Lithium-induced malaise does not interfere with adaptation of the hypothalamic-pituitary-adrenal axis to stress

    Lithium-induced malaise does not interfere with adaptation of the hypothalamic-pituitary-adrenal axis to stress Maria Sanch´ıs-Oll´e, Juan Antonio Ortega-S´anchez, Xavier Belda, Humberto Gagliano, Roser Nadal, Antonio Armario PII: DOI: Reference:

S0278-5846(17)30031-3 doi:10.1016/j.pnpbp.2017.01.006 PNP 8998

To appear in:

Progress in Neuropsychopharmacology & Biological Psychiatry

Received date: Revised date: Accepted date:

1 August 2016 7 December 2016 12 January 2017

Please cite this article as: Sanch´ıs-Oll´e Maria, Ortega-S´ anchez Juan Antonio, Belda Xavier, Gagliano Humberto, Nadal Roser, Armario Antonio, Lithium-induced malaise does not interfere with adaptation of the hypothalamic-pituitary-adrenal axis to stress, Progress in Neuropsychopharmacology & Biological Psychiatry (2017), doi:10.1016/j.pnpbp.2017.01.006

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Lithium-induced malaise does not interfere with adaptation of the hypothalamic-pituitaryadrenal axis to stress

IP

T

Maria Sanchís-Ollé 1, 2, 3* Juan Antonio Ortega-Sánchez 1, 2, 3*, Xavier Belda 1, 2, 3, +

SC R

Humberto Gagliano 1, 2, 3, Roser Nadal 1, 2, 4, Antonio Armario 1, 2, 3

1

Institut de Neurociències, 2Red de trastornos adictivos and CIBERSAM, 3Animal Physiology

NU

Unit, School of Biosciences, 4Psychobiology Unit, School of Psychology, Universitat Autònoma

D

* Co-authors, equal contribution.

MA

de Barcelona, Cerdanyola del Vallès 08193, Barcelona, Spain.

TE

+ Correspondence to: Dr. Antonio Armario Animal Physiology Unit, School of Biosciences, Universitat Autònoma de Barcelona,

CE P

Cerdanyola del Vallès 08193, Barcelona, Spain. E-mail: [email protected]

AC

Phone: 34-935811840 FAX: 34-935812390

1

ACCEPTED MANUSCRIPT Abstract We have recently demonstrated that adaptation of the hypothalamic-pituitary-adrenal (HPA) axis to repeated exposure to a stressor does not follow the rules of habituation and can be

T

fully expressed after a single experience with severe stressors. In the present work we tested

IP

the hypothesis that adaptation could be impaired if animals experience malaise during initial

SC R

exposure to the stressor. To this end, animals were allowed to drink saccharin for 30 minutes before being exposed for 3 hours to immobilization on boards (IMO), a severe stressor; then they were given either saline or lithium ip after the first hour of IMO. Stress-naïve rats followed exactly the same procedure except IMO. Exposure to IMO caused a strong activation

NU

of the HPA axis whereas the effect of lithium was modest. Both IMO and lithium administration resulted in conditioned taste aversion to saccharin when evaluated 4 days later.

MA

When all animals were exposed to IMO 6 days later, reduced HPA response and less impact on body weight was observed in the two groups previously exposed to IMO as compared with stress-naïve rats. Therefore, lithium administration during the first IMO exposure did not affect

D

adaptation of the HPA axis and weight gain. These results indicate that malaise per se only

TE

weakly activated the HPA axis and argue against the hypothesis that signs of physical malaise

CE P

during exposure to the stressor could impair HPA adaptation.

Key words: hypothalamic-pituitary-adrenal axis, adaptation, conditioned taste aversion,

AC

lithium chloride, immobilization stress, stress-induced malaise.

2

ACCEPTED MANUSCRIPT 1. Introduction

T

Acute exposure to stressors elicit important behavioral changes and alters many physiological

IP

parameters, with the aims of better coping with the present situation and preparing the organisms for future encounters with similar or different stressors. The most extensively

SC R

studied physiological system in stress is the hypothalamic-pituitary-adrenal (HPA) axis (Armario, 2006). Stress activates different brain areas, depending on the particular nature of stressors (Pacák and Palkovits, 2001, Ulrich-Lai and Herman, 2009), but this information

NU

converges at the paraventricular nucleus of the hypothalamus (PVN), where neurons forming various types of neuropeptides are located. These include the corticotrophin-releasing hormone (CRH), vasopressin (VP) and other neuropeptides that are released into the pituitary

MA

portal-blood to stimulate the synthesis and release of the adrenocorticotropic hormone (ACTH). ACTH acts on the adrenal cortex to activate the synthesis and release of

D

glucocorticoids (corticosterone in rodents).

TE

The activation of the stress response has a cost. Therefore, when animals are daily exposed to a stressor, a reduction of certain physiological and behavioral responses can be observed when

CE P

confronted again with the same (homotypic) stressor, but not when exposed to novel (heterotypic) stressors. This reduced homotypic response is usually observed with emotional or mixed nature stressors, even with those of high intensity, and less frequently with those stressors having a systemic component (Martí and Armario, 1998). For instance, daily exposure

AC

to immobilization on boards (IMO) resulted in HPA adaptation whereas exposure to foot-shock did not, despite the former being more severe as evaluated by classical markers of stress (Rabasa et al., 2011), probably because of the influence of inflammatory signals associated with foot-shock. Although it has been assumed that adaptation to daily repeated stress is a habituation-like phenomenon (Grissom and Bhatnagar, 2009), we have recently challenged this assumption regarding the response of the HPA axis to predominantly emotional stressors (Rabasa et al., 2015). In brief, adaptation can be achieved with a single exposure to a severe stressor such as IMO and it is directly proportional to the intensity of the stressors and the interval (days) between successive exposures (Rabasa et al., 2015). The behavioral and physiological responses elicited by emotional stressors are considered to be anticipatory in that such situations have a certain probability to be followed by an actual serious challenge to homeostasis. For instance, detection of a predator odor or the predator itself represents a 3

