were recorded in each case distinguishable by the size in the direct recording (A). The bars mark the duration of the CO₂ pulse. (B, C) represent the impulse frequency histograms of the discharge (imp/s) of the two different units recorded in (A). The latency of the response and the 30 seconds of postdischarge are marked by the arrows.

Recording of Corneal Cold Nerve Terminals.

Corneas were excised with a circular cut around the limbus and pinned to the bottom of a recording chamber continuously superfused with the physiological solution at 34°C. Nerve terminal impulse (NTI) activity was recorded as described elsewhere, ^3,18,19 using a glass micropipette applied onto the surface of the cornea with a micromanipulator and attached by slight suction. Pipettes had a tip diameter of approximately 50 μm and were filled with the physiological saline solution. Signals were recorded with respect to an Ag/AgCl pellet in the bath. Electrical activity was amplified (AC preamplifier NL 103; Digitimer, Welwyn, UK), filtered (high pass 150 Hz, low pass 5 KHz; filter module NL 125; Digitimer), transferred to a PC with a CED micro-1401 acquisition system (Cambridge Electronic Design), and analyzed with the indicated software. The corneal surface was explored with the tip of the micropipette until a site in which spontaneous activity of a single nerve terminal was detected. Nerve impulses originating at single cold-sensitive nerve endings were identified by their regular ongoing discharge, which increased with cooling and silenced during warming. ^16,18–20 For thermal stimulation, the temperature of the physiological solution bathing the cornea was changed with the Peltier device, from the basal temperature of 34°C down to 20°C (average cooling rate −0.5°C/s) or up to 50°C (average heating rate +0.4°C/s). Spontaneous NTI activity at basal temperature was recorded for at least 1 minute before the application of a cooling ramp to 20°C, followed by rewarming to the basal temperature. After recovering a stable baseline impulse activity, the solution bathing the cornea was heated to 50°C and immediately cooled again to basal temperature. 

The following parameters were measured to quantify the cooling and heating response ³ (Fig. 2): (a) Mean frequency of the ongoing NTI activity: NTI frequency in imp/s measured during 1 minute at the basal temperature of 34°C; (b) Cooling threshold from 34°C: decrement in temperature in °C (measured in a cooling ramp from 34°C to 20°C) required to increase by 25% the mean ongoing frequency of discharge; (c) Peak frequency: maximal value of the firing frequency in imp/s reached during a cooling ramp; (d) Temperature change for peak frequency: temperature change in °C required to reach the peak frequency value during the cooling ramp; (e) Cooling threshold to resume NTI activity: temperature decrement in °C required to resume NTI firing when decreasing temperature from 50°C to 34°C; (f) Peak frequency during cooling from 50°C to 34°C: maximal value of the firing frequency in imp/s reached during cooling from 50°C to 34°C; (g) Temperature change for peak frequency from 50°C to 34°C: temperature change in °C required to reach the peak frequency value during the cooling from 50°C to 34°C; (h) Silencing threshold: temperature change in °C required for silencing the terminal during a heating pulse; and (i) Incidence of paradoxical response: percentage of terminals firing in response to heating pulses. 

Figure 2

 

Example of the change in NTI activity in a cold thermoreceptor terminal evoked by thermal stimulation. Upper trace: Temperature change. Lower trace: Histogram of the firing frequency in imp/s. The following parameters (see Methods) were used to quantify the cooling response: (a) mean frequency of the ongoing NTI activity at 34°C, (b) cooling threshold from 34°C to 20°C, (c) peak frequency from 34°C to 20°C, (d) temperature change for peak frequency from 34°C to 20°C, (e) cooling threshold to resume NTI activity, (f) peak frequency from 50°C to 34°C, (g) temperature change for peak frequency from 50°C to 34°C, (h) silencing threshold, and (i) incidence of paradoxical response (none in this case).

