Sweet taste potentiates the reinforcing effects of e-cigarettes

JID: NEUPSY

ARTICLE IN PRESS

[m6+;August 6, 2018;13:13]

European Neuropsychopharmacology (2018) 000, 1–14

www.elsevier.com/locate/euroneuro

Sweet taste potentiates the reinforcing effects of e-cigarettes Nils B. Kroemer a,b,c,d,∗, Maria G. Veldhuizen a,b, Roberta Delvy b,e, Barkha P. Patel a,b,f, Stephanie S. O’Malley a, Dana M. Small a,b,g,∗∗ a

Department of Psychiatry, Yale University School of Medicine, New Haven, CT 06511, USA The John B. Pierce Laboratory, New Haven, CT 06519, USA c Department of Psychiatry and Psychotherapy, University of Tübingen, 72076 Tübingen, Germany d Department of Psychiatry and Neuroimaging Center, Technische Universität Dresden, 01187 Dresden, Germany e School of Nursing, Yale University, Orange, CT 06477, USA f Division of Endocrinology, The Hospital for Sick Children, Toronto, ON M5G 1X8, Canada g Department of Psychology, Yale University, New Haven, CT 06520, USA b

Received 19 March 2018; received in revised form 16 July 2018; accepted 17 July 2018 Available online xxx

KEYWORDS Flavor conditioning; fMRI; Reward; Addictive liability; Electronic cigarettes; Smoking

Abstract Electronic cigarettes (e-cigarettes) are becoming increasingly popular. The popularity of fruit flavors among e-cigarette users suggests that sweet taste may contribute to e-cigarette appeal. We therefore tested whether sweet taste potentiates the reinforcing effects of nicotine. Using a conditioning paradigm adapted to study e-cigarettes, we tested whether exposure to flavored e-cigarettes containing nicotine plus sweet taste would be more reinforcing than unsweetened e-cigarettes. Sixteen light cigarette smokers smoked 4 distinctly colored e-cigarettes containing sweetened and unsweetened flavors with or without nicotine for 2 days each. Brain response was then assessed to the sight and smell of the 4 exposed e-cigarettes using fMRI. After exposure, sweet-paired flavors were wanted (p = .024) and tended to be liked (p = .053) more than nicotine-paired flavors. Moreover, sweet taste supra-additively increased liking for nicotinepaired flavors in individuals who did not show increased liking for nicotine alone (r = −.67, p = .005). Accordingly, cues predicting sweet compared to non-sweet flavors elicited a stronger response in the nucleus accumbens (NAcc, pSVC = .050) and the magnitude of response to the sight (pSVC = .022) and smell (pSVC = .017) of the e-cigarettes correlated with changes in liking.

∗ Corresponding

author at: Department of Psychiatry and Psychotherapy, University of Tübingen, 72076 Tübingen, Germany. author at: Department of Psychiatry, Yale University School of Medicine, New Haven, CT 06511, USA. E-mail addresses: [email protected] (N.B. Kroemer), [email protected] (D.M. Small).

∗∗ Corresponding

https://doi.org/10.1016/j.euroneuro.2018.07.102 0924-977X/© 2018 Elsevier B.V. and ECNP. All rights reserved.

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

2

[m6+;August 6, 2018;13:13]

N.B. Kroemer et al. By contrast, the sight and smell of cues predicting nicotine alone failed to elicit NAcc response. However, the sight and smell of e-cigarettes paired with sweet+nicotine (pSVC = .035) produced supra-additive NAcc responses. Collectively, these findings demonstrate that sweet taste potentiates the reinforcing effects of nicotine in e-cigarettes resulting in heightened brain cue-reactivity. © 2018 Elsevier B.V. and ECNP. All rights reserved.

1.

Introduction

Organisms must procure energy to survive, and, as such, mechanisms have evolved to promote feeding. For example, sweet taste perception evolved to signal the availability of energy and promote its intake by producing a pleasant sensation and motivating ingestive behavior (de Araujo, 2011; Sclafani, 1987; Sheffield and Roby, 1950). The circuits that orchestrate such behaviors are therefore tuned to integrate sensory perceptions with post-ingestive reinforcing signals conveying nutritional information. In turn, organisms learn to like the available energy sources, leading to greater intake (de Araujo, 2016; de Araujo et al., 2013; Tellez et al., 2016). Whether sweet taste can interact with other reinforcers, such as nicotine, to promote consummatory behavior is unknown. However, this is an important gap in knowledge because tobacco products, including e-cigarettes, cigars, hookahs, and smokeless tobacco, are often consumed with sweet taste (Miao et al., 2016). Alternative tobacco products such as e-cigarettes are becoming increasingly popular (Drummond and Upson, 2014). One appeal of e-cigarettes is that they often combine nicotine with sweet taste and flavors. In an online survey involving more than 5000 regular “vapers”, the top-ten flavors were characterized by sweet or fruity taste (http:// ecigarettereviewed.com/best- e- juice- flavors). This raises the possibility that sweet flavors enhance the reinforcement potency of nicotine to promote e-cigarette use. Supporting this possibility, for smokers, the aroma of nicotine-predictive tobacco is a potent cue that can promote smoking (Carpenter et al., 2014). Likewise, cross-sectional studies suggest that flavors increase the subjective value of e-cigarettes. For example, among an adolescent sample (Ambrose et al., 2015), most reported first using flavored products. Similarly, in a study of representative US 8th, 10th, and 12th graders, taste was found to be the second most important reason for e-cigarette use (Patrick et al., 2016) and was associated with a higher frequency of use (Patrick et al., 2016). Among adults, smallscale studies also point to the importance of sweet flavors, such as fruit, although to a lesser degree than in adolescents (Kim et al., 2016; Morean et al., 2017). With continued use, nicotine becomes a highly potent reinforcer primarily acting via nicotinic acetylcholine receptors that regulate dopamine release in the mesolimbic circuit (Changeux, 2010; Maskos et al., 2005; Tolu et al., 2013; Zhang et al., 2012). Nicotine also exerts long-lasting effects on reward sensitivity (Kenny and Markou, 2006). For example, nicotine amplifies firing of ventral tegmental area (VTA; Clark and Little, 2004; Tizabi et al., 2002) and NAcc shell neurons (Tizabi et al., 2007) in response to alcohol suggesting

that nicotine can potentiate the reinforcement of other reinforcers. It is unknown whether this generalizes to sweet taste or, vice versa, if sweet taste amplifies the reinforcing effect of nicotine by increasing flavor liking or wanting. However, nicotine use is associated with differential responses in feeding circuits to food-related stimuli in humans (Geha et al., 2013; Kroemer et al., 2013) and activation of nicotinic acetylcholine receptors in feeding circuits alters appetite in rodents (Mineur et al., 2011). Taken together, data suggest that taste contributes to the reinforcing effects of nicotine-containing e-cigarettes and raise the possibility that the combination of two reinforcers may enhance the appeal of the product. If such interactions occur, the NAcc is a likely substrate. Both food and smoking cues reliably induce BOLD responses in the NAcc (Tang et al., 2012). NAcc mu-opioid hedonic “hotspots” are implicated in subjective pleasure (Berridge and Kringelbach, 2015; Castro and Berridge, 2014; Kelley et al., 2002) and striatal dopamine plays a key role in rewardrelated learning (Kroemer and Small, 2016; O’Doherty et al., 2004; Pessiglione et al., 2006; Schonberg et al., 2007; Steinberg et al., 2013; Valentin and O’Doherty, 2009; Veldhuizen et al., 2011). In line with animal studies, smoking induces dopamine release in the NAcc (Brody et al., 2004) and short-term abstinence from smoking increases BOLD responses to cigarette puffs as reward in the caudate head and NAcc (Sweitzer et al., 2014). Moreover, Pavlovian cues have been found to increase instrumental responding via dopamine D1 and D2 receptors in the NAcc (Lex and Hauber, 2008), thus entailing the potential to modulate behavior such as e-cigarette use. Collectively, these studies suggest that the addition of sweet taste could potentiate the addictive liability of tobacco products because both reinforcers act via shared neural pathways. To test whether the reinforcing effects of sweet taste and nicotine interact to enhance reward from e-cigarettes, we adapted a flavor-nutrient conditioning paradigm for use with e-cigarettes (de Araujo et al., 2013). Participants were exposed to four novel e-cigarette flavors that varied in sweetness (sweetened or unsweetened) and nicotine content (with or without; 2 × 2 design). The reinforcing effects of the nicotine and/or sweet predicting conditioned flavors were then measured by assessing liking and wanting ratings and brain responses to the sight and the smell of the predictive e-cigarette flavors (i.e. in the absence of sweet taste and nicotine). Liking reflects hedonic aspects of conditioning, whereas wanting has been shown to be a good predictor of actual drug consumption (Ostafin et al., 2010). In addition, brain response to food and nicotine-related cues is a reliable predictor of behavioral outcomes reflecting reinforcement such as weight gain (Sun et al., 2015),