ACCEPTED MANUSCRIPT potential risk for survival that will require intense activity to escape or fight and the possibility of being wounded. The stress response would contribute to prepare the organisms for it. We have hypothesized that during exposure to emotional stressors a “safety signal” is

T

progressively elaborated if no actual severe threat to homeostasis occurs that is directly

IP

proportional to the intensity of the stressful situation and the length of exposure provided that no sign of actual physical damage appears. This “safety signal” could contribute to the

SC R

progressive reduction of the response to acute prolonged emotional stressors, but also to modify the response when the same stressor is further encountered thanks to the memory about the situation. When confronted with the same stressor, this “safety signal” is triggered

NU

and opposes to the HPA activation caused by the stressor. Unfortunately, the nature of the alterations actually detected to favor or impede adaptation has not been explored. One possibility is that detection of malaise while the animals are being exposed for the first time to

MA

a stressor might constitute a signal that such exposure is potentially dangerous for the organism and could interfere with adaptation.

D

Given that a single exposure to a severe stressor such as IMO resulted in the same degree of

TE

reduction of the HPA response to a second exposure to IMO as daily repeated exposure (Rabasa et al., 2015), this paradigm offers important theoretical (easy manipulation) and

CE P

practical (less time-consuming) advantages in order to study possible processes and mechanisms involved in adaptation of the HPA axis. Thus, in the present study, the possible contribution of malaise to impair adaptation to a severe stressor (IMO) was studied using lithium chloride, which is well-known to cause malaise that induces conditioned taste aversion

AC

(CTA) to novel substances taken before lithium administration (e.g. Nachman and Ashe, 1973, Smotherman et al., 1976). CTA was firstly described by García et al. (1955) after exposing rats to radiation and constitutes a learning phenomenon thereby animals avoid novel tastes if during the next hours they have signs of nausea or gastrointestinal discomfort (see for review Mediavilla et al., 2005; Lin et al., 2016). This is an associative learning based on classical conditioning that can be easily acquired after a single trial and represents a biologically-programmed protection from novel foods containing toxins. In the present study malaise/discomfort was induced using saccharin as the conditioned stimuli and lithium as the unconditioned stimulus. To enhance the probability that animals associate malaise to IMO, lithium was administered after a first hour of IMO but the animals were maintained under stress for two additional hours. Simultaneous evaluation of saccharin aversion in all groups could be used not only as a positive control for the effect of lithium but also to test whether or not the mere exposure to IMO causes CTA. 4

ACCEPTED MANUSCRIPT 2. Methods

T

2.1. Animals

IP

Forty-five male Sprague–Dawley rats obtained from the breeding center of the Universitat

SC R

Autònoma de Barcelona were used. Rats were 2 month old at the beginning of the experiment. Upon arrival to the animal facility, animals were housed individually under standard conditions of temperature (22±1 ºC) in a 12 hour light/dark schedule (lights on at 7:00) with ad libitum access to food and water. They were allowed at least one week to acclimate themselves to the

NU

animal room before starting the experiment. The experimental protocol was approved by the Committee of Ethics of the Universitat Autònoma de Barcelona and the Generalitat de

MA

Catalunya, and it was carried out in compliance with the European Council Directive (2010/63/EU) and Spanish legislation (RD 53/2013).

TE

D

2.2. Experimental procedures

The experimental procedures were always done in the morning. One week after arrival,

CE P

animals were handled 3 times for approximately 2 minutes on different days across a period of another week. During this week, animals were subjected once to tail nick and once to the saline intraperitoneal (ip) injection to habituate animals to these procedures. CTA was induced using lithium as the unconditioned stimulus. Lithium (63 mg/Kg) was given ip as the chloride

AC

salt (Sigma Aldrich) at isotonic concentration (0.15 M) in a volume of 10 ml/Kg to maintain osmolality and avoid possible painful properties of hypertonic solutions lithium. This dose has been found to induce CTA (Bourne et al 1992). Saccharin was used as the conditioned stimuli. Rats have strong preference for sweet solutions such as those containing sucrose or saccharin, but the latter has not caloric value. As IMO results in reduced food intake (Martí et al., 1994; Pastor-Ciurana et al., 2014) we wanted to avoid the problems of interpreting changes in a sweet solution having caloric properties. Then we chose to measure saccharin intake (Saccharin sodium salt, Sigma, S1002-500G, 0.1% w/v) always under ad libitum food condition. Animals were allowed to habituate to special 250 ml bottles and stoppers (to prevent leakage) filled with water before starting the stress/injection protocol (Figure 1). The day before CTA induction, all bottles were removed in order to cause water deprivation for 20 hours. On the next day (day 1), bottles filled with saccharin were placed in the homecages and all animals

5

ACCEPTED MANUSCRIPT were allowed to drink for 30 minutes. After that, the bottles were removed and the stress

NU

SC R

IP

T

protocol/injection started.

MA

Figure 1. Experimental procedures. Panel A depicts the general procedure. In all rats, conditioned taste aversion (CTA) acquisition and exposure to the first IMO took place on day 1; CTA testing took place on day 5 and a second exposure to IMO (1h) on day 7. Panel B depicts the specific procedures followed on day 1. All rats were blood sampled (BS), then given access to saccharin in their homecages; 30 min later those to be stressed were exposed to

D

IMO for 3 h; 1 h after starting IMO, all rats (including controls) were sampled again and immediately injected with

TE

vehicle or lithium; two additional blood samples were obtained from all rats just after the 3 h IMO exposure and at 60 minutes post-IMO (or corresponding times in controls, R60).

