A Novel Class of Neurons at the Trigeminal Subnucleus Interpolaris/Caudalis Transition Region Monitors Ocular Surface Fluid Status and Modulates Tear Production

Surgical preparation

Male rats (270-440 gm; Sprague Dawley, Harlan) were anesthetized initially with pentobarbital sodium (70 mg/kg, i.p.) before surgery. The left femoral artery (blood pressure monitor) and jugular vein (anesthesia and drug infusions) were catheterized, and after tracheotomy, animals were artificially respired with oxygen-enriched room air. Anesthesia was maintained by a continuous infusion of methohexital sodium (35-40 mg · kg^-1 · hr^-1) and switched to a mixture of methohexital sodium (26-40 mg · kg^-1 · hr^-1) and the paralytic agent, gallamine triethiodide (14-32 mg · kg^-1 · hr^-1), after completion of all surgical procedures and just before the recording session. The animal was placed in a stereotaxic frame, and portions of the occipital bone and C1 vertebra were removed to expose the dorsal surface of the medulla. The brainstem surface was covered with warm mineral oil. Expiratory end-tidal CO₂ was monitored continuously and kept at 4-5% by adjusting tidal volume. Mean arterial pressure (MAP) remained above 100 mmHg throughout the experiment. Body temperature was maintained at 38°C with a heating blanket and thermal probe.

Electrophysiology recording techniques

Single neurons were recorded at the rostral Vi/Vc transition and laminas I-II at the caudal Vc/C1 junction region extracellularly using tungsten electrodes (9-15 MΩ; FHC, Bowdoinham, ME) as described previously (Meng et al., 1997; Hirata et al., 1999, 2003) (Fig. 1). Light mechanical (fine camel-hair brush) stimulation of the corneal surface was used to search for responsive neurons. Unit activity was amplified, displayed on a digital oscilloscope to monitor spike shape and amplitude, and passed through a window discriminator. Discriminated neural spikes, MAP, and a marker for CO₂ stimulus pulses were acquired and displayed online with an Apple computer (G3) through a DAQ interface board using LabVIEW software (National Instruments, Austin, TX). Data were also recorded on a four-channel DAT/SCSI-based acquisition system (Cygnus Technology) for further off-line analyses.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Experimental setup to measure tear volume and single-unit activity in response to ocular stimulation. The timing and concentration of CO₂ stimulus pulses as well as data collection were controlled by a computer and LabVIEW software. Drugs were delivered by pressure injection monitored visually by the movement of the meniscus. C, Cornea; Vc/C1, trigeminal subnucleus caudalis-cervical spinal cord junction region; Vi/Vc, trigeminal subnucleus interpolaris-caudalis transition region.

Units were tested for several general response properties and then for A- and C-fiber type corneal input (electrical stimuli applied from a bipolar electrode, 2 mm separation, mounted on the ear bar). Electrical evoked responses occurring after 30 msec were assumed to indicate C-fiber input (Hu, 1990; Meng et al., 1997). Convergent cutaneous receptive field (RF) properties were determined by applying innocuous mechanical stimulation and then noxious pinch and deep pressure to the ipsilateral face. Corneal units with an excitatory convergent cutaneous RF were classified as either low-threshold mechanoreceptive (LTM), wide dynamic range (WDR), or nociceptive-specific (NS) units. In addition, many corneal units at the Vi/Vc transition were classified as “complex,” defined by an inhibitory convergent cutaneous RF, which was contiguous with the cornea-conjunctiva border and occurred alone or in addition to an excitatory cutaneous RF. The inhibitory RF of complex units was mapped by light pinch stimulation with a small forceps and often included the ipsilateral intranasal mucosa. Neurons with no apparent cutaneous RF were classified as cornea only (CO) units. Although the term “corneal unit” is used throughout the text, no effort was made to distinguish between units responsive to corneal stimulation alone and those responsive to corneal plus conjunctival stimulation.