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

Sweet taste potentiates the reinforcing effects of e-cigarettes snacking in the absence of hunger (Lawrence et al., 2012), and smoking cessation (Versace et al., 2014). We hypothesized that sweet taste (1) increases the appeal of flavors (i.e., main effect) and (2) potentiates the reinforcing effects of nicotine reflected by enhanced liking and wanting ratings and neural response in the NAcc to the sight and smell of e-cigarettes with sweet+nicotine compared to sweet or nicotine alone (i.e., supra-additive interaction effect).

2. 2.1.

Experimental procedures Participants

We enrolled 16 occasional/intermittent to light smokers (4 female; M = 27 years, range 19–45; body mass index, MBMI = 24.9 kg/m2 , range 19.7–30.1; categories according to Schane et al., 2010) from New Haven, CT, who reported having smoked at least 100 cigarettes in one’s lifetime, currently smoking for the last three months ≥once per month and ≤10 cigarettes per day and a score of ≤2 on the Fagerström Test for Nicotine Dependence (FTND; 7 cases FTND = 0, 6 cases FTND = 1, 3 cases FTND = 2; Heatherton et al., 1991). Participants were not required to have used e-cigarettes, and none regularly used. Intermittent to light smokers were selected because we reasoned that their consumption would be more strongly guided by instrumental aspects (e.g., Vollstädt-Klein et al., 2011), such as reinforcing potential of a flavor. Participants were screened to be 18–45 years of age and free from psychiatric or eating disorders and chronic medical conditions, head injury with loss of consciousness or chemosensory impairments, history of drug or alcohol dependence, current drug use other than tobacco and alcohol, daily medications (except for monophasic birth control), and ferromagnetic material that would preclude an MRI. At their first visit, all participants provided written informed consent. The study was approved by the Yale Human Investigations Committee.

2.2.

Experimental procedure

Participants were asked to refrain from smoking for 10h and to arrive at the laboratory in a fasted state for the screening/intake session (for procedural details, see SI). Smoking abstinence was verified by an expired air CO level ≤10 ppm (median CO = 7.5 ppm, range 2–10). Breath alcohol levels were measured (Alcohawk Elite Breathalyzer) and urine toxicology screens for illicit drugs were conducted using the Integrated E-Z Split Key Cup II (Innovacon Inc., San Diego, CA). Females of childbearing potential were given a urine pregnancy test. Participants who tested positive on any of these screens were informed of the result and excluded from further participation. Thereafter, participants were familiarized with our psychophysical scales and trained in their use. Visual analogue scales assessed hunger, wanting, and familiarity (left anchor: “Not at all” right anchor: “Extremely”). Perceived intensity was assessed using the general Labeled Magnitude Scale (gLMS; Green et al., 1996). Liking was assessed with the Labeled Hedonic Scale, LHS (Lim et al., 2009), an empirically derived rating scale that produces ratio-like data and enables participants to rate how much they like (0–100) or dislike (−100 to 0) a stimulus, with the following labels slightly asymmetrically spaced around “neutral” (at 0): (dis)like slightly, (dis)like moderately, (dis)like very much and (dis)like extremely. To familiarize participants with the procedures, the screening comprised a mock fMRI run including the vapor delivery device (Small et al., 2008; Veldhuizen et al., 2010). To reduce exposure

[m6+;August 6, 2018;13:13]

3 to the e-cigarette odorants used throughout the study, participants received vapors different from pre-test odorants. To select equally-liked flavors for exposure, participants then completed a pre-exposure evaluation of all flavors without the addition of sweetener or nicotine (Fig. 1). Participants were handed 10 distinctly colored e-cigarettes one at a time and instructed to puff each vaporized flavor three times before rating the flavor for harshness, sweetness, coolness, intensity, familiarity, wanting, and liking. Participants rinsed their mouths with water between flavors to minimize carry-over effects. Each flavor was rated three times leading to a total of 30 trials. These ratings were then used to select equally liked flavors by inspecting ratings displayed in a boxplot with labeled scale anchors for reference. Potential participants (N = 6) were excluded if we could not identify five similarly liked flavors within the range of neutral to moderately liked or if ratings fell consistently outside the predefined range. Eligible participants were invited for exposure sessions. To this end, the five selected flavors were randomized to the following conditions: (1) flavor only, (2) flavor+sweetener, (3) flavor+nicotine, (4) flavor+sweetener+nicotine, and (5) unexposed control flavor. Moreover, we permuted the order of the exposure conditions and ratings indicated no order confounds (see SI). While four flavors were used during the exposure sessions to assess the main effects and interaction of the 2 two-level factors sweetness and nicotine, the fifth flavor served as an initially equally-liked control flavor for fMRI testing. Participants completed 2 exposure days per e-cigarette condition for a total of 8 exposure days within 2–3 weeks. Participants first sampled the e-cigarette and rated the flavor and their wanting and liking in the laboratory. Then, they took home the e-cigarette plus the filled and uniquely colored cartridge for use throughout the next two days to use the e-cigarettes in a real-life setting. Participants were asked to take a minimum of 20 puffs per day and to use the e-cigarettes approximately equally often. After the twoday exposure, participants attended a laboratory session in which they sampled and rated the exposed flavor again. This procedure was repeated for each e-cigarette condition (i.e., 4 flavor-additive pairings, each exposed for 2 days). After the exposure period, participants completed fMRI testing. They were outfitted with the nasal mask for odor delivery and inserted into the scanner (see fMRI paradigm). Lastly, participants completed a post-test that was identical to the pre-test. Critically, all flavors were sampled again without added sweetener or nicotine, but with the corresponding color bands to assess flavor conditioning. Participants were then debriefed and received a completion bonus payment.

2.3.

E-cigarettes

The Smokio vaping system and Evod2 clearomizers filled with 1ml of e-liquid were used. We chose 10 novel non-sweet flavors, including star fruit, lemongrass, dill pickle, anise, allspice, Cuban gold tobacco, lavender, jasmine, black pepper, and neroli. These flavors were selected after extensive evaluation by the study team to ensure that flavors were sufficiently distinct (Fig. S.1), not perceived as sweet, and not associated with calories. All e-liquids were made from a base of 50% propylene glycol (PG), 50% vegetable glycerin (VG), a standard base for many e-liquids (e.g., Audrain-McGovern et al., 2016). The nicotine content for nicotine conditions was 12 mg/ml (moderate level; Audrain-McGovern et al., 2016). Without the addition of sweetener, most ratings fell below a moderate intensity of sweetness (71%) regardless of the flavor (Fig. S.2). The e-liquids were sweetened with the addition of 20 drops/ml of ethyl maltol in 50%PG/50%VG base (ECBlends), which is the most common sweet-associated component in e-cigarette liquids (Miao et al., 2016). In pre-tests, this amount produced a noticeable sweet taste when vaped like commercially available e-liquids, without

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

4

ARTICLE IN PRESS

[m6+;August 6, 2018;13:13]

N.B. Kroemer et al.