CE P

Animals were assigned to four experimental groups: (i) Control - Vehicle (n=11), rats given ip saline (isotonic 0.15 M sodium chloride; 10 ml/kg) and then returned to their homecages, (ii) Control - Lithium (Control - Li, n=11), rats given ip 10 ml/kg of isotonic lithium chloride and

AC

then returned to their homecages (iii) IMO – Vehicle (n=11), rats exposed to 3 hours of IMO and given ip saline 1 hour after starting IMO exposure; and (iv) IMO - Li (n=12), rats exposed to 3 hours of IMO and given ip lithium chloride 1 hour after starting IMO. The IMO procedure involved immobilizing the rats on boards, as previously described (i.e. Muñoz-Abellán et al., 2011). On day 1, blood samples were taken at different times (Figure 1): just before stress (PRE-IMO), 1 hour after starting exposure to IMO (TIME 1), just after 3 hours of IMO (TIME 2) and 1 hour after the termination of stress (TIME 3). Blood samples were taken by tail nick, which consisted of gently wrapping the animals with a cloth, making a 2 mm incision at the end of one of the tail veins, and then massaging the tail while collecting blood into ice-cold EDTA capillary tubes (Sarstedt SA, Granollers, Spain). This procedure is extensively used in our lab and others because true resting levels of hormones are obtained (Belda et al., 2004; Vahl et al., 2005).

6

ACCEPTED MANUSCRIPT From the end of IMO exposure until day 4 all groups had ad libitum free-access to food and water (regular bottles). On day 4, all animals were water deprived for 20 hours. On day 5, the special bottles containing saccharin were placed in every homecage and all animals were

T

allowed to drink for 1 hour in order to assess the possible development of CTA. On day 7, all

IP

animals were exposed to 1 hour of IMO and blood samples were taken just before stress (baseline), just after the first hour of IMO and at 45 and 90 minutes after stress (R 45 and R

SC R

90).

Body weight was measured the day before the start of the experimental procedure in order to

NU

assess that there were not differences between groups. Then, body weight changes with respect to body weight on the day prior to IMO were measured on day 2 and day 4. Rats were weighed again on day 6 to study body weight changes after the IMO of day 7 from day 8 to 12.

MA

Rats were not weighed on the same days of exposure to IMO in order to not interfere with the response to the stressor.

ACTH

and

TE

Plasma

D

2.3. Biochemical analysis

corticosterone

levels

were

determined

by

double-antibody

CE P

radioimmunoassay, as previously described (Muñoz-Abellán et al., 2011). All samples to be compared were run in the same assay to avoid inter-assay variability. The intra-assay coefficient of variation was less than 6% for ACTH and corticosterone. The sensitivity was 12.5

AC

pg/ml for ACTH and 1 ng/ml for corticosterone.

2.4. Statistical analysis

Statistical Package for Social Sciences (SPSS, version 24 for Windows) was used. Plasma hormone concentrations were log-transformed. For the statistical analysis of hormone dynamics (days 1 and 7, separately) and body weight changes, generalized linear models with repeated measures (generalized estimating equations model, GEE; Hardin and Hilbe, 2003) were used. For the hormone dynamics, TIME was the within-subjects factor (three levels for DAY 1 and four levels for DAY 7). For body weight gain, DAY was the within-subjects factor (two levels after Day 1 IMO or five levels after Day 7 IMO). In every case, STRESS and DRUG on day 1 (STRESS-D1 and DRUG-D1) were the between-subjects factor (each factor with two levels). Body weight gain before the experimental phase, basal hormone levels and saccharin intake on day 1 and day 5, were analyzed with generalized linear models (GzLM; McCulloch 7

ACCEPTED MANUSCRIPT and Searle, 2001), with the already described between-subjects factors. GzLM and GEE are more flexible statistical tools than the standard general lineal model (GLM) because several types of distribution and different covariance structures of the repeated measures data could

T

be chosen. In addition, the generalized linear model does not require homogeneity of

IP

variances and admit missing values without removing all data subject. The maximum likelihood (ML) as a method of estimation and normality distribution and identity as a link function were

SC R

always used. The significance of the effects was determined by the Wald chi-square statistic followed by pairwise comparisons. Additional sequential Bonferroni comparisons were made

AC

CE P

TE

D

MA

NU

when needed. The criterion for significance was set at p < 0.05.

8

ACCEPTED MANUSCRIPT 3. Results

IP

T

3.1. Saccharin intake

SC R

The GzLM analysis of saccharin intake (Figure 2) on day 1 showed no group effect [Wald Χ2 (3) = 4.02; NS], indicating that basal intake was similar in all groups. The GzLM analysis of saccharin intake on day 5 showed significant effect of STRESS-D1 [Wald Χ2 (1) = 39.4; p < 0.001], DRUG-D1 [Wald Χ2 (1) = 86.6; p < 0.001] and the interaction STRESS-D1 X DRUG-D1

NU

[Wald Χ2 (1) = 30.6; p < 0.001]. Further comparisons showed that both IMO and lithium, given separately, significantly reduced saccharin intake as compared with Control – Vehicle (p <

MA

0.001 in the two cases). The combination of IMO and lithium reduced saccharin intake versus Control – Vehicle group (p < 0.001) and IMO – Vehicle (p = 0.007), but not below the levels of

TE

20 15

AC

CE P

Saccharin intake (mL)

D

Control – Li group (NS).