In some cases an electrode array was positioned at the SSN-facial motor nucleus region (SSN/VII) (Toth et al., 1999) to test for efferent projections. The array was angled rostral 2° off vertical at the following coordinates: 4 mm caudal to λ, 2 mm lateral, and 5-7 mm ventral to the cerebellar surface. Tests for antidromic activation were done after all other testing was complete to avoid possible persistent effects from high-intensity and high-frequency stimulation. Stimuli were presented from an array of two (1-2 mm rostrocaudal separation) concentric bipolar stimulating electrodes (SNE-100, Rhodes Medical Instruments). Antidromically evoked spikes were driven at a constant latency (<0.1 msec jitter), followed by high frequency (0.1 msec pulse, 200-300 Hz, 20 msec train duration), and were collided with orthodromically driven spikes occurring within a critical time window (Lipski, 1981). A stimulus intensity of 500 μA was defined as the maximum allowable current for specific activation. Sites of lowest current for antidromic activation were marked electrolytically (30 μA, 30 sec).

Corneal stimulation by carbon dioxide

The general setup for the experimental design is shown in Figure 1, and a detailed description of the delivery of CO₂ pulses to the corneal surface has been reported previously (Hirata et al., 2000). Briefly, variable concentrations (0-95%) of CO₂ gas were obtained by mixing the outflow from tanks of 100% CO₂ and air through a proportional gas mixer as monitored from the bleeder valve output by an infrared detector (Cap-Star 100, CWE, Ardmore, PA). Humidified CO₂ gas mixtures were delivered at a constant flow rate to the left cornea. The timing and duration of CO₂ pulses (40 sec duration, minimum of 15 min between pulses) were computer-controlled by LabVIEW software. In most experiments only the responses to 0% (air) and 80% CO₂ pulses were compared. In some cases, however, a range of CO₂ concentrations (0, 30, 60, 80, and 95%) was used to confirm that the threshold and slope of the neural stimulus-response curves were similar to those reported previously (Hirata et al., 1999). The cornea was kept moist with a pH-balanced artificial tear fluid during surgery and the interstimulus intervals throughout the experiment.

Drug administration

Drugs were injected into the Vi/Vc transition and Vc/C1 junction regions using a single- or dual-barrel thick-walled micropipette (40-80 μm, outer diameter). The selective GABA_A receptor agonist muscimol hydrobromide (Sigma, St. Louis, MO; 0.5 mm, pH 6.7) or the GABA_A antagonist bicuculline methiodide (BMI) (Sigma; 0.5 mm, pH 6.1) was delivered by pressure microinjection (Neuro-Phore, Medical Systems, Greenvale, NY) in a volume of 300 nl over 5 min. Microinjections of l-glutamate (10 mm, pH 6.8, 300 nl) were made over 2 min. Drugs were dissolved in artificial CSF containing (in mm): 150 NaCl, 2.6 KCl, 1.3 CaCl₂, 1.8 MgCl₂, and mixed with a 10% solution of Evan's Blue dye to visualize the location of injection. In experiments in which unit activity was recorded and muscimol injected, the drug was delivered from a separate glass micropipette angled stereotaxically to a target position 0.5-0.8 mm from the recording electrode. This proximity was confirmed histologically after the experiment from the deposition of Evan's Blue dye and a lesion at the recording site.

Experimental design

Tear volume. Tear volume was measured using three experimental designs in separate preparations. In each experiment, tear volume was estimated by the change in weight of a filter paper (∼5 × 8 mm) in contact with the cornea-conjunctiva for 2 min per sample. The filter paper was positioned at the inferior-lateral edge of the cornea-conjunctiva interface, which allowed tear volume to be determined while CO₂ pulses were applied to the center of the cornea. First, tears were evoked by a graded series of CO₂ pulses (40 sec duration) and measured at the start of the experiment. Each filter paper was weighed and then placed at the cornea-conjunctiva border 1 min before and removed 1 min after stimulus onset. The cornea was kept moist with artificial tears during the interstimulus intervals (15 min), and excess fluid was removed just before placement of the filter paper for subsequent samples. Second, the effect of local synaptic blockade of the rostral Vi/Vc transition or caudal Vc/C1 junction regions on tears evoked by 80% CO₂ pulses was assessed. Local synaptic blockade by muscimol injection began 4 min before presentation of the 80% CO₂ test pulse. Subsequent 2 min samples were collected at 5, 20, and 35 min after muscimol, and then the selective GABA_A receptor antagonist BMI was injected through the second barrel of the dual pipette 4 min before presentation of the last 80% CO₂ test pulse. In the third design, microinjections of l-glutamate (10 mm, 300 nl) were made into the rostral Vi/Vc transition or caudal Vc/C1 junction regions to determine the effect of local synaptic excitation on tear formation. Tear volume samples were taken 2 min before glutamate, for 2 min beginning at the onset of injection and from 3 to 5 min after injection. Filter paper weight was measured accurately to ±0.1 mg on an electronic balance. In preliminary measurements it was determined that deposition of 1 μl of artificial tear fluid was equal to a 1.1 mg change in filter paper weight.