Fig. 1 Schematic of the study design. a: Participants completed a pre-exposure evaluation of all flavors without the addition of sweetener or nicotine. b: For exposure sessions, we selected five equally liked flavors and randomized them to the e-cigarette conditions. Participants completed 2 exposure days per condition for a total of 8 exposure days within 2–3 weeks. c: Paradigm used to study brain response to the sight and the smell of the conditioned e-cigarettes. To reduce the statistical dependency of the regressors, the onset of the e-cigarettes odor was jittered (truncated exponential distribution, mu = 3, max = 16) and enabled us to decompose the brain response correlates. VG = vegetable glycerin, PG = propylene glycol, dur = duration. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

[m6+;August 6, 2018;13:13]

Sweet taste potentiates the reinforcing effects of e-cigarettes diluting the intensity. Colored bands were added to each clearomizer with a small band of heat-shrink tubing to distinguish the different e-cigarettes. Flavor–condition–color pairings were constant within but counterbalanced across individuals. Moreover, both experimenter and participants were blinded to the pairings.

2.4.

fMRI paradigm

Brain response to the sight and smell of the 4 exposed flavors and 1 unexposed e-cigarette flavor was measured with fMRI. First, an image of the e-cigarette was presented with the corresponding colored sticker superimposed. To prepare for the delivery of an odor, participants heard the instruction “Sniff” 0.5s before the onset of the vaporized flavor. Flavors were delivered for 3s to the nasal mask fitted over each participant’s nose using our fMRI-compatible olfactometer (Veldhuizen et al., 2010). Each vaporized flavor was presented 7 times in random order per block (Fig. 1) for a total of 21 repeats per stimulus. Participants completed 3 blocks in total. Two blocks had to be excluded from the analysis for one participant due to a synchronization error.

2.5.

fMRI data acquisition and preprocessing

Imaging data were acquired on a Siemens 3.0T TIM Trio Scanner at the Yale University Magnetic Resonance Research Center (see SI). To collect BOLD responses, we used a high-resolution multiband echoplanar imaging (EPI) sequence. Imaging data were preprocessed and analyzed with SPM12 (Wellcome Department of Imaging Neuroscience, London, UK). Preprocessing included coregistration, unwarping, slice-time correction, and realignment. The anatomical scan was coregistered to the EPI data before segmenting and normalizing it to MNI space. Normalization was then applied to the functional data, resampled with a voxel size of 2 × 2 × 2 mm3 and smoothed using an isotropic Gaussian kernel (6 mm full-width at half-maximum).

3. 3.1.

Data analysis

5 the exposure sessions. More specifically, we performed a reduced mixed-effects analysis as implemented in the univariate ANOVA in SPSS (fixed factor: pre/post exposure, random factor: participant) on the sessions involving e-cigarettes that contained nicotine only. Lastly, we did not include session order effects in the statistical model since we found no evidence for a confounding effect during exposure sessions (see SI).

3.2.

fMRI data analysis

For the first-level analysis, we modeled the conditions “sight” (as event), “smell” (as epoch, duration = 3s), rating (as epoch, duration = 5s–response time), and motor (as an event occurring when the rating was submitted) as explanatory variables within the context of the general linear model on a voxel-by-voxel basis. Then, we used parametric modulators to decompose the variance of the five flavors using a GLM approach. We dummy-coded and effect-centered (i.e., conditions were coded with 0.5 and −0.5) the regressors in the following order: nicotine, sweet, sweet × nicotine, and exposed. The “exposed” regressor reflected all flavors that were used during the exposure phase to control for a potential effect of mere repeated exposure to a flavor. However, we observed no differences in brain responses that survived correction for multiple comparisons indicating that mere exposure is unlikely to have a strong effect on the perception of conditioned flavors within our paradigm. For the current paper, we focus on sight and smell as conditions of interest for second-level analysis (see SI). Due to our a priori hypothesis on the correspondence between conditioned changes in liking and brain response, we focused only on correlational analyses with liking. Notably, the high correlation of conditioned changes in liking and wanting precludes a dissociation at the individual level (see Fig. 3). Realignment parameters (3 translation, 3 rotation) were included in the model as nuisance regressors.

Perceptual ratings

To account for inter-individual differences in flavor liking and wanting, we included random effects for the flavor only control (intercept) and main effects of sweet, nicotine as well as the interaction sweet × nicotine in a hierarchical linear model. To calculate interaction effects with exposure sessions (pre vs. post) at the group level, we used only fixed effects because significance tests of the variance components indicated that the random effects could be omitted in favor of a simpler statistical model (4 out of 6 had p > .1). Consistent with this observation, model comparisons indicated that the more complex models were not considerably better in explaining the data (wanting: χ (21)= 31.17, p = .071; liking: χ (21) = 37.27, p = .016). Thus, to derive the best estimates, we used the more parsimonious models for inference of group effects. In addition, we fitted models with random effects for every parameter as estimates of individual effects and used them to explore potential correlations of changes in ratings across conditions that might be suggestive of common mechanisms. To determine if liking and wanting changed over time during exposure sessions (reflecting conditioning effects with nicotine present), we conducted post hoc analyses on ratings collected during

3.3.

Statistical threshold and software

For behavioral and ROI analyses, we used α ≤.05 (twotailed) as significance threshold. A priori hypothesis tests or exploratory analyses are specified throughout the manuscript. Based on prior work implicating the NAcc in conditioned flavor learning (de Araujo et al., 2013) and in keeping with our published investigations of NAcc function, we used a mask derived from the brainmap database (Kroemer et al., 2014, 2013, 2016, 2015) to correct for multiple comparisons across voxels within the small volume. For exploratory fMRI whole-brain analyses, we used onesided contrast maps (mass-univariate t-tests) thresholded at p < .001 and k > 20 for display and reported clusters exceeding a family-wise error correction at the cluster level. Moreover, we also report one effect that failed to exceed cluster-level correction, while surviving correction at the peak level. Since this correlation with liking in the mPFC was congruently seen across the sight and smell phases of the design, it adds support that this finding is unlikely to have arisen by chance alone. To analyze and plot data, we used SPSS v23-24, R v3.3.2 (R Core Team, 2015), R_Deducer

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

6

[m6+;August 6, 2018;13:13]

N.B. Kroemer et al.

Fig. 2 Sweet taste compared to nicotine increases wanting ratings of the flavors with which it is paired. a: Pre–post changes in flavor ratings (without additives). b: Pre–post changes in flavor ratings during the initial evaluation and after exposure (including additives). No conditioning occurs for e-cigarette flavors paired with nicotine although wanting and liking ratings increase after repeated use. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

(Fellows, 2012), HLM v7 (Raudenbush et al., 2011), Mango v4.0, and MATLAB v2016b.

4.

