3-Iodothyronamine Induces Tail Vasodilation Through Central Action in Male Mice

Abstract

3-Iodothyronamine (3-T1AM) is an endogenous thyroid hormone (TH)–derived metabolite that induces severe hypothermia in mice after systemic administration; however, the underlying mechanisms have remained enigmatic. We show here that the rapid 3-T1AM–induced loss in body temperature is a consequence of peripheral vasodilation and subsequent heat loss (e.g., over the tail surface). The condition is subsequently intensified by hypomotility and a lack of brown adipose tissue activation. Although the possible 3-T1AM targets trace amine-associated receptor 1 or α2a-adrenergic receptor were detected in tail artery and aorta respectively, myograph studies did not show any direct effect of 3-T1AM on vasodilation, suggesting that its actions are likely indirect. Intracerebroventricular application of 3-T1AM, however, replicated the phenotype of tail vasodilation and body temperature decline and led to neuronal activation in the hypothalamus, suggesting that the metabolite causes tail vasodilation through a hypothalamic signaling pathway. Consequently, the 3-T1AM response constitutes anapyrexia rather than hypothermia and closely resembles the heat-stress response mediated by hypothalamic temperature-sensitive neurons. Our results thus underline the well-known role of the hypothalamus as the body’s thermostat and suggest an additional molecular link between TH signaling and the central control of body temperature.

Thyroid hormones (THs) are well-known for their metabolic actions that increase body temperature (1, 2). Increased levels of THs as in hyperthyroid patients are associated with elevated energy expenditure, tachycardia, and heat sensitivity, whereas the opposite effects are seen in hypothyroidism, a disease characterized by lower TH levels.

In contrast to these thermogenic effects of THs, 3-iodothyronamine (3-T1AM), a decarboxylated and deiodinated form of TH, has been reported to induce rapid hypothermia in mice (3) and hamsters (4), as determined by rectal measurements. 3-T1AM administration also alters metabolic function causing hyperglycemia and hypophagia (5). In addition, it increases learning and memory (6), suggesting that it can impact the brain.

Although it seems obvious that 3-T1AM is derived from TH, its biosynthesis is not yet clear. However, different pathways and enzymes are likely to be involved, such as deiodinases (Dio1, Dio2, and Dio3) and intestinal ornithine decarboxylase (7). 3-T1AM can subsequently also be converted to other TH derivatives such as 3-iodothyroacetic acid, which has no thermoregulatory functions, or the iodine-free thyronamine (3-T0AM), which has less pronounced cardiac or thermoregulatory effects (8, 9).

Likewise, little is known about the molecular targets of 3-T1AM action. It has been proposed that 3-T1AM can interact with two G protein–coupled receptors, trace amino-associated receptor (TAAR1) and α2a-adrenergic receptor (3, 10). However, the hypothermic effect of 3-T1AM does not seem to be mediated by TAAR1 because it persists in TAAR1 knockout mice (11). Therefore, it remains enigmatic how 3-T1AM modulates body temperature regulation.

Because recent findings in mice suggest that body temperature regulation by THs depends on intact tail vasoconstriction to regulate heat loss (12), we speculated that 3-T1AM might also act on tail vasodilation to induce hypothermia. In this study, we show that in mice, intraperitoneal (IP) administration of 3-T1AM leads to an initial short warming flush in the tail, followed by a substantial decrease in interscapular brown adipose tissue (iBAT) and body temperature. We did not find any direct effects of 3-T1AM on tail artery or aorta, but observed a similar tail heat loss effect after intracerebroventricular (ICV) injection. Together with the observed c-FOS activation in the hypothalamus, our findings suggest that, through centrally mediated vasodilation, 3-T1AM induces a condition that should be classified as anapyrexia (centrally coordinated downregulation of body temperature) rather than hypothermia (13).

Materials and Methods

Animal husbandry

C57BL/6 wild-type male mice at the age of 2 to 5 months were single-housed at 22°C on a 12-hour light/12-hour dark cycle for 1 week before starting the experiments. During the entirety of the study, they had ad libitum access to standard-diet chow and water. All animal procedures were approved by the University of Santiago de Compostela, the Djurförsöksethiska Nämnd in Stockholm, or the Ministerium für Energiewende, Landwirtschaft, Umwelt und ländliche Räume (University of Lübeck).