Corneal unit activity. One corneal unit was studied in each experiment and recorded at either the Vi/Vc transition or Vc/C1 junction region. After the general RF properties were determined as described above, excess tears were removed by gently placing a small piece of filter paper at the lower lateral cornea-conjunctiva junction, and then units were tested for possible CO₂ responsiveness with a pulse of 0% CO₂ at a flow rate that had a minimal effect on spontaneous activity followed by two pulses of 80% CO₂ presented at 15 min intervals. Generally, only CO₂-responsive units were studied in detail, whereas units not responsive to CO₂ were classified as “mechanical only” and excluded from further analyses. Units displaying at least a 50% increase in activity above background to 80% CO₂ were tested further for responses to moisture and drying of the corneal surface. Excess spontaneous tears were removed, and then background unit activity was recorded continuously for 5 min before and after application of 20 μl of artificial tears to the corneal surface and after tear removal. All additional testing such as for CO₂ intensity-response relationships or sensitivity to drugs followed the initial collection of results for moisture and drying. In cases during which the effects of GABAergic drugs were tested, drug injection began 6 min after the second control 80% CO₂ pulse and continued for 5 min. Then 4 min later (i.e., 15 min after the previous 80% CO₂ pulse), another series of test pulses was presented followed by additional pulses of 80% CO₂ every 15 min for 1 hr.

C-fos immunocytochemistry

A separate group of male rats was used to determine whether acute drying (2 hr duration) of the ocular surface was sufficient to cause widespread excitation at different levels of the trigeminal brainstem complex in regions not assessed by electrophysiology. Under barbiturate anesthesia (pentobarbital sodium, 65 mg/kg), the upper and lower eyelids of the left eye were retracted with small adhesive strips, and spontaneous tears were removed by filter paper placed at the lateral edge of the ocular surface throughout the 2 hr survival period. Care was taken to avoid directly touching the ocular surface. Control animals were anesthetized, and the eyelids were allowed to close and tears to accumulate. After 2 hr, animals were perfused (fixative: 4% paraformaldehyde, 0.1 m phosphate buffer, pH 7.3) and postfixed overnight. Transverse brainstem sections were cut on a vibratome at 50 μm and incubated successively in 5% donkey serum (30 min), affinity-purified rabbit polyclonal anti-Fos antibody (Ab) (Ab-5, Oncogene Science; 1:15,000, 40 hr at 4°C), biotinylated donkey anti-rabbit secondary Ab (1:300, 105 min; Chemicon, Temecula, CA), and avidin-biotin-peroxidase complex (60 min; Vector Laboratories, Burlingame, CA). Fos-like immunoreactivity (Fos-LI) was visualized after 2-4 min incubation in activated nickel-cobalt-diaminobenzidine solution. Fos-LI was observed under bright-field illumination and distinguished as a homogenous gray-black precipitate restricted to cell nuclei. Controls for Fos Ab specificity were processed as above except that the primary antiserum was omitted.

Brainstem sections were categorized according to their approximate rostrocaudal location at 500 μm intervals from 2 mm rostral to 7 mm caudal to the obex. The obex is a surface landmark defined by the caudal end of the fourth ventricle ∼500 μm rostral to the most caudal tip of Vi (Yoshida et al., 1991). Sixty to 70 sections per animal were counted at 100× magnification without previous knowledge of the experimental treatment. The number of Fos-positive neurons per section was compared across treatments at different rostrocaudal levels of the Vsp by two-way ANOVA. Individual comparisons were made by the Newman-Keuls test after ANOVA (Winer, 1971).

