Lipopolysaccharide does not alter behavioral response to successive negative contrast in mice

Abstract

Reduced motivation is one of the main symptomatic features of inflammation-induced depression. However, the exact nature of inflammation-induced alterations in motivation remains to be fully defined. As inflammation has been shown to increase sensitivity to negative stimuli, the present series of experiments was initiated to determine whether systemic inflammation induced by infra-septic doses of lipopolysaccharide (LPS) in mice influences consummatory and instrumental responding to successive negative contrast. Successive negative contrast is operationally defined by a shift to a lower value reward than mice were trained with. Mice trained to drink a high sucrose concentration solution responded to an acute shift to a lower concentration of sucrose by a reduction in the volume of sucrose consumed and a decrease in lick numbers and bout durations. LPS administered 24 h before the shift did not modify these effects. In the same manner, mice trained to nose poke for chocolate pellets according to a fixed reinforcement schedule 10 (10 nose pokes for the food reinforcement) and shifted to a lower reward value (decreased number of chocolate pellets or replacement of chocolate pellets by less preferred grain pellets) altered their instrumental responding in a manner that did not vary consistently under the effect of LPS administered 24 h before. These results indicate a high variability in the effects of LPS on successive negative contrast and fails to provide evidence in favor of the hypothesis that LPS increases sensitivity to decreases in expected rewards.

1. Introduction

Within psychiatry it is now well accepted that inflammation can contribute to the pathophysiology and treatment resistance of at least some subtypes of depression ¹. However, this recognition has yet to lead to significant improvements in the management of depression. One reason for this is that, despite its diversity of symptoms, depression still tends to be studied as a categorical rather than a multidimensional disorder. A few dimensional investigations of inflammation-associated depression show a predominance of somatic (e.g., fatigue, reduced motivation, sleep disorders, reduced appetite) over psychological symptoms (e.g., depressed mood, suicidal ideation) ^2–5. This clinical presentation is consistent with the observation that inflammation induces mainly motivational and motor deficits, which potentially correspond to anhedonia and psychomotor retardation. It is yet unknown if there are more symptoms that might distinguish inflammation-associated depression from other forms of depression. Based on the results of experimental studies in rodents and in volunteers, there is evidence that the motivational changes that are observed in inflamed individuals are more nuanced than what is usually described as an inflammation-induced reduction in motivation ^6,7. In particular, we have shown that inflammation shifts effort toward the most valuable reward despite its higher cost in inflamed mice or volunteers subjected to effort valuation tasks in which different cost-benefit reward ratios are concurrently available ^8,9. This shift of preference toward the most valuable reward independently of the effort it requires for its procurement appears to be specific to inflammation as depressed patients in a general psychiatric population only show reduced motivation in the effort valuation task ¹⁰. The present study was carried out to test this interpretation.

Based on the aforementioned findings and the accumulating literature on the enhancing effects of inflammation on sensitivity to negative emotional stimuli, e.g., monetary loss versus monetary gain or negative versus positive social feedback ^6,11, we predicted that in addition to reducing motivated behavior directed toward food rewards, inflammation should increase the behavioral consequences of a successive negative contrast defined as a temporary reduction in responding to a smaller reward by mice previously exposed to a larger reward ¹². Our prediction was that inflammation should enhance the behavioral response to negative contrast. To make sure our expected findings would be independent of the nature of the behavioral response, these predictions were tested on responding in an instrumental negative contrast in which mice are trained to nose poke for a food reward and in a consummatory negative contrast in which mice are trained to lick a drinking tube for a sucrose solution. In both cases, inflammation was induced by a systemic injection of lipopolysaccharide (LPS) and mice were tested 24 hours after, when the major signs of sickness induced by this treatment had dissipated and LPS-induced depressive behavior is observed ¹³.

