Involvement of Neuropeptide Galanin Receptors 2 and 3 in Learning, Memory and Anxiety in Aging Mice

The neuropeptide galanin (GAL), which is expressed in limbic brain structures, has a strong impact on the regulation of mood and behavior. GAL exerts its effects via three G protein-coupled receptors (GAL1–3-R). Little is known about the effects of aging and loss of GAL-Rs on hippocampal-mediated processes connected to neurogenesis, such as learning, memory recall and anxiety, and cell proliferation and survival in the dorsal dentate gyrus (dDG) in mice. Our results demonstrate that loss of GAL3-R, but not GAL2-R, slowed learning and induced anxiety in older (12–14-month-old) mice. Lack of GAL2-R increased cell survival (BrdU incorporation) in the dDG of young mice. However, normal neurogenesis was observed in vitro using neural stem and precursor cells obtained from GAL2-R and GAL3-R knockouts upon GAL treatment. Interestingly, we found sub-strain differences between C57BL/6J and C57BL/6N mice, the latter showing faster learning, less anxiety and lower cell survival in the dDG. We conclude that GAL-R signaling is involved in cognitive functions and can modulate the survival of cells in the neurogenic niche, which might lead to new therapeutic applications. Furthermore, we observed that the mouse sub-strain had a profound impact on the behavioral parameters analyzed and should therefore be carefully considered in future studies.


Introduction
In mice and humans, adult neurogenesis takes place in the sub-granular zone of the dorsal dentate gyrus (dDG) and the subventricular zone of the lateral ventricles [1][2][3]. Neurogenic events such as neural stem and progenitor cell proliferation, survival and integration are orchestrated by the neurogenic niche. The connection between learning and neurogenesis is complex and a matter of ongoing debate. Early work showed, for example, that cognitive tasks stimulate the generation of new neurons [4], whereas disruption of neurogenesis leads to cognitive impairment, which is also observed in aging [5]. Neuropeptides such as galanin (GAL) regulate various crucial functions of the central nervous system (CNS). However, direct involvement of GAL signaling in neurogenesis and learning, especially during aging, has not been investigated.
GAL is widely distributed in the central and peripheral nervous systems and regulates a variety of biological and physiological functions, ranging from energy homeostasis, repro-duction and feeding to cognition and learning [6,7]. The GAL peptide is highly conserved throughout evolution and found in many species. GAL is processed from a 123 amino acid (aa) pre-pro-peptide, which gives rise to the mature GAL (29 aa in rodents, 30 aa in humans) and the GAL-message-associated peptide. GAL co-localizes with noradrenaline in the locus coeruleus of rats [8][9][10] and humans [11][12][13]. Additionally, GAL co-localizes with 5-hydroxytryptamine (serotonin) in the dorsal raphe nucleus of rats [9,14]. In mice, GAL mRNA is expressed at high levels in the amygdala, hypothalamus, locus coeruleus and the dentate gyrus of the hippocampus [15], which are crucial structures for the regulation of anxiety and stress [16,17]. GAL inhibits functions mediated by norepinephrine, serotonin, dopamine, glutamate and acetylcholine, which could be relevant for the pathophysiology of psychiatric disorders [18,19].
Three G-protein-coupled receptors (GPCRs) have been identified for GAL, namely GAL 1 -R, GAL 2 -R and GAL 3 -R. The GAL-Rs differ substantially in their functional coupling and signaling activities, thereby contributing to the diverse effects attributed to GAL [7]. GAL 1 -R mRNA is expressed in the mammalian CNS. In rodents, GAL 1 -R expression is especially high in the olfactory regions, subregions of the amygdala, hypothalamus, thalamus, medulla and spinal cord [20][21][22]. Expression of both GAL 1 -R and GAL 2 -R has been detected in the subventricular zone and the rostral migratory stream [23,24]. GAL 2 -R is found in the forebrain, midbrain and hindbrain, and it is highly expressed in the DG and hypothalamus. In the hindbrain, GAL 2 -R mRNA is located in the spinal trigeminal tract and the dorsal vagal complex as well as the paraventricular, ventromedial and dorsomedial nuclei of the hypothalamus [20,21,25]. Moderate expression of GAL 2 -R was found in large alpha motor neurons located in the ventral horn of the rat spinal cord. Spinal neurons in the gray matter, the ventral horns, in the dorsal horn and intermediate lateral cells (paraand sympathetic) were also found to express the GAL 2 -R [20,25]. GAL 3 -R expression in the CNS appears to be restricted to the hypothalamus [26][27][28].
Studies on various rodent models revealed an important role of GAL 3 -R in behavior. For example, Scheller and colleagues reported that administration of the GAL 3 -R antagonist SNAP 37,889 suppresses alcohol drinking and morphine self-administration in mice [29]. However, treatment of wild-type (WT) mice with GAL 3 -R antagonists such as SNAP 37,889 and SNAP 398,299 had an anxiolytic effect [30]. In contrast, Brunner and colleagues reported an anxiety-like phenotype in GAL 3 -knockout(KO) mice [31]. Furthermore, GAL 2 -KO mice displayed an anxiogenic-like phenotype [32].
The survival-and growth-promoting activities of GAL on different types of neurons in the peripheral and central nervous systems, including dorsal root ganglion (DRG) sensory neurons, have been well documented [33,34]. Several studies suggest that GAL 2 -R, and not GAL 1 -R, is primarily involved in the pro-regenerative activity of GAL on DRG and hippocampal neurons [35][36][37]. Because correlations between anxiety levels and neurogenesis have been suggested [38], and GAL-R signaling was shown to be involved in anxietyrelated behaviors [31,32], we hypothesized that mice lacking GAL2-R or GAL3-R would behave differently from each other in anxiety-related tests and have different numbers of proliferating and newly integrated cells in the dDG.
In the present study, we analyzed neurogenesis-related behavioral and cognitive functions in young (3-month-old) and middle-aged (12-14-month-old) mice lacking GAL2-R or GAL3-R in comparison to the respective WT mice and quantified the numbers of proliferating and integrating cells in the dDG.

