Psychoneuroendocrinology 125 (2021) 105127 Contents lists available at ScienceDirect Psychoneuroendocrinology journal homepage: www.elsevier.com/locate/psyneuen Individual differences in the neuroendocrine response of male rats to emotional stressors are not trait-like and strongly depend on the intensity of the stressors Roser Nadal a, c, d, Marina Gabriel-Salazar a, b, María Sanchís-Ollé a, b, Humberto Gagliano a, b, Xavier Belda a, b, Antonio Armario a, b, d, * a Institut de Neurociències, Universitat Autònoma de Barcelona, Spain b Animal Physiology Unit (Department of Cellular Biology, Physiology and Immunology), Faculty of Biosciences, Universitat Autònoma de Barcelona, Spain c Psychobiology Unit, Faculty of Psychology, Universitat Autònoma de Barcelona, Spain d CIBERSAM, Spain A R T I C L E I N F O A B S T R A C T Keywords: Biological response to stressors is critical to understand stress-related pathologies and vulnerability to psychiatric Stress markers diseases. It is assumed that we can identify trait-like characteristics in biological responsiveness by testing Individual differences subjects in a particular stressful situation, but there is scarce information on this issue. We then studied, in a Hypothalamic-pituitary-adrenal axis normal outbred population of adult male rats (n = 32), the response of well-characterized stress markers (ACTH, Prolactin Reliability corticosterone and prolactin) to different types of stressors: two novel environments (open-field, OF1 and OF2), Consistency an elevated platform (EP), forced swim (SWIM) and immobilization (IMO). Based on both plasma ACTH and prolactin levels, the OF1 was the lowest intensity situation, followed by the OF2 and the EP, then SWIM and finally IMO. When correlations between the individual responses to the different stressors were studied, the magnitude of the correlations was most dependent on the similarities in intensity rather than on other charac­ teristics of stressors, with good correlations between similar intensity stressors and no correlations at all were found between stressors markedly differing in intensity. In two additional confirmatory experiments (n = 37 and n = 20) with HPA hormones, we observed good correlation between the response to restraint and IMO, which were close in intensity, and no correlation between OF1 and SWIM. The present results suggest that individual neuroendocrine response to a particular stressor does not predict the response to another stressor greatly differing in intensity, thus precluding characterization of low or high responsive individuals to any stressor in a normal population. The present data have important implications for human studies. 1. Introduction extensively studied biological variables are those associated with the activation of the hypothalamic-pituitary-adrenal (HPA) axis (e.g. plasma Exposure to stress has been associated to a wide range of pathologies, levels of ACTH and glucocorticoids) and the autonomic nervous system including immune suppression, anxiety, depression and susceptibility to (ANS), mainly the sympathetic branch (Armario et al., 2020). ANS drug addiction. Nevertheless, it is acknowledged that the particular activation is mainly reflected in the release of adrenaline from the ad­ consequences of exposure to stress are markedly dependent on genetic renal medulla and that of noradrenaline from the adrenal medulla and or environmentally acquired individual differences in susceptibility. It is sympathetic terminals, together with important cardiovascular (CV) thus critical to characterize such individual differences in vulnerability changes. Other well-studied physiological functions under stress are the as a means to predict the detrimental impact of stress and prevent, if endocrine axis of the anterior pituitary (e.g. prolactin) and the immune possible, exposure to severe stressors or its consequences. system. It is reasonable to assume that individual differences in sus­ Although stress alters numerous physiological systems, the most ceptibility to stress-related pathologies are linked to differences in the * Correspondence to: Animal Physiology Unit, Faculty of Biosciences, Universitat Autònoma de Barcelona, Campus Bellaterra, Cerdanyola del Vallès, 08193 Barcelona, Spain. E-mail address: Antonio.armario@uab.es (A. Armario). https://doi.org/10.1016/j.psyneuen.2021.105127 Received 7 August 2020; Received in revised form 18 November 2020; Accepted 3 January 2021 Available online 4 January 2021 0306-4530/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). R. Nadal et al. Psychoneuroendocrinology 125 (2021) 105127 biological response to stressors, particularly in those systems whose were handled at least three times on different days for approximately 2 response is related to the intensity and duration of stressors (Char­ min a day. In addition, one blood sample was taken by tail-nick as mandari et al. 2005; Armario et al., 2020). described previously (Belda et al., 2004), in order to habituate the an­ However, before establishing a putative relationship between the imals to the procedure. Tail-nick is extensively used in our lab and others HPA or other stress responsive systems and particular behavioral traits because very low resting levels of hormones are obtained under or individual differences in susceptibility to stress, it is important to appropriate conditions (Belda et al., 2004; Vahl et al., 2005). All know whether individual differences in biological responsiveness to experimental procedures were done in the morning. Cage-mates were stress are reliable, particularly regarding stressors having a major sampled simultaneously (two experimenters were sampling at the same emotional component. The study of the