Infrared thermography

Infrared (IR) cameras [(Compact-IR-Thermal-Imaging-Camera FLIR T355; Termisk Systemteknik, Linköping, Sweden) or (Thermografiesystem VarioCAM hr, head 680s #730 mm; InfraTec, Dresden, Germany)] were used for detecting surface temperatures of mice tails, backs, inner ears, and iBATs. All images were taken while animals were moving freely in the cage at room temperature and analyzed using FLIR Systems or IRBIS 3 plus software. Several pictures were taken for analyses and the maximum temperatures in the selected area of the images were chosen. For the IR videos, mice were placed in the recording cage, monitored for baseline values for 10 minutes, then injected with 3-T1AM (Santa Cruz sc-209626, 98% purity, 50 mg/kg IP in 60% dimethyl sulfoxide (DMSO) in phosphate-buffered saline (PBS), 4 μL/g body weight) or the same volume of solvent, and recorded for 60 minutes with one picture per second.

Radiotelemetry

Implantable radio transmitters and receiver plates (Mini Mitter Respironics; Hugo-Sachs Elektronic, March-Hugstetten, Germany) were used to measure heart rate and body temperature in awake and freely moving animals (1, 14). 3-T1AM was dissolved in 30% ethanol in PBS. Mice received a single IP injection of 3-T1AM (50 mg/kg) or solvent only.

Contractile response studies using a myograph

A wire myograph (520A-DMT; AD Instruments, Oxford, UK) was used to measure the contractility of vessels as described previously (12). Short pieces of aorta or tail artery were dissected from wild-type male mice and mounted into the wire myograph using stainless steel wires (40 µm diameter). The chamber of the wire myograph was filled with Krebs-Ringer buffer (123 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgCl₂, 20 mM NaHCO₃, 2.5 mM CaCl₂, 5.5 mM glucose) and bubbled with carbogen gas to reach pH 7.4 at 37°C. Vessels were allowed to recover for 1 hour; then, using increasing forces and recording the corresponding contraction, the optimal tension was determined. Vessels were prestimulated with either 1 µM T1AM or 100 nM T3 (Sigma Aldrich, Darmstadt, Germany) for 5 minutes, which were both dissolved in DMSO (subsequently diluted in Krebs-Ringer buffer; final DMSO concentration in chamber: 0.01%). To perform a dose-response curve, increasing doses of phenylephrine [(PE) Sigma Aldrich] (10⁻⁸ to 10⁻² M) were added in 3-minute intervals into the chamber. The induced force was detected and normalized by the previously recorded KCl stimulation force using Labchart software 8.1.

ICV cannulation

ICV cannulation was performed under ketamine-xylazine (50 mg/kg, IP injection) or isoflurane anesthesia as described previously (15,–18). After cannula implantation, the animals were allowed to recover for at least 3 days. 3-T1AM was dissolved in 60% DMSO in PBS, and 40, 80, or 200 ng were injected in 2-μL volume. Controls received the same volume of vehicle.

Western blot

Western blot analysis was performed as described previously (15,–18). All used antibodies are listed in Table 1.

Table 1.

Antibody Table

Peptide/Protein Target	Name of Antibody	Manufacturer, Catalog #, and/or Name of Individual Providing the Antibody	Species Raised in; Monoclonal or Polyclonal	Dilution Used	RRID
Primary antibody
Acetyl-CoA-carboxylase	Antiacetyl-CoA-carboxylase	Cell Signaling Technology, #3676	Rabbit; monoclonal	1:1000	AB_2219397
Phospho-acetyl-CoA carboxylase	Anti-phospho-acetyl-CoA carboxylase	Cell Signaling Technology, #3661	Rabbit; polyclonal	1:1000	AB_330337
Phosphorylated AMPK	Anti-AMPK	Cell Signaling Technology, #2535	Rabbit; monoclonal	1:2000	AB_331250
Phosphorylated AMPK α1	Anti-AMPK α1	Millipore, #07-350	Rabbit; polyclonal	1:1000	AB_310542
Phosphorylated AMPK α2	Anti-AMPK α2	Millipore, #07-363	Rabbit; polyclonal	1:1000	AB_310553
Fatty acid synthase	Anti-FAS	BD Biosciences, #610962	Mouse; monoclonal	1:1000	AB_398275
B-actin	Anti-B-actin	Sigma, #A5316	Mouse; monoclonal	1:5000	AB_476743
c-FOS (Western blot)	c-FOS	Santa Cruz, #sc-52	Rabbit; polyclonal	1:1000	AB_2106783
c-FOS (immunohistochemistry staining)	c-FOS	Tocris Bioscience, # 2066	Rabbit; polyclonal	1:5000	AB_10285944
Secondary antibody
Rabbit immunoglobulins (Western blot)	Rabbit immunoglobulin	Dako, #P0448	Goat; polyclonal	1:5000	AB_2617138
Mouse immunoglobulins (Western blot)	Mouse immunoglobulin	Dako, #P0260	Rabbit; polyclonal	1:5000	AB_2636929
Biotinylated antirabbit (immunohistochemistry staining)	Biotinylated antirabbit	Vector Laboratories, #BA1000	Goat	1:250	AB_2313606