Data analysis

Neural recording data were acquired and displayed by LabVIEW as peristimulus time histograms of spikes per 1 sec bins, exported to a spreadsheet, and analyzed off-line. The responses to moisture and drying of the cornea-conjunctiva surface were determined from the average spontaneous activity rate (spikes per bin) sampled 30 sec immediately before and after application of artificial tears to the cornea (wet state) and then after an additional 10 min, 30 sec before and after tears were removed (dry state). As described previously (Hirata et al., 1999), two main classes of corneal units at the Vi/Vc transition region were activated by CO₂ pulses. Type I units displayed a late excitatory phase (latency = 7-22 sec) that was proportional to CO₂ concentration. Type II units displayed an early inhibitory phase (latency <10 sec) in addition to the late excitatory phase. Examples of type I and type II unit responses to CO₂ stimulation are shown in Figure 5. To confirm the early-onset inhibitory period of type II units, only neurons with a background discharge of >1 Hz were included for further analyses. Accordingly, the average background discharge rate for Vi/Vc units in this study was somewhat elevated compared with previous reports in which ∼30% of Vi/Vc units had a low (< 0.5 Hz) rate (Hirata et al., 1999). All CO₂-responsive units recorded in laminas I-II at the caudal Vc/C1 junction region displayed a type I-like response pattern consisting of a single late excitatory response proportional to CO₂ concentration. Excitatory responses to CO₂ pulses for type I and type II units were quantified by calculating the total response magnitude (total R_mag) of the late excitatory component. The total R_mag for a given stimulus was defined as the cumulative sum of spikes for those contiguous bins in which the spike count exceeded the mean + 2 SD of the background activity. The total R_mag was calculated for each CO₂ stimulus period and can be thought of as equivalent to the “area under the curve.” Units classified as type II displayed an initial inhibitory phase within 10 sec after the onset of 80% CO₂ stimulation that was at least 50% below background activity. The maximum initial inhibitory response for type II units was defined as the maximum percentage decrease from background activity averaged over five consecutive bins after the onset of 80% CO₂ pulses. Similarly, a maximum inhibitory response for type II units to wetting the ocular surface was defined as the maximum percentage decrease from background activity averaged over five consecutive bins after the application of artificial tear fluid. Many type II units were classified further as complex cells, i.e., having a convergent inhibitory cutaneous RF. Inhibitory RF areas of complex units were mapped onto standardized drawings of the rat face, digitized, and quantified by a planimetric method using NIH Image software (v. 1.61). Unit activity recorded before and after wetting and drying the cornea, total R_mag, initial inhibitory phase (for type II units only) to CO₂ stimulation, and inhibitory RF areas of complex units were assessed statistically by ANOVA corrected for repeated measures (Winer, 1971) and individual comparisons by Newman-Keuls. Spearman rank-order analysis determined whether changes in inhibitory phase magnitude to CO₂ and wetting of the cornea were correlated. Changes in tear volume and mean arterial pressure to CO₂ stimulation were assessed by ANOVA corrected for repeated measures.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Examples of peristimulus time histograms of corneal units recorded at the rostral Vi/Vc transition region and classified by the response to CO₂ stimulation. Type I units were characterized by a late excitatory phase alone, whereas type II units displayed an initial inhibitory phase (asterisk) followed by a late excitatory phase.

Histology

At the end of the experiment, the animal was given a bolus of pentobarbital sodium (60 mg/kg, i.v.) and perfused through the heart with saline followed by 10% formalin plus saturated potassium ferrocyanide. Brainstem sections were cut at 40 μm on a freezing microtome and stained with cresyl violet. Recording sites at the rostral Vi/Vc transition and caudal Vc/C1 junction regions were marked electrolytically (5 μA, 10 sec). Antidromic stimulation sites in SSN/VII were reconstructed and drawn on a standardized series of brain outlines adapted from the atlas of Paxinos and Watson (1997). Microinjection sites for muscimol, BMI, and glutamate were identified from the deposit of Evan's Blue dye.