2. Materials and methods

2.1. Animals and treatments

Adult male C57BL/6J purchased from Jackson Laboratory (Ellsworth, Maine) were group-housed (n=5/group) in a temperature- and humidity-controlled environment on a 12-h light:dark cycle (lights on at 07:00 am) with food and water available ad libitum. Mice were maintained on PicoLab® Rodent Diet 20 (20% protein, 10.6% fat (ether extract + acid hydrolysis), and 4.7% fiber, with a physiological fuel value of 3.41 kcal/gm). Each cage was provided with wood chip bedding and nesting material (Nestlets™). Once fully habituated to handling mice were single housed, food restricted to 80–85% of their body weight, and trained in the reward tasks. Once they achieved a stable behavioral performance, mice received intraperitoneal (i.p.) injections of LPS (L-3129, serotype 0127:B8, Sigma-Aldrich) or sterile saline. Based on a power analysis, we determined that 5–7 mice/group would be sufficient to detect an effect with 0.80 power with a 5% type 1 error rate. The effects of LPS on behavior were tested 24 h after injection. All procedures performed on the mice including level of food restriction were approved by the University of Texas MD Anderson Cancer Center Institutional Animal Care and Use Committee.

2.2. Successive negative contrast procedures

2.2.1. Consummatory successive negative contrast

To assess the effects of LPS on consummatory behavior in response to different reward values, mice were trained to drink a sucrose solution with sucrose diluted in tap water. The first experiment was carried out with mice that were food restricted to 80–85% of their body weight and trained to drink either a 12% or a 2% sucrose solution during 7 daily sessions (n = 6/group). Drinking behavior was measured by the volume of sucrose solution ingested. Mice were injected with 0.5 mg/kg LPS or saline 1 h after their last training session and tested on the next day. On the test day mice trained on a 12% sucrose solution were shifted to a 2% sucrose solution while the testing condition did not change for mice trained on the 2% sucrose solution.

For next two experiments mice were placed in an operant conditioning chamber (Med Associates, St. Albans, VT) equipped with a single standard water bottle connected to a contact lickometer that recorded the number and timing of licks (Med Associates, St. Albans, VT). The second experiment was carried out with a different group of mice restricted to 80–85% of their body weight. Mice (n=7/group) were trained daily to drink a 4% or 12% sucrose solution over 10 days, until drinking behavior, as assessed by total number of licks, stabilized. Mice were tested 24 h after they were injected with LPS 0.5 mg/kg or saline. On the test day, the sucrose concentration in mice exposed to 12% was shifted to 4% (12–4 condition) while mice trained to drink 4% sucrose did not undergo a switch (4–4 condition). Mice were then crossed over (12% to 4%; 4% to 12%) and retrained for 7 days. After stabilization of their drinking behavior, mice were retested under the alternative treatment, such that each mouse only received LPS once.

A third experiment was carried out in the mice from experiment 2, under free feeding conditions (n = 7/group). Mice were again trained daily for 10 days to drink either 4% or 12% sucrose solution. Mice were then re-injected with LPS (1 mg/kg) or saline 2 h after their last training session. An increased dose of LPS (1 mg/kg) was used as mice display blunted response to subsequent LPS treatments. During testing, all mice were shifted to a 1% sucrose solution (12–1 or 4–1 condition).

2.2.2. Instrumental successive negative contrast

To assess the effects of LPS on instrumental behavior maintained by food reward of different values, mice restricted to 80–85% of their body weight were trained to nose poke for chocolate-flavored or grain dustless Precision Pellets (20 mg, BioServ, Frenchtown, NJ) in operant conditioning chambers equipped with a single nose-poke response unit and food dispenser (Med Associates, St. Albans, VT) (n = 6–8/group). Behavior was measured by the total number of nose pokes per session. Baseline performance was calculated as the average number of nose pokes per session during the 3 sessions immediately before treatment.

In a first experiment (shift in reward size), mice were trained on a fixed ratio (FR)1 schedule (1 nose poke for 1 or 4 grain pellets) for 3 days, FR5 schedule (5 nose pokes for 1 or 4 grain pellets) for 2 days, and then FR10 schedule (10 nose pokes for 1 or 4 grain pellets) for 10 days. On days 15 and 22, mice were injected with LPS (0.5 mg/kg) or saline (in a counterbalanced fashion with 1-week interval) 2 h after training. During testing that took place 24 h later, all mice were provided 1 pellet for each bout of 10 responses, such that mice trained to work for 4 chocolate pellets were submitted to a devaluation test.