Control - Vehicle Control - Li IMO - Vehicle IMO - Li



10

***

5 0

***

Day 1 (Baseline)

***

Day 5 (CTA)

Figure 2. Conditioned taste aversion induced by IMO stress and lithium administration. Saccharin intake was measured on day 1 before any treatment; then, animals were assigned to four different groups: Control – Vehicle (n=9); Control – Li (n=10); IMO – Vehicle (n=8); IMO – Li (n=12). On day 5, groups exposed to IMO, lithium or both stimuli showed a marked conditioned taste aversion (CTA) when compared with the Control – Vehicle group. Note that some animals with problems to drink saccharin from the special bottles were eliminated from the statistical analysis. All values are expressed as mean ± SEM. *** p < 0.001 vs Control-Vehicle, ∆∆ p < 0.01 between the signaled groups

9

ACCEPTED MANUSCRIPT 3.2. ACTH and corticosterone

T

The GzLM analysis of basal ACTH levels (Figure 3) on day 1 showed no group effect [Wald Χ2 (3)

IP

= 3; NS], indicating no baseline differences between groups at the start of the experimental procedure. The GEE analysis of ACTH response on day 1 showed significant effect of STRESS-D1

SC R

[Wald Χ2 (1) = 642.7; p < 0.001], DRUG-D1 [Wald Χ2 (1) = 22.8; p < 0.001], TIME [Wald Χ2 (2) = 355.2; p < 0.001], and the interactions STRESS-D1 X TIME [Wald Χ2 (2) = 355.0; p < 0.001], DRUG-D1 X TIME [Wald Χ2 (2) = 21.3; p < 0.001] and STRESS-D1 X DRUG X TIME [Wald Χ2 (2) =

NU

32.3; p < 0.001]. Due to the triple interaction, pair-wise comparisons were performed. IMO increased plasma ACTH levels at all sampling times, in both saline- and lithium-injected animals with respect to Control – Vehicle group (p < 0.001 in all cases). Administration of lithium

MA

(Control – Li group) was only able to increase plasma ACTH measured 2 hours after its administration (TIME 2, p < 0.001 vs Control-Vehicle). Moreover, the IMO – Li group did not differ from the IMO –Vehicle group at TIME 2, but at TIME 3 higher levels were observed in the

D

former group (p = 0.03). Finally, the IMO – Li group showed higher levels of plasma ACTH at

TE

TIME 2 and TIME 3 than the Control – Li group (p < 0.001 in both cases).

750

+++

250



*** ***

+++

500

*** *** ***

ACTH (pg/ml)

1000

***

750

+++

*** ***

+++

600

*** ***

450

***

300

0

***

150 pre - IMO

TIME 1

TIME 2

***

corticosterone (ng/ml)

0

*** ***

AC

Control - Vehicle Control - Li IMO - Vehicle IMO - Li

***

CE P

1250 DAY 1

TIME 3

Figure 3. Plasma ACTH and corticosterone levels under the various treatments of day 1. Mean ± SEM are represented. Insets represent hormone levels before access to saccharin and any treatment on day 1. The groups were as follows: Control – Vehicle (n=11); Control –Li (n=11); IMO – Vehicle (n=11); IMO – Li (n=12). Arrows indicate the time of vehicle or Li injection. Time 1 corresponds to 1 h of IMO, Time 2 to 3 h of IMO and Time 3 to 45 min after the termination of IMO. *** indicates significant difference (p < 0.001) versus Control-Vehicle, +++ p < 0.001 versus Control – Li, in all cases within the same sampling time; ∆ indicates significant difference (p < 0.05) between the signaled groups.

10

ACCEPTED MANUSCRIPT

Control - Vehicle Control - Li IMO - Vehicle IMO - Li

1000 750 500

$$$

250

$$$

$$

MA

600 450

$$$

300 150

D

corticosterone (ng/ml)

0 750

BASAL

1 H IMO

TE

0

SC R

DAY 1 treatments

DAY 7

NU

ACTH (pg/ml)

1250

IP

T

The GEE analysis of ACTH on day 7 (Figure 4) showed significant effect of STRESS-D1 [Wald Χ2 (1) = 36.3; p < 0.001], TIME [Wald Χ2 (3) = 3070.2; p < 0.001]; and STRESS-D1 X TIME interaction [Wald Χ2 (3) = 74.8; p < 0.001], but no effect of DRUG-D1 or its interactions. The decomposition of the interaction STRESS-D1 X TIME showed no effect of prior IMO exposure on day 1 on the ACTH levels observed just after the second IMO, but significantly lower levels at R45 and R90 (p < 0.001).

$$$

R 45

R 90

CE P

Figure 4. Plasma ACTH and corticosterone levels in response to IMO on day 7. Mean ± SEM are represented. All rats were exposed to 1 h of IMO and blood sampled immediately after IMO and again 45 and 90 min after the termination of IMO (R45 and R90). The groups indicate the treatment on day 1 and were as follows: Control – Vehicle (n=11); Control –Li (n=11); IMO – Vehicle (n=11); IMO – Li (n=12). $$$ indicate significant effect (p <0.001) of

AC

IMO exposure on day 1, regardless of lithium treatment, always within the same sampling time.

The GzLM analysis of basal corticosterone (Figure 3) on day 1 showed no baseline differences among the groups [Wald Χ2 (3) = 1.8; NS]. The GEE analysis of corticosterone response to treatments on day 1 showed significant effect of STRESS-D1 [Wald Χ2 (1) = 501.4; p < 0.001], DRUG-D1 [Wald Χ2 (1) = 40.9; p < 0.001], TIME [Wald Χ2 (2) = 17.9; p < 0.001], and the interactions STRESS-D1 X DRUG-D1 [Wald Χ2 (1) = 34.5; p < 0.001], STRESS-D1 X TIME [Wald Χ2 (2) = 41.9; p < 0.001], DRUG-D1 X TIME [Wald Χ2 (2) = 15.9; p < 0.001] and STRESS-D1 X DRUGD1 X TIME [Wald Χ2 (2) = 19.8; p < 0.001]. Due to the triple interaction, pair-wise comparisons were made. Higher plasma corticosterone levels in the two IMO groups compared with the Control-Vehicle group at all sampling times (p < 0.001 in all cases). Lithium administration increased plasma corticosterone response at 2 and 3 hours after its administration (TIME 2 and TIME 3) as compared with Control - Vehicle group (p < 0.001 in the two cases). Moreover, IMO 11

ACCEPTED MANUSCRIPT – Li and IMO – Vehicle groups did not differ at any time. Finally, the IMO – Li group showed higher levels of plasma corticosterone at TIME 2 and TIME 3 than the Control – Li group (p < 0.001 in all cases).

