Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Studies examining the role of spinal cholinergic receptor systems have demonstrated that spinal nicotinic acetylcholine receptors appear to play a role in modulating the animal's response to noxious stimuli (Damaj et al., 1998; Khan et al., 1998; Lawand et al., 1999; Xu et al., 2000). Systemic administration of nicotinic receptor agonists, such as nicotine, cytisine, and epibatidine, can induce analgesia as measured by the escape indices in models of thermal, mechanical, and chemical nociception (Bannon et al., 1998; Damaj et al., 1998). The mechanisms underlying these responses to nicotinic receptor agonists are complex, but reflect in part actions at the spinal level. Thus, delivery of these agents to segmental regions of the spinal cord can also produce significant antinociception with its pharmacological response reflecting local stimulation of spinal nicotinic acetylcholine receptors (nAChRs) (Khan et al., 1998; Lawand et al., 1999). This antinociceptive or analgesic response, produced by either systemic or spinally delivered drug is, however, accompanied by significant cardiovascular responses and signs of behavioral agitation (Buccafusco, 1996; Khan et al., 1997, 1998). In recent studies, it has been hypothesized that these cardiovascular and nociceptive responses may be mediated, in part, by distinct populations of spinal nAChRs (Khan et al., 1994a,b,c, 1996b, 1997, 1998; Damaj et al., 1998).

Molecular cloning of nAChRs from rat brain indicates that neuronal nAChRs belong to a multigene family, which includes at least nine -subunits and four -subunits (non-). Transfection of various combinations of these - and -subunit genes results in the assembly of functional surface receptors (Boulter et al., 1987; Bertrand et al., 1990). Ligand recognition is associated with an interface of the -subunit, when associated with a juxtaposing -subunit interface. In the heteromeric nAChRs, two -subunits and three of the non-- or -subunits associate to form functional pentameric receptors in neuronal tissue (Lindstrom, 2000). Even though not all combinations of - and -subunits can form functional receptors, the number of permutations and thus the potential for assembly of various subtypes of neuronal nAChRs is very large.

We have been interested in characterizing the several behavioral and physiological actions of spinal nAChRs with respect to their binding properties, localization, and subunit composition. Recent studies reported that (+)-epibatidine is the most potent nicotinic receptor agonist in eliciting spinal systemic actions. Based on intrathecal delivery studies, epibatidine appears to exert its action through multiple spinal nAChRs systems (Khan et al., 1998). Moreover, specificities of nicotinic receptor antagonists indicate that different subtypes of neuronal nAChRs elicit antinociceptive and nociceptive responses to spinal epibatidine. This is consistent with the binding properties of [³H]epibatidine to spinal cord membranes where more than a single class of receptors with distinct affinities is revealed (Khan et al., 1997). Recently, a synthetic analog of epibatidine, so-called A-85380, has been shown to possess significant analgesic activity when administered systemically (Curzon et al., 1998). This nicotinic receptor agonist is thought to be specific for the 42 subtype of neuronal nicotinic receptor (Curzon et al., 1998). Given this subtype specificity and the reported selectivity of effects, we sought to characterize the pharmacology of the physiological and behavioral actions of spinally administered A-85380 in relation to other agonists. The well defined spinal pathways and defined pharmacological end points might enable us to distinguish the sites of action and receptor selectivity of these nicotinic receptor agonists.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Experimental Animals

Male Sprague-Dawley rats (300-350 g) were purchased from Harlan Co. (Indianapolis, IN). Animals were housed in the University of California, San Diego (UCSD) animal facility and were maintained on 12-h light/dark cycles. They received standard Purina Rat Chow and water ad libitum. All studies were carried out according to protocols approved by the UCSD Institutional Animal Care Committee.

For spinal drug delivery, intrathecal (i.t.) catheters were implanted as described previously (Khan et al., 1998). Briefly, rats were anesthetized with halothane (2-3%), placed in a stereotaxic frame, and the atlanto-occipital membrane was exposed. A 9-cm saline-filled polyethylene (PE-10) tubing was placed into the intrathecal space through the atlanto-occipital membrane and passed down to the rostral edge of the lumbar enlargement. The catheter was externalized on top of the skull and sealed with a piece of stainless steel wire. The incision was closed and the rats were allowed to recover for at least 5 days before further study. Only animals exhibiting normal motor behavior were used in the study. Animals with impaired motor function or an elevated sensory threshold were euthanized.

