It is widely believed that cigarette smoking stems largely from nicotine addiction. Reinforcing effects of nicotine have been extensively investigated using the nicotine intravenous self-­administration (IVSA) procedure in adult rats. After describing methodological aspects of the technique, we highlight its main applications in this species, particularly related to pharmacotherapy development. Finally, we discuss further refinements of the procedure that may provide a closer model of human cigarette smoking. (Please note that we will use the terms “dependence” and “addiction” interchangeably in this chapter.) Important research has also been performed in other species, notably squirrel monkeys (1, 2), rhesus monkeys (3, 4), and in wild type and genetically modified mice (5–7). These species fall outside the scope of this review and the reader is referred to recent reviews by others (8, 9).

Tobacco dependence is a major preventable cause of death. In developed countries, tobacco smoking is estimated to cause about 30% of all deaths from cancer, and an even greater number of deaths from other diseases (10). Globally, smoking accounts for 4–5 millions deaths each year (10). The primary problem facing smokers who attempt to quit is the high rate of relapse and the poor effectiveness of current pharmacological and non-pharmacological therapies (11). A better understanding of tobacco dependence appears necessary in order to provide more effective pharmacotherapies.

Cigarette smoke contains about 4,000 identified chemicals, the great majority of which have never been characterized pharmacologically or behaviorally. Nevertheless, the 1988 US Surgeon General’s Report concluded that “nicotine is the drug in tobacco that causes addiction” (12). This widely held view has provided the rationale for well over 100 published reports of nicotine ­self-administration. Indeed, nicotine IVSA is currently the dominant animal model related to cigarette smoking. Recently, nicotine’s central role in tobacco dependence has been vigorously questioned (13, 14), and a more nuanced view appears to be emerging (15, 16).

Most nicotine self-administration studies employ intravenous delivery, but there are also several published reports of oral self-administration. The latter approach offers several advantages: catheterization surgery is avoided, extensive training is not required, 24-h access is convenient, and long-term testing is straightforward. One obvious drawback is that nicotine absorption is slow and potentially variable. Moreover, since nicotine solutions are bitter, rats need to be encouraged to consume ­nicotine solutions, either by schedule-induced polydipsia (17), ­sweetening with sucrose (18), or by gradually increasing the ­nicotine ­concentration (19). Results to date are somewhat ­equivocal. For example, in one study, rats lever-pressed more for oral solutions of nicotine + sucrose than of sucrose alone, yet drank less when consumption was unrestricted. In a further experiment, rats could choose to respond for either aqueous nicotine solutions or water; weaker nicotine solutions were mildly preferred, whereas stronger solutions were avoided (18).

Cigarette smoking is a complex behavior modulated by pharmacological, psychological, social, and economic variables. Which aspects of this behavior, then, does nicotine self-administration model? In general terms, the self-administration procedure has been used to model the acquisition, maintenance, extinction, and reinstatement of nicotine-taking behavior. Self-administration behavior, tested in limited-access sessions, is presumably controlled by the acute positive consequences of drug administration. With more extended daily exposure, withdrawal signs can emerge (20, 21); here, negative reinforcement may play an important role. Importantly, nicotine self-administration behavior tends to be weak unless the drug is paired with sensory cues, and it has been proposed that animals self-administer intravenous ­nicotine primarily because the drug makes these ancillary cues more reinforcing (see below).

Although compulsion is a hallmark of drug addiction, nicotine self-administration studies to date have provided little if any evidence of compulsive drug seeking (but see (20, 21)). However, most studies of nicotine self-administration are quite possibly not long enough for compulsive behavior to manifest itself. It is worth noting that even with cocaine, compulsive drug-seeking behavior can require several months of repeated exposure and/or testing before motivated drug seeking increases sufficiently to be regarded as compulsive (22, 23).

Self-administration studies have been instrumental in identifying molecular and neuronal mechanisms that are critical to the acute reinforcing effects of nicotine. Nicotinic receptors are widely expressed in both the periphery and CNS, but it is the central action of nicotine that appears important to self-administration (24). Several brain mechanisms have been identified that are critical to nicotine IVSA, as described below. In the search for more effective pharmacotherapies for cigarette smoking, numerous drugs have been tested in animals self-administering intravenous nicotine. However, more testing will be necessary to determine whether the standard nicotine self-administration procedure can usefully predict clinical efficacy (25, 26).