Abbreviations: AMPK, adenosine 5′-monophosphate kinase; RRID, research resource identifier.

Quantitative real-time polymerase chain reaction

RNA isolation was performed using RNeasy Lipid Tissue Mini Kit (QIAGEN, Hilden, Germany) for iBAT as well as soleus muscle and RNeasy Mini Kit (QIAGEN) for liver samples according to the manufacturer’s instructions. All samples were from snap-frozen tissues. For complementary DNA synthesis, the Molecular Biology RevertAid strand cDNA Kit (Thermo Fisher Scientific, Darmstadt, Germany) was used according to the manufacturer’s instructions. Quantitative polymerase chain reaction (PCR) was performed using Quantstudio Applied Biosystems (Thermo Fisher Scientific) and SYBR green PCR master mix (Roche, Mannheim, Germany). Primer sequences are available on request. Hprt was used as housekeeping gene. Standard curves were used to correct for PCR efficiency.

Glycogen determination

Glycogen measurement in liver tissue was performed as previously described (19). For gastrocnemius muscle, 150 mg of tissue was homogenized and 300 µL of supernatant was taken for precipitation.

Immunohistochemistry

The mice were euthanized 90 minutes after the 3-T1AM IP injection and their brains removed and fixed in 4% paraformaldehyde overnight followed by 2 days in 30% sucrose in PBS and frozen at −80°C. Immunohistochemistry was performed as described previously (20). The antibodies used were anti-c-FOS (Tocris Bioscience, Bristol, UK, #2066, 1:500) and biotinylated antirabbit (Vector Laboratories, Peterborough, UK, BA1000, 1:250).

Statistical analysis

GraphPad Prism 5 software (GraphPad Software, Inc., La Jolla, CA) was used to analyze the data. The results are shown as mean ± standard error of the mean (SEM). Statistical significance was determined using an unpaired two-tailed Student t test for simple comparisons and corrected for multiple comparisons where appropriate, two-way analysis of variance with repeated measurements for the time course studies and nonlinear fit models for dose-response curves. Statistical significance was defined as *P < 0.05, **P < 0.01, and ***P < 0.001.

Result

Hypothermia effect of 3-T1AM

A single IP injection of 50 mg/kg 3-T1AM has been reported to induce a rapid loss of about 9°C in body temperature within minutes after injection, accompanied by bradycardia (3). Using radiotelemetry and a different solvent, we also observed the same effect; however, with a 4°C difference, the effect was smaller [Fig. 1(a); Supplemental Fig. 1A].

Figure 1.

Effect of 3-T1AM IP injection on body temperature. (a) 3-T1AM lowers body temperature after 50 mg/kg IP injection as analyzed by radiotelemetry in conscious and freely moving mice (n = 4 per group). (b) Selected images from IR videography of control and 3-T1AM–treated mice before (top), shortly after (middle), and 60 minutes after IP 50 mg/kg injection of 3-T1AM (bottom). The arrow indicates the heat flush in the tail of 3-T1AM–treated mice, which was not observed in controls. (c–f) Quantification of the changes in back, iBAT, and tail temperatures from representative IR images during the course of the experiment as well as iBAT/back ratio. The values depict mean ± SEM of 3 mice, which were analyzed in a paired crossover study. Statistical significance was tested with a t test corrected for repeated measurements with the Holm-Sidak method and considered significant at ***P < 0.001.