Measurement of Arterial Blood Pressure

Five to 6 days following i.t. catheter implantation, rats were reanesthetized with halothane (2-3%) and their tail arteries were catheterized with a PE-50 tube filled with heparin containing saline (1 U/ml). The wound was covered with gauze-tape, and the rats placed in a plastic cylindrical restraining cage. The cage was constructed to enable the animal to maintain a typical crouching posture. The tail artery was connected to a blood pressure transducer coupled to a Gould polygraph. Heart rates were measured using a cardiotachometer triggered from pressure pulses. Rats were allowed to recover for at least 30 min after placement in the cage for administration of drugs. Blood pressure and heart rate were monitored continuously for the duration of the experiment. The cardiovascular parameters were recorded with a Gould polygraph, and the data further analyzed with the aid of Ponemah Physiology Platform-3 (Gould Instruments Systems, Valley View, OH).

Drugs and Their Administration

The following chemicals were obtained from Sigma Chemical Co. (St. Louis, MO): A-85380, cytisine, mecamylamine, -lobeline, (+)-epibatidine hydrochloride, dihydro--erythroidine, methyllycaconitine (MLA), and 2-amino-5-phosphopentanoic acid (AP-5). Sterile saline was used as the vehicle to dissolve the drugs for i.t. administration. Before all testing, an intrathecal catheter was connected to a motor-driven microinjection pump via a length of calibrated PE-50 tubing with each 10-µl compartment separated from the other by a small air bubble to avoid mixture of drugs. For i.t. delivery, drugs were injected in a volume of 10 µl followed by 10 µl of normal saline to flush the catheter over a 10-s interval (pump delivery rate of 60 µl/min.).

Test Measurements

Antinociception. Analgesia or antinociception was measured according to the experimental protocol described previously (Khan et al., 1998). Briefly, unanesthetized rats were placed in Plexiglas cages (9 × 22 × 25 cm) on top of a glass plate. A thermal stimulus was positioned under the glass to focus the projection bulb on the plantar surface. Initiation of the current to the bulb started a timer. Bulb current and time were automatically terminated when paw elevation was sensed by photodiodes or when an interval of 20 s (cut-off time) had passed. The surface under the glass was maintained at 30°C by a feedback-controlled heater fan. A focused stimulus (stimulus current = 4.8 amp with an escape latency of approximately 9 s) was reliably accomplished by a mirror attached to the stimulus, which permitted visualization of the undersurface of the paw. Light beam intensity was monitored by measurement of bulb current, and the stimulus intensity was calibrated daily by assessing the temperature change after 10 s sensed by an underglass thermocouple (t_1/2 = 0.2 s). After placing the rat in the plastic cages for a 20-min adaptation period, the first measurement was conducted on both hind paws, and the response latencies averaged and counted as baseline score (time zero). Tests were then made at 3, 6, 9, 15, 20, 30, 40, 50, and 60 min after injection or for repeated intrathecal injections of drugs at 3, 6, 9, 15, 20, and 30 min after the first injection. After the second injection, testing was performed again at 3-, 6-, 9-, 15-, 20-, and 30-min intervals. Animals remained connected to the injection pump and were not handled during the test sequence.

Nociception and Agitation Behavior

Spontaneous agitation/spontaneous vocalization (SA/SV) was scored on a scale of 0 to 4.5. Scores were measured according to the following scale: 0.5, slight movement of the paws; 1, whole body movement; 1.5, slight twitching; 2, licking of the paws, moderate twitching; 2.5, moderate ambulation; 3, severe twitching, more pronounced ambulation; 3.5, hunchback posture, limited escape behavior; 4, rolling over, intense ambulation, and escape behavior; and 4.5, high-pitched squeaking, frantic ambulation, and escape behavior. Score rankings were directly proportional to the dose of the agonist, and each score represented the corresponding behavioral responses, including those from the lower scores.