T

The GEE analysis of corticosterone on day 7 (Figure 4) showed significant effect of STRESS-D1

IP

[Wald Χ2 (1) = 16.9; p < 0.001], TIME [Wald Χ2 (3) =1353.2; p < 0.001] and STRESS-D1 X TIME

SC R

interaction [Wald Χ2 (3) = 80.8; p < 0.001], but no effect of DRUG-D1 or its interactions. Further comparisons showed higher plasma corticosterone levels just after 1 hour of IMO in those groups with previous experience with IMO (p = 0.003), whereas lower levels were observed in

NU

both post-IMO periods (p < 0.001).

MA

3.3. Body weight changes

D

The GzLM analysis of body weight gain during the pre-experimental period showed no group

TE

differences [Wald Χ2 (3) = 1.9; NS]. The GEE analysis of body weight changes after day 1 treatments (Figure 5A) showed significant effect of STRESS-D1 [Wald Χ2 (1) = 120.4, p < 0.001],

CE P

DAY [Wald Χ2 (1) = 83.7, p < 0.001] and STRESS-D1 X DAY interaction [Wald Χ2 (1) = 35.5, p < 0.001] but no effect of DRUG-D1 or its interactions. Further comparisons revealed a significant reduction of body weight in both IMO groups compared with controls on day 2 and day 4 (p <

AC

0.001).

The GEE analysis of body weight changes after the acute IMO on day 7 (Figure 5B) indicated significant effect of STRESS-D1 [Wald Χ2 (1) = 21.4, p < 0.001], DAY [Wald Χ2 (4) = 200.6, p < 0.001] and STRESS-D1 X DAY interaction [Wald Χ2 (4) = 20.3, p < 0.001], but no effect of DRUGD1 or its interactions. Further comparisons revealed that in animals with previous experience with IMO, a lower impact of the second IMO was found from days 9 to 12 (p < 0.001 in all cases), but not on day 8.

12

ACCEPTED MANUSCRIPT A Control - Vehicle Control - Li IMO - Vehicle IMO - Li

15

IP

T

5

SC R

-5 -15

$$$

$$$

-25

Day 2

Day 4

NU

Body Weight changes (g/rat)

25

B

MA

10

-5

D

$$$

TE

0

$$$

$$$

5

$$$

CE P

Body Weight changes (g/rat)

15

-10 -15

AC

-20

Day 8

Day 9 Day 10 Day 11 Day 12

Day 7 IMO 1h

Figure 5. Body weight changes throughout all the experimental period. Mean ± SEM are represented. The groups indicate the treatment on day 1 and were as follows: Control – Vehicle (n=8); Control –Li (n=11); IMO – Vehicle (n=9); IMO – Li (n=11). Panel A represents the changes in body weight from day -1 to 2 (day 2) and day -1 to 4 (day 4). Panel B represents the changes in body weight after exposure of all rats to 1 h of IMO on day 7. Groups with a previous experience with IMO on day 1 showed a higher body weight gain throughout following days. Note that some animals with problems to drink saccharin from the special bottles were eliminated from the statistical analysis because body weight was affected by problems with access to drinking fluid. $$$ indicate significant effect of IMO exposure on day 1 (p <0.001), regardless of lithium administration.

13

ACCEPTED MANUSCRIPT 4. Discussion

T

The aim of the present work was to test the hypothesis that experiencing malaise during

IP

exposure to a severe (predominantly) emotional stressor (IMO) could interfere with adaptation of the HPA axis and body weight to this stressor. We exposed animals to 3 hours of

SC R

IMO and induced malaise injecting lithium chloride just after the first hour of IMO. We studied not only adaptation of the HPA axis to a second exposure to the stressor six days later, but also CTA development. Lithium and also IMO induced strong CTA. However, the reduction of the

NU

HPA axis response after a second exposure to IMO was not affected by lithium administration, suggesting that the experience with a systemic stressor did not interfere with the process of

MA

HPA axis adaptation to IMO.

A single prior exposure to IMO causes a reduction of the response of HPA hormones to a second exposure to the same stressor that is quite similar to that observed after daily repeated

D

exposure for at least one week (Rabasa et al., 2015). This supports that adaptation to daily

TE

repeated stress and the homotypic reduction of the HPA response after a single exposure is the same phenomenon. We have hypothesized that this adaptation to a stressor is dependent

CE P

on the elaboration of a “safety signal” that appears if there is no actual strong homeostatic alteration. We then exposed animals to IMO for 3 hours and after the first hour we injected lithium, which it is well-known to cause malaise and CTA, to test whether induction of malaise during the first day of exposure to IMO could interfere with the process of adaptation to the

AC

stressor. Adaptation was evaluated by the response of the HPA axis and changes in body weight after a second IMO days later (Dal-Zotto et al., 2004). To confirm that in our conditions lithium administration caused the expected CTA, all animals were given access to saccharin for 30 minutes; then, after an additional period of 30 minutes, different groups of animals were exposed or not to IMO and were given or not lithium. In accordance with previous reports, lithium administration per se caused a marked aversion for saccharin when preference was tested four days after lithium administration. Importantly, IMO exposure also resulted in strong reduction of saccharin preference. This was unlikely to be due to a residual nonconditioned effect of the stressor as we have previously observed that acute IMO exposure causes only partial reduction of saccharin preference that almost completely recover at postIMO day 4 (Pastor-Ciurana et al., 2015). The IMO-induced CTA is in accordance with the available literature about conditioned stressinduced saccharin aversion. Thus, Dess et al., (1988) firstly reported evidence for CTA to 14