consistency of individual dif­ time and a third was gently holding the two rats). Blood was centrifuged ferences in response to stress has two main steps. This first one is to at 4930× g (15 min, 4 ºC), and plasma was frozen (− 20 ºC) until assay. know whether the response to a particular type of stressors is reliable Animals were assigned at random to the different experimental groups when measured in more than two occasions, days, months or years later. in function of their date of birth and body weight. The second is whether low or high responsiveness is maintained when exposed to stressors differing in nature or intensity. More precisely, 2.2. Experimental designs (Fig. 1) whether or not we can actually define universal stress hypo- or hyper- responsive phenotypes based on the data obtained from specific stress 2.2.1. Main experiment (Exp. 1) situations. Rats were assigned to control (n = 10) and stress (n = 32) groups. In humans, reliability of the neuroendocrine, cardiovascular and The stress group rats were sequentially exposed to various stressors for immune response to stressors has attracted great interest, although re­ 20 min: open-field 1 (OF1) on day 1, elevated platform (EP) on day 4, sults are not conclusive particularly when comparing different situations forced swim (SWIM) on day 8, OF2 on day 11 and immobilization on on different days (Parati et al., 1988; Cohen et al., 2000; Halpern et al., boards (IMO) on day 15. Controls rats were only exposed to OF2 on day 2002; Hankin et al., 2015). A more recent study showed a good corre­ 11. Both control and stress rats were blood sampled immediately after lation in salivary cortisol response to real-life academic examination and the stressors or after being taken from their home-cages if not stressed. the laboratory trier social stress test (Henze et al., 2017), although the In the particular case of IMO, rats were again sampled 30 and 60 min two stressors were likely to be of similar intensity. Surprisingly, there are after the termination of the stressor (R30 and R60), because this is a to our knowledge very scarce studies on this topic in rodents despite the severe stressor and the overall HPA response is better evaluated vast literature dealing with the relationship between behavioral traits following the post-stress period (García et al., 2000; Márquez et al., and endocrine responsiveness. In a normal population of adult outbred 2002). male rats, individual differences in the catecholamine response to The OF1 consisted of a grey rectangular box (56 × 36 × 31 cm) immobilization (Taylor et al., 1989) or the corticosterone response to a placed in a room with dim light. The EP consisted of a non-protected same novel environment (Cavigelli et al., 2009) appeared to be quite white small platform (15 × 15 cm) 100 cm above the floor. The OF2 stable when assessed several months apart. Similarly, we have previ­ consisted of a rectangular box (68 × 56 × 42 cm) with a white floor and ously reported good correlations in ACTH, corticosterone and prolactin black walls, placed in a room with high intensity light. OFs and EP were responsiveness to different novel environments (circular corridor, cleaned carefully between animals with a tap water solution containing elevated plus maze and hole-board), which represent stressors of similar ethanol (5% v/v). nature and relatively low intensity (Márquez et al., 2005; 2006). How­ SWIM was done in a transparent cylindrical tanks (height: 40 cm, ever, there is no evidence that such correlation might be maintained internal diameter: 19 cm) with 24 cm of water (36–37 ◦ C) and water was after exposure to stressors differing in nature and intensity. changed between animals (Rabasa et al., 2015). If there is no consistent individual HPA response across different IMO rats were immobilized on boards as previously described types of emotional stressors and that response is critically dependent on (Rabasa et al., 2015). Rats were restrained in a prone position by the particular stressful situation, this inconsistency could at least in part attaching their four limbs to metal mounts with adhesive tape. Head explain the controversial results regarding HPA activity and particular movements were restricted by means of two metal loops around the behavioral traits (Armario and Nadal, 2013). Therefore, the aim of the neck, and the body was subjected to the board by means of a piece of present work was to characterize the correlation of individual differ­ plastic cloth (10 cm wide) attached with Velcro® which surrounded all ences in the response of adult male rats to a set of predominantly the trunk. stressors chosen because there is evidence that they greatly differ in nature and intensity on the basis of well-known biological indexes, 2.2.2. Complementary experiments including HPA hormones, prolactin and food intake (Belda et al., 2016; In Exp. 2, thirty-seven rats were exposed for 15 min to the OF1 and Márquez et al., 2002; Pace et al., 2005; Rabasa et al., 2015; Rotllant two days later to SWIM. In Exp. 3 twenty rats were firstly exposed to 30 et al.,2007), all of them good markers of stressor intensity in animals and min restraint in tubes and 12 days later to 30 min IMO. Plexiglas cy­ humans (Armario et al., 2020). lindrical restrainers (WPI, UK, Ref. STR554) were used, measuring 6 cm in diameter and 21.5 cm in length, with several holes in the walls of the 2. Materials and methods cylinder to provide fresh air (Rabasa et al. 2015). 