Measurement of Nocifensive Response, Tearing Rate, and Clinical Signs of Inflammation

All parameters were measured before and 24 or 48 hours after UV radiation. The presence of conjunctival hyperemia was scored in arbitrary units from 0 (absence) to 4 (maximal vasodilation). ³⁶ The apparition of epithelial defects in the ocular surface was assessed with the slit lamp. Fluorescein staining was performed in some animals to confirm the presence of corneal epithelial damage. ³⁶  

Blinking rate, time of eyelid closure, and number of wiping movements directed to the eye were explored as a quantitative assessment of pain (nocifensive response). These parameters can be easily determined in the awake guinea pig and reflect the eye irritation–related behavior of the animal. ³ Nocifensive response parameters were assessed in real time, watching the animal freely moving in an open-ceiling cage during a period of 5 minutes. Immediately afterward, tearing rate was measured in one eye (the irradiated eye in UV-treated animals) using 30-mm-length phenol red threads (Zone-Quick; Menicon, Nagoya, Japan) placed in the nasal side of the lower lid for 30 seconds. 

In a set of experiments, one eye of control and UV-irradiated animals (the irradiated eye in this case) was treated with a 10-μL drop of a solution containing the TRPV1 agonist capsaicin (100 μM), the TRPA1 agonist allyl isothiocyanate (AITC, 10 mM), or the TRPM8 agonist menthol (200 μM; all from Sigma-Aldrich, St. Louis, MO, USA). Blinking frequency, time of eyelid closure in seconds, and tearing rate were measured immediately after the test solution instillation as described above. 

Data Analysis

Data were collected and processed for statistical analysis (SigmaStat, v3.5; Systat Software, Point Richmond, CA, USA). Data are expressed as mean ± SEM, being n the number of explored nerve terminals, fibers, or animals. Changes of firing in response to CO₂ pulses or temperature changes (cooling or heating) in control and UV-irradiated eyes were compared using parametric and nonparametric statistical tests, as indicated. The z-test or Fisher's exact test were used to compare proportions, as indicated. A P value below 0.05 was considered significant. Statistical analysis did not show differences in any of the parameters measured between albino and pigmented animals; therefore data from both types of animals were pooled. 

Results

Corneal Nerve Impulse Activity

A total of 191 corneal nerve fibers (103 mechanonociceptors and 88 polymodal nociceptors) and 53 cold-sensitive nerve terminals (cold thermoreceptors) were recorded from control eyes, and 170 corneal nerve fibers (86 mechanonociceptors and 84 polymodal nociceptors) and 57 nerve terminals (cold thermoreceptors) were recorded from UV-irradiated eyes (Tables 1 1552–3). 

Table 1

 

Spontaneous Activity of Corneal Nociceptive Fibers Recorded From Control and UV-Irradiated Eyes at 24 and 48 Hours After UV Radiation With Different Energies

	Control Eyes	UV-Irradiated Eyes
		24 h After		48 h After
		500 mJ/cm2	1000 mJ/cm2	500 mJ/cm2	1000 mJ/cm2
Mechanonociceptors
Mean discharge of active units, imp/s	0.5 ± 0.1	0.04	0.7 ± 0.0	0.0 ± 0.0	0.7 ± 0.5
Spontaneously active units/recorded units	4/103	1/31	2/17	0/11	5/27
% of active units	3.9	3.1	11.7	0	18.5*
Polymodal nociceptors
Mean discharge of active units, imp/s	0.5 ± 0.2	0.9 ± 0.5	0.0 ± 0.0	1.2	0.3 ± 0.2
Spontaneously active units/recorded units	8/88	2/26	0/9	1/8	9/41
% of active units	9.1	8.3	0	12.5	22
No. of eyes	34	11	5	3	11

*   P < 0.05, z-test, compared with control.

Table 2

 

Characteristics of the Impulse Response of Polymodal Nociceptors From Control and UV-Irradiated Eyes to Chemical Stimulation With 30-Second Pulses of 98.5% CO₂ Applied to the Corneal Receptive Field

Response to Chemical Stimulation, CO2 Pulses	Control Eyes	UV-Irradiated Eyes
		24 h After		48 h After
		500 mJ/cm2	1000 mJ/cm2	500 mJ/cm2	1000 mJ/cm2
Latency, s	13.4 ± 1.4	12.1 ± 2.0	8.9 ± 3.3	9.7 ± 2.8	8.1 ± 1.4†
Firing response, imp/s	1.7 ± 0.2	2.4 ± 0.6	3.0 ± 1.0	2.1 ± 0.3	3.4 ± 0.5*
Postdischarge, imp/s	1.8 ± 0.5	1.4 ± 0.7	2.4 ± 1.1	1.4 ± 0.3	1.6 ± 0.3
No. of fibers	37	19	8	8	45

*   P < 0.05, † P < 0.01, Mann-Whitney Rank Sum test, differences from control.