A second experiment (shift in reward type) was carried out 2 weeks later only in mice that had been trained to receive 1 chocolate pellet. Mice were injected with LPS (0.33 mg/kg) or saline (in a counterbalanced fashion with 1-week interval) 2 h after training. On the day of testing, mice in the non-shift condition continued to receive a chocolate pellet reward (C-C condition) while mice in the shift condition received the less preferred grain pellet (C-G condition).

2.3. Statistical analysis

Data were calculated as volume consumed in milliliters or total number of licks per session for consummatory behavior and total number of nose pokes per session for instrumental responding. To analyze the microstructure of drinking we also calculated the size of licking bursts, with each burst defined as a rapid succession of licks separated from the previous burst by a minimum of 1 second ¹⁴. Data were analyzed using IBM SPSS® Statistics (version 26) and are presented as mean ± standard error of the mean (SEM). Baseline differences were analyzed using independent samples t-tests. Two-way ANOVAs with time as a repeated measure were used to analyse the effects of LPS on instrumental and consummatory negative contrast measured before and 24 h after treatment. Individual raw data are presented in figures 1 to 4 and results of statistical analysis are summarized in supplementary tables S1 and S2.

3. Results

3.1. Effects of inflammation on consummatory successive negative contrast

In the first experiment, mice previously trained to drink a 12% or a 2% sucrose solution were all tested with a 2% sucrose solution 24 h after LPS. Before LPS, mice trained on 12% sucrose drank approximately twice as much sucrose as mice trained on 2% sucrose ((0.9 +/− 0.04 vs. 0.44 +/− 0.02 ml, t(22)=11.9, p<0.001; Fig. 1). A 2 (+/− LPS) by 2 (+/− shift) repeated measures ANOVA from baseline to testing revealed significant main effects of time (F(1,20)=186, p<0.001) and shift (F(1,20)=149, p<0.001) as well as LPS x time (F(1,20)=38.6, p<0.001) and shift x time (F(1,20)=95.5, p<0.001) interactions. These interactions indicate both LPS and the shift in reward value resulted in a reduction in sucrose intake. While the shifted group treated with LPS showed the most marked drop in sucrose intake the 3-way interaction (LPS x shift x time) was not significant.

Fig. 1 –

Baseline performance and response to successive negative contrast (shift) measured by volume consumed in mice trained to drink a 12% or 2% sucrose solution and treated or not with LPS 24 h before being tested with a 2% sucrose solution. Baseline corresponds to the average performance during the last 3 days before LPS treatment. Individual values, n=6 mice/group

In the second experiment food-restricted mice were trained to drink either a 12% or a 4% sucrose solution and were all tested with a 4% sucrose solution. Similar to the previous experiment, at baseline mice trained on 12% sucrose had approximately twice as many lick responses than those trained on 4% sucrose (772 +/− 46.6 vs. 365 +/− 32.0 licks, t(26)=7.2, p<0.001; Fig. 2A). A 2 (+/− LPS) x 2 (+/− shift) repeated measures ANOVA from baseline to testing revealed a significant main effect of time (F(1,24)=10.2, p<0.01) and shift (F(1,24)=28.9, p<0.001) and a time x shift interaction (F(1,24)=14.6, p<0.001) (Fig. 2A). These data indicate that shift in the reward value decreased lick numbers and this effect was not modified by LPS which did not affect the number of licks. Similarly, licking bursts were longer in mice trained to drink 12% sucrose compared to mice trained to drink 4% (365 +/− 24.0 vs. 41.4 +/− 9.86, t(26)=5.10, p<0.001; Fig. 2B). The apparent decrease in burst durations in the shifted group did not reach significance (F(1,24)=4.08, p=0.055, Fig. 2B). There was no effect of LPS on burst duration.

Fig. 2 -.