2.1. Animals and general procedure 2.3. Hormone analysis Male Sprague–Dawley rats obtained from the breeding centre of the Plasma ACTH and corticosterone levels were determined by double- Universitat Autònoma de Barcelona were used. They were 2 months-old antibody radioimmunoassay (RIA) following our general procedures at the beginning of the experiments. The animals were housed in pairs (Muñoz-Abellán et al., 2011). In brief, ACTH RIA used 125I-ACTH (Perki­ under standard conditions of temperature (21 ± 1 ºC) in a 12:12 h light/ nElmer Life Science, Boston, USA) as the tracer, rat synthetic ACTH1–39 dark schedule (lights on at 07:00 h), with food and water ad libitum. The (Sigma, Barcelona, Spain) as the standard and an antibody raised against experimental protocol was approved by the Committee of Ethics of the rat ACTH (rb7) kindly provided by Dr. W.C. Engeland (Department of Universitat Autònoma de Barcelona and by the Generalitat de Catalunya Surgery, University of Minnesota, Minneapolis, USA). The characteristics and was carried out in accordance to the European Communities Council of the antibody have been described previously (Engeland et al., 1989) and Directive (2010/63/EU) and Spanish legislation (BOE53-2013). we followed a non-equilibrium procedure. Corticosterone RIA used 125 Starting at least two days after their arrival to the facility, all animals I-corticosterone-carboximethyloxime-tyrosine-methylester (ICN-Biolink 2 R. Nadal et al. Psychoneuroendocrinology 125 (2021) 105127 2000, Barcelona, Spain), synthetic corticosterone (Sigma, Barcelona, same pattern was observed with corticosterone, except that corticoste­ Spain) as the standard and an antibody raised in rabbits against cortico­ rone levels after IMO were lower than after SWIM. Finally, prolactin sterone–carboximethyloxime-BSA kindly provided by Dr. G. Makara follows the same pattern as ACTH, although the response to the EP was a (Institute of Experimental Medicine, Budapest, Hungary). The character­ bit lower than that to OF2. Accordingly with the overall response the istics of the antibody and the basic RIA procedure have been described stressors were ordered in Figures as follows: OF1, OF2, EP, SWIM and previously (Zelena et al., 2003) and we followed an equilibrium procedure. IMO. Prolactin was determined by RIA using 125I-prolactin (NEN, Boston, MA, In response to the OF2, no differences were observed between the USA) as the tracer, rat prolactin (rat PRL-RP-3) as the standard and an control group (only exposed to blood-sampling) and the stress group antibody raised against rat prolactin (anti-rPRL-S-9), kindly provided by (previously exposed to OF1, EP and SWIM), suggesting that prior stress Dr. A. F. Parlow (NIDDK National Hormone and Peptide Program, CA, experience did not alter the response to the new novel environment USA). All samples were run in the same assay to avoid inter-assay vari­ (Fig. 3). As expected, exposure to IMO resulted in a HPA response that ability. The intra-assay coefficient of variation was 3.3% for ACTH, 7.8% still persisted 60 min after the termination of IMO (Fig. 4). for corticosterone and 4% for prolactin. The sensitivity of the assay was 25 Correlations of the hormonal responses between the various stressors pg/ml for ACTH, 2 ng/ml for corticosterone and 0.5 ng/ml for prolactin. can be seen in Table 1. A heat-map representing correlations are pre­ Samples were run at least in duplicates. No data lower than the minimum sented as Supplementary Fig. S1. The AUCs of ACTH and corticosterone detection level of the assay was found. responses to IMO are also included to rule out that a ceiling effect could have determined the lack of correlations. The pattern of correlations was 2.4. Statistical analysis similar, although not identical, for the three hormones: moderate to good between stressors of similar intensity and poor between stressors Data were analyzed by means of the Statistical Program for Social greatly differing in intensity. When correlations between ACTH and Sciences (SPSS-IBM for Windows, version 24, Armonk, NY, IBM Cor­ corticosterone were calculated for each particular stressor, they were as poration). Hormonal data were log-transformed to achieve normality follows: OF1 (r = 0.80, p < 0.001), EP (r = 0.62, p < 0.001), OF2 (Shapiro-Wilk). A repeated-measures analysis of variance was used (r = 0.62, p < 0.001), SWIM (r = − 0.25, NS), IMO-post (r = 0.36, (within-subjects factor: time or type of stressor), followed by additional p = 0.02) and IMO-AUCs (r = 0.51, p = 0.003). Correlations between pair-wise comparisons. In other cases, when only two observations were HPA hormones and prolactin for each particular stressors only yielded compared, t-tests for independent o dependent means were performed. significance regarding ACTH and prolactin after IMO (r = 0.38, Pearson coefficient (two-tailed) was used to assess correlations between p = 0.03). the different hormones. The area under the curve (AUC) for each animal Classical studies on individual differences classified animals in two was calculated with Graph Pad Prism 5 (GraphPad Software, La Jolla, or more groups in function of a given variable and studied the conse­ CA, USA). ACTH and corticosterone concentrations (pg/ml and ng/ml, quences on other variables. Although we expected that information respectively) were plotted in the y-axis versus time (minutes) in the x- given by this approach will not essentially differ that derived from the axis. The area is computed connecting a straight line between every set correlations, we tested this classifying rats in low (n = 10), intermediate of adjacent points defining the curve, and calculating the area beneath (n = 12) and high responders (n = 10) for each variable and the two these lines. The criterion for significance was set at p < 0.05. Data are stressors most differing in intensity (OF1 and IMO). The results were in available upon request. accordance with the correlational data and can be seen in the Supple­ mentary Table S1. 