Experimental Design

Each rat was used in one or two experiments with a minimum interval of 5 days between injections. Animals were randomly assigned to receive a single dose of the agonist, the antagonist followed by the agonist, or the antagonist or vehicle (normal saline) alone. The antagonist was injected 10 to 15 min before agonist (A-85380, 5 µg) administration followed by a 10-µl saline flush. The doses of the antagonists were selected from previously described studies (Khan et al., 1994b, 1996a, 1997, 1998).

Data Presentation

For thermal antinociception, data are represented as means ± S.E.M. For the time course, each time point represents the latency period in seconds after agonist or vehicle delivery. Thermal latencies were converted to the percentage of the maximum possible effect (%MPE) according to the following formula:
Dose-response curves are presented as the %MPE observed within the testing interval for the particular measure. For all drugs, dose-response relations were analyzed as described by Tallarida and Murray (1986). Changes in the thermal latency with and without antagonist pretreatment were tested for significance using unpaired Student's t test (two-tail). For all comparisons between treatment groups, the highest %MPE observed in each rat within 9 min following agonist administration was used as the data point. Differences between multiple groups were compared using an ANOVA. Values of P < 0.05 were considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Intrathecal A-85380-Elicited Responses

Similar to other nicotinic receptor agonists, i.t. A-85380 also elicited a dose-dependent nociceptive or agitation response (Fig. 1B). Figure 2B depicts the temporal agitation response following a 5-µg dose of A-85380. The SA/SV response was of rapid onset and it lasted for about 10 min (Fig. 2B). After this time period, only intermittent agitation responses were observed in the rats. In addition, to the nociceptive response, i.t. A-85380 also exhibited a transient antinociceptive response. This response was also dose-dependent (Fig. 1A). As shown in Fig. 2A, at a 5-µg dose, A-85380 elicited approximately 70% of the MPE. The peak response was observed within 3 min following agonist administration and had a duration of approximately 9 min. The ability of higher doses of A-85380 to elicit analgesia could not be examined, because they would produce unacceptable levels of irritation in the rats.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   A-85380 elicits dose-dependent increases in paw-withdrawal latencies (as represented by %MPE) (A), nociceptive response (SA/SV) (B), and pressor response upon intrathecal administration (C). Data for A and B were obtained at 3 to 6 min and 1 to 2 min, respectively, after injection from the same rat for a specific dose of A-85380. Data for C were generated from a separate group of rats. The number in parentheses represents the number of rats used to generate that data point. Each value represents a mean ± S.E.M. The ordinate in each figure represents nanomoles of A-85380 injected for the corresponding doses of 0.005 µg (0.021 nmol), 0.05 µg (0.21 nmol), 0.5 µg (2.1 nmol), and 5.0 µg (21 nmol) of A-85380. **P < 0.001 for 21-nmol compared with 2.1- and 0.21-nmol treated rats, and *P < 0.01 for 2.1-nmol compared with 0.21-nmol treated rats (A and B). P < 0.01 for 21-nmol compared with 0.021-nmol treated rats, and P < 0.01 for 2.1-nmol compared with 0.021-nmol treated rats (C).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Time course for the antinociceptive (A) and nociceptive (B) responses for the various doses of intrathecal A-85380. Each data point represents a mean ± S.E.M. **P < 0.001 for 5 versus 0.5 and 0.05 µg of A-85380-treated rats, and *P < 0.01 for 5 versus 0.5 µg of A-85380-treated rats for the antinociceptive response (A); P < 0.005 for 5 versus 0.5 and 0.05 µg of A-85380-treated rats, and P < 0.01 for 5 versus 0.5 µg A-85380-treated rats for the nociceptive response (B) at the specific time points.

In addition to the antinociceptive and nociceptive responses, intrathecal A-85380 also resulted in dose-dependent increases in arterial blood pressure (Fig. 1C). The data presented in Fig. 1C are from a different group of rats than those presented in Fig. 1, A and B. As seen for other spinal nicotinic receptor agonists (Khan et al. 1994b, 1996a, 1998), spinal A-85380 also resulted in dose-dependent increases in heart rate (data not shown). The onset of the cardiovascular response was also rapid. It appeared within 1 min following injection and lasted for about 10 to 12 min.