Table 3

 

Characteristics of the Impulse Response to Cooling and Heating Pulses Measured in Cold Nerve Terminals From Control and UV-Irradiated Eyes

	Control Eyes	UV Irradiated Eyes
		24 h After		48 h After
		500 mJ/cm2	1000 mJ/cm2	500 mJ/cm2	1000 mJ/cm2
Ongoing activity at 34°C, imp/s	8.6 ± 0.5	10.2 ± 1.2	9.2 ± 1.0	7.5 ± 0.8	6.1 ± 0.9*
Cooling from 34°C to 20°C
Cooling threshold, change in temperature, in °C	−2.6 ± 0.3	−3.7 ± 1.5	−2.9 ± 0.3	−1.9 ± 0.3	−2.6 ± 0.4
Peak frequency, imp/s	29.8 ± 1.3	25 ± 2.8	24 ± 1.5*	26.0 ± 4.1	18.9 ± 1.8‡
Temperature decrease to reach peak frequency, change in temperature, in °C	−6.1 ± 0.4	−4.8 ± 1.6	−4.7 ± 1.2	−4.3 ± 0.8	−5.6 ± 0.7
Cooling from 50°C to 34°C
Cooling threshold to resume NT activity, change in temperature, in °C	−14 ± 0.5	−16.1 ± 0.9	−15.9 ± 1.1	−11.6 ± 0.8	−10.6 ± 1.8‡
Peak frequency, imp/s	19.7 ± 1.3	16.6 ± 2.7	18.6 ± 1.7	11.8 ± 1.5†	12.2 ± 1.6**
Temperature decrease to reach peak frequency, change in temperature, in °C	−16.7 ± 0.4	−19.7 ± 1.5	−18.8 ± 1.8†	−13.8 ± 0.5†	−13.5 ± 0.7‡
Heating from 34°C to 50°C
Silencing threshold, change in temperature, in °C	4.1 ± 0.3	4.1 ± 0.9	4.0 ± 0.7	1.9 ± 0.3‡	1.8 ± 0.3‡
Incidence of paradoxical response, % of terminals	15.0	12.5	0	22.0	21.0
No. of terminals	53	13	16	9	19
No. of eyes	18	4	4	3	6

*   P < 0.05, † P < 0.01, ‡ P < 0.001, t-test or Mann-Whitney Rank Sum test, differences from control.

Mechano- and Polymodal Nociceptors.

Spontaneous Activity.

The percentage of mechanonociceptors and polymodal nociceptors exhibiting spontaneous activity at the beginning of the recording increased (P = 0.026 and P = 0.083 for mechano- and polymodal nociceptors, respectively, z-test) 48 hours after UV irradiation at the maximal intensity (Table 1). The mean firing frequency of the spontaneously active mechano- and polymodal nociceptive fibers was not significantly different 24 or 48 hours after UV radiation (Table 1). 

Response to Mechanical Stimulation.

The mechanical threshold of mechanonociceptors was significantly reduced after UV radiation (P = 0.016; Mann-Whitney Rank Sum test, Fig. 3). In contrast, mechanical threshold of polymodal nociceptors was not significantly modified after UV radiation and even showed a slight increase 24 hours after exposure to 500 mJ/cm² (P = 0.045; Mann-Whitney Rank Sum test, Fig. 3). 

Figure 3

 

Mechanical threshold measured in mechano- and polymodal nociceptors from control eyes and 24 hours or 48 hours after exposure to different UV radiation energies. Data from control, nonirradiated animals are shown in the “0” UV intensity column. *P < 0.05, Mann-Whitney Rank Sum test, differences from control (0 intensity).

Response to CO₂ .

After UV, the mean firing response (imp/s) of polymodal nociceptors elicited by CO₂ pulses augmented in comparison with nonirradiated corneas, as illustrated in the example shown in Figure 1, an increment that was proportional to the intensity of irradiation. Differences became significant 48 hours after 1000 mJ/cm² (Table 2; Fig. 4). Likewise, the latency of the impulse discharge evoked by CO₂ also decreased after exposure to UV, becoming significantly shorter 48 hours after 1000 mJ/cm² UV radiation (Table 2; Fig. 1). The firing frequency of the postdischarge was not significantly modified by UV exposure (Table 2; Fig. 1). 