3. Results In Exp. 2, ACTH and corticosterone responses to the OF1 (Fig. 5A) were much lower than to SWIM [ACTH: t (36) = 17.9, p < 0.001; In Exp. 1, the repeated-measures ANOVAs revealed significant effect corticosterone: t(36) = 10.1, p < 0.001], and no correlation between of the type of stressor for ACTH [F(4, 124) = 287.6, p < 0.001], corti­ the response to the two stressors was found for ACTH [r(35) = +0.16], costerone [F(4, 124) = 38.2, p < 0.001] and prolactin [F(4, 124) = although for corticosterone it approached significance [r(35) = +0.32, 158.8, p < 0.001]. Further pair-wise comparisons (detailed statistical p = 0.056]. differences can be seen in Fig. 2) showed that the order of intensity of the In Exp. 3, ACTH response to restraint was lower than that to IMO [t stressors in terms of ACTH was OF1 < OF2 = EP < SWIM < IMO. The (19) = 4.8, p < 0.001] and a good correlation was found [r(18) = + Fig. 1. Study design in Experiments (EXP) 1, 2 and 3. Male Sprague-Dawley adult rats were exposed to each stressor on the indicated days (D) and blood sampled (BS) to analyze hormone response. The duration of the stressors was 20 min in EXP1, 15 min in EXP2 and 30 min in EXP3. The stressors were two different open- fields (OF1, OF2), elevated platform (EP), forced swim (SWIM), restraint in tubes (RE­ STRAINT) and immobilization on boards (IMO). Additional details in the Materials and methods Section. 3 R. Nadal et al. Psychoneuroendocrinology 125 (2021) 105127 Fig. 3. Comparison of the neuroendocrine response to the second open-field (OF2) in stress-naïve (control, n = 10) and previously stressed (n = 32) rats. Means and SEM of plasma levels of hormones are represented (ACTH in (A), Fig. 2. Neuroendocrine response to different stressors (20 min exposure). corticosterone in (B) and prolactin in (C)). No significant group differences Means and SEM (n = 32) of plasma levels of ACTH (A), corticosterone (B) and were found. prolactin (C) are represented. Bars with different letters are statistically different. Horizontal lines indicates average basal values obtained in controls Corticosterone followed a similar pattern, but, surprisingly, corticoste­ rats sampling in parallel. The stressors used were open-field-1 (OF1), open-field rone levels just after the stressor were higher after SWIM than after IMO 2 (OF-2), elevated platform (EP), forced swim (SWIM) and immobiliza­ tion (IMO). despite lower ACTH response to the former stressor. We expected similar corticosterone levels after SWIM and IMO, as the ACTH levels achieved with both were enough to saturate the adrenal 0.64, p = 0.002] between the two stressors (Fig. 5B). Corticosterone cortex (Keller-Wood et al., 1983). We have no clear explanation for this response to the two stressors was similar, with a good correlation be­ result that has nevertheless been replicated in another study from our tween the stressors [r(18) = +0.54, p = 0.013]. lab (unpublished data). Interestingly, whereas significant positive cor­ relations between ACTH and corticosterone were found with all 4. Discussion stressors, including IMO (particularly AUCs), a non-significant (nega­ tive) correlation was found after SWIM, suggesting that some factor The present data indicate that we cannot characterize trait-like stress specifically associated with this stressor modulates the adrenal response hypo- or hyper-responsive subjects in a normal population of adult male to ACTH. SWIM at 36 ◦ C did not change body temperature so that a rats in terms of HPA and prolactin response. Instead, individual differ­ putative factor could be muscular activity associated to swim, which ences in responsiveness appear to be markedly dependent on the in­ might modulate the adrenal responsiveness to ACTH. We are not aware tensity of the stressor chosen to evaluate them. of any experimental data directly supporting this possibility, but chronic When the same animals were exposed to different types of stressors, voluntary running wheel exercise in male rats has been found to alter all of them having a strong emotional component, we observed (as ex­ corticosterone response to acute stressors independently of ACTH levels pected), that the lowest ACTH and prolactin response corresponded to (Droste et al., 2007). Interestingly, the direction of the changes in the small OF1 and the highest one to IMO. The hormonal response to the stress-induced corticosterone in exercise rats with respect to controls EP was similar to that of the OF2 for ACTH and corticosterone, but a bit were opposite in response to novel environment and to forced swim, higher for prolactin, although the hormonal responses to the EP and OF2 suggesting that the modulatory role of extra-ACTH factors were were clearly lower than SWIM. ACTH and prolactin give rise to a quite dependent of the type of stressor. Evidence for an extra-ACTH regulation similar order of classification of the stressors, thus supporting previous of the adrenal cortex has been accumulating over the years since the studies demonstrating that the two hormones have been found to be pioneering research of Dallman`s laboratory with the study of the good markers of such an intensity (Armario et al. 2012, 2020). circadian rhythm of HPA hormones (Dallman et al., 1978; Engeland 4 R. Nadal et al. Psychoneuroendocrinology 125 (2021) 105127 exposed the animals first to presumably low intensity stressors (OF1, EP), then to SWIM and finally to the most severe stressor (IMO). We also included in the study two different novel environments, OF1 and OF2, which represent qualitatively similar situations. As exposure to the OF2 was done after exposure to SWIM, we wanted to rule out that such response was