The nociceptive and cardiovascular responses occurred at lower doses of A-85380 than the antinociceptive response. This rank order of responses was similar to that of epibatidine (Khan et al., 1998); however, A-85380 appears to be less potent than epibatidine.

Antagonism of Intrathecal A-85380-Elicited Responses

Nicotinic Antagonists. To distinguish the receptor specificities of spinal A-85380-elicited responses, various nicotinic receptor antagonists were evaluated on A-85380-evoked responses. As reported previously, the nicotinic receptor antagonists, when administered alone, did not increase the thermal threshold (Khan et al., 1998). Intrathecal administration of the nicotinic receptor channel blocker mecamylamine (50 µg), 10 min before A-85380, completely blocked the antinociceptive and nociceptive responses to 5 µg of A-85380 (Fig. 3, A and B). The pressor (Fig. 3C) and heart rate (data not shown) responses to 0.5 µg of A-85380 were also significantly blocked by mecamylamine pretreatment.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Effects of various nicotinic antagonists on A-85380-elicited antinociceptive (A), nociceptive SA/SV (B), and pressor (C) responses. Individual antagonists (50 µg i.t., each) were administered 10 to 15 min before a single dose of A-85380 administration in a separate group of rats. The control rats received saline before A-85380 administration. The rats used to generate the data in A and B received 5 µg of A-85380, whereas the rats represented in C received 0.5 µg of the agonist. All responses shown denote the near maximal responses following specified doses of A-85380 administration in the presence or absence of nicotinic antagonists. Each value represents the mean ± S.E.M. The numbers in the boxes indicate the number of animals for the corresponding data point. ***P < 0.001 and **P < 0.01 compared with saline-pretreated rats.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Temporal analysis of the antinociceptive and SA/SV responses to A-85380 (5 µg i.t.) in DE (50 µg i.t.)-pretreated and saline-pretreated control rats. The rats for the two treatment groups and the experimental protocol are the same as those used to generate the data in Fig. 3, A and B. Each data point represents the mean ± S.E.M. **P < 0.01 and *P < 0.05 compared with control rats for that specific time point.

Desensitization and Cross-Desensitization

Intrathecal pretreatment with 5 µg of A-85380 resulted in desensitization of antinociceptive, nociceptive, and cardiovascular responses to a subsequent 5-µg dose of A-85380 administered 30 min later (Fig. 5, A-C). Moreover, two consecutive doses of A-85380 significantly desensitized the nociceptive responses to epibatidine. In contrast, the antinociceptive response to epibatidine was marginally reduced by prior doses of A-85380 and not significantly different from saline-treated control rats (Fig. 6, A and B). On the contrary, two consecutive doses of 0.5 µg of epibatidine 30 min apart significantly desensitized both the antinociceptive and nociceptive responses to a subsequent 5-µg dose of A-85380 30 min later (Fig. 6, C and D). Similar pretreatment with cytisine (5 µg each) only desensitized the nociceptive response to A-85380 (Fig. 6, E and F). The antinociceptive response to A-85380 appeared to be modestly attenuated following repeated dosing with cytisine, but it was not statistically significant. Analysis of analgesia over time revealed that the peak antinociceptive response to A-85380 (at 3 min) was significantly desensitized by cytisine pretreatment. However, the antinociceptive response, while modest, is sustained for a substantially longer period (Fig. 7). This presumably arises from desensitization of nociception and unmasking of analgesia.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Effect of prior treatment by A-85380 (5 µg i.t.) on subsequent A-85380 (5 µg i.t.)-elicited antinociceptive (A), nociceptive (B), and pressor responses (C). The second dose of the agonist was administered 30 min after the first dose of A-85380. All responses shown denote the maximal values in each category of response. Each value represents a mean ± S.E.M. For A and B, n = 12 and for C, n = 8. ¶P < 0.001, **P < 0.01, and *P < 0.05 compared with first dose.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   Effect of repeated administration of A-85380 (5 µg i.t., A and B), epibatidine (0.5 µg i.t., C and D), and cytisine (5 µg i.t., E and F) on the antinociceptive (top) and nociceptive SA/SV (bottom) responses elicited by a subsequent dose of epibatidine (0.5 µg i.t.), A-85380 (5 µg i.t.), and A-85380 (5 µg i.t.), respectively. In each experimental group, the first agonist was injected at 0 min and then again at 25 to 30 min following the initial injection. The test agonist was administered 55 to 60 min following the second injection. Control rats received two injections of saline followed by the injection of the test agonist at the specific time points. All responses shown denote the maximal values in each category of response. Each value represents mean ± S.E.M. *P < 0.01 compared with control rats for the final agonist.