Baseline performance and response to successive negative contrast (shift) in mice trained to drink a 12% or 4% sucrose solution and treated or not with LPS 24 h before being tested with a 4% sucrose solution. (A) Number of licks and (B) lick bouts duration in seconds. Baseline corresponds to the average performance during the last 3 days before LPS treatment. Individual values, n=7 mice/group

These first two experiments indicate that shifting the reward value to a lower value results in a decrease in volume consumed and lick numbers but not in lick burst durations. However. LPS treatment does not significantly modify these effects.

In the third experiment mice were trained to drink either a 12% or a 4% sucrose solution before being all tested with a 1% sucrose solution (Fig. 3). There was no significant difference in the total number of licks in mice trained on 12% compared to 4% sucrose (221 +/− 34.7 vs. 145 +/− 23.4 licks, t(26)=1.83, p<0.1; Fig. 3A). A 2 (+/− LPS) x 2 (4–1 vs. 12–1 shift) repeated measures ANOVA on number of licks did not reveal any main effect of shift (F(1,24)=3.46, p=0.075). However, lick burst durations were more sensitive to detecting changes. As previously observed mean licking burst durations were longer in mice trained to drink 12% compared to 4% sucrose (52.3 +/− 2.60 vs. 43.7 +/− 2.97, t(26)=2.20, p<0.05; Fig. 3B). There was a day by shift (F(1,24)=7.47, p<0.05) and a day by LPS (F(1,24)=5.41, p<0.05) interaction. The day by shift by LPS interaction did not reach significance (F(1,24)=4.06, p=0.057). These results indicate that the decrease in burst duration in the 12% trained saline group was blunted in the 12% trained LPS treated group. LPS had no effect on the response to the shift in reward value whether measured by total number of licks or licking burst duration.

Fig. 3 –

Baseline performance and response to successive negative contrast (shift) in mice trained to drink a 12% or 4% sucrose solution and treated or not with LPS 24 h before being tested with a 1% sucrose solution. (A) Number of licks and (B) lick bouts duration in seconds. Baseline corresponds to the average performance during the last 3 days before LPS treatment. Individual values, n=7 mice/group

3.2. Effects of inflammation on instrumental successive negative contrast

We then evaluated negative contrast using an instrumental task in which food-restricted mice were trained to nose pose for different value rewards. Mice were trained on an FR10 schedule of food reinforcement for either 1 or 4 grain pellets. Mice trained with 1 grain pellet poked approximately four times more than mice trained to work for 4 grain pellets (203 +/− 11.9 vs 60.8 +/− 4.23 nose pokes, t(24)=10.5, p<0.001; Fig. 4A). A 2 (+/− LPS) x 2 (+/− shift) repeated measures ANOVA from baseline to testing revealed a significant main effect of time (F(1, 22)= 7.92, p<0.05), LPS (F(1,22)= 14.3, p<0.01) and shift (F(1,22)= 18.1, p<0.001), as well as a time by shift interaction (F(1,22)=30.1, p<0.001) and time by LPS (F(1,22)=52.4, p<0.001). These results indicate that the shift from 4 to 1 grain pellet increased the total number of nose pokes emitted by saline-treated mice to a value close to that of the number of nose pokes emitted by mice trained to nose poke for only 1 grain pellet. Further, they indicate that the shift enhancing effect of nose poking was abrogated by LPS.

Fig. 4 –

Baseline performance and response to successive negative contrast (shift) in mice trained to work for a food reinforcement according to a fixed ratio 10 and treated or not with LPS 24 h before negative contrast. (A) Total number of nose pokes emitted by mice trained to obtain 4 or 1 grain pellets as a reinforcement in response to 10 nose pokes and shifted or not to 1 grain pellet. (B) Total number of nose pokes emitted by mice trained to obtain 1 chocolate pellet (C) or a less preferred grain pellet (G) and submitted or not to a shift to 1 grain pellet. Baseline corresponds to the average performance during the last 3 days before LPS treatment. Individual values, n=6–8 mice/group.