altered by prior experience of the rats with the other stressors by comparing stress-experienced rats with a group of control (stress-naive) rats. No differences between the groups were observed, indicating that the response to the OF2 was not altered by the prior history of stress. Behavioral response to the OF2 was also not affected by prior stress experience (not shown). On the basis of the overall endocrine data, we can assume that the ACTH and prolactin responses, and corticosterone with some limita­ tions, appear to similarly classify the predominantly emotional stressors used in the present study in terms of intensity. We are aware of the difficulty of objectively determining the intensity of an emotional stressor and consequently a high degree of circularity is involved when using the biological response to assign a level of intensity to a stressor. However, we consider that this is a reasonable approach considering that all parameters that has been found to be sensitive to intensity change show always the same pattern in an important number of experimental studies (see Armario et al., 2020). Regarding individual consistency across stressors, with plasma ACTH levels, the highest correlations were between the OF1 and the OF2, and between OF2 and EP, whereas that between the OF1 and the EP was statistically significant but lower. The OF1 and the OF2 represent qualitatively similar situations involving novelty and potential danger as Fig. 4. The dynamics of the response to IMO as compared to that of unstressed well as free active and exploratory behavior. Nevertheless, the higher animals. Means and SEM of plasma levels of hormones are represented (control n = 10; IMO n = 32). Samples were taken just after 20 min IMO (post) and at response to the OF2 suggests that the latter was more aversive than the 30 and 60 min after the termination of IMO (R30 and R60, respectively). Basal OF1, probably because the latter was smaller and done under dim light values are the average of various previous sampling days (control group). conditions. The EP represents a potential risk for predation and falling Differences between control and IMO rats were always p < 0.001 and are out, and there is no possibility for free activity of the animals. Therefore, not indicated. the EP has important differences with the OF2, but both were of similar intensity. These data tentatively suggest that being of similar magnitude et al., 1977) and has also been found comparing inbred rat strains and sharing certain characteristics are important to elicit a similar ACTH (Gómez and Lahmame, 1996). It is unlikely that the lower ACTH levels response. This hypothesis is supported by the correlations that included after SWIM than IMO reflects a greater negative glucocorticoid feedback the two most severe stressors: (i) correlations of the low intensity rather than a lower intensity of the former stressor. When comparing stressors with SWIM and IMO were low and non-significant; (ii) corre­ both, the impact of IMO in terms of the post-stress recovery time, impact lations between the latter two stressors was moderate. Correlations of on food intake in the next 24 h and heterotypic HPA sensitization is corticosterone did not parallel ACTH and were better between stressors stronger (Belda et al., 2016; Rabasa et al., 2015). of intermediate intensity (EP, OF2, SWIM), which poorly correlated with To avoid carry over effects of prior exposure to other stressors, we those of low intensity (OF1) or high intensity (IMO). When the corre­ lation of the AUCs of the ACTH and corticosterone response to IMO with Table 1 Correlations of the response of ACTH, corticosterone and prolactin to the different stressors. Stressor Hormone OF1 OF2 EP SWIM IMOpost IMO-AUC OF1 ACTH – +0.65*** +0.42* +0.22 +0.25 − 0.02 Corticosterone – +0.32# +0.14 − 0.02 − 0.10 − 0.07 Prolactin – +0.52** +0.38* +0.13 +0.01 NA OF2 ACTH – +0.64*** +0.27 +0.15 +0.02 Corticosterone – +0.65*** +0.45*** +0.28 +0.34# Prolactin – +0.53** +0.62** +0.29 NA EP ACTH – +0.18 +0.14 +0.18 Corticosterone – +0.56** +0.21 +0.16 Prolactin – +0.51*** +0.25 NA SWIM ACTH – +0.33# − 0.02 Corticosterone – +0.69*** +0.68*** Prolactin – +0.58*** NA IMOpost ACTH – +0.56*** Corticosterone – +0.85*** Prolactin – NA Pearson coefficient correlations (two-tailed) are indicated (n = 32). ACTH: adrenocorticotropic hormone; AUC: area under the curve; EP: elevated platform; IMO: immobilization; OF1: open field 1; OF2: open-field 2. * At least p < 0.05. ** At least p < 0.01. *** At least p < 0.001. # Marginal; NA: non-applicable. 5 R. Nadal et al. Psychoneuroendocrinology 125 (2021) 105127 Fig. 5. Comparison of the neuroendocrine response of the same rats to two different stressors. Means and SEM of plasma levels of hormones are represented: in (A) (n = 37) the ACTH and corticosterone responses to an open-field (OF1) and SWIM; in (B) (n = 20) the ACTH and corticosterone responses to restraint and immobilization (IMO). Correlations between the two stressors are also indicated. ***p < 0.001 versus the other stressor. the other stressors was introduced, no improvement was found, sug­ HAB rats) did not show consistent differences in the HPA response to gesting that is a characteristic of the stressor rather than a problem of stressors as they are markedly dependent on the type of stressor (Liebsch catching the integrated response to the situation. Prolactin followed the et al. 1998; Landgraf et al. 1999; Frank et al. 2006). More consistent same pattern as ACTH, although correlations were in general higher response to different types of stressors are expected if animals have been than those of ACTH. Therefore, the overall results suggest that the genetically selected on the