Impulse Activity in Corneal Sensory Nerve Fibers after Photorefractive Keratectomy

Abstract

purpose. To evaluate the changes in spontaneous and stimulus-evoked nerve impulse activity of corneal polymodal and mechanonociceptor sensory fibers of the cornea after photorefractive keratectomy (PRK).

methods. A central corneal ablation 6 mm in diameter and 70 μm in depth was performed with an excimer laser in both eyes of three anesthetized cats, after removal of the corneal epithelium. Single nerve fiber activity was recorded in these animals 12 to 48 hours after surgery. Activity in corneal nerve fibers with receptive fields (RFs) within and/or close to the wound, as well as with RFs far from the lesioned area, was studied. Incidence and frequency of spontaneous discharges and nerve impulse firing responses to mechanical (Cochet-Bonet esthesiometer) and chemical (CO₂ gas pulses) stimuli were studied.

results. The incidence of nociceptor fibers exhibiting ongoing activity (15/35 vs. 1/9) and the frequency of their spontaneous firing (0.25 ± 0.09 impulses [imp]/s versus 0.08 ± 0.08 imp/s) was higher in fibers with RFs within and/or bordering the wounded area than in those with RFs far away from the wound. Mechanical responsiveness of fibers with RFs within or nearby the ablated area was often reduced. In these fibers, CO₂ pulses evoked a lower-frequency impulse discharge (0.9 ± 0.2 imp/s inside, 2.3 ± 0.7 imp/s outside the wound). CO₂-evoked discharges recorded from fibers innervating the intact wound border were similar to those recorded in corneal fibers of intact cats.

conclusions. The spontaneous impulse activity and the abnormal responsiveness shown by a part of the corneal nerve fibers innervating the injured cornea are presumably the neurophysiological substrate of the pain sensations experienced by human patients hours after PRK surgery.

Photorefractive surgery, and other procedures such as phototherapeutic keratectomy (PTK) performed with excimer laser, has become a widely used procedure for treating myopia, hypermetropia, astigmatism, scars, and other diseases of the cornea. ^{1 2} This type of surgery—in particular photorefractive keratectomy (PRK)—is accompanied by severe ocular pain that becomes strongest 24 hours after PRK and is described as a throbbing, burning, and/or stinging pain usually accompanied by nasal congestion, tearing, and photophobia. ³ Postsurgical acute pain is often followed by less-intense discomfort sensations that may persist for weeks or months after surgery. ^{4 5}  

Photorefractive surgery causes injury to the epithelial and stromal cells of the cornea and to corneal sensory nerve branches running in the lesioned tissues and causes various degrees of local inflammatory reaction. ^{6 7 8} The cornea is innervated by sensory fibers that have their origins in different functional types of trigeminal ganglion neurons: mechanonociceptor fibers activated by mechanical forces; polymodal nociceptor fibers that respond to mechanical, thermal, and chemical noxious stimuli; and cold receptor fibers that respond primarily to temperature reductions on the corneal surface. ^{9 10 11 12 13} After injury, nociceptor fibers of the skin and other somatic tissues are often the source of spontaneous pain sensations. ¹⁴ In addition, release by the injured tissues of inflammatory mediators activates and/or sensitizes intact and damaged nociceptor endings, further contributing to spontaneous pain and to development of hyperalgesia and neurogenic inflammation. ^{15 16} It is conceivable that a similar process takes place in the cornea subsequent to PRK and other surgical procedures, as a result of the accompanying injury of epithelial and subepithelial corneal nerves and of epithelial and stromal cells. In fact, an enhancement of mass-impulse activity in corneal sensory fibers in the rabbit after photorefractive surgery has been reported. ¹⁷ However, the effects of surgical injury on the response characteristics of the various functional classes of corneal sensory fibers are unknown. The purpose of this work was to study the change in spontaneous and stimulus evoked activity of nociceptor fibers that innervate the cornea in cats, 12 to 48 hours after PRK. 