To study the hypothermia effect in greater detail, we used IR videography, in which we could nicely visualize the cooling effect after 3-T1AM injection [Fig. 1(b), top and bottom panels]. Of note, one of the animals did not respond to the 3-T1AM administration twice on consecutive days and was excluded from further analyses. When the cooling was quantified, we observed that the temperature of the iBAT and the back of the mice decreased in parallel upon 3-T1AM injection [Fig. 1(c), 1(d), and 1(f)], suggesting that the decline in brown fat thermogenesis is not the primary cause of the hypothermia. Interestingly, the tissue was also not activated during the course of the experiment to compensate for the hypothermia. The overall decline in temperature, moreover, was accompanied by reduced locomotor activity as seen in the overlay of all IR images (Supplemental Fig. 1B), which also occurred after the initial hypothermia.

Most interestingly, we observed a very short warming flush of the tail a few minutes after the 3-T1AM injection, which was not seen in the control group [Fig. 1(b), middle panel; Supplemental Fig. 1C]. When tail temperature was analyzed over the entire time of the experiment, we observed that it remained high for a longer period and was only significantly lower at the end of the experiment [Fig. 1(e)], suggesting that the expected vasoconstriction occurring upon hypothermia to reduce heat loss over the tail surface was impaired.

Testing the peripheral effect of 3-T1AM

Given the observed heat flush in the tail and the impairments in vasoconstriction upon the 3-T1AM–induced hypothermia, we speculated that 3-T1AM might directly induce vasodilation because α2a-adrenergic receptor and taar1 messenger RNA was detected by reverse transcriptase PCR in aorta and tail artery, respectively (Supplemental Fig. 1E). To test for a possible vasodilatory effect, we used wire myography with isolated abdominal aorta and tail artery segments and recorded a dose-response curve to the vasoconstrictor PE in the presence or absence of 1 µM 3-T1AM. Notably, 3-T1AM had no effect on vasoconstriction in either vessel [Fig. 2(a) and 2(b), left panel], whereas 100 nM T3 used as a positive control induced vasodilation at low concentrations of PE and enhanced the vasoconstriction at high PE concentrations as expected (12, 21).

Figure 2.

Dose-response curves of (a) abdominal aorta and (b) tail artery from wild-type mice to increasing concentrations of PE, prestimulated with either 1 µM 3-T1AM (left) or 100 nM T3 (right). The same volume of DMSO was used as control. Depicted are induced forces (i.e., vessel contractions normalized to 2 M KCl response) as mean ± SEM, n = 5 per group with nonlinear curve fits used to calculate the EC₅₀ values. Statistical significance was defined as *P < 0.05. EC₅₀, median effective concentration.

Testing the centrally mediated effect of 3-T1AM

Because our results demonstrated no direct influence of 3-T1AM on vasocontractility and we observed increased neuronal activation in the paraventricular nucleus of the hypothalamus after 3-T1AM IP administration (Supplemental Fig. 1D), we hypothesized that the hypothermia occurring after 3-T1AM injection might be centrally mediated. We therefore administered 3-T1AM directly into the brain using ICV injection and monitored the effect on body temperature, iBAT thermogenesis, and tail heat loss through IR photography.

Although we observed a tendency toward reduced iBAT and body temperature and elevated tail temperature after the injection of 40 ng 3-T1AM ICV only (Supplemental Fig. 2A, left), an injection of 80 ng 3-T1AM ICV significantly decreased iBAT temperature and elevated the tail to body temperature ratio at 15 minutes after injection (Supplemental Fig. 2A, right). To test whether we could acquire an even more pronounced phenotype, which would be comparable to the IP injection, we increased the dose to 200 ng 3-T1AM ICV and recorded the temperature changes using high-resolution IR video (1 picture per second). Here, we observed a significantly decreased back temperature at the end of the experiment [Fig. 3(a)] without activation of iBAT [Fig. 3(b)] and a significantly elevated tail temperature within the first 4 minutes after the injection, which subsequently declined in parallel with the back temperature toward the end of the experiment [Fig. 3(c) and 3(d)].

Figure 3.