We also tested negative contrast using an instrumental task in which reward value was shifted by reducing quality rather than quantity of reward. Mice were trained on a FR10 schedule of food reinforcement to receive either a chocolate (C) or a less preferred grain (G) pellet and all mice were tested with a grain pellet reward. A 2 (+/− LPS) x 2 (+/− shift) repeated measures ANOVA from baseline to testing on the total number of nose pokes revealed a significant main effect of time (F(1,22)=15.8, p<0.001) and LPS (F(1,22)=4.55, p<0.05) as well as a time x shift (F(1,22)=13.6, p<0.01) and time x LPS (F(1,22)=24.6, p<0.001) interactions. These results indicate that LPS did not affect the negative effect of the shift in reward type on nose-poking.

Taken together, the results of these two experiments on negative instrumental contrast indicate that while LPS negatively impacts instrumental responding for food reward at 24 h post-treatment it does not significantly affect the consequences of reward devaluation under these conditions.

4. Discussion

The present experiments were carried out to better understand how inflammation affects motivated behavior. Specifically, we sought to determine whether LPS modulates the response to a successive negative contrast in mice trained to obtain a reward either by an instrumental response or by a consummatory response. The results we obtained were mixed, suggesting that the effects of inflammation depend upon the response modality (instrumental vs. consummatory) and the nature of the shift.

Following the pioneering experiment by Crespi on the behavioral effects of decreasing the size of the expected reward in animals trained to run an alley to get their food ¹⁵, negative contrast has been studied by psychologists in the 1970–90s ¹². Most of the experiments using this approach have been done in rats, although the generality of this phenomenon has been assessed in various species. A second wave of research on negative contrast has taken place in the last two decades. It differs from the first wave in terminology (reward or outcome devaluation rather than negative contrast) and emphasis (switch to the contrast between habit behavior and goal-directed behavior) ¹⁶. The general take-away is that insensitivity to changes in reward value after training under a fixed reward value is characteristic of habitual behavior while behavioral flexibility in response to changes in the response outcome value is a signature of goal-directed behavior. The problem with reward devaluation procedures is that they usually rely on satiety-induced devaluation or on conditioned taste aversion to the food reward presented. This is not ideal for testing the effects of LPS as its administration is sufficient to reduce food intake and to induce conditioned taste aversion ¹⁷. This explains why we opted to use traditional negative contrast approaches for the purpose of the present study.

One of the conditions for successive negative contrast to take place is that there is a reduction in responding when subjects are shifted to a lower value reward than the one they had been trained to expect ¹². This pattern of results was not observed in the present experiments. Mice shifted to lesser rewards decreased their instrumental or consummatory responding compared to their previous responding, but their new responding did not differ from that of mice already trained with this smaller reward. This appears to be in contrast to results from previous studies in which mice that were shifted from a 32% to a 4% sucrose solution displayed the predicted negative contrast ^18,19. However, a close examination of these results shows that the expected depression effect is very short-living or appears only in certain conditions. For instance, female mice maintained at 85% of their free feeding body weight and shifted from a 32 to a 4% sucrose solution did not lick less nor consumed a different volume of sucrose than control mice trained with the 4% sucrose solution and unshifted ¹⁹. The only difference was in the lick cluster size during the first 2 minutes of the test. To the best of our knowledge, there are no similar studies on instrumental successive negative contrast in mice. In addition, it is worth noting that follow-up studies on the Crespi effect have revealed that the ‘depression’ effect that is supposed to occur when animals are shifted to a smaller reward has been more difficult to replicate than the ‘elation’ effect which develops when animals are shifted to a larger reward ²⁰.