Materials and Methods

Photorefractive Keratectomy

The standard PRK procedure applied to human corneas was performed by an experienced ocular surgeon (ARG) on both eyes of three adult cats (two male, one female; 2.8–3.7 kg), anesthetized with pentobarbital sodium (40 mg/kg; intraperitoneal [IP] injection). The animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The corneas of both eyes were additionally anesthetized by topical instillation of 0.1% tetracaine and 0.4% oxybuprocaine chlorhydrate. The epithelium of the central cornea was removed manually with a sterile microsponge, and an ablation (6.0-mm diameter and 70-μm depth) was made in the center of the cornea with a single-beam excimer laser with 193-nm emission wavelength, 10-Hz fixed-pulse repetition rate, and 180-mJ/cm² radiance exposure. 

Nerve Recordings

Twelve to 48 hours after surgery, the cats were anesthetized again with an IP injection of pentobarbital sodium (40 mg/kg) and kept in an areflexic state throughout the experiment by continuous intravenous infusion of diluted pentobarbital (5 mg/kg) through the saphenous vein. Animals breathed spontaneously through a tracheal cannula. End-tidal CO₂, rectal temperature, and blood pressure were continuously monitored and maintained within physiological limits. At the end of the experiment, the cats were killed with an overdose of the anesthetic. 

Using Ag-AgCl electrodes and conventional electrophysiological equipment, we made extracellular recordings from single corneal afferent fibers dissected from the mixed ciliary nerves in the orbital cavity of the cat's eye, as described elsewhere. ¹⁸ Corneal sensory fibers were first identified by their response to slight mechanical stimulation of the corneal surface with a fine wet brush. The mechanical threshold was subsequently measured with a Cochet-Bonnet esthesiometer provided with a no. 12 filament (0.1–1.9 mN) or with calibrated von Frey hairs (0.002–2.0 N). Receptive fields were then mapped with suprathreshold force levels. Sensitivity to chemical stimulation was ascertained by applying on the receptive field jets of gas containing 98% CO₂ for 30 seconds at a flow of 80 mL/min. ¹⁹  

The conduction velocity (CV) of the recorded fibers was calculated, measuring the latency of the nerve impulse evoked by an electric shock (0.1–0.5 ms duration, 0.5–3 mA) applied on the receptive field area with a pair of silver electrodes spaced 3 to 5 mm apart. Conduction distance was estimated by placing an 8.0-gauge thread along the trajectory of the nerve. 

Neural discharge and stimulating pulses were recorded on an FM magnetic tape for off-line computer analysis by the appropriate software (CED 1401 plus and Spike2; Cambridge Electronic Design Ltd. Cambridge, UK). The impulse frequency of the spontaneous activity (mean discharge rate, in impulses/second) and the changes in the firing response evoked by CO₂ pulses were measured. The parameters measured included the mean discharge rate during the 30-second stimulation period (in impulses/second), the peak frequency value (in impulses/second), and the mean discharge rate during a 30-second period after the CO₂ pulse (postdischarge frequency, in impulses/second). To avoid the unnecessary deaths of a large number of animals, data from nonsurgical adult cats collected in two previous studies ^{9 18} performed under similar experimental conditions have been used to compare the functional properties of fibers innervating surgical and intact corneas. The data represent the “control corneas group.” 

Data are presented as mean ± SEM. Both parametric and nonparametric statistical tests were applied, as indicated in Results. 

Results

General Properties of Corneal Nerve Fibers

The receptive fields (RFs) of 50 units recorded in eyes with ablated corneas were mapped with suprathreshold mechanical stimulation and their size measured (Table 1) . Based on the location of the points that evoked an impulse response, fibers were classified into four groups: group A fibers respond exclusively within the ablated area in the central cornea (n = 1); group B fibers respond both inside and outside the ablated area (n = 11); group C fibers respond to stimulation outside the ablated area, with the RF borders reaching the limits of the ablation (n = 29); and group D fibers have RFs located in the peripheral cornea, distant from the ablated area (n = 9; Fig. 1 ). 