Results

4.1. Effects of sweetener and nicotine on liking and wanting ratings To assess the conditioning effects of sweet, nicotine, and sweet × nicotine in our factional design, we used a hierarchical linear model (see Methods) entering the pre- and post-exposure liking and wanting ratings of the distinctly colored and flavored e-cigarettes (note that there are no

sweeteners or nicotine present during either testing session). Changes in wanting (࢞wanting = 6.31; t = 1.18) and liking (࢞liking =7.47; t = 1.28) of sweet-paired e-cigarettes versus control e-cigarettes were not significantly higher (p > .20; Figs. 2a and S.3). However, post hoc effect-size estimates (wanting r = .29; liking r = .30) suggested that sweet conditioning yielded moderately-sized increases compared to control e-cigarettes. In addition, increases in wanting for the sweet-paired e-cigarette after conditioning were significantly higher compared with the nicotine-paired ecigarette, χ (1) = 4.97, p = .024. Likewise, we observed that the sweet-paired e-cigarettes tended to become more liked than the nicotine-paired e-cigarettes after conditioning,

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

Sweet taste potentiates the reinforcing effects of e-cigarettes but this difference did not reach significance, χ (1) = 3.63, p = .053. Moreover, there was no supra-additive effect of combining sweet+nicotine on ratings of wanting and liking (sweet × nicotine, ps >.5). Since we did not observe the hypothesized conditioned increase in liking or wanting for the e-cigarettes previously paired with nicotine, we examined the ratings collected during the exposure sessions when sweetness and nicotine were present. Wanting of the e-cigarette with nicotine increased over exposure sessions, F(1,15) = 9.3, p = .008, and a trend towards an increase in liking, F(1,15) = 4.2, p = .060, was observed (Fig. 2b). This indicates that nicotine was reinforcing after repeated use of the e-cigarette but failed in generalizing to liking and wanting for the nicotinepredictive e-cigarette flavors. Next, we explored if changes in flavor liking and wanting were correlated across e-cigarettes. Intriguingly, we observed that participants who disliked (r = −.67, p = .005) and did not want (r = −.53, p = .034) the e-cigarettes containing only nicotine showed the strongest increase in liking for the e-cigarettes paired with sweet+nicotine. In other words, the addition of sweetness over-compensated for the negative effect of nicotine on liking and wanting in participants who disliked nicotine alone (Fig. 3).

4.2. Increased NAcc response to the sight but not the aroma of the sweet-paired e-cigarettes We hypothesized that the sight of the e-cigarettes that had been paired with sweet, nicotine, and sweet+nicotine would elicit stronger NAcc responses compared to the control e-cigarettes. Therefore, we analyzed the effects of the conditions using a factorial coding of the design within the GLM analysis (see Methods). In line with behavioral results, we observed a stronger BOLD response in the NAcc to the sight of the sweet-paired (pSVC = .050), but not to the nicotine-paired (p ≥.05 uncorrected) e-cigarettes. Furthermore, as predicted a supra-additive response was observed in the NAcc to the sight of the e-cigarette paired with sweet+nicotine (pSVC = .035; Table 1, Fig. 4; for additional information and the unmodulated cue response, see SI and Fig. S.4, respectively). Next, we explored the effects of conditioning on e-cigarette aromas. We observed robust responses in chemosensory regions including the piriform cortex, the lateral orbitofrontal cortex, and the insula (Fig. S.5). However, we observed no effect of sweet, nicotine, or a sweet × nicotine interaction in the NAcc that survived correction within the small volume. Lastly, we also explored the effects of conditioning on response to the sight and the smell at a whole-brain level but failed to see effects that exceeded whole-brain correction for multiple comparisons.

4.3. Change in e-cigarette liking is correlated with BOLD response to the sight and smell of sweet e-cigarettes Since the neural effects of conditioning in the NAcc mirrored the behavioral results, we tested if the strength of increases in e-cigarette liking were associated with BOLD re-

[m6+;August 6, 2018;13:13]

7 sponses. To this end, we set up two additional group statistics, each including the contrast images of sweet, nicotine, and sweet × nicotine and covariates for the increases in liking after conditioning. We focused on conditioned liking due to our a priori hypothesis and did not conduct additional regression analyses for wanting ratings, which were strongly correlated with changes in liking. Critically, we observed that the magnitude of BOLD responses in the NAcc to the sight (pSVC = .022) and the smell (pSVC = .017; Table 1, Fig. 5) of the sweet-paired e-cigarettes correlated with changes in liking. Correlations with conditioned changes in liking for the nicotine-paired e-cigarettes or with the supra-additive effect of sweet+nicotine were not significant. Furthermore, at the whole-brain level, significant correlations between liking and the sight and smell of the sweet-paired e-cigarette survived whole-brain correction for multiple comparisons. For the sight of the e-cigarette, the effects were in the mid cingulate gyrus, the anterior cingulate gyrus (ACC) and the medial prefrontal gyrus (mPFC; Table 1). For the aroma of the e-cigarette, effects were in the ACC/mPFC and the olfactory cortex (piriform cortex/amygdala; Table 1). Whereas the significant correlations with changes in liking for activation elicited by the sight and the smell of sweet-paired e-cigarettes were mostly non-overlapping at the given whole-brain threshold, we observed convergence within the mPFC (12/54/−4; Fig. S.6). Collectively, these results suggest that the strength of the conditioned response in the brain is closely associated with increases in conditioned flavor liking.

5.

Discussion

Sweet taste and nicotine both exert reinforcing effects on behavior by acting upon NAcc circuits (Tang et al., 2012). However, it was unknown if these reinforcers interact to potentiate the reinforcing effect of e-cigarettes rendering them potentially more addictive. To test this hypothesis, we adapted a flavor-nutrient conditioning paradigm to e-cigarettes. As predicted, we found that added sweetener can increase the reinforcing potential of an e-cigarette with and without nicotine via conditioning. Effective conditioning was reflected in moderate, but non-significant, increases of wanting and liking compared to the flavor only control, and large effects on increased NAcc response to the sight of e-cigarettes paired with sweet taste, relative to the control e-cigarettes. Moreover, the magnitude of change in liking was associated with the magnitude of NAcc response to the sight and smell of the sweet-paired e-cigarettes. In contrast, conditioning did not occur for nicotine-paired e-cigarettes. However, supra-additive effects of sweet+nicotine were observed in the NAcc as predicted. Here, response to the sight of the sweet+nicotine e-cigarette was more than the additive response to the sight of the e-cigarettes paired with sweet or nicotine alone. Although supra-additive effects were not observed for liking, sweet taste did increase liking of the sweet+nicotine e-cigarette in those who disliked e-cigarettes associated with nicotine alone. Taken together, our results suggest that sweet taste potentiates the reinforcing value of nicotine-containing e-cigarettes and that a mesocorticolimbic mechanism subserves this effect. More specifically, the

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

8

ARTICLE IN PRESS

[m6+;August 6, 2018;13:13]

N.B. Kroemer et al.

Fig. 3 Heatmap depicting correlations of changes in liking and wanting ratings after exposure. Boxes outlined in white and showing numbers indicate significant correlations (p < .05). S = sweet, Nic = Nicotine, ∗∗ = p < .01. Color indicates direction and magnitude of the correlation. Only the upper triangle of the matrix is color-coded to improve clarity. Individuals reporting a decrease in liking for the nicotine-paired flavor show greater changes in liking for the flavor paired with sweet+nicotine. No association is observed between changes in liking for the flavor paired with sweet and the flavor paired with sweet+nicotine. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

ARTICLE IN PRESS

JID: NEUPSY

[m6+;August 6, 2018;13:13]

Sweet taste potentiates the reinforcing effects of e-cigarettes

Table 1 Contrast

9

Summary of fMRI results. Brain region

MNI X/Y/Z

tmax

pFWE_peak

Main and interaction effects of conditioning: Sight: Sweet NAcc −12/18/−10∗a 3.71 .050 Sight: Nicotine NAcc −10/8/−6∗a 1.68 ≥.05 unc. Sight: SxNic NAcc −10/6/−12∗a 3.81 .035 No significant main or interaction effects for the smell of e-cigarettes Correlations with conditioned increases in liking for sweet-paired e-cigarettes: Sight: r(Sweet, NAcc −16/14/−10∗a 3.89 .022 conditioned liking) mid cingulate 26/12/34 6.36 .006 ACC 18/40/10 6.31 .007 mPFC 16/58/2 6.03 .015 Smell: r(Sweet, NAcc −14/2/−6∗a 4.01 .017 conditioned liking) ACC/mPFC −12/48/2 5.62 .049 olfactory cortex 34/6/−12 5.10 .188

k

pFWE_cluster

234 162 127

.005 .029 .072

273 166

.002 .025

S = sweet, Nic = Nicotine, NAcc = nucleus accumbens, unc. = uncorrected, ACC = anterior cingulate cortex, mPFC = medial prefrontal cortex. a∗ = peak voxel within region of interest.