basis of their neuroendocrine response to endocrine response to stressor markedly differing in intensity poorly stressors. This has been done regarding corticosterone response to re­ correlates and this is not restricted to a particular endocrine system as straint stress in mice that resulted in low, intermediate and high were obtained with ACTH and prolactin, two independent neuroendo­ responsive lines (Touma et al., 2008). However, only two stressors have crine system. been tested (restraint vs open field plus forced swim), which showed a To demonstrate that the results can be consistent across studies we similar pattern (Mattos et al., 2013). Interestingly, in the mouse lines calculated, in two independent experiments, the correlation of the discussed above (Knapman et al., 2010; Touma et al., 2008) as well as in response of HPA hormones after exposure to OF1 vs SWIM (two days rat lines also genetically selected for the corticosterone response to apart), and after exposure to restraint vs IMO (12 days apart). The stressors (Walker et al., 2017; Walker and Sandi, 2018), differences in conclusions were similar to the main experiment: no correlation be­ coping as well as aggressiveness were also observed, although the di­ tween stressors greatly differing in intensity (OF1 vs SWIM) and good rection of the changes were not concordant. correlation between those close in intensity (restraint vs IMO). This The present data bear important implications regarding character­ consistency was found despite certain differences in the duration of ization of putative phenotypes differing in responsiveness to stress in stressors between the experiments: 20 min (Exp. 1), 15 min (Exp. 2) and normal populations of animals and humans in that they indicate that we 30 min (Exp. 3). Therefore, minor differences in the procedure cannot cannot identify individuals characterized by a generalized hypo or affect the conclusions. hyper-responsiveness to stressors. The results clearly illustrate the We have previously reported good correlations between the response importance of paying attention to the intensity of stressors, but they do of HPA hormones and prolactin across different novel environments (e. not rule out that behavioral traits such as coping style could interact g. elevated plus-maze, hole-board), but in those studies the endocrine with qualitative aspects of stressors (e.g. controllability and unpredict­ response was very similar in all environments (Márquez et al., 2005; ability) to determine the biological response (Koolhaas et al., 2010). In 2006). Therefore, the latter results are compatible with the suggestion this regard, availability of coping (presence of bedding material allow­ that good correlations are observed when the situations are similar in ing burying) has been found to alter the plasma corticosterone and magnitude and also qualitatively. catecholamine response to an electrified prod in the cage (De Boer et al., To our knowledge, there is no previous similar study in normal 1990). It would be extremely interesting to study how coping style in­ populations of rats, but some results obtained in rats genetically selected teracts with certain characteristics of stressors. for anxiety give support to the present conclusions. For instance, in rats Although considerable attention has been paid in human research to genetically selected for low or high avoidance in a shuttle-box, pre­ study reliability of the biological response to stressors, most of it has sumably related to high vs low anxiety, respectively, differences in HPA focused on repetition of the same situation/task or tasks of similar in­ and prolactin response has been observed but only in response to low tensity (see for instance Carroll et al., 1984; Parati et al., 1988; Kirsch­ intensity but not high intensity stressors (Gentsch et al., 1982). Also, rat baum et al., 1995; Cohen et al., 2000; Hawkley et al., 2001; Burleson lines selected for anxiety behavior in the elevated plus maze (LAB and et al., 2003; Henze et al., 2017; Bachmann et al., 2019), and studies 6 R. Nadal et al. Psychoneuroendocrinology 125 (2021) 105127 specifically comparing stressor markedly differing in intensity are Armario, A., Labad, J., Nadal, R., 2020. Focusing attention on biological markers of acute stressor intensity: empirical evidence and limitations. Neurosci. Biobehav. Rev. 111, lacking. This is of major relevance when we want to know whether the 95–103. endocrine response to laboratory stressors, usually of low intensity, are Bachmann, P., Finke, J.B., Rebeck, D., Zhang, X., Larra, M.F., Koch, K.P., Turner, J.D., related to the vulnerability to severe stressors (i.e. development of Schäschinger, H., 2019. Test-retest reproducibility of a combined physical and post-traumatic stress disorder). Although we are aware of ethical con­ cognitive stressor. Biol. Psychol. 148, 107729. Belda, X., Márquez, C., Armario, A., 2004. Long-term effects of a single exposure to stress cerns, information regarding the response to real-life (non-provoked), in adult rats on behavior and hypothalamic-pituitary-adrenal responsiveness: relatively severe stressors can contribute to shed lights to this issue in comparison of two outbred rat strains. Behav. Brain Res. 154 (2), 399–408. humans. Belda, X., Nadal, R., Armario, A., 2016. Critical features of acute stress-induced cross- sensitization identified through the hypothalamic-pituitary-adrenal axis output. Sci. In conclusion, the present data indicate that testing individual dif­ Rep. 6, 31244. ferences in neuroendocrine responsiveness to a particular, predomi­ Burleson, M.H., Poehlmann, K.M., Hawkley, L.C., Ernst, J.M., Berntson, G.G., nantly emotional, stressor does not predict individual differences in the Malarkey, W.B., Kiecolt-Glaser, J.K., Glaser, R., Cacioppo, J.T., 2003. Neuroendocrine and cardiovascular reactivity to stress in mid-aged and older response to other emotional stressors markedly differing in intensity. women: long-term temporal consistency of individual differences. Psychophysiology Therefore, we cannot identify individuals characterized by a consistent 40 (3), 358–369. and generalized hypo- or hyper-responsiveness to any type of emotional Carroll, D., Turner, J.R., Lee, H.J., Stephenson, J., 1984. Temporal consistency of individual differences in cardiac response to a video game. Biol. Psychol. 