Most of the sensory units identified by mechanical stimulation of the center of the cornea with a wet brush had their RFs both outside and inside the ablation or only outside but limited by the wounded area (groups B and C). Six of the 40 corneal units of these groups responded only to mechanical stimulation (mechanonociceptor units) whereas the remaining 34 were also activated by CO₂ (polymodal units). The only fiber included in group A responded exclusively to mechanical stimulation. Nine fibers had RFs distant from the wound (group D), three of them were only mechanosensitive, and six responded to mechanical and CO₂ stimulation. 

CV was measured in 43 units and ranged from 0.40 to 14.0 m/s. Twenty-seven of the units were A-delta fibers (CV > 2 m/s) and the rest (n = 16) C fibers (CV < 2.0 m/s; Table 1 ). 

Ongoing Activity

Of the fibers in groups A, B, and C, 43% showed a variable degree of irregular spontaneous activity at rest. The ongoing activity in the fibers of these groups was higher than in those with RFs located far from the wounded area (group D), in which only one fiber fired spontaneously and at a lower frequency than did fibers innervating RFs inside or near the wounded area (Table 2 , Fig. 2A ). 

Responses to Mechanical Stimulation

Nineteen (59%) of 32 of the fibers in the central cornea (groups A, B, and C) were activated by mechanical stimulation with the brush, although they did not respond to punctate stimulation applied with the Cochet-Bonnet esthesiometer within the working range of the instrument (Table 2 , Fig. 2B ). Some units in group B (with RFs inside and outside the ablated area) responded to the esthesiometer but only in the portion of the RFs located outside the wound. About half (4/10) of the group-D fibers responded to mechanical stimulation with the Cochet-Bonnet esthesiometer, with the same mean threshold as the fibers with RFs close to the wounded area (group B). The remaining five fibers in group D exhibited mechanosensitivity to application of the brush but did not respond to the Cochet-Bonnet esthesiometer. 

Response to Chemical Stimulation

In 14 fibers with RFs covering areas inside and outside the wound (group B) a pulse of 98% CO₂ was applied on the RFs inside and outside the ablated area (Fig. 3) . The mean impulse frequency of the response to CO₂ inside the wound was significantly lower than when the stimulus was applied in the intact portion of the RF (Table 3 , Fig. 2C , Fig. 3 ). It was also below the mean firing frequency of the response to the same stimulus in polymodal nociceptors of intact cat's corneas recorded in experiments performed previously by our research group (Fig. 2C) . 

Discussion

Our results show that 24 to 48 hours after PRK, corneal sensory fibers innervating the wounded area and their surroundings appeared to be functionally altered. The number of fibers displaying ongoing activity was higher than in fibers with RFs far from the wound or in fibers innervating intact corneas. ^{9 10 18} Moreover, fibers responding to stimulation of the wounded area exhibited an abnormal response to mechanical stimulation and a lower activation by CO₂. Even units with RFs far from the wounded area displayed mechanical thresholds that were comparatively higher than those of healthy corneas. ¹⁸ Some corneal nociceptors responded to the mechanical stimulus with the brush, but not to the Cochet-Bonnet esthesiometer. It is well known that corneal sensory fibers respond more effectively to the sliding of a brush on the surface than to a punctate mechanical stimulus. ^{20 21} The differing responses may be due to the need to recruit a minimal number of nerve terminal branches of a single parent axon to evoke a propagated action potential. This recruitment is achieved more effectively with a moving mechanical stimulus. The Cochet-Bonnet esthesiometer has a tip surface of 0.0113 mm² and therefore causes indentation of a very small corneal area. A stronger force would thus be required to produce enough deformation to stimulate the number of nerve terminals necessary to evoke a propagated nerve impulse. 