Fig. 4 Sweet taste increases the brain response to conditioned cues during the anticipatory phase. a: In the a priori region-ofinterest nucleus accumbens (red outline), we observed increased BOLD response to sweet flavors after conditioning (tmax = 3.71, pSVC = .050). b: Whereas there was no main effect of nicotine on anticipatory brain response in the nucleus accumbens, we observed a significantly positive sweet x nicotine interaction (tmax = 3.81, pSVC = .035) indicative of a supra-additive effect. The images are displayed at t > 2 and k > 60 (uncorrected) and highlighted clusters exceed small-volume correction (SVC) for multiple comparisons. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

10

[m6+;August 6, 2018;13:13]

N.B. Kroemer et al.

Fig. 5 Increased liking of sweet flavors is associated with increased brain response. a: During the anticipatory phase, increased BOLD response in the mid cingulate gyrus, the anterior cingulate gyrus (ACC), and the medial prefrontal gyrus (mPFC) was associated with increased liking of sweet flavors after conditioning. b: When the odor was being delivered, increased BOLD response in the ACC/mPFC and the piriform cortex/amygdala was associated with increased liking of sweet flavors after conditioning. The images are displayed at p < .001 and k > 20 (uncorrected) and highlighted regions exceed whole-brain correction for multiple comparisons. Scatterplots depict increases in brain response to the sight and the smell of the sweet-paired flavors as a function of increases in liking. Selection of the clusters was dependent on the group analysis of the correlations and the plots are just used as illustrations. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

correspondence between liking and NAcc response supports the hypothesis that the activation is indicative of changes in reinforcement value and increased addictive potential.

5.1. Sweet taste increases e-cigarette liking and brain cue-reactivity In line with our hypothesis, we found that sweet taste increased liking, wanting, and NAcc response to e-cigarettes. These results extend prior work showing that liking of ecigarette flavors is correlated with perceptions of sweetness (Kim et al., 2016) by demonstrating the reinforcing capability of sweet taste. Intriguingly, in contrast to our previ-

ous study where NAcc response to a flavor previously paired with calories did not correlate with changes in flavor liking (de Araujo et al., 2013), here we found that NAcc response to the sight and smell of the sweet-paired e-cigarette was correlated with change in liking. This suggests that the underlying neurobiological mechanisms for sweet conditioning differ from nutrient conditioning. More specifically, sweet taste appears to be more effective than nutritional value at conditioning flavor liking via the NAcc (Tellez et al., 2016). In addition to NAcc, changes in liking for the sweet-paired e-cigarettes correlated with other regions in the “valuation network” (Bartra et al., 2013), namely the ACC/mPFC. Likewise, a correlation with change in liking was observed in response to the aroma of sweet-paired e-cigarettes in the

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

Sweet taste potentiates the reinforcing effects of e-cigarettes ACC and the piriform cortex/amygdala. The piriform corresponds to primary olfactory cortex (Howard et al., 2009; Veldhuizen et al., 2010) and its projections have been shown to be involved in identity-specific and predictive signaling of reward (Howard et al., 2015). Collectively, these correlated changes in brain response at the whole-brain level reflect the changes in liking induced by conditioning with sweet taste within the mesocorticolimbic circuit.

5.2. Sweet taste potentiates nicotine reinforcement signals Supporting our primary hypothesis, we observed a supraadditive response in the NAcc to the sight of the e-cigarette paired with sweet+nicotine versus the e-cigarettes paired with sweet or nicotine alone. However, we failed to see a supra-additive effect on liking and wanting. One possibility is that the neural effect might be driven partly by the negative correlation between changes in liking for the nicotine-paired e-cigarette and the sweet+nicotine ecigarette. A similar compensatory effect has been observed for food intake in rodents, where sweet taste can largely counteract the negative effect of nicotine (Grunberg et al., 1985). To the best of our knowledge, we are the first to report such a supra-additive interaction effect of sweet taste and nicotine on the NAcc response to a conditioned stimulus. Notably, the main effect of sweetener and the sweet × nicotine interaction effect were not overlapping within the NAcc and occurred along the tilted anterior– posterior axis, which differentiates NAcc core and shell in humans (Baliki et al., 2013). Whereas we observed stronger effects of conditioning on the response in the left NAcc, effects in the right NAcc were qualitatively similar and larger studies are needed to establish lateralization of the effects.

5.3. E-cigarettes containing only nicotine did not condition liking or brain response In contrast to our initial hypothesis, we failed to see a robust conditioning effect of nicotine on pre–post flavor ratings and brain response to the sight and smell of the e-cigarettes. While unanticipated, this finding is consistent with prior reports that non-sweet e-cigarettes are liked less and perceived as more harsh and bitter (i.e., qualities of nicotine; Kim et al., 2016). However, it is also possible that the rapid action of nicotine on the brain conditioned the sight and smell of the e-cigarettes to generate positive prediction signals which would have been followed by a negative prediction error signal at the post-test when the effect of nicotine on central circuits would be predicted but absent. If so, this could account for decreased liking and the absence of NAcc response. Several other explanations are also possible, but less likely. First, it is possible that nicotine did not exert reinforcing effects in our group of intermittent and light smokers. However, increases in wanting and liking ratings during exposure sessions argue against this interpretation. Likewise, previous smoking-cue reactivity studies indicate that lower severity of nicotine dependence is associated with stronger reactivity and cue-induced craving (Vollstädt-Klein et al., 2011; Watson et al., 2010). Sec-

[m6+;August 6, 2018;13:13]

11 ond, nicotine’s conditioning effects for flavors might follow a different learning trajectory compared to sweet taste perception such that more extensive exposures are required. However, the large body of evidence on the addictive potency (Benowitz, 1988; Pontieri et al., 1996; Russell, 1990) and cue-induced reinstatement of nicotine use (Fowler and Kenny, 2011; Liu et al., 2008, 2007) argue against these explanations. Third, smokers might have been able to differentiate between a flavor that is paired with nicotine and a flavor that has been paired with nicotine in the past, unlike perceived sweetness which could become fused with the flavor. Finally, other studies have shown that nicotine conditioning in humans critically depends on explicit contingency knowledge (Hogarth and Duka, 2006). Thus, if participants did not expect to receive nicotine with the conditioned flavors during the fMRI scan or could easily tell if nicotine’s contribution to the flavor was missing, this might explain the absence of positive flavor-conditioning effects. Although future studies are needed to test these suggestions, our results indicate that switching from a flavor previously paired with nicotine to the same flavor without nicotine is less reinforcing than expected and possibly even associated with negative outcome learning.

5.4.