The Surgeon General’s 1988 conclusions relied quite heavily on two published reports of intravenous nicotine self-administration in human subjects (27, 28). However, these studies have been strongly criticized on a number of grounds (13). First, few subjects were tested in any condition; several had a history of drug abuse. Second, few subjects appeared to self-administer more nicotine than saline. Third, several reported observations about nicotine self-administration were not sufficiently supported by data. Fourth, subjects were told that they might receive nicotine, but nevertheless identified the drug as cocaine. This last observation might be related to the rapidity of drug delivery (9 s), or the use of unit doses (approx. 10–40 μg/kg) that far exceed the nicotine yield from single cigarette puffs (approx. 1–2 μg/kg) (29).

Two recent reports provide stronger evidence of IVSA in human smokers (30, 31). In one of these studies (31) subjects were willing to respond 1,600 times for each nicotine infusion and responded very little for saline infusions. However, several subjects were past or present drug users, and it is not clear how far these results would generalize to the average cigarette smoker (31). A further complication in human IVSA studies is that nicotine produces a recognizable cue, and many subjects believe that nicotine is addictive. Therefore, nicotine self-administration, where observed, could potentially be generated by expectation, unless this possibility is explicitly ruled out (13, 32).

Human subjects have also been tested for self-administration of nicotine in the form of nasal spray and polacrilex gum. In several acute studies, Perkins et al. have compared self-administration of nicotine versus placebo nasal spray in a forced choice task; overall, fewer than half of the smokers chose nicotine consistently over saline, even when abstinent from cigarettes (33, 34). However, a longer-term study suggested that reinforcing effects of nicotine nasal spray may take several days to emerge (35). Similarly, most subjects prefer placebo chewing gum to nicotine gum in acute tests (36, 37), whereas nicotine gum is preferred with prolonged access (35).

In summary, a clear demonstration of nicotine self-administration in humans remains elusive (32), although animals will clearly self-administer the drug. Animal models of drug self-administration offer some significant benefits; for example, more rigorous controls are possible, and the problems of past drug history and biases based on expectations are avoided. The following sections deal with self-­administration of nicotine in animals and will be focused primarily on one animal – the adult rat.

The IVSA procedure represents the gold standard for studying the motivational aspects of a drug in animals. This procedure ­possesses significant face validity (38), and it results in patterns of intake that closely match human drug consumption (39–42). In addition, the progressive ratio schedule of reinforcement provides a measure of how motivated the animal is to obtain drug infusions (43). In this procedure, the amount of work required to receive each successive infusion of a drug increases exponentially. At some point, the response requirement becomes too high and the animal ceases to respond for the drug, and thus, the amount of responding required to obtain the last infusion is termed the “breakpoint.” Other drugs can be tested within this procedure to measure any change in the motivation to self-administer a ­particular drug (see (44) for a review).

The standard nicotine IVSA procedure in the rat is essentially the same as that for other drugs of abuse. Under general anesthesia, the animal is catheterized in the right jugular vein and the catheter is passed subcutaneously to a cannula that is either fixed to the skull or positioned to exit between the scapulae. Cathe­terization surgery is described in detail elsewhere (45). In order to obtain nicotine infusions, the animal makes an operant response. This response may be a lever press, a nose-poke entry into a receptacle, or movement of a wheel manipulandum. During the self-­administration session, a houselight may be illuminated. The infusion is almost invariably coupled with a sensory cue, usually a light placed above the lever or an auditory signal (i.e., a tone). Such sensory cues play a critical role in nicotine self-­administration, as discussed below. Following the infusion, a timeout period is typically imposed, lasting from seconds to minutes, during which no further infusion is available; this period serves to minimize aversive drug effects and prevent overdose. Usually, a second, “inactive” lever is provided; responses on this lever do not result in drug delivery and instead serve as a control for general increases in ­activity caused by the drug or by some other experimental manipulation.

The first demonstrations of intravenous nicotine self-­administration in rats employed a variety of methods. For example, Singer et al. obtained high levels of nicotine self-administration by using a procedure normally associated with schedule-induced polydipsia (46). Accordingly, rats were food-deprived and placed on a fixed interval 60 s (FI 60) schedule of food delivery. Cox and coworkers provided an early demonstration of IVSA using a more conventional approach (47). These authors showed that rats would press a lever to obtain intravenous nicotine at doses of 10 and 30 μg/kg (but not 3 μg/kg), and would redirect their responding when nicotine was reassigned to a previously inactive lever. Nicotine self-administration did not require schedule induction, pre-exposure or food restriction. However, only a subset of the rats responded preferentially for nicotine in a consistent manner.