19 (2), stress on the basis of results obtained with a particular situation. In­ 81–93. tensity, rather than qualitative aspects of stressors, appear to be critical Cavigelli, S.A., Ragan, C.M., Michael, K.C., Kovacsics, C.E., Bruscke, A.P., 2009. Stable to detect consistent individual differences. Clearly studies in humans behavioral inhibition and glucocorticoid production as predictors of longevity. Physiol. Behav. 98 (1–2), 205–214. dealing with this problem are needed. Charmandari, E., Tsigos, C., Chrousos, G., 2005. Endocrinology of the stress response. Annu Rev. Physiol. 67, 259–284. Ethical statement Cohen, S., Hamrick, N., Rodriguez, M.S., Feldman, P.J., Rabin, B.S., Manuck, S.B., 2000. The stability of and intercorrelations among cardiovascular, immune, endocrine, and psychological reactivity. Ann. Behav. Med. 22 (3), 171–179. The experimental protocol was approved by the Committee of Ethics Dallman, M.F., Engeland, W.C., Rose, J.C., Wilkinson, C.W., Shinsako, J., Siedenburg, F., of the Universitat Autònoma de Barcelona and by the Generalitat de 1978. Nycthemeral rhythm in adrenal responsiveness to ACTH. Am. J. Physiol. 235 (5), R210–R218. Catalunya and was carried out in accordance to the European Commu­ De Boer, S.F., Slangen, J.L., Van der Gugten, J., 1990. Plasma catecholamine and nities Council Directive (2010/63/EU) and Spanish legislation (BOE53- corticosterone levels during active and passive shock-prod avoidance behavior in 2013). rats: effects of chlordiazepoxide. Physiol. Behav. 47 (6), 1089–1098. Droste, S.K., Chandramohan, Y., Hill, L.E., Linhorst, A.C.E., Reul, J.M.H., 2007. Voluntary exercise impacts on the rat hypothalamic-pituitary-adrenocortical axis CRediT authorship contribution statement mainly at the adrenal level. Neuroendocrinology 86 (1), 26–32. Engeland, W.C., Shinsako, J., Winget, C.M., Vernikos-Danellis, J., Dallman, M.F., 1977. Circadian patterns of stress-induced ACTH secretion are modified by corticosterone AA obtained funding, designed the experiments, analyzed and responses. Endocrinology 100 (1), 138–147. interpreted the data and wrote the draft of the manuscript; RN obtained Engeland, W.C., Miller, P., Gann, D.S., 1989. Dissociation between changes in plasma funding, contributed to the design, performed statistical analysis and bioactive and immunoreactive adrenocorticotropin after hemorrhage in awake dogs. Endocrinology 124 (6), 2978–2985. participated in the elaboration of the manuscript; MG-S did the great Frank, E., Salchner, P., Aldag, J.M., Salomé, N., Singewald, N., Landgraf, R., Wigger, A., part of the experimental work of the main experiment in collaboration 2006. Genetic predisposition to anxiety-related behavior determines coping style, with MS-O; HG and XB did the complementary experiments; MG-S, HG neuroendocrine responses, and neuronal activation during social defeat. Behav. Neurosci. 120 (1), 60–71. and XB did the hormone analyses. García, A., Martí, O., Vallès, A., Dal-Zotto, S., Armario, A., 2000. Recovery of the hypothalamic-pituitary-adrenal response to stress. Effect of stress intensity, stress duration and previous stress exposure. Neuroendocrinology 72 (2), 114–125. Declaration of Competing Interest Gentsch, C., Lichtsteiner, M., Driscoll, P., Feer, H., 1982. Differential hormonal and physiological responses to stress in Roman high- and low-avoidance rats. Physiol. Behav. 28 (2), 259–263. None. Gómez, F., Lahmame, A., De Kloet, E.R., Armario, A., 1996. Hypothalamic-pituitary- adrenal response to chronic stress in five inbred rat strains: differential responses are mainly located at the adrenocortical level. Neuroendocrinology 63 (4), 327–337. Acknowledgements Halpern, C.T., Campbell, B., Agnew, C.R., Thompson, V., Udry, J.R., 2002. Associations between stress reactivity and sexual and nonsexual risk taking in young adult human The laboratory was supported by Spanish grants to AA and/or RN males. Horm. Behav. 42 (4), 387–398. Hankin, B.L., Badanes, L.S., Smolen, A., Young, J.F., 2015. Cortisol reactivity to stress from Ministerio de Ciencia e Innovación (SAF2017-83430-R) and Gen­ among youth: stability over time and genetic variants for stress sensitivity. eralitat de Catalunya (SGR2017-00457). RN was a recipient of an ICREA J. Abnorm. Psychol. 124 (1), 54–67. Academia recognition (Generalitat de Catalunya, 2015-19). Our group is Hawkley, L.C., Burleson, M.H., Poehlmann, K.M., Berntson, G.G., Malarkey, W.B., Cacioppo, J.T., 2001. Cardiovascular and endocrine reactivity in older females: part of the CIBERSAM network, Instituto de Salud Carlos III (CB19/09/ intertask consistency. Psychophysiology 38 (6), 863–872. 00029). The Universitat Autònoma de Barcelona received funding from Henze, G.I., Zänkert, S., Urschler, D.F., Hiltl, T.J., Kudielka, B.M., Pruessner, J.C., the Generalitat de Catalunya (2015 FEDER7S-20IU16-0019) for the Wüst, S., 2017. Testing the ecological validity of the Trier Social Stress Test: association with real-life exam stress. Psychoneuroendocrinology 75, 52–55. construction of the Animal Facility. Keller-Wood, M.E., Shinsako, J., Dallman, M.F., 1983. Integral as well as proportional adrenal responses to ACTH. Am. J. Physiol. 