PRK produces a direct injury of the nerves of the treated corneal area including those of the epithelium and upper stromal layers. ^{22 23} Although the dynamics of the degeneration and regeneration processes in the cat's corneal nerves after PRK has not been studied in detail, experimental damage of corneal nerves in this species through surgical or freezing wounds caused degeneration of the distal segment of the injured nerve branches and formation of nerve bulbs and growth cones at the axotomized branches. Shortly after the lesion, intact nerves in the vicinity of the wound begin to send sprouts toward the injured area. ^{12 24 25 26} Changes also occur in the soma of injured corneal neurons, including enhanced expression of the proto-oncogen c-Jun whereas expression of neuropeptides is not markedly altered. ^{12 27 28}  

In surgical corneas, mechanical threshold in fibers with central and peripheral RFs were similar. The thresholds were in general higher than in the corresponding fibers of control, healthy corneas (see Fig. 2 ), although differences did not reach significance. Most corneal nerve fibers in the cat exhibit large RFs that may cover a quadrant of the cornea. ²¹ It is conceivable that surgical damage still affects some of the peripheral branches of axons whose receptive field center is in the periphery of the cornea, thereby reducing their mechanical sensitivity due to an overall reduction of their total arborization. 

The development of spontaneous activity in the central cut end of injured sensory axons is well documented in skin nerves where ongoing activity starts within hours after the lesion and reach a maximum 1 day afterward. ^{29 30 31} Also, responsiveness of these injured fibers to mechanical stimulation and their sensitivity to exogenous chemicals and endogenous inflammatory mediators change markedly. ^{31 32 33} These functional alterations are the result of the up- or downregulation in axotomized sensory neurons of various subclasses of ion channels. For instance, Nav1.3 mRNA is upregulated in adult rat dorsal root ganglion (DRG) neurons 16 hours after axotomy or spinal nerve ligation in coincidence with the development of ectopic discharges in these neurons, whereas other Na⁺ channels (Nav 1.7, 1.8, and 1.9) decline after injury. ^{34 35 36 37 38} Altered responsiveness of DRG neurons after axotomy persists for several weeks after injury. ³⁹  

Likewise, the enhanced spontaneous activity and altered responsiveness to natural stimuli observed in PRK corneas are presumably due to changes in the excitability of corneal trigeminal ganglion neurons consecutive to the damage of the peripheral axons. The laser beam destroys the terminal portion of the corneal sensory nerves containing the molecules that transduce physical and chemical stimuli into propagated nerve impulses. ⁴⁰ The loss of stretch-activated channels explains the lack of responsiveness to mechanical forces of axotomized nerve fibers at the wounded area. In turn, loss of ion channels involved in acidic transduction such as TRPV1 ⁴¹ and ASICs ⁴² predictably eliminates the response to CO₂ of cut nerve fibers. Activation by protons of some of these channels and of other voltage-gated channels in the central stump of axotomized nerve fibers could be the cause of the nonspecific, residual responsiveness to CO₂ in injured fibers of the ablated area. In agreement with the altered responsiveness in feline corneal nerve fibers after PRK, esthesiometry of human corneas subjected to PRK evidence a marked reduction of mechanical and chemical sensitivity of the injured region. ⁴³  

In parallel to the decreased ability to transduce natural stimuli, corneal fibers with RFs inside the PRK wound, exhibited a pronounced increase in ongoing activity compared with the level found in the intact cornea of the cat. ^{9 18} There is abundant experimental evidence that sustained, low-frequency firing of peripheral nociceptor fibers consecutive to tissue injury and/or peripheral inflammation evokes spontaneous pain sensations. ^{44 45 46} This activity is also augmented when the receptive field of the fiber is exposed to external stimuli such as dryness and cooling caused by evaporation between blinks, air currents, or even sliding of the eyelid over the corneal surface. These responses would also explain the beneficial effect of contact lenses in preventing postsurgical discomfort and pain as lenses reduce the exposure of injured nerve endings to these stimuli. 

Therefore, the intense pain sensations that appear hours after PRK in humans are attributable to the enhanced spontaneous activity developed by injured nerve fibers. Moreover, the local inflammatory reaction in the surgical tissues releases endogenous mediators that may contribute to the maintenance of ongoing activity at the cut nerve fibers of the injured area but also to sensitization of intact nerve terminals located in neighboring areas. ^{16 47} No obvious sensitization of fibers innervating the periphery of the injury was observed, suggesting that the contribution of inflammation to spontaneous activity was less important than the direct effect of nerve injury. 