ACCEPTED MANUSCRIPT saccharin in animals which were exposed to a session of uncontrollable tail-shocks (typical of the learned helplessness paradigm) 15 minutes after removing the saccharin bottle. Testing for CTA took place 2 days after the shocks. They also demonstrated that this was not an

T

unconditioned consequence of the stressor and that control over shocks did not diminish the

IP

effects. After that, other emotional stressors have been found to induced CTA, although higher number of sessions is required (Howell et al., 1999; Masaki and Nakajima, 2004; Revusky and

SC R

Reilly, 1989), probably because the magnitude of CTA is dependent on the severity and duration of exposure to the stressors (Masaki and Nakajima, 2005, 2006, Brand et al., 2008). In our hands, exposure to both IMO and lithium resulted in CTA for saccharin that was similar

NU

to that observed after lithium. These data conflict with results by Dess et al., (1988) who observed that despite tail-shocks induced CTA, the stressor reduced lithium-induced CTA.

MA

Using different stressors and more than one session, several laboratories have reported similar interference of stress with lithium-induced CTA (Bourne et al., 1992, Revusky and Reilly, 1989). This interference is intriguing as it can be observed when exposure to stress occurs before

D

saccharin intake and also shortly (15 minutes) after lithium administration (Bourne et al.,

TE

1992). The latter authors discussed differed explanations for the results, including that exteroceptive signals interfere with interoceptive ones, that prior stress could have reduced

CE P

the aversive effects of lithium, differences in the state between conditioning and testing, or overshadowing, but there are other possibilities. First, stress before saccharin could have diminished CTA by reducing saccharin consumption during the conditioning session. Although this was not directly tested, there is evidence for a reduction of saccharin intake when mice

AC

had access to saccharin 30 minutes after stress (Rademacher and Hillard, 2007). However, this possibility does not apply to results by Dess et al. (1988) as stress exposure always occurred after saccharin intake. Second, stress exposure after lithium might interfere with the capability of the animals to associate saccharin exclusively with lithium administration. If severe emotional stressors per se can generate some kind of malaise having similarities with that induced by lithium, animals will have difficulties to specifically associated malaise with lithium. However, how can we explain the discrepancies between our data and previous studies? A major difference is that we gave lithium to the animals while immobilized, whereas all other authors gave it before or after stress exposure. In the latter case, it is possible the release of the animals from stress was acting as a rewarding stimulus (Shen et al., 2010, 2011), thus partially counteracting the aversive properties of lithium. In contrast, in our conditions lithium likely potentiated the aversive interoceptive signals associated with IMO, although no additive effects could be found due to the potent effect of each individual stimulus. 15

ACCEPTED MANUSCRIPT Regardless of the reasons for the discrepant results, the present and previous results strongly suggest that exposure to stressors in which the emotional component predominates appears to elicit interoceptive cues that are quite similar to those caused by lithium and probably by

T

physical stressors such as endotoxin and cytokines such as interleukin-1 (Tazi et al., 1988),

IP

particularly when they are severe.

SC R

Plasma corticosterone levels prior to saccharin availability on day 1 were somewhat higher than those observed after saccharin consumption in those rats not exposed to IMO, suggesting moderately elevated levels. These results are consistent with a prior report showing that one day of water deprivation increased basal levels of corticosterone, and these levels declined

NU

after access to water (Arnold et al., 2006, see the paper for further discussion on this subject). More importantly, lithium administration increased plasma levels of ACTH and corticosterone,

MA

supporting previous results indicating that this procedure is capable of activating the HPA axis (Jacobs, 1978). The main effect of lithium is probably located within the brain as it induces cfos activation in the PVN (Gu et al., 1993; Lamprecht and Dudai, 1995; Spencer and Houpt,

D

2001), but an additional direct effect on the anterior pituitary cannot be ruled out (Zatz and

TE

Reisine, 1985). It is of note that the effect of lithium was very modest as compared to that of IMO, despite the former elicited strong CTA. It could be argued that the modest activation of

CE P

the HPA axis could not reflect activation of other physiological markers of stressor intensity. However, food intake is sensitive to stressor intensity (Márquez et al., 2002; Martí et al., 1994) and was not affected by lithium, in contrast to the strong effect of IMO (not shown). Similarly, in the present study IMO caused a marked hyperglycemia, whereas lithium did not (data not

AC

shown), supporting the modest stressful properties of lithium. It thus appears that activation of the HPA axis and other physiological responses by IMO-associated signals might be much stronger than that elicited by lithium-derived interoceptive signals. Unfortunately, the specific signals involved in the activation of the physiological stress response by different types of stressors have not been thoroughly characterized. In this regard, IMO-induced c-fos expression in several brain areas, including the PVN, was reduced after treatment of young rats with capsaicin that damage type C fibers and thus impairs interoceptive signal entering the nucleus of the solitary tract (Pan et al., 1997). However, C fibers carry out not only all types of visceral signals but also pain-related signals, making it difficult to know the precise type of signals contributing to HPA activation. Prior exposure to 3 hours of IMO resulted in reduced ACTH and corticosterone response to a second IMO that was evident not just after the stressor but during the post-IMO period, in accordance with most of our previous results (Dal-Zotto et al., 2003, Martí et al., 2001, Rabasa 16

ACCEPTED MANUSCRIPT et al., 2015), although an effect on the initial response is often observed (Rabasa et al., 2015). Also, prior IMO diminished the impact of the last IMO on body weight changes over the following days (Dal-Zotto et al., 2004). Importantly, lithium administration during the first

T

exposure to the stressor did not interfere with the reduced response to a second IMO. There

IP

are at least two possible explanations for this. First, if IMO elicited similar interoceptive cues as lithium, and the effect of IMO was strong, there would be no room for an increased signal after

SC R

lithium administration to stressed rats. Second, lithium did not interfere with HPA adaptation to IMO because the kind of (safety) signal participating in adaptation could have a predominant somatosensorial rather than visceral origin. This differential influence of

NU

exteroceptive versus interoceptive signals might also contribute to explain the lack of effect of

MA

lithium to interfere with HPA adaptation to IMO.