Figure 4

 

Increment of the firing activity (imp/s) of polymodal nociceptors in response to chemical stimulation of the cornea with a 30-second duration CO₂ pulse in control eyes and 24 hours or 48 hours after exposure to UV radiation of different intensities. *P < 0.05, t-test or Mann-Whitney Rank Sum test, differences from control (0-intensity column).

Cold Thermoreceptors.

The effects of UV radiation on ongoing activity of corneal cold thermoreceptors at basal temperature (34°C) were more prominent 48 hours after UV radiation with 1000 mJ/cm², and consisted in a significant decrease of impulse frequency (Table 3; Figs. 5, 6A), although the firing pattern was not modified, as evidenced by the similarity of the NTI interval distribution before and after UV exposure (data not shown). Cooling threshold was not modified after application of different intensities of irradiation (Table 3; Fig. 6B). The peak frequency value of the response to cooling ramps from 34°C to 20°C was significantly lower 24 and 48 hours after UV radiation of 1000 mJ/cm² (Table 3; Fig. 6C); also, 48 hours after UV exposure to 500 or 1000 mJ/cm², the temperature required to silence cold receptor activity in response to a heating pulse to 50°C was significantly lower than in intact corneas (Table 3; Fig. 6D). 

Figure 5

 

Response to a cooling pulse of cold thermoreceptors recorded from a control cornea (left) and from a UV-irradiated cornea (right) 48 hours after 1000 mJ/cm². (A) Instantaneous frequency of discharge (1/interval duration) in Hz. (B) Impulse frequency histogram of discharge in impulses per second. (C) Temperature of the bath solution. (D) Recording of the NTI activity. Insets represent expanded views of the recording during 1 second in indicated moments of the recording.

Figure 6

 

Values of the different parameters of the response of cold thermoreceptors to cooling and heating stimulation, 24 to 48 hours after exposure to different intensities of UV radiation. (A) Ongoing activity at basal temperature (34°C). (B) Decrement in temperature required to increase by 25% the basal firing frequency of discharge at 34°C (cooling threshold). (C) Peak frequency of discharge in response to cooling ramps from 34°C to 20°C. (D) Increase of temperature needed to silence NTI firing (silencing threshold) during heating ramps from 34°C to 50°C. *P < 0.05, †P < 0.01, ‡P < 0.001, t-test or Mann-Whitney Rank Sum test compared with control (0 mJ/cm² UV radiation intensity).

Nocifensive Response and Clinical Signs of Inflammation

Conjunctival hyperemia increased with higher UV radiation intensities (see Table 4 for details). Punctate fluorescein staining scattered throughout the surface of the cornea and exposed conjunctiva, was also observed in irradiated animals (see Table 4 for details). No behavioral responses suggestive of irritation or pain (eyelid closure, eye-wiping movements) were observed 24 and 48 hours after exposure to UV. Also, blinking frequency and tearing rate were not modified 24 hours after 1000 mJ/cm² UV-irradiation (n = 8, paired t-test; data not shown). Forty-eight hours after 1000 mJ/cm² UV exposure, blinking rate was significantly increased (0.7 ± 0.2 vs. 1.4 ± 0.3 blinks/min, before and after UV exposure, respectively, n = 24; P < 0.05, paired t-test), whereas tear secretion rate was not modified (13.1 ± 1.1 vs. 13.9 ± 1.2 mm, before and after, n = 27; P = 0.68, paired t-test). 

Table 4

 

Hyperemia and Fluorescein Staining Before and After UV Irradiation With Different Intensities

	Before UV	After UV Exposure
		24 h After		48 h After
		500 mJ/cm2	1000 mJ/cm2	1000 mJ/cm2
Hyperemia	0 ± 0	2.3 ± 0.3	2.8 ± 0.5	2.1 ± 0.4
	0/14	3/13	4/11	7/12
		23%	36%	58%
Fluorescein staining	0 ± 0	5/5	3/3	1/2
	0/10	100%	100%	50%

 Data shown in each row are mean ± SEM; eyes with clinical sign of inflammation/total number of explored eyes; % of eyes showing hyperemia or staining.

Nocifensive Response to Ocular Instillation of TRPV1, TRPA1, and TRPM8 Agonists.