Limitations

While our study has several unique strengths, it also has limitations. We adapted an established flavor-nutrient conditioning paradigm (de Araujo et al., 2013; Veldhuizen et al., 2017) to e-cigarettes which helped us to make several novel observations. Given the rapidly increasing prevalence of ecigarette use, our design enabled us to address a largely unaddressed question of great relevance for public health: Does the addition of sweetener enhance the reinforcing effects of an e-cigarette? Our results provide preliminary support for such an enhancement at the behavioral and neural level. We demonstrate that the conditioned brain-response to cues predicting sweet flavors was greater in the NAcc. Critically, the strength of the conditioned brain-response to cues and the associated odors was associated with the strength of increases in flavor liking lending support to the hypothesis that conditioning effects are mediated via the mesocorticolimbic circuit. However, the major limitation of the study is the small sample-size which limits the study’s power to large effects only (for more detailed calculations and discussions, see SI). Therefore, the observed medium-sized increases in flavor liking and wanting for sweet-tasting flavors alone were not strong enough to provide conclusive evidence on the presence or absence of conditioned changes in reinforcing potential. Moreover, we only tested one a priori ROI as in previous studies using flavor conditioning (de Araujo et al., 2013) although other brain regions are involved in cue reactivity as well (Engelmann et al., 2012; Tang et al., 2012). While the risk exists that this finding could be a false positive, the correspondence between increases in liking after exposure and conditioned brain responses (a conjunction) argue against this possibility. A second limitation is that we failed to see a conditioned increase in e-cigarette liking and wanting after the addition of nicotine only, which was unexpected. To isolate the

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

12

N.B. Kroemer et al.

effects of sweet taste, we selected e-cigarette flavors that were not perceived as sweet. It is conceivable that nicotine only would condition liking and wanting for flavors like candy and fruit that are expected to be sweet (even without sweeteners). However, the absence of positive conditioning effects might also be accounted for by an absence of contingent nicotine self-administration (Hogarth and Duka, 2006) since we observed increases in liking and wanting ratings for nicotine flavors during the exposure sessions when nicotine was present in the e-cigarettes or by the generation of a negative error signal associated with the absence of the expected nicotine reinforcer. Third, although we observed a supra-additive effect of sweet+nicotine flavors on NAcc response, we did not observe supra-additive effects on our behavioral outcomes. Thus, the extent to which the reported supra-additive effect affects reinforcement at a behavioral level remains to be determined. Fourth, the extent to which our broad inclusion criteria for intermittent and light smokers may have contributed to heterogeneity in the elicited effects of the conditioning remains an open question, especially since every participant received the same dose of nicotine. Our results suggest that conditioning with sweet taste plays a bigger role than conditioning with nicotine in this group of smokers, but the results might be qualitatively different in moderate to strongly dependent smokers. This calls for future research in bigger samples including daily smokers to improve the generalizability of findings.

5.5.

[m6+;August 6, 2018;13:13]

Summary and conclusions

To conclude, as predicted, we found that sweet taste and nicotine interact to promote NAcc reactivity to e-cigarette cues. In addition, we observed that sweet-paired e-liquids were wanted and liked more compared to nicotine-paired e-liquids. The increased appeal of the sweet-paired flavors was echoed in an increased BOLD response to the sight of an e-cigarette in the NAcc. In line with the hypothesis that NAcc response tracks the reinforcing value of e-cigarettes, stronger increases in flavor liking after conditioning with sweet taste were associated with greater BOLD response to the sight and the smell of an e-cigarette in the NAcc and the ACC/mPFC. Furthermore, our findings suggest that the addition of sweet taste might play a bigger role in a subset of smokers who dislike flavors that are paired with nicotine only. Collectively, these findings indicate that sweet taste increases the reinforcing potential of an e-cigarette. In general, sweeteners may increase the liability for abuse of alternative tobacco products (Henningfield et al., 2011; Miao et al., 2016). Consequently, it could be beneficial to regulate the addition of sweetener to e-cigarettes to reduce their overall appeal and reinforcing value. Such regulation of sweeteners may help to reduce continued use of alternative tobacco products following experimentation, which may ultimately reduce the incidence rates of nicotine dependence.

Acknowledgment We thank Yuko Nakamura, Caroline Burrasch, Christina Bui, and Mollie Rogers for help in collecting the data.

Financial disclosure Although not related to the current work, Stephanie O’Malley, Ph.D. is a member of the American Society of Clinical Psychopharmacology’s (ASCP’s) Alcohol Clinical Trials Initiative, supported by Alkermes, Amygdala, Arbor Pharma, Ethypharm, Indivior, Lundbeck, Otsuka; Consultant/advisory board member, Alkermes, Amygdala, Indivior, Mitsubishi Tanabe, Opiant; Medication supplies, Astra Zeneca, Novartis, Pfizer.

Author contributions BPP, SSO, and DMS were responsible for the study concept and design. NBK, MGV, RD, BPP, and DMS conducted initial evaluations of the stimuli. NBK, RD, and MGV collected data. NBK performed the data analysis and MGV contributed to analyses. NBK drafted the manuscript and NBK and DMS wrote the manuscript. All authors contributed to the interpretation of findings, provided critical revision of the manuscript for important intellectual content and approved the final version for publication.

Role of the funding source Research reported in this publication was supported by grant number P50DA036151 from the National Institute on Drug Abuse and FDA Center for Tobacco Products (CTP). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the Food and Drug Administration. Salary support for NBK was provided by the Deutsche Forschungsgemeinschaft, grant DFG KR 4555/1-1 and the University of Tübingen fortune grant #2453-0-0, and for BPP by the Canadian Institutes of Health Research Postdoctoral Fellowship, grant MFE-127387.

Conflict of interest All other authors declare no potential conflict of interest.

Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.euroneuro. 2018.07.102.

References Ambrose, B.K., Day, H.R., Rostron, B., Conway, K.P., Borek, N., Hyland, A., Villanti, A.C., 2015. Flavored tobacco product use among US youth aged 12-17 years, 2013-2014. JAMA 314, 1871–1873. Audrain-McGovern, J., Strasser, A.A., Wileyto, E.P., 2016. The impact of flavoring on the rewarding and reinforcing value of e-cigarettes with nicotine among young adult smokers. Drug Alcohol Depend. 166, 263–267.