Effect of 3-T1AM ICV on body temperature. (a–c) Quantification of the changes in back, iBAT, and tail temperatures compared with baseline after the ICV administration of 200 ng 3-T1AM in 2 μL 60% DMSO/40% PBS or the same volume of vehicle using IR videography. Values depict mean ± SEM of 4 mice, which were analyzed in a paired crossover study. Statistical significance was tested with a t test corrected for repeated measurements with the Holm-Sidak method and considered significant at ***P < 0.001. Right, area under the curve for tail temperature in the first 240 seconds after the ICV injection of 3-T1AM compared with vehicle (1 picture per second). Statistical significance was tested with a paired t test and considered significant at *P < 0.05. (d) Quantification of hypothalamic proteins using Western blot analysis given as mean ± SEM of n = 5 controls and n = 6 animals treated with 80 ng 3-T1AM ICV; statistical significance was defined as *P < 0.05; **P < 0.01.

To test whether this central effect was mediated through hypothalamic adenosine 5′-monophosphate kinase (AMPK) signaling, an important key player in temperature homeostasis (22), we euthanized the mice 60 minutes after ICV injection of 80 ng of 3-T1AM and performed a Western blot of dissected hypothalami. The results revealed increased c-FOS in the 3-T1AM–treated group compared with controls, again demonstrating neuronal activation in the hypothalamus [Fig. 3(e); Supplemental Fig. 2B]. For the hypothalamic AMPK signaling pathway, however, we did observe a slight albeit not substantial increase in AMPKa1 and AMPKa2 and phosphorylated AMPK per total AMPK [Fig. 3(e); Supplemental Fig. 2B], which concurs with the lack of brown adipose tissue activation found in these animals. Likewise, gene expression in iBAT, muscle, or liver as well as hepatic and muscular glycogen was not altered within the timeframe of this experiment (Supplemental Fig. 2A–C).

Discussion

Our data show that 3-T1AM administration in mice leads to an initial warming flush in the tail, which precedes a strong decrease in body temperature. Subsequently, the iBAT temperature declines and the mice gradually move less. From these observations, we conclude that the initial cause for the 3-T1AM–induced temperature effect is heat loss over the body surface resulting from vasodilation. Because iBAT is not activated to counteract the heat loss and the hypomotility further reduces heat production from the muscles, the condition severely aggravates until the effect wears off after about 90 minutes.

Although we did not find any direct effect of 3-T1AM on vessel contractility in our myograph studies, IP injection and ICV infusion activated neurons in the hypothalamus as seen by higher levels of c-FOS. Moreover, the ICV administration replicated the phenotype after systemic administration including tail vasodilation within 4 minutes and a decline in back temperature. This suggests that 3-T1AM can alter tail vasocontraction through hypothalamic signaling pathways, which is in line with the well-known role of the hypothalamus as the body’s thermostat featuring warm- and cold-sensitive neurons, for instance in the preoptic (23) or the anterior hypothalamic (14) areas. Interestingly, we did not observe an activation of iBAT thermogenesis despite the lower body temperature. Accordingly, the hypothalamic AMPK signaling cascade was tendentially even elevated by 3-T1AM, concurring with previous studies associating reduced AMPK signaling with iBAT activation (16, 22). Consequently, we conclude that 3-T1AM acts centrally to lower body temperature by elevating peripheral heat loss (conductance) and preventing iBAT activation. This mechanism would be classified as anapyrexia rather than hypothermia (24) because it is characterized by a lower central setpoint of the desired body temperature rather than an incapacity to maintain body temperature (13). Requiring 2 to 3 hours to return to normal body temperature after 3-T1AM administration is largely consistent with the recorded half-life of this compound in various tissues, ranging from 19 minutes to a few hours (3, 25, 26).

Unfortunately, little is known on 3-T1AM concentrations that reach the brain after systemic application; however, it could be speculated that local differences in concentrations because of the route of administration are the underlying reasons why the effects were more pronounced after systemic administration than after ICV injection. Nevertheless, it was conclusively demonstrated in several previous studies that 3-T1AM exerts central mechanisms (6) and is consequently capable of crossing the blood-brain barrier. However, the precise molecular mechanism how it enters the brain and exerts its actions remains elusive. Although we detected the possible 3-T1AM targets taar1 and the α2a-adrenergic receptors (3, 10) in tail artery and aorta, respectively, neither vessel responded to high doses of 3-T1AM, indicating that at least in these tissues, the receptors are not bona fide targets of 3-T1AM action. That TAAR1 is not underlying the hypothermic effect of 3-T1AM was recently demonstrated by an elegant study using TAAR1 knockout mice, where the anapyrexic effect persisted (11).