Tear volume

Tear volume was low (0.3 ± 0.1 mg/2 min; n = 15) in the absence of overt stimulation; however, pulses of room air (0% CO₂) were sufficient to cause a small but significant increase (1.2 ± 0.2 mg/2 min; n = 19) compared with the unstimulated condition (p < 0.005; F = 12.6; df = 1,30). Increasing concentrations of CO₂ gas applied to the ocular surface evoked progressive increases in tear volume as seen in Figure 2A. After 60% CO₂, tear volume was increased significantly above the response to 0% CO₂, with a maximum increase observed after 80% CO₂. This was a consistent finding because at least a 65% increase in tear volume occurred after stimulation by 80% CO₂ pulses in 16 of 19 experiments. Stimulation with 80% CO₂ did not cause a significant change in tear volume produced by the contralateral eye (n = 3; change in volume <0.3 mg). Mean arterial pressure increased transiently after 80% CO₂ (+2.7 ± 0.6 mmHg; p < 0.01) but not after lower concentrations of CO₂.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

A, Graded concentrations of CO₂ pulses applied to the ocular surface increased tear volume. Tear volume was increased above spontaneous baseline levels after 60% CO₂. ^**p < 0.01 versus 0% CO₂ stimulus; n = 19. B, Microinjection of the selective GABA_A receptor agonist muscimol (Musc) into the rostral Vi/Vc transition region inhibited CO₂-evoked tear formation, whereas injection into the caudal Vc/C1 junction region had no effect. Muscimol-induced inhibition of CO₂-evoked tear formation was reversed after injection of these lective GABA_Areceptor antagonist bicuculline methiodide (0.5 mm). Muscimol (0.5 mm) and BMI were injected in 300 nl over 5 min. ^**p < 0.01 versus 0% CO₂ stimulus; Vi/Vc, n = 11; Vc/C1, n = 7. Test stimuli of 80% CO₂ were given every 15 min.

The selective GABA_A receptor agonist muscimol was injected into the rostral Vi/Vc transition or caudal Vc/C1 junction region to determine whether local synaptic activity in either region was necessary for CO₂-evoked tear formation. Muscimol injection into the Vi/Vc transition region significantly (p < 0.01) reduced tear volume evoked by 80% CO₂ pulses, an effect reversed by injection of the selective GABA_A receptor antagonist (Fig. 2B). In contrast, muscimol injection into the caudal Vc/C1 junction region did not inhibit CO₂-evoked tear formation. Examples of injection sites in the rostral Vi/Vc transition and caudal Vc/C1 junction regions are shown in Figure 3. Note that at the Vi/Vc transition (Fig. 3A) dye was restricted to the ventral portion of the nucleus, whereas at the Vc/C1 junction region dye spread throughout the lateral dorsal horn (Fig. 3B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Examples of muscimol injection sites at the rostral Vi/Vc transition (A) and caudal Vc/C1 junction (B) regions as identified by the deposit of Evan's Blue dye. Asterisk indicates center of injection; dashed lines indicate the limits of visible dye spread. Dotted lines indicate approximate borders of the Vi/Vc and laminar portion of the medullary dorsal horn in A and B, respectively. Scale bar, 300 μm.

Glutamate was injected into the rostral Vi/Vc transition or Vc/C1 junction region to determine whether local excitation of either region altered tear production. As seen in Figure 4, glutamate excitation of the Vi/Vc transition, but not the Vc/C1 junction region, significantly increased tear volume (F = 34.1; df = 2,54; p < 0.001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Microinjection of glutamate (10 mm) into the rostral Vi/Vc transition but not the caudal Vc/C1 junction region caused a prompt increase in tear volume. Tear volume was measured over 2 min before injection, during the first 2 min after the start of injection (Post 1), and 3-5 min after injection (Post 2). Sample size: Vi/Vc = 18; Vc/C1 = 11. ^**p < 0.01 versus prestimulus tear volume; b = p < 0.01, Vi/Vc versus Vc/C1 sites.