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

Sweet taste potentiates the reinforcing effects of e-cigarettes Baliki, M.N., Mansour, A., Baria, A.T., Huang, L., Berger, S.E., Fields, H.L., Apkarian, A.V., 2013. Parceling human accumbens into putative core and shell dissociates encoding of values for reward and pain. J. Neurosci. 33, 16383–16393. Bartra, O., McGuire, J.T., Kable, J.W., 2013. The valuation system: a coordinate-based meta-analysis of BOLD fMRI experiments examining neural correlates of subjective value. Neuroimage 76, 412–427. Benowitz, N.L., 1988. Drug therapy. Pharmacologic aspects of cigarette smoking and nicotine addiction. N. Engl. J. Med. 319, 1318–1330. Berridge, K.C., Kringelbach, M.L., 2015. Pleasure systems in the brain. Neuron 86, 646–664. Brody, A.L., Olmstead, R.E., London, E.D., Farahi, J., Meyer, J.H., Grossman, P., Lee, G.S., Huang, J., Hahn, E.L., Mandelkern, M.A., 2004. Smoking-induced ventral striatum dopamine release. Am. J. Psychiatry 161, 1211–1218. Carpenter, M.J., Saladin, M.E., LaRowe, S.D., McClure, E.A., Simonian, S., Upadhyaya, H.P., Gray, K.M., 2014. Craving, cue reactivity, and stimulus control among early-stage young smokers: effects of smoking intensity and gender. Nicotine Tob. Res. 16, 208–215. Castro, D.C., Berridge, K.C., 2014. Opioid hedonic hotspot in nucleus accumbens shell: mu, delta, and kappa maps for enhancement of sweetness “liking” and “wanting”. J. Neurosci. 34, 4239–4250. Changeux, J.P., 2010. Nicotine addiction and nicotinic receptors: lessons from genetically modified mice. Nat. Rev. Neurosci. 11, 389–401. Clark, A., Little, H.J., 2004. Interactions between low concentrations of ethanol and nicotine on firing rate of ventral tegmental dopamine neurones. Drug Alcohol Depend. 75, 199–206. de Araujo, I.E., 2011. Sweet taste signaling and the formation of memories of energy sources. Front. Syst. Neurosci. 5, 99. de Araujo, I.E., 2016. Circuit organization of sugar reinforcement. Physiol. Behav. 164, 473–477. de Araujo, I.E., Lin, T., Veldhuizen, M.G., Small, D.M., 2013. Metabolic regulation of brain response to food cues. Curr. Biol. 23, 878–883. Drummond, M.B., Upson, D., 2014. Electronic cigarettes. Potential harms and benefits. Ann. Am. Thorac. Soc. 11, 236–242. Engelmann, J.M., Versace, F., Robinson, J.D., Minnix, J.A., Lam, C.Y., Cui, Y., Brown, V.L., Cinciripini, P.M., 2012. Neural substrates of smoking cue reactivity: a meta-analysis of fMRI studies. Neuroimage 60, 252–262. Fellows, I., 2012. Deducer: a data analysis GUI for R. J. Stat. Softw. 49, 1–15. Fowler, C.D., Kenny, P.J., 2011. Intravenous nicotine self-administration and cue-induced reinstatement in mice: effects of nicotine dose, rate of drug infusion and prior instrumental training. Neuropharmacology 61, 687–698. Geha, P.Y., Aschenbrenner, K., Felsted, J., O’Malley, S.S., Small, D.M., 2013. Altered hypothalamic response to food in smokers. Am. J. Clin. Nutr. 97, 15–22. Green, B.G., Dalton, P., Cowart, B., Shaffer, G., Rankin, K., Higgins, J., 1996. Evaluating the ’labeled magnitude scale’ for measuring sensations of taste and smell. Chem. Senses 21, 323–334. Grunberg, N.E., Bowen, D.J., Maycock, V.A., Nespor, S.M., 1985. The importance of sweet taste and caloric content in the effects of nicotine on specific food consumption. Psychopharmacology 87, 198–203. Heatherton, T.F., Kozlowski, L.T., Frecker, R.C., Fagerstrom, K.O., 1991. The Fagerstrom test for nicotine dependence: a revision of the Fagerstrom tolerance questionnaire. Br. J. Addict. 86, 1119–1127. Henningfield, J.E., Hatsukami, D.K., Zeller, M., Peters, E., 2011. Conference on abuse liability and appeal of tobacco products:

[m6+;August 6, 2018;13:13]

13 conclusions and recommendations. Drug Alcohol Depend. 116, 1–7. Hogarth, L., Duka, T., 2006. Human nicotine conditioning requires explicit contingency knowledge: is addictive behaviour cognitively mediated. Psychopharmacology 184, 553–566. Howard, J.D., Gottfried, J.A., Tobler, P.N., Kahnt, T., 2015. Identity-specific coding of future rewards in the human orbitofrontal cortex. Proc. Natl. Acad. Sci. USA 112, 5195–5200. Howard, J.D., Plailly, J., Grueschow, M., Haynes, J.D., Gottfried, J.A., 2009. Odor quality coding and categorization in human posterior piriform cortex. Nat. Neurosci. 12, 932–938. Kelley, A.E., Bakshi, V.P., Haber, S.N., Steininger, T.L., Will, M.J., Zhang, M., 2002. Opioid modulation of taste hedonics within the ventral striatum. Physiol. Behav. 76, 365–377. Kenny, P.J., Markou, A., 2006. Nicotine self-administration acutely activates brain reward systems and induces a long-lasting increase in reward sensitivity. Neuropsychopharmacology 31, 1203–1211. Kim, H., Lim, J., Buehler, S.S., Brinkman, M.C., Johnson, N.M., Wilson, L., Cross, K.S., Clark, P.I., 2016. Role of sweet and other flavours in liking and disliking of electronic cigarettes. Tob. Control 25, ii55–ii61. Kroemer, N.B., Guevara, A., Ciocanea Teodorescu, I., Wuttig, F., Kobiella, A., Smolka, M.N., 2014. Balancing reward and work: anticipatory brain activation in NAcc and VTA predict effort differentially. Neuroimage 102 (2), 510–519 Pt. Kroemer, N.B., Guevara, A., Vollstadt-Klein, S., Smolka, M.N., 2013. Nicotine alters food-cue reactivity via networks extending from the hypothalamus. Neuropsychopharmacology 38, 2307–2314. Kroemer, N.B., Small, D.M., 2016. Fuel not fun: reinterpreting attenuated brain responses to reward in obesity. Physiol. Behav. 162, 37–45. Kroemer, N.B., Sun, X., Veldhuizen, M.G., Babbs, A.E., De Araujo, I.E., Small, D.M., 2016. Weighing the evidence: variance in brain responses to milkshake receipt is predictive of eating behavior. Neuroimage 128, 273–283. Kroemer, N.B., Wuttig, F., Bidlingmaier, M., Zimmermann, U.S., Smolka, M.N., 2015. Nicotine enhances modulation of food-cue reactivity by leptin and ghrelin in the ventromedial prefrontal cortex. Addict. Biol. 20, 832–844. Lawrence, N.S., Hinton, E.C., Parkinson, J.A., Lawrence, A.D., 2012. Nucleus accumbens response to food cues predicts subsequent snack consumption in women and increased body mass index in those with reduced self-control. Neuroimage 63, 415–422. Lex, A., Hauber, W., 2008. Dopamine D1 and D2 receptors in the nucleus accumbens core and shell mediate Pavlovian-instrumental transfer. Learn. Mem. 15, 483–491. Lim, J., Wood, A., Green, B.G., 2009. Derivation and evaluation of a labeled hedonic scale. Chem. Senses 34, 739–751. Liu, X., Caggiula, A.R., Palmatier, M.I., Donny, E.C., Sved, A.F., 2008. Cue-induced reinstatement of nicotine-seeking behavior in rats: effect of bupropion, persistence over repeated tests, and its dependence on training dose. Psychopharmacology 196, 365–375. Liu, X., Caggiula, A.R., Yee, S.K., Nobuta, H., Sved, A.F., Pechnick, R.N., Poland, R.E., 2007. Mecamylamine attenuates cue-induced reinstatement of nicotine-seeking behavior in rats. Neuropsychopharmacology 32, 710–718. Maskos, U., Molles, B.E., Pons, S., Besson, M., Guiard, B.P., Guilloux, J.P., Evrard, A., Cazala, P., Cormier, A., Mameli-Engvall, M., Dufour, N., Cloez-Tayarani, I., Bemelmans, A.P., Mallet, J., Gardier, A.M., David, V., Faure, P., Granon, S., Changeux, J.P., 2005. Nicotine reinforcement and cognition restored by targeted expression of nicotinic receptors. Nature 436, 103–107. Miao, S., Beach, E.S., Sommer, T.J., Zimmerman, J.B., Jordt, S.E.,