Loss of GAL 3 -R Induces Anxiety in Middle-Aged Mice
We assessed the role of GAL-R in neurogenesis in relation to aging, which should correlate with anxiety-like behaviors in mice. The GAL 2 -KO and GAL 3 -KO mouse lines used in this study were bred on different genetic backgrounds, C57BL/6J and C57BL/6N, respectively. Therefore, statistical evaluation of differences of the KO-animals was restricted to the corresponding WT group unless otherwise stated. We tested 'young-adult' (3 months old) and 'middle-aged' (12-14 months old) GAL 3 -KO and WT mice for anxiety-related behavior using the elevated plus maze (EPM). In the EPM test, time spent in the closed arms of the apparatus, as opposed to the open arms or the central area, is considered an indicator of anxiety-related behavior. In the young cohort, EPM analysis revealed no significant difference in time spent in either the open or closed arms between GAL 3 -KO mice and WT controls (Figure 1a, time in open arms 24.2 ± 13.52 s (WT, n = 6) vs. 98.1 ± 96.2 s (GAL 3 -KO, n = 11), p = 0.08; time in closed arms: 199.4 ± 21.1 s (WT, n = 6) vs. 141.8 ± 80.3 s (GAL 3 -KO, n = 11), p = 0.11). We extended the same analyses to GAL2-KO mice. In the EPM, young GAL2-KO mice stayed longer in the closed arms compared to their age-matched WT controls, but no differences for the open arms were detected (closed arms: 180.5 ± 31.6 s (WT, n = 11) vs. 216.7 In contrast, for the older cohort, GAL 3 -KO mice stayed for significantly shorter lengths of time in the open arms and for significantly longer periods in the closed arms compared to the WT controls ( Figure 1b, open arms: 141.5 ± 104.2 s (WT, n = 11) vs. 19.34 ± 21.74 s (GAL 3 -KO, n = 10), p = 0.002; closed arms: 94.8 ± 93.7 s (WT, n = 11) vs. 215.3 ± 29.9 s (GAL 3 -KO, n = 10), p = 0.001). Thus, older GAL 3 -KO mice presented significantly higher anxiety-like behavior compared to their age-matched WT counterparts. The frequency of center entrances (Figure 1c) did not differ between GAL 3 -KO and WT mice in their respective age groups (number of center entrances in the 3-month-old group: 15.9 ± 5.1 (WT, n = 7) vs. 14.7 ± 5.9 (GAL 3 -KO, n = 11), p = 0.683; 12-14-month-old group: 11.5 ± 5.5 (WT, n = 11) vs. 14.2 ± 6.9 (GAL 3 -KO, n = 11), p = 0.319).