D

5. Conclusions

TE

In sum, the present results confirm that a single exposure to a severe stressor caused a strong CTA for saccharin. Also, demonstrate that adaptation of the HPA axis to the homotypic

CE P

stressor, after a single prior experience with the stressor, is not affected by lithium-induced malaise while the animals were exposed for the first time to the stressor. Other kind of signals

AC

might participate in the elaboration of a “safety signal” leading to adaptation.

6. Conflicts of interest

There was no conflict of interest for this study.

7. Acknowledgements

The work in the laboratory was supported by Spanish grants to AA and RN from “Ministerio de Economía

y

Competitividad”

(SAF2014-53876R),

“Instituto

de

Salud

Carlos

III”

(RD12/0028/0014, “Redes Temáticas de Investigación Cooperativa en Salud”, “Ministerio de 17

ACCEPTED MANUSCRIPT Sanidad y Consumo” and “Generalitat de Catalunya” (SGR2014-1020). RN is the recipient of an ICREA-ACADEMIA award (Generalitat de Catalunya).

SC R

IP

T

8. References

Armario, A. 2006. The hypothalamic-pituitary-adrenal axis: what can it tell us about stressors? CNS Neurol. Disord. Drug Targets. 5, 485-501.

NU

Arnhold, M.M., Wotus, C., Engeland, W.C. 2007. Differential regulation of parvocellular neuronal activity in the paraventricular nucleus of the hypothalamus following single vs

MA

repeated episodes of water restriction-induced drinking. Exp. Neurol. 206, 126-36. Belda, X., Márquez, C., Armario, A. 2004. Long-term effects of a single exposure to stress in adult rats on behavior and hypothalamic-pituitary-adrenal responsiveness: comparison of two

D

outbred rat strains. Behav. Brain Res. 154, 399-408.

TE

Bourne, M.J., Calton, J.L., Gustavson, K.K., Schachtman, T.R. 1992. Effects of acute swim stress

CE P

on LiCl-induced conditioned taste aversions. Physiol. Behav. 51, 1227-34. Brand, L., Groenewald, I., Stein, D.J., Wegener, G., Harvey, B.H. 2008. Stress and re-stress increases conditioned taste aversion learning in rats: possible frontal cortical and hippocampal

AC

muscarinic receptor involvement. Eur. J. Pharmacol. 586, 205-11. Dal-Zotto, S., Martí, O., Armario, A. 2003. Glucocorticoids are involved in the long-term effects of

a

single

immobilization

stress

on

the

hypothalamic-pituitary-adrenal

axis.

Psychoneuroendocrinology. 28, 992-1009. Dal-Zotto, S., Martí, O., Delgado, R., Armario, A. 2004. Potentiation of glucocorticoid release does not modify the long-term effects of a single exposure to immobilization stress. Psychopharmacology. 177, 230-7. Dess, N.K., Raizer, J., Chapman, C.D., García, J. 1988. Stressors in the learned helplessness paradigm: effects on body weight and conditioned taste aversion in rats. Physiol. Behav. 44, 483-90. Garcia, J., Kimeldorf, D.J., Koelling, R.A. 1955. Conditioned aversion to saccharin resulting from exposure to gamma radiation. Science 122, 157-158. 18

ACCEPTED MANUSCRIPT Grissom, N., Bhatnagar, S. 2009. Habituation to repeated stress: get used to it. Neurobiol. Learn. Mem. 92, 215-24. Gu, Y., González, M.F., Chin, D.Y., Deutsch, J.A. 1993. Expression of c-fos in brain subcortical

T

structures in response to nauseant lithium chloride and osmotic pressure in rats. Neurosci.

IP

Lett. 157, 49-52.

SC R

Hardin, J., Hilbe, J.M. 2003. Generalized Estimating Equations. Champman and Hall: CRC. Howell, L.A., Harris, R.B., Clarke, C., Youngblood, B.D., Ryan, D.H., Gilbertson, T.A. 1999. The effects of restraint stress on intake of preferred and nonpreferred solutions in rodents. Physiol.

NU

Behav. 65, 697-704.

Jacobs, J.J. 1978. Effect of lithium chloride on adrenocortical function in the rat. Proc. Soc. Exp.

MA

Biol. Med. 157, 163-7.

Lamprecht, R., Dudai, Y. 1995. Differential modulation of brain immediate early genes by

D

intraperitoneal LiCl. Neuroreport. 7, 289-93.

TE

Lin, J.-Y., Arthurs, J., Reilly, S. 2016. Conditioned taste aversions: from poison to pain to

CE P

drugs of abuse. Psychon. Bull. Rev. (in press). Márquez, C., Belda, X., Armario, A. 2002. Post-stress recovery of pituitary-adrenal hormones and glucose, but not the response during exposure to the stressor, is a marker of stress

AC

intensity in highly stressful situations. Brain Res. 926, 181-5. Martí, O., Armario, A. 1998. Anterior pituitary response to stress: time-related changes and adaptation. Int. J. Dev. Neurosci. 16, 241-60. Martí, O., García, A., Vallès, A., Harbuz, M.S., Armario, A. 2001. Evidence that a single exposure to aversive stimuli triggers long-lasting effects in the hypothalamus-pituitary-adrenal axis that consolidate with time. Eur. J. Neurosci. 13, 129-36. Martí, O., Martí, J., Armario, A. 1994. Effects of chronic stress on food intake in rats: influence of stressor intensity and duration of daily exposure. Physiol. Behav. 55, 747-53. Masaki, T., Nakajima, S. 2004. Taste aversion learning induced by delayed swimming activity. Behav. Processes. 67, 357-62. Masaki, T., Nakajima, S. 2005. Further evidence for conditioned taste aversion induced by forced swimming. Physiol. Behav. 84, 9-15. 19

ACCEPTED MANUSCRIPT Masaki, T., Nakajima, S. 2006. Taste aversion in rats induced by forced swimming, voluntary running, forced running, and lithium chloride injection treatments. Physiol. Behav. 88, 411-6. McCulloch, C., Searle, S. 2001. Generalized, linear and mixed models. Wiley and Sons: New

IP

T

York.