Despite these issues, we used changes in responding to reward shifts as a baseline to assess the effects of LPS on the consequences of what we are still calling for convenience a negative contrast, operationally defined by a shift to a lower value reward than the one mice were trained with. In order to do so, we chose to administer LPS 24 h before the shift. This time interval corresponds to the time necessary for dissipation of LPS-induced anorexia and fatigue. It is commonly used to assess the effects of LPS on depression-like behavior. The dose of LPS we used varied from 0.33 to 1 mg/kg. This variation was not random. In our experience, food-restricted mice are more resistant than non-deprived mice to the behavioral effects of LPS. When mice are repeatedly administered LPS they develop tolerance, therefore the need to increase the dose from the first to the second injection. Our results did not show a consistent effect of LPS on responding to shifts in reward value. In mice trained to work for 4 chocolate pellets the shift to 1 pellet resulted in an increase in responding in saline-treated mice. This was not an elation or overshoot effect in the Crespi terminology but likely related to a shorter ingestion time, allowing for more time to nose poke than when 4 pellets are delivered in succession. The fact that LPS-treated mice did not respond in the same way could be interpreted as inflammation-induced insensitivity to changes in reward value. However, the interpretation that LPS-treated mice persist in habit responding cannot explain why LPS did not interfere with the decrease in instrumental responding observed in mice shifted from 1 chocolate pellet to 1 grain pellet. Independently of these two observations, the only experimental result in favor of a possible enhancing effect of LPS on negative contrast was the decreased sucrose consumption of LPS-treated mice when shifted from 12 to 2% sucrose. However, this effect did not materialize in other conditions of consummatory negative contrast. In other words, the variability in the effects of LPS on responding to shift in reward values in different conditions of shift and modality of response clearly shows that a unitary concept such as increased sensitivity to change in outcome values is inadequate to account for the behavioral effects of LPS.

In the present experiments we did not assess the effect of LPS on responding to different reward values independently of the shift as this was outside of the scope of the current investigation. Experiments on consummatory behavior using different concentrations of sucrose already indicate that LPS is more effective to decrease sucrose intake when the concentration of sucrose is low than when it is higher than 5% ²¹. In the same manner, low concentrations of sucrose discriminate better mice with an autoimmune disorder compared to control mice than concentrations higher than 4% ²². This explains why current protocols to study anhedonia in mice induced by various stressors are most effective when they use low concentrations (i.e., 1–2%) sucrose solutions ^23,24. No similar study has been carried out on the sensitivity of instrumental responding to LPS with different reward values. In a different paradigm, we had already reported that mice trained in an effort valuation procedure in which they could choose between a low effort-low reward mode of response (1 grain pellet for 1 nose poke) and a high effort-high reward mode of response (1 chocolate pellet for 10 nose pokes) responded to LPS by a decrease in incentive motivation measured by the total number of nose pokes ⁸. However, their preference for working for the chocolate pellet increased despite the higher effort it required. Taken together with the present findings, these data indicate that LPS could be more effective to decrease behavior maintained by low than by high incentive value. However, this interpretation appears to be contradicted by the negative results obtained with mice trained to drink different concentrations of sucrose, especially when their performance is measured by the number of licks rather than by the volume of sucrose solution that is ingested. This apparent contradiction can be explained by the fact that licks are a better measure of palatability than reward value of a solution of sucrose and that LPS does not affect palatability ^25,26. The consideration of durations of licking burst sizes confirms this observation.

There are a few limitations to our current study. Our general approach for this series of studies was to run multiple complementary studies to determine whether an effect was worthy of further investigation. While we think that obtaining consistent results using a combination of different consummatory or operant conditioning procedures increases our confidence in our data, we acknowledge that the sample sizes for each experiment is relatively low. A further limitation is related to our LPS dose escalation procedure. As mentioned previously, tolerance develops in response to LPS limiting the number of studies for which a mouse can be studied. To overcome this, we increased the second dose of LPS to a level that produces a comparable behavioral and inflammatory response. However, we acknowledge that the first and second inflammatory responses might be quantitatively different.

In conclusion, the results obtained in the present series of experiments carried out in mice trained to consume rewarding sucrose solutions or to nose poke for food reinforcement in operant conditioning cages confirm that systemic inflammation induced by LPS tends to result in decreased motivation to engage in the corresponding rewarded behavior. However, inflamed mice show a high variability on sensitivity to negative changes in outcome values, which is not consistent with an interpretation in terms of higher sensitivity to negative contrast.