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Temporal analysis of the antinociceptive response to A-85380 (5 µg) following repeated dosing with cytisine (5 µg, open circles) or following saline pretreatment (closed circles). The data represent the same group of rats and experimental protocol as presented in Fig. 6, E and F. Each data point represents mean ± S.E.M. *P < 0.05 compared with control rats.

Adrenergic Receptor Antagonists on Intrathecal Nicotinic Agonist Responses

Pretreatment with phentolamine (25 µg i.t.), an -adrenergic receptor antagonist, significantly inhibited the antinociceptive and nociceptive responses to A-85380 (Fig. 8, A and B). In contrast, phentolamine pretreatment had no significant effects on epibatidine-elicited antinociceptive or nociceptive responses (Fig. 8, C and D). Phentolamine had no effect on the cardiovascular responses to A-85380 (data not shown).

View larger version (38K): [in this window] [in a new window]   Fig. 8.   Effects of phentolamine (25 µg i.t.) on A-85380 (5 µg i.t.)-elicited antinociceptive (A) and nociceptive SA/SV (B) responses. Phentolamine was administered 10 to 15 min before single i.t. administration of A-85380 administration. The control rats received saline before A-85380 administration. A similar dose of phentolamine did not have any significant effect on the antinociceptive (C) and SA/SV (D) responses elicited by epibatidine (0.5 µg i.t.). All responses shown denote the maximal values in each category of response following A-85380 administration in the presence or absence of phentolamine. Each value represents a mean ± S.E.M. The numbers in the boxes indicate the number of animals for the corresponding data points. **P < 0.001 and *P < 0.01 compared with saline-pretreated rats.

N-Methyl-D-aspartate Antagonism and A-85380 Activity

AP-5 (5 µg i.t.) administered before A-85380 (5.0 µg) did not alter the antinociceptive response of the agonist, but did depress the nociceptive responses (Fig. 9, A-C). Although AP-5 has been demonstrated to block the cardiovascular responses to nicotine, cytisine, and epibatidine (Khan et al., 1996, 1997, 1998), it did not have any significant effect on the A-85380-elicited pressor response (Fig. 9C).

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Effects of the N-methyl-D-aspartate antagonist AP-5 (5 µg i.t.) on A-85380-elicited antinociceptive (A), SA/SV (B), and pressor (C) responses. AP-5 was administered 10 to 15 min before a single administration of A-85380. The control rats received saline before A-85380 administration. The rats used to generate the data in A and B received 5 µg of A-85380, whereas the rats represented in C received 0.5 µg of the agonist. All responses shown denote near maximal responses in each category following A-85380 administration in the presence or absence of AP-5. Each value represents a mean ± S.E.M. The numbers in the boxes indicate the number of animals for the corresponding data points. **P < 0.01 compared with saline-pretreated rats.

Competitive Displacement of [³H]Epibatidine Binding by A-85380

Similar to other nicotinic receptor agonists, A-85380 displaces [³H]epibatidine binding from spinal cord membranes in a dose-dependent manner (Fig. 10). The inhibitory dissociation constant of A-853380 for competitively displacing [³H]epibatidine binding from the high-affinity sites was 0.23 nM (the mean from two independent K_i measurements of 0.19 and 0.28 nM). The competitive curve for A-85380 shows a shallower slope, suggesting two or more classes of binding sites whose relative affinities for epibatidine and A-85380 are not identical. When compared with the K_i values of other nicotinic receptor agonists (Khan et al., 1994, 1997), A-85380 appears to be more potent than cytisine (K_i = 0.56 nM), but less potent than epibatidine (K_i = 0.05 nM) in binding to high-affinity sites of the spinal nAChR.