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102

JID: NEUPSY

ARTICLE IN PRESS

14 2016. High-intensity sweeteners in alternative tobacco products. Nicotine Tob. Res. 18, 2169–2173. Mineur, Y.S., Abizaid, A., Rao, Y., Salas, R., DiLeone, R.J., Gundisch, D., Diano, S., De Biasi, M., Horvath, T.L., Gao, X.B., Picciotto, M.R., 2011. Nicotine decreases food intake through activation of POMC neurons. Science 332, 1330–1332. Morean, M.E., Butler, E., Bold, K., Kong, G., Camenga, D., Cavallo, D., Simon, P., O’Malley, S., Krishnan-Sarin, S., 2017. E– cigarette flavor preferences are associated with vaping frequency among adolescents but not among adults. PLoS One 13, e0189015. O’Doherty, J., Dayan, P., Schultz, J., Deichmann, R., Friston, K., Dolan, R.J., 2004. Dissociable roles of ventral and dorsal striatum in instrumental conditioning. Science 304, 452–454. Ostafin, B.D., Marlatt, G.A., Troop-Gordon, W., 2010. Testing the incentive-sensitization theory with at-risk drinkers: wanting, liking, and alcohol consumption. Psychol. Addict. Behav. 24, 157–162. Patrick, M.E., Miech, R.A., Carlier, C., O’Malley, P.M., Johnston, L.D., Schulenberg, J.E., 2016. Self-reported reasons for vaping among 8th, 10th, and 12th graders in the US: nationally-representative results. Drug Alcohol Depend. 165, 275–278. Pessiglione, M., Seymour, B., Flandin, G., Dolan, R.J., Frith, C.D., 2006. Dopamine-dependent prediction errors underpin reward-seeking behaviour in humans. Nature 442, 1042–1045. Pontieri, F.E., Tanda, G., Orzi, F., Di Chiara, G., 1996. Effects of nicotine on the nucleus accumbens and similarity to those of addictive drugs. Nature 382, 255–257. R Core Team, 2015. R: A Language and Environment for Statistical Computing, 3.2.2 ed. R Foundation for Statistical Computing, Vienna, Austria. Raudenbush, S.W., Bryk, A.S., Cheong, Y.F., Congdon, R.T., Du Toit, M., 2011. HLM 7. Scientific Software International Inc, Lincolnwood, IL. Russell, M.A., 1990. The nicotine addiction trap: a 40-year sentence for four cigarettes. Br. J. Addict. 85, 293–300. Schane, R.E., Ling, P.M., Glantz, S.A., 2010. Health effects of light and intermittent smoking: a review. Circulation 121, 1518–1522. Schonberg, T., Daw, N.D., Joel, D., O’Doherty, J.P., 2007. Reinforcement learning signals in the human striatum distinguish learners from nonlearners during reward-based decision making. J. Neurosci. 27, 12860–12867. Sclafani, A., 1987. Carbohydrate taste, appetite, and obesity: an overview. Neurosci. Biobehav. Rev. 11, 131–153. Sheffield, F.D., Roby, T.B., 1950. Reward value of a non-nutritive sweet-taste. J. Comp. Physiol. Psychol. 43, 471–481. Small, D.M., Veldhuizen, M.G., Felsted, J., Mak, Y.E., McGlone, F., 2008. Separable substrates for anticipatory and consummatory food chemosensation. Neuron 57, 786–797. Steinberg, E.E., Keiflin, R., Boivin, J.R., Witten, I.B., Deisseroth, K., Janak, P.H., 2013. A causal link between prediction errors, dopamine neurons and learning. Nat. Neurosci. 16, 966–973. Sun, X., Kroemer, N.B., Veldhuizen, M.G., Babbs, A.E., de Araujo, I.E., Gitelman, D.R., Sherwin, R.S., Sinha, R., Small, D.M., 2015. Basolateral amygdala response to food cues in the absence of hunger is associated with weight gain susceptibility. J. Neurosci. 35, 7964–7976.

[m6+;August 6, 2018;13:13]

N.B. Kroemer et al. Sweitzer, M.M., Geier, C.F., Joel, D.L., McGurrin, P., Denlinger, R.L., Forbes, E.E., Donny, E.C., 2014. Dissociated effects of anticipating smoking versus monetary reward in the caudate as a function of smoking abstinence. Biol. Psychiatry 76, 681–688. Tang, D.W., Fellows, L.K., Small, D.M., Dagher, A., 2012. Food and drug cues activate similar brain regions: a meta-analysis of functional MRI studies. Physiol. Behav. 106, 317–324. Tellez, L.A., Han, W., Zhang, X., Ferreira, T.L., Perez, I.O., Shammah-Lagnado, S.J., van den Pol, A.N., de Araujo, I.E., 2016. Separate circuitries encode the hedonic and nutritional values of sugar. Nat. Neurosci. 19, 465–470. Tizabi, Y., Bai, L., Copeland Jr., R.L., Taylor, R.E., 2007. Combined effects of systemic alcohol and nicotine on dopamine release in the nucleus accumbens shell. Alcohol Alcohol. 42, 413–416. Tizabi, Y., Copeland Jr., R.L., Louis, V.A., Taylor, R.E., 2002. Effects of combined systemic alcohol and central nicotine administration into ventral tegmental area on dopamine release in the nucleus accumbens. Alcohol. Clin. Exp. Res. 26, 394–399. Tolu, S., Eddine, R., Marti, F., David, V., Graupner, M., Pons, S., Baudonnat, M., Husson, M., Besson, M., Reperant, C., Zemdegs, J., Pages, C., Hay, Y.A., Lambolez, B., Caboche, J., Gutkin, B., Gardier, A.M., Changeux, J.P., Faure, P., Maskos, U., 2013. Co-activation of VTA DA and GABA neurons mediates nicotine reinforcement. Mol. Psychiatry 18, 382–393. Valentin, V.V., O’Doherty, J.P., 2009. Overlapping prediction errors in dorsal striatum during instrumental learning with juice and money reward in the human brain. J. Neurophysiol. 102, 3384–3391. Veldhuizen, M.G., Babbs, R.K., Patel, B., Fobbs, W., Kroemer, N.B., Garcia, E., Yeomans, M.R., Small, D.M., 2017. Integration of sweet taste and metabolism determines carbohydrate reward. Curr. Biol. 27, 2476–2485. Veldhuizen, M.G., Douglas, D., Aschenbrenner, K., Gitelman, D.R., Small, D.M., 2011. The anterior insular cortex represents breaches of taste identity expectation. J. Neurosci. 31, 14735–14744. Veldhuizen, M.G., Nachtigal, D., Teulings, L., Gitelman, D.R., Small, D.M., 2010. The insular taste cortex contributes to odor quality coding. Front. Hum. Neurosci. 4, 1–11. Versace, F., Engelmann, J.M., Robinson, J.D., Jackson, E.F., Green, C.E., Lam, C.Y., Minnix, J.A., Karam-Hage, M.A., Brown, V.L., Wetter, D.W., Cinciripini, P.M., 2014. Prequit fMRI responses to pleasant cues and cigarette-related cues predict smoking cessation outcome. Nicotine Tob. Res. 16, 697–708. Vollstädt-Klein, S., Kobiella, A., Bühler, M., Graf, C., Fehr, C., Mann, K., Smolka, M.N., 2011. Severity of dependence modulates smokers’ neuronal cue reactivity and cigarette craving elicited by tobacco advertisement. Addict. Biol. 16, 166–175. Watson, N.L., Carpenter, M.J., Saladin, M.E., Gray, K.M., Upadhyaya, H.P., 2010. Evidence for greater cue reactivity among low-dependent vs. high-dependent smokers. Addict. Behav. 35, 673–677. Zhang, L., Dong, Y., Doyon, W.M., Dani, J.A., 2012. Withdrawal from chronic nicotine exposure alters dopamine signaling dynamics in the nucleus accumbens. Biol. Psychiatry 71, 184–191.

Please cite this article as: N.B. Kroemer et al., Sweet taste potentiates the reinforcing effects of e-cigarettes, European Neuropsychopharmacology (2018), https://doi.org/10.1016/j.euroneuro.2018.07.102