Mediavilla, C., Molina, F., Puerto, A. 2005. Concurrent conditioned taste aversion: a learning

substances. Neurosci. Biobehav. Rev. 29, 1107-18.

SC R

mechanism based on rapid neural versus flexible humoral processing of visceral noxious

Muñoz-Abellán, C., Rabasa, C., Daviu, N., Nadal, R., Armario, A. 2011. Behavioral and endocrine

NU

consequences of simultaneous exposure to two different stressors in rats: interaction or independence? PLoS One. 6, e21426.

MA

Nachman, M., Ashe, J.H. 1973. Learned taste aversions in rats as a function of dosage, concentration, and route of administration of LiCl. Physiol. Behav. 10, 73-8.

D

Pacák, K., Palkovits, M. 2001. Stressor specificity of central neuroendocrine responses:

TE

implications for stress-related disorders. Endocr. Rev. 22, 502-48. Pan, B., Castro-Lopes, J.M., Coimbra, A. 1997. Chemical sensory deafferentation abolishes

CE P

hypothalamic pituitary activation induced by noxious stimulation or electroacupuncture but only decreases that caused by immobilization stress. A c-fos study. Neuroscience. 78, 1059-68. Pastor-Ciurana, J., Rabasa, C., Ortega-Sánchez, J.A., Sanchís-Ollé, M., Gabriel-Salazar, M.,

AC

Ginesta, M., Belda, X., Daviu, N., Nadal, R., Armario, A. 2014. Prior exposure to repeated immobilization or chronic unpredictable stress protects from some negative sequels of an acute immobilization. Behav. Brain Res. 265, 155-62. Rabasa, C., Muñoz-Abellán, C., Daviu, N., Nadal, R., Armario, A. 2011. Repeated exposure to immobilization or two different foot shock intensities reveals differential adaptation of the hypothalamic-pituitary-adrenal axis. Physiol. Behav. 103, 125-33. Rabasa, C., Gagliano, H., Pastor-Ciurana, J., Fuentes, S., Belda, X., Nadal, R., Armario, A. 2015. Adaptation of the hypothalamus-pituitary-adrenal axis to daily repeated stress does not follow the rules of habituation: A new perspective. Neurosci. Biobehav. Rev. 56, 35-49. Rademacher, D.J., Hillard, C.J. 2007. Interactions between endocannabinoids and stressinduced decreased sensitivity to natural reward. Prog. Neuropsychopharmacol. Biol. Psychiatry. 31, 633-41. 20

ACCEPTED MANUSCRIPT Revusky, S., Reilly, S. 1989. Attenuation of conditioned taste aversions by external stressors. Pharmacol. Biochem. Behav. 33, 219-26. Shen, Y.L., Chen, J.C., Liao, R.M. 2011. Place conditioning and neurochemical responses elicited

T

by the aftereffect of acute stressor exposure involving an elevated stand. Neurosci. Lett. 504,

IP

156-9.

SC R

Shen, Y.L., Chen, Y.C., Liao, R.M. 2010. Dopamine receptor antagonists impair place conditioning after acute stress in rats. Behav. Pharmacol. 21, 77-82.

Smotherman, W.P., Hennessy, J.W., Levine, S. 1976. Plasma corticosterone levels as an index

NU

of the strength of illness induced taste aversions. Physiol. Behav. 17, 903-8. Spencer, C.M., Houpt, T.A. 2001. Dynamics of c-fos and ICER mRNA expression in rat forebrain

MA

following lithium chloride injection. Mol. Brain Res. 93, 113-26. Tazi, A., Dantzer, R., Crestani, F., Le Moal, M. 1988. Interleukin-1 induces conditioned taste

D

aversion in rats: a possible explanation for its pituitary-adrenal stimulating activity. Brain Res.

TE

473, 369-71.

Ulrich-Lai, Y.M., Herman, J.P. 2009. Neural regulation of endocrine and autonomic stress

CE P

responses. Nat. Rev. Neurosci. 10, 397-409. Vahl, T.P., Ulrich-Lai, Y.M., Ostrander, M.M., Dolgas, C.M., Elfers, E.E., Seeley, R.J., D´Alessio, D.A., Herman, J.P. 2005. Comparative analysis of ACTH and corticosterone sampling methods

AC

in rats. Am. J. Physiol Endocrinol. Metab. 289, E823-8. Zatz, M., Reisine, T.D. 1985. Lithium induces corticotropin secretion and desensitization in cultured anterior pituitary cells. Proc. Natl. Acad. Sci. U S A. 82, 1286-90.

21

ACCEPTED MANUSCRIPT Highlights Lithium chloride (ip) causes conditioned taste aversion (CTA) for saccharin.

IP

Immobilization (IMO) stress causes strong HPA activation and CTA.

T

Lithium weakly activates the hypothalamic-pituitary-adrenal (HPA) axis.

AC

CE P

TE

D

MA

NU

SC R

Lithium given during a first IMO does not affect HPA adaptation to the same stressor.

22