Corneal unit activity

General properties

A total of 111 neurons were recorded at the Vi/Vc transition region (n = 69) and in laminas I-II at the Vc/C1 junction region (n = 42) and subsequently tested for responsiveness to wetting and drying of the ocular surface. All units included in these analyses were responsive to mechanical (brush) and chemical (80% CO₂) stimulation of the cornea. Units were classified by the response to CO₂ pulses and cutaneous RF properties (Table 1) as described in detail previously (Meng et al., 1997; Hirata et al., 1999) (Fig. 5). Units with a type I-like pattern were found at the Vi/Vc transition and Vc/C1 junction regions, whereas units with a type II response pattern to CO₂ were found only at the Vi/Vc transition region. The majority of type II units also had a convergent inhibitory cutaneous RF (see Fig. 8A; listed as complex units in Table 1). Mechanical stimulation of facial skin or the nasal cavity, or both, inhibited 22 of 25 type II units and none were excited, whereas 5 of 18 type I units were inhibited and 13 were excited. A total of 36 Vi/Vc units had an inhibitory cutaneous RF (type I = 8; type II = 28). The average inhibitory cutaneous RF area for all complex units was 2.98 ± 0.29 cm² (mean ± SE; n = 34) and was similar for type I and type II complex units. Two additional type II units had a cutaneous inhibitory RF that included the entire body surface. The recording sites for type I (n = 17) and type II (n = 18) units that were marked electrolytically and could be verified histologically at the rostral Vi/Vc transition region are shown Figure 6. The recording depth was similar for type I and type II units at the Vi/Vc transition region (2.04 ± 0.09 vs 2.05 ± 0.04 mm); however, χ² analysis indicated that type II units were located more rostrally than type I units (χ² = 5.04; df = 1; p < 0.05). Corneal units recorded in laminas I-II at the caudal Vc/C1 junction region displayed only a type I-like response to 80% CO₂, and with one exception, all Vc/C1 units received excitatory convergent input from ipsilateral periorbital skin and were classified as WDR- or NS-like units. Each of the 11 Vc/C1 units recorded in laminas I-II was excited by mechanical stimulation of the ipsilateral nasal cavity.

View this table:

• View inline
• View popup

Table 1.

Summary of cells and their classification tested for responsiveness to acute moisture and drying of the ocular surface

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

Type II units were inhibited by either wetting of the ocular surface or application of 80% CO₂ pulses. A, Example of a type II unit displaying a pronounced initial inhibitory phase to 80% CO₂ and after wetting of the corneal surface. Note that this unit also had an inhibitory RF on the ipsilateral face (complex unit; inset at top right). B, Scattergram for all type II units (n = 31) and the regression line indicating a significant relationship between the magnitude of the initial inhibitory phase to 80% CO₂ pulses and inhibition to wetting of the ocular surface.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Recording sites at the rostral Vi/Vc transition region for type I and type II units. TypeI and typeII units were segregated rostrocaudally, with typeI units being more likely found at more caudal levels. ION, Inferior olive; NTS, nucleus tractus solitarius; Vc, trigeminal subnucleus caudalis; Vi, trigeminal subnucleus interpolaris. Numbers to bottom left of each outline refer to the distance in millimeters caudal to the obex. Scale bar, 0.5 mm.