245 (1), R53–R59. Kirschbaum, C., Prüssner, J.C., Stone, A.A., Federenko, I., Gaab, J., Lintz, D., Appendix A. Supporting information Schommer, N., Hellhammer, D.H., 1995. Persistent high cortisol responses to repeated psychological stress in a subpopulation of healthy men. Psychosom. Med. Supplementary data associated with this article can be found in the 57 (5), 468–474. online version at doi:10.1016/j.psyneuen.2021.105127. Knapman, A., Heinzmann, J.M., Hellweg, R., Holsboer, F., Landgraf, R., Touma, C., 2010. Increased stress reactivity is associated with cognitive deficits and decreased hippocampal brain-derived neurotrophic factor in a mouse model of affective References disorders. J. Psychiatr. Res. 44 (9), 566–575. Koolhaas, J.M., de Boer, S.F., Coppens, C.M., Buwalda, B., 2010. Neuroendocrinology of coping styles: towards understanding the biology of individual variation. Front. Armario, A., Nadal, R., 2013. Individual differences and the characterization of animal Neuroendocrinol. 31 (3), 307–321. models of psychopathology: a strong challenge and a good opportunity. Front. Landgraf, R., Wigger, A., Holsboer, F., Neumann, I.D., 1999. Hyper-reactive Pharm. 4, 137. hypothalamo-pituitary-adrenocortical axis in rats bred for high anxiety-related Armario, A., Daviu, N., Muñoz-Abellán, C., Rabasa, C., Fuentes, S., Belda, X., behaviour. J. Neuroendocrinol. 11 (6), 405–407. Gagliano, H., Nadal, R., 2012. What can we know from pituitary-adrenal hormones Liebsch, G., Linthorst, A.C., Neumann, I.D., Reul, J.M., Holsboer, F., Landgraf, R., 1998. about the nature and consequences of exposure to emotional stressors? Cell Mol. Behavioral, physiological, and neuroendocrine stress responses and differential Neurobiol. 32 (5), 749–758. 7 R. Nadal et al. Psychoneuroendocrinology 125 (2021) 105127 sensitivity to diazepam in two Wistar rat lines selectively bred for high- and low- Rabasa, C., Gagliano, H., Pastor-Ciurana, J., Fuentes, S., Belda, X., Nadal, R., Armario, A., anxiety-related behavior. Neuropsychopharmacology 19 (5), 381–396. 2015. Adaptation of the hypothalamus-pituitary-adrenal axis to daily repeated stress Márquez, C., Belda, X., Armario, A., 2002. Post-stress recovery of pituitary-adrenal does not follow the rules of habituation: a new perspective. Neurosci. Biobehav. Rev. hormones and glucose, but not the response during exposure to the stressor, is a 56, 35–49. marker of stress intensity in highly stressful situations. Brain Res. 926 (1–2), Rotllant, D., Nadal, R., Armario, A., 2007. Differential effects of stress and amphetamine 181–185. administration on Fos-like protein expression in corticotropin releasing factor- Márquez, C., Nadal, R., Armario, A., 2005. Responsiveness of the hypothalamic-pituitary- neurons of the rat brain. Dev. Neurobiol. 67 (6), 702–714, 702-214. adrenal axis to different novel environments is a consistent individual trait in adult Taylor, J., Weyers, P., Harris, N., Vogel, W.H., 1989. The plasma catecholamine stress male outbred rats. Psychoneuroendocrinology 30 (2), 179–187. response is characteristic for a given animal over a one-year period. Physiol. Behav. Márquez, C., Nadal, R., Armario, A., 2006. Influence of reactivity to novelty and anxiety 46 (5), 853–856. on hypothalamic-pituitary-adrenal and prolactin responses to two different novel Touma, C., Bunck, M., Glasl, L., Nussbaumer, M., Palme, R., Stein, H., Wolferstätter, M., environments in adult male rats. Behav. Brain Res. 168 (1), 13–22. Zeh, R., Zimbelmann, M., Holsboer, F., Landgraf, R., 2008. Mice selected for high Mattos, G.E., Heinzmann, J.M., Norkowski, S., Helbling, J.C., Minni, A.M., Moisan, M.P., versus low stress reactivity: a new animal model for affective disorders. Touma, C., 2013. Corticosteroid-binding globulin contributes to the neuroendocrine Psychoneuroendocrinology 33 (6), 839–862. phenotype of mice selected for extremes in stress reactivity. J. Endocrinol. 219 (3), Vahl, T.P., Ulrich-Lai, Y.M., Ostrander, M.M., Dolgas, C.M., Elfers, E.E., Seeley, R.J., 217–229. D’Alessio, D.A., Herman, J.P., 2005. Comparative analysis of ACTH and Muñoz-Abellán, C., Rabasa, C., Daviu, N., Nadal, R., Armario, A., 2011. Behavioral and corticosterone sampling methods in rats. Am. J. Physiol. Endocrinol. Metab. 289 (5), endocrine consequences of simultaneous exposure to two different stressors in rats: E823–E828. interaction or independence? PLoS One 6 (6), e21426. Walker, S.E., Sandi, C., 2018. Long-term programing of psychopathology-like behaviors Pace, T.W.W., Gaylord, R., Topczewski, F., Girotti, M., Rubin, B., Spencer, R.L., 2005. in male rats by peripubertal stress depends on individual’s glucocorticoid Immediate-early gene induction in hippocampus and cortex as a result of novel responsiveness to stress. Stress 21 (5), 433–442. experience is not directly related to the stressfulness of that experience. Eur. J. Walker, S.E., Zanoletti, O., Guillot de Suduiraut, I., Sandi, C., 2017. Constitutive Neurosci. 22 (7), 1679–1690. differences in glucocorticoid responsiveness to stress are related to variation in Parati, G., Pomidossi, G., Casadei, R., Ravogli, A., Groppelli, A., Cesana, B., Mancia, G., aggression and anxiety-related behaviors. Psychoneuroendocrinology 81, 1–10. 1988. Comparison of the cardiovascular effects of different laboratory stressors and Zelena, D., Mergl, Z., Foldes, A., Kovács, K.J., Tóth, Z., Makara, G.B., 2003. Role of their relationship with blood pressure variability. J. Hypertens. 6 (6), 481–488. hypothalamic inputs in maintaining pituitary-adrenal responsiveness in repeated restraint. Am. J. Physiol. Endocrinol. Metab. 285 (5), E1110–E1117. 8