Nicotine self-administration was further refined and characterized by DeNoble and Mele, in a paper whose publication was involuntarily delayed by 20 years (48). Meanwhile, in 1989, Corrigall and Coen described a nicotine IVSA procedure that has served as a template for many subsequent studies (49). Their seminal paper described robust self-administration of nicotine across a range of doses and across various schedules of reinforcement (see Fig. 1).

The refinements associated with the Corrigall and Coen (1989) procedure are as follows:

Several intrinsic factors affect rates of nicotine IVSA in rodents, including genetic strain (57, 60, 65, 76) and sex (63, 70, 77). Developmental stage also plays a role, and several groups have reported that adolescent rats self-administer intravenous nicotine more than adult rats (58, 76, 78). However, when response demands are increased, adolescent rats may actually self-administer less than adult subjects (79).

The widely used Corrigall–Coen procedure permits limited, 1-h daily access to nicotine. However, some researchers provide more extended access. For example, Sharp and associates permitted rats to lever-press for infusions of nicotine over a 23-h period (51). This procedure has better face validity, and provides nicotine plasma levels closer to those found in human smokers (80). Interestingly, total daily nicotine intake in the unlimited access procedure is not significantly different than in the limited access procedure. Extended-access nicotine IVSA has been studied with respect to strain (57), age (81), extinction behavior (55) and pregnancy (70). Somewhat disheartening, from a treatment standpoint, is the finding that reduction in access to nicotine results in a compensatory increase in daily intake that rebounds above baseline levels upon resumption of normal access (68).

Many drugs have been tested in rodent nicotine IVSA procedures, providing some insight into the neurobiological mechanisms underpinning this behavior. Typically, drug challenges are presented acutely to rats that have already acquired nicotine IVSA in limited-access sessions on a fixed-ratio schedule of reinforcement. This behavioral assay is also widely used to screen for novel drugs that might inhibit cigarette smoking, although the predictive validity of this procedure has not been established (26, 38). Here, we highlight some of the main findings in the drug-testing arena; more detailed reviews are available elsewhere (82–84). Unless otherwise stated, studies refer to acute drug effects on established nicotine IVSA.

Relapse is a recurring problem in the treatment of smoking. Indeed, of smokers who quit without pharmacological or psychological support, only 3–5% remain abstinent after 6–12 months (151). In abstinent subjects, drug seeking can be triggered by a variety of stimuli including drug-associated cues, exposure to the drug itself, or stress. Different relapse scenarios are modeled in the rodent reinstatement procedure of drug self-administration, pioneered by de Wit and Stewart (152). In this procedure, rats are first trained to self-administer a drug. Self-administration behavior is then extinguished over multiple sessions, during which drug-associated cues may or may not be available. Animals are then acutely exposed to three types of stimuli: the drug itself, stressors, or drug-associated cues provided the latter have not been already extinguished. The resultant increase in (unreinforced) responding provides a measure of drug seeking. Brain mechanisms underlying reinstatement are highly dependent on the nature of the triggering stimulus (153). This behavioral procedure is now widely used to screen drugs for therapeutic potential.

Reinstatement related to nicotine IVSA has recently been reviewed in detail (9). Reinstatement has been induced by intravenous and systemic injections of nicotine (154, 155), footshock stress (156) and re-exposure to nicotine-associated stimuli (157). Nicotine-induced reinstatement was inhibited by antagonists of dopaminergic (SB277011A), glutamatergic (MPEP, EMQMCM) (158, 159), and cannabinoid (AM251) (139) receptors; the mGluR1 antagonist EMQMCM also reduced reinstatement of food seeking (159). In contrast, stress-induced reinstatement was attenuated by both the corticotropin-releasing factor (CRF) receptor 1/2 antagonist D-Phe CRF (12–41) (administered icv) and the α2 adrenergic receptor agonist clonidine (administered systemically) (160).

Studies of cue reactivity in cigarette smokers have largely focused on craving and physiological measures rather than drug seeking behavior itself (161). Indeed, there is surprisingly little direct evidence that smoking cues actually promote the occurrence of smoking behavior (161–163). Nevertheless, various drugs have been screened in animals for their ability to block reinstatement of nicotine seeking triggered by nicotine-associated cues. Drugs that inhibit reinstatement include the nAChR antagonist mecamylamine (164, 165), and modulators of GABA (CGP44532) (127), glutamate (MPEP, EMQMCM) (159, 166) and cannabinoid receptors (SR141716, AM251) (139, 167). In contrast, bupropion inhibited nicotine self-administration prior to extinction, but acutely increased reinstatement triggered by nicotine-associated cues (157). Speculatively, the latter result may be related to the limited efficacy of bupropion in smoking cessation (but see above).