Effects of wetting and drying the ocular surface on unit activity

Spontaneous activity was monitored continuously for 5 min before and 10 min after application of artificial tears to the ocular surface. At 10 min after wetting of the surface, all excess fluid was gently removed with a tissue paper by the ventrolateral corner of the eye, and unit activity was averaged for the next 1 min. The average firing rate before and after wetting and drying of the ocular surface for the three main classes of cells (defined by the response to 80% CO₂ pulses and recording location) are shown in Figure 7. All units were spontaneously active before wetting of the ocular surface. At the Vi/Vc transition, type II units had a slightly higher average firing rate (21.1 ± 4.1 spikes/sec) than type I units (17.4 ± 2.2 spikes/sec); however, this difference was not significant. By 1 min after wetting of the cornea, type II units displayed a marked and persistent decrease in firing rate (>75% from background; p < 0.01), whereas type I units at the Vi/Vc transition or those in laminas I-II at the Vc/C1 junction region were not affected consistently. In addition, several (n = 4) corneal units were recorded from deep laminas at the Vc/C1 junction (depth >600 μm), and none were affected by wetting of the ocular surface (11.7 ± 5.6 vs 12.4 ± 6.2 spikes/sec, before vs after wetting). Drying of the ocular surface caused a prompt increase in activity of type II units to levels similar to the pre-wet condition. Moisture sensitivity among type II units was a highly consistent finding because 30 of 31 units were inhibited by wetting of the cornea (Table 1). The thermal sensitivity of type II units was not studied systematically; however, in a few cases (n = 3), the artificial tear fluid was warmed to 40°C and found to have a similar inhibitory effect as tear fluid applied at 25°C. Regression analyses indicated that the magnitude of inhibition after wetting of the ocular surface was proportional to the magnitude of the initial inhibitory phase to 80% CO₂ pulses (Fig. 8B). An example of a type II unit with a complex inhibitory cutaneous RF displaying an initial inhibitory phase to 80% CO₂ pulses and a marked persistent inhibition to wetting of the ocular surface is shown in Figure 8A. Correlation by Spearman rank-order analyses revealed no significant relationship between the area of the inhibitory cutaneous RF and the magnitude of inhibition evoked by wetting the cornea or the initial inhibitory phase to 80% CO₂ pulses (r < 0.1; n = 28).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Artificial tears applied to the corneal surface (upward arrow) caused a significant decrease in spontaneous activity of corneal type II units, whereas acute removal of tears (downward arrow) caused an increase in activity toward resting levels. Type I units recorded at the Vi/Vc transition or caudal units found at the Vc/C1 junction region displayed only minor changes in spontaneous activity after wetting of the ocular surface. Sample sizes: Vi/Vc units type I (n = 38),Vi/Vc units typeII(n=31), and Vc/C1(n=42). ^**p<0.01 versus pre-wet condition; b= p < 0.01 versus other cell classes.

Several units (n = 13) were tested for efferent projections to the SSN region. All CO₂-responsive corneal units recorded at the Vi/Vc transition (type I, two of two; type II, four of four) were antidromically activated, whereas only three of eight units in laminas I-II at the Vc/C1 junction were driven by sites in the SSN region. The locations of effective antidromic activation sites in SSN were similar to those reported previously (Hirata et al., 2000) (Fig. 7, plane -10.8).

C-fos immunocytochemistry

Fos-LI was used to determine whether acute drying of the ocular surface caused widespread neural activation in the Vsp, especially in regions outside the Vi/Vc transition and Vc/C1 junction. The upper and lower eyelids of the left eye were held open for 2 hr, and spontaneous tears were gently removed by tissue paper every few minutes. The greatest number of Fos-positive neurons was produced at the level of the ventral Vi/Vc transition region ipsilateral to drying, although a small number of cells were found in the corresponding contralateral region (Fig. 9). No Fos-positive cells were found within the Vsp rostral to the Vi/Vc transition region. Fos-LI produced at the Vi/Vc transition was restricted to the ventrolateral pole (Fig. 10A), whereas at the caudal Vc/C1 junction region a small, but significant, increase in Fos-positive neurons was seen in the superficial laminas ipsilateral to drying (Fig. 10B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 9.

Rostrocaudal distribution of Fos-positive neurons within the Vsp ipsilateral and contralateral to drying of the ocular surface. The upper and lower eyelids of the left eye were gently retracted by adhesive strips, and tears were removed throughout the 2 hr survival period (Dry). The no-stimulus controls (NS) also were anesthetized and survived for 2 hr; however, eyelids were allowed to close naturally and tears were allowed to accumulate. ^**p < 0.01, ipsilateral versus contralateral; a = p < 0.05, b = p < 0.01 versus NS controls; n = 4 per treatment.

• Download figure
• Open in new tab
• Download powerpoint

Figure 10.

Examples of Fos-positive neuronal nuclei at the Vi/Vc transition (A) and caudal Vc/C1 junction (B) regions ipsilateral to the dry ocular surface. Note that Fos-LI was restricted to the ventrolateral portion of the Vi/Vc transition and to the superficial laminas at the Vc/C1 junction region. Scale bar, 0.2 mm.