The presence of spontaneous impulse activity and altered responsiveness of nerve fibers innervating PRK-treated corneas can explain the pain sensations described by patients within hours after ocular surgery. ^{48 49} It is well established that pain that follows acute tissue injury is caused by the build-up of a sustained, low-frequency impulse in peripheral nociceptor fibers, which causes sensitization of the higher-order neurons within the central nervous system (CNS) involved in nociception and pain. ⁴⁶ Disturbances of nerve excitability presumably persist in injured and regenerating nerve fibers of surgically treated corneas for long periods and would be the cause of the altered sensitivity and aberrant sensations experienced by some patients, weeks and even months after corneal surgery. ^{3 5 48}  

 Supported by SAF2005-07277 (JG), BFU2005-08741 (CB) and Fundación Marcelino Botín, Spain.

 Submitted for publication January 5, 2007; revised April 11, 2007; accepted June 26, 2007.

 Disclosure: J. Gallar, None; M.C. Acosta, None; A.R. Gutiérrez, None; C. Belmonte, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Functional Properties of Corneal Nociceptive Units Recorded after PRK

	Central Cornea (Groups A, B, C)	Peripheral Cornea (Group D)
CV (m/s)	3.5 ± 0.5 [0.6–14] (n = 36)	3.6 ± 1.1 [0.4–8] (n = 7)
RF diameter (mm)	7.3 ± 0.6 [3–12] (n = 23)	6.5 ± 0.5 [6–7] (n = 2)

 Data are expressed as the mean ± SEM; the range is shown in brackets.

Figure 1.

 

Location of the RFs of the units recorded after PRK. Shaded circle: wounded area in the center (dotted circle) of the cornea and the sclera (solid circle), respectively. A: RF inside the laser-ablated area. B: RF inside and outside the laser ablated area. C: RF outside but near the ablated area. D: RF in the peripheral cornea, distant from the ablated area. RFs in some cases extended to the sclera (6/9 in group D and 6/40 in groups B and C).

Table 2.

 

Ongoing Activity and Response to Mechanical Stimulation of Corneal Nociceptive Units Recorded in PRK Corneas

	Central Cornea (Groups A, B, C)	Peripheral Cornea (Group D)
Ongoing activity (imp/s)	0.25 ± 0.09 (n = 35) 15/35	0.08 ± 0.08 (n = 9) 1/9
Mechanical threshold (mN)	2.17 ± 0.84 [0.22–9.24] 13/32	2.33 ± 1.25 [0.66–6.22] 4/9

 Data are expressed as the mean ± SEM. The number of responding units/number of explored units and range (in brackets) are shown. No statistically significant differences were found between mean mechanical thresholds in central and peripheral RF.

Figure 2.

 

(A) Ongoing activity, (B) mechanical threshold, and (C) response to CO₂ in central and peripheral areas of wounded corneas compared with control corneas from previous control experiments (n = 41 in A; n = 27 in B; n = 68 in C). ^{9 18} *P < 0.05, one-way ANOVA and Holm-Sidak test; differences from control.

Figure 3.

 

Response of polymodal corneal nociceptor fibers to a 30-second pulse of 98% CO₂ applied inside (A) and outside (B) the wound. Horizontal bar: duration of the stimulus. The unit recorded in (A) responded more when the stimulus was applied outside the wound, and in these conditions new polymodal nociceptor units were recruited by the acidic stimulus.

Table 3.

 

Response of Polymodal Nociceptors with RFs Inside and Outside the Wound (Group B) to CO₂ Applied Inside and Outside the Wound

Response to CO2	Inside the Wound	Outside the Wound
Mean discharge rate (imp/s)	0.9 ± 0.2*	2.3 ± 0.7
Peak frequency (imp/s)	5.9 ± 1.7	8.6 ± 2.1
Postdischarge (imp/s)	0.2 ± 0.1	0.6 ± 0.2

 Data are mean ± SEM (n = 14).

*  Paired t-test or Wilcoxon test, P < 0.005.

The authors thank Alfonso Perez-Vegara for providing technical assistance.