The effect of capsaicin, AITC, and menthol on tear secretion, eyelid closure, and blinking rate was measured in control and UV-irradiated animals 48 hours after 1000 mJ/cm². Phenol red threads became saturated (that is, fully wetted in less than 30 seconds) in 50% of the measurements performed after capsaicin treatment both in control and UV-irradiated eyes, in 33% of the measurements after AITC in irradiated eyes and in 8% of the measurements after menthol in control eyes. 

Nocifensive responses to instillation of the TRPV1 and TRPA1 agonists, capsaicin, and AITC were increased in UV-irradiated animals compared with controls, whereas tearing rate in response to TRPM8 agonist menthol was significantly decreased in UV-irradiated eyes (Fig. 7). 

Figure 7

 

Time of eyelid closure, blinking, and tearing rates produced by ocular instillation of the TRP channel agonists capsaicin (100 μM), AITC (10 mM), and menthol (200 μM) in control animals (nonirradiated) and in UV-irradiated animals 48 hours after exposure to 1000 mJ/cm² UV radiation. Normal basal values of eyelid closure, blinking, and tearing rate in control, untreated guinea pigs are shown in black columns for comparison. *P < 0.05, §P < 0.005, t-test or Mann-Whitney Rank test, differences between UV-irradiated and control animals.

Discussion

The present results provide evidence that exposure of the corneal surface to UV-C radiation causes moderate ocular surface inflammation signs and modifies the spontaneous and stimulus-evoked activity of corneal sensory receptors. Altogether, responsiveness of polymodal- and mechanonociceptors augmented, whereas activity of cold thermoreceptors was reduced. The magnitude of impulse firing disturbances was associated with the intensity of the UV radiation and varied for each receptor type. 

Human subjects exposed to UV-C radiation with an absorbed energy of 700 mJ/cm² reported ocular symptoms 2 to 19 hours after irradiation; 77% of the individuals developed ocular burning sensation and approximately 30% reported tearing and pain sensations lasting 2 to 4 days. ³⁰ Likewise, in our guinea pig model, as also occurs in rabbits, ³⁶ conjunctival hyperemia (present in approximately 30% of the animals) and augmented nocifensive responses developed 24 to 48 hours after UV irradiation, although tear secretion did not increase significantly. 

Ultraviolet radiation is known to produce the sensitization of polymodal nociceptors in the skin, ^37,38 increasing spontaneous activity and responsiveness to heat and chemical stimuli and causing thermal and mechanical hyperalgesia and allodynia. A difference in mechanical threshold between mechano- and polymodal nociceptors after UV radiation has been reported in the rat's skin, ^37,38 although no significant change in spontaneous firing of nociceptors was observed following UV exposure. ³⁹ In the present experiments, we also observed differences in mechanical threshold between mechano- and polymodal corneal nociceptors and sensitization to mechanical stimulation after UV radiation, proportional to the intensity of exposure. Sensitization was evidenced by the decrease in mechanical threshold and the increase in the percentage of spontaneous activity incidence among mechanonociceptors. Sensitization of polymodal nociceptors to chemical stimuli also developed after UV radiation. Although no significant increase of the mean frequency of spontaneous activity of individual polymodal nociceptors was observed, the percentage of nociceptor fibers displaying spontaneous activity was slightly higher. This represents an overall higher input of peripheral nociceptors to second-order corneal trigeminal nucleus neurons, thus explaining the enhanced background impulse activity recorded in such neurons after ocular exposure to UV radiation. ⁴⁰ The enhanced nociceptor activity, although not very prominent, is expectedly sustaining the spontaneous discomfort/pain sensations experienced by human subjects exposed to UV radiation; recordings in human sensory fibers have shown that conscious sensation of pain is evoked when background nociceptor fiber firing frequency overpasses 0.4 imp/s. ⁴¹ Furthermore, the higher incidence of ongoing firing and augmented responsiveness of polymodal nociceptors, combined with the reduced threshold of mechanonociceptors, also explains the presence of mechanical and chemical hyperalgesia observed 24 hours after UV exposure. ^39,42  