View larger version (19K): [in this window] [in a new window]   Fig. 10.   Inhibition of [3H]epibatidine binding to spinal cord membranes by A-85380. Membranes (400-600 µg of protein) were incubated with 0.08 nM [3H]epibatidine and various concentration of competing A-85380 for 60 min at 4°C.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Neuronal nAChRs and Analgesia. As early as 1932, it was shown that systemic nicotine administration elicits analgesia (Davis et al., 1932). The low efficacy of nicotine in eliciting analgesia and the accompanying side effects precluded its utility as an analgesic agent. Recently, systemic delivery of epibatidine, a neurotoxin obtained from Ecuadorian frog skin, was reported to evoke a robust analgesic response (Badio and Daly, 1994). Importantly, epibatidine elicits significant analgesia after spinal delivery at doses that are not systemically active (Khan et al., 1998; Lawand et al., 1999). However, this antinociceptive response to spinal epibatidine is transient, and accompanied by cardiovascular and nociceptive responses (Khan et al., 1996a, 1997, 1998). The multiple responses produced by (+)-epibatidine appear to be mediated through distinct subpopulations of central nervous system neuronal nAChRs (Khan et al., 1997, 1998; Bannon et al., 1998; Donnelly-Roberts et al., 1998). The diverse responses arising from distinct sites are consistent with binding studies, which suggest nicotinic receptor agonists interact with multiple nicotinic sites.

Intrathecal A-85380-Elicited Responses in Rats. Systemic A-85380 has been reported to be equally effective as, but less potent than, epibatidine in producing analgesia (Curzon et al., 1998). No other behavioral or cardiovascular responses were reported. Although direct evidence is lacking, it is suggested that systemic A-85380 elicits analgesia via stimulation of neurons in the region of the nucleus raphe magnus in the brain stem (Curzon et al., 1998).

Antagonism of A-85380-Elicited Responses. The antinociceptive response to A-85380 was blocked by DE, an 42-specific nAChR antagonist (McIntosh, 2000). Although this observation is consistent with the report that A-85380 is 42 subtype-specific nicotinic receptor agonist, DE did not block the nociceptive or cardiovascular responses to i.t. A-85380. Thus, A-85380 also interacts with other spinal nAChR subtypes. Similar to mecamylamine, MLA significantly blocked all the responses to A-85380. As we indicated in earlier studies (Khan et al., 1994b,c, 1997, 1998), MLA behaves more like a channel blocker than a competitive antagonist in the spinal system.

Cross-Desensitization. It appears that both A-85380 and epibatidine bind to a common receptor to elicit the nociceptive response since both of the agonists cross-desensitize each others' nociceptive response. Moreover, the nociceptive response to A-85380, like epibatidine, is significantly inhibited by intrathecal AP-5 treatment. This indicates that, similar to other nicotinic receptor agonists, a significant portion of the nociceptive response to A-85380 is mediated via spinal release of excitatory amino acids (Khan et al., 1996a, 1997, 1998).

Subunit Composition. Recently, binding studies with tissue from 2-deficient mice indicate that 42 is the primary nAChR subtype expressed in the brain (Zoli et al., 1998); however, 32, 34, and 44 subunit compositions appeared also to be present in the brain. 3 subunits can be detected in the trigeminal neurons, sympathetic ganglia, and the dorsal root ganglia (Wada et al., 1990; Boyd et al., 1991; Lukas, 1993; Flores et al., 1996) as 34 (Flores et al., 1996). Complementary to studies of Wada et al. (1990), our findings reveal that 5 transcripts and expressed subunits are present in the dorsal root ganglion as well as spinal interneurons (I. Khan, unpublished observations). Incorporation of 5 subunit into 32, 34, and 44 subunit compositions is known to modulate the activities, ligand binding affinities, and desensitization properties of these receptors in vitro (Ramirez-Lattore et al., 1996; Gerzanich et al., 1998). Therefore, it seems likely that more than one subtype of nAChR is expressed on afferent terminals. Such an expression pattern might confer selectivity of epibatidine over A-85380 for a particular subtype of nAChR to elicit afferent terminal desensitization.