Most intriguingly, in a recent publication, the activation of the warm-sensitive ion channel transient receptor potential M2 led to a similar thermoregulatory response observed in our animals (27). Given that transient receptor potential M8 has already been identified as a potential target of 3-T1AM in corneal epithelial cells (28), it is tempting to speculate that 3-T1AM could act on temperature-sensitive transient receptor potential (TRP) channels, which subsequently trigger hyperthermia defense mechanisms, including tail vasodilation. This could even include peripheral TRP channels mediating thermal nociception responses (29, 30), which would explain the very rapid and pronounced effect after IP administration of 3-T1AM. Because of the plethora of possible TRP channel targets, the precise mechanism remains yet to be identified; however, our immunohistochemistry data suggest that the site of 3-T1AM action might be the paraventricular nucleus rather than the preoptic area, which is in agreement with a previous study observing no rapid effects of 3-T1AM after direct injection into the preoptic area (31). This also limits the number of possible target TRP channels, likely excluding those predominantly found in the preoptic area (27).

In this context, DMSO itself has some vasodilatory effects if given at volumes larger than 4 μL per g body weight (32). Hence, we used 50 mg/kg 3-T1AM in 30% ethanol for the initial radiotelemetry study, which also induced anapyrexia, although the effects were less pronounced than previously published (3). For the IR videos, we subsequently used a mixture of 60% DMSO and 40% PBS, which resulted in a total of 2.4 μL DMSO per gram of body weight, a volume that induced only a very mild hypothermia in the vehicle treatment group in accordance with previous studies (32). Although it cannot be excluded that 3-T1AM and DMSO act synergistically to induce anapyrexia, our ICV studies using the same vehicle did not show any solvent effect, suggesting that 3-T1AM might be the sole driver of the temperature effect.

Taken together, our results revealed that the TH metabolite 3-T1AM induces anapyrexia through tail vasodilation, an effect not mediated by direct vascular effects but rather through central mechanisms potentially involving warm-sensitive neurons. Thus, we have identified a metabolite with central vascular and thermoregulatory actions that could constitute a previously unknown feedback mechanism from peripheral tissues producing or inactivating 3-T1AM to alter the central thermostat set point. Although it is currently not possible to distinguish between physiology and pharmacology, our findings add another link between TH metabolism and body temperature regulation and prompt further studies to investigate the role of 3-T1AM in thyroid pathologies known to be accompanied by altered thermosensation.

Abbreviations:

 
• AMPK

  adenosine 5′-monophosphate kinase

 
• DMSO

  dimethyl sulfoxide

 
• iBAT

  interscapular brown adipose tissue

 
• ICV

  intracerebroventricular

 
• IP

  intraperitoneal

 
• IR

  infrared

 
• PBS

  phosphate-buffered saline

 
• PCR

  polymerase chain reaction

 
• PE

  phenylephrine

 
• SEM

  standard error of the mean

 
• TH

  thyroid hormone

 
• 3-T1AM

  3-iodothyronamine

 
• TAAR1

  trace amino-associated receptor

 
• TRP

  transient receptor potential.

Acknowledgments

We thank Jaafar Al-Hasani for fruitful discussions and help with the manuscript.

This study was generously funded by the German Research Council DFG (Heisenberg Programme MI1242/2-1, SPP Thyroid TransAct MI1242/4-1, and GRK1957 Adipocyte-Brain-Crosstalk to J.M.; HO5096/1-1 to C.S.H.), the European Community's Seventh Framework Programme (FP7/2007-2013) under Grant agreement 281854 (to M.L.), the ObERStress project (to M.L.), Xunta de Galicia (2015-CP079) (to M.L.), and Ministry of Economy, Industry and Competitiveness (MINECO), Spain, cofunded by Spanish Federation for Rare Diseases (FEDER) (SAF2015-71026-R and BFU2015-70454-REDT/Adipoplast) and Atresmedia Corporación (to M.L.).

Disclosure Summary: The authors have nothing to disclose.

Author notes

Address all correspondence and requests for reprints to: Jens Mittag, PhD, Universität zu Lübeck, Center of Brain, Behavior and Metabolism (CBBM), Ratzeburger Allee 160, 23562 Lübeck, Germany. E-mail: [email protected].