There is ample evidence of epithelial damage by UV radiation, with the generation of reactive oxygen species and oxidative stress ^43,44 that are detrimental to corneal epithelial cells and may lead ultimately to apoptosis. Proinflammatory molecules, such as interleukins, cytokines, matrix metalloproteinases, nuclear factor-κB, or nitric oxide, are locally released in the cornea ^12,45–50 accompanied by a significant elevation of interleukins and TNF-α concentrations in tears. ^43,44 Ultraviolet radiation may cause direct damage to ion channels expressed by ocular nerve axons and terminals affecting their excitability. ^51–53 Although such disturbances could contribute to the altered responsiveness of corneal nerves observed in our experiments, this appears to be primarily the result of the sensitizing effect of inflammatory mediators released by cell damage, with neurogenic inflammation further contributing to the overall inflammatory reaction evoked by UV radiation. ^36,54,55  

Nociceptor sensitization involves modification in the expression and activity of different membrane ion channels. TRPV1, the capsaicin receptor primarily expressed by nociceptive sensory ganglion neurons, is directly activated and sensitized by chemical mediators released during tissue injury, such as protons, ATP, bradykinin, prostaglandins, serotonin, or NGF. TRPV1 is particularly important for signaling increased sensitivity to non-noxious (allodynia) or noxious (hyperalgesia) stimuli after tissue damage. ^56–58 The enhanced response to capsaicin observed after UV irradiation, in particular the increased time of eyelid closure and tearing flow, and the augmentation of polymodal nerve impulse discharges evoked by CO₂ in UV-irradiated eyes suggest that UV radiation increases the expression and/or the activity of TRPV1 channels in corneal sensory nerve terminals, as it has been already reported in skin cells. ^34,59  

TRPA1 is another ion channel involved in chemical and mechanical nociception ^60–63 contributing to enhanced pain after tissue injury caused by free radical molecules. ⁶⁴ Increased expression of TRPA1 during inflammation has been associated with the development of mechanical hyperalgesia. ^60,63,65 Thus, TRPA1 activation could explain the decrease in mechanical threshold of mechanonociceptors seen in this study and the augmented nocifensive response to its agonist AITC (increased time of eyelid closure and reflex tearing) observed in irradiated animals. Nevertheless, the contribution of other ion channels involved in mechanical transduction or of voltage-activated channels modulating neuronal excitability cannot be excluded. ^66–68  

TRPM8, the cold-sensing channel, ⁶⁹ is inhibited by bradykinin, histamine, and other inflammatory agents, which decrease the gating probability of the channel through a direct inhibitory effect mediated by the G protein subunit Gαq. ²⁵ Expectedly, a similar effect is produced by the inflammatory mediators released after UV exposure. This mechanism can explain the inhibition of cold sensory receptor activity observed in our experiments, also seen in other inflammatory models ^3,25,70 and the reduced effect of the TRPM8 agonist menthol on tearing rate that we observed in UV-irradiated animals. 

Perceptual characteristics of ocular discomfort and pain are determined in initial stages by the convergence of nociceptive and non-nociceptive (mechanical, thermal) input on medullary dorsal horn, second-order neurons. ^71,72 There, neurons experience profound functional changes after peripheral inflammation or nerve injury. ⁵⁶ Our data do not provide information about the relative contribution of sensory inflow from peripheral polymodal and cold thermoreceptors to the final perceptual aspects of ocular inflammatory pain. Innocuous cooling of the skin produced analgesia in healthy mice but caused cold allodynia when inflammation is present; this effect was markedly attenuated in animals where TRPM8-expressing sensory neurons had been selectively ablated. ^73,74 Hence, it can be speculated that during corneal inflammation, reduction of TRPM8 cold thermoreceptor input helps to attenuate cold allodynia. 

Altogether, the present results support the interpretation that discomfort sensations experienced by humans after exposure to UV radiation are the consequence of sensitization of mechano- and polymodal nociceptors by locally released inflammatory agents, possibly mediated through changes in the activity of TRPA1 and TRPV1 channels in parallel with an inhibition of TRPM8 channels, decreasing the activity of cold thermoreceptors. 

Acknowledgments

The authors thank Manuel Bayonas and Alfonso Pérez-Vegara (deceased) for skillful technical assistance. 

Supported in part by Grants GV/2007/030 (MCA) from Generalitat Valenciana, Spain, and SAF2011-22500 (JG), and by Grant BFU2008-04425 (CB) from the Ministerio de Economía y Competitividad, Spain. 

Disclosure: M.C. Acosta, None; C. Luna, None; S. Quirce, None; C. Belmonte, None; J. Gallar, None 

Footnotes

 CL and SQ contributed equally to the work presented here and should therefore be regarded as equivalent authors.