Age-Dependent Deficit in Spatial Learning in Mice Lacking GAL3-R
We assessed spatial learning with the Morris Water Maze (MWM) test, in which mi have to find and remember the position of a hidden platform on five consecutive day

Age-Dependent Deficit in Spatial Learning in Mice Lacking GAL 3 -R
We assessed spatial learning with the Morris Water Maze (MWM) test, in which mice have to find and remember the position of a hidden platform on five consecutive days. Young GAL 3 -KO mice needed more time to find the platform than age-matched WT mice on day 2 and day 3 (day 2: p = 0.021, day 3: p = 0.034, Figure 3a). However, no differences were observed for the older GAL 3 -KO cohort compared to WT (Figure 3c) or the young or older GAL 2 -KO cohorts compared to WT (Figure 3b,d).

Mouse Embryonic Forebrain Cells Express GAL System mRNA But Do Not Respond to GAL Treatment In Vitro
We detected expression of GAL and all three GAL-Rs in mouse embryonic forebrain (MEF) cultures derived from C57BL/6N WT mice (Figure 5a). Therefore, we analyzed the effect of signaling through the GAL-Rs on neurogenic processes in mouse E16.5 MEF cultures expressing the firefly luciferase reporter gene under the control of the doublecortin (DCX) promoter, an early neuronal marker [40,41]. Direct application of GAL in concentrations ranging from 0.01 to 10 µM did not induce DCX promoter activity in WT MEF cells (Figure 5b). We also treated MEF cells derived from GAL2-KO and GAL3-KO mice with GAL and did not observe an increase in the luciferase signal. Treatment of the  1420.0 ± 249.5 (GAL 2 -KO, n = 9), p = 0.001). As expected, less BrdU + cells were detected in the aged mice, but no genotype-associated difference was observed (65.0 ± 43.1 (WT, n = 12) vs. 60.0 ± 40.0 (GAL 2 -KO, n = 10), p = 0.782). Strikingly, the GAL 2 -KO and WT mouse cohorts, both on the C57BL/6J background, showed more than twice as many BrdU + cells in the dDG compared to mice on the C57BL/6N background, independent of age (Figure 4c,d).
To discriminate the influence on proliferation from a change in survival of the newly generated cells in the dDG, the number of proliferating neural stem cells was estimated based on the presence of proliferating cell nuclear antigen (PCNA). Lack of GAL 3 -R did not change the number of PCNA + cells in the dDG of the 3-month-old mice (Figure 4a,e, 1960.0 ± 599.9 (WT, n = 9) vs. 1841.1± 666.1 (GAL 3 -KO, n = 10), p = 0.686). Also, in the older mice, the loss of GAL 3 -R did not lead to a difference in the number of proliferating cells (Figure 4e, 175.6 ± 83.8 (WT, n = 9) vs. 141.1 ± 81.1 (GAL 3 -KO, n = 9), p = 0.388).
In the 3-month-old GAL 2 -KO mice and age-matched WT mice, no differences in the number of PCNA + cells were measured in the dDG (Figure 4f, 2469.0 ± 539.8 (WT, n = 11) vs. 2393.0 ± 320.6 (GAL 2 -KO, n = 9), p = 0.725). Fewer PCNA + cells were found in the older mouse brains of both WT and GAL 2 -KO mice (Figure 4f, 186.7 ± 86.06 (WT, n = 12) vs. 203.0 ± 95.5 (GAL 2 -KO, n = 10), p = 0.677). Hence, the increase in the number of BrdU + cells observed in the 3-month-old GAL 2 -KO mice resulted from an improved survival of newly generated cells, rather than from a higher proliferation rate in the dDG.

Mouse Embryonic Forebrain Cells Express GAL System mRNA but Do Not Respond to GAL Treatment In Vitro
We detected expression of GAL and all three GAL-Rs in mouse embryonic forebrain (MEF) cultures derived from C57BL/6N WT mice (Figure 5a). Therefore, we analyzed the effect of signaling through the GAL-Rs on neurogenic processes in mouse E16.5 MEF cultures expressing the firefly luciferase reporter gene under the control of the doublecortin (DCX) promoter, an early neuronal marker [40,41]. Direct application of GAL in concentrations ranging from 0.01 to 10 µM did not induce DCX promoter activity in WT MEF cells (Figure 5b). We also treated MEF cells derived from GAL 2 -KO and GAL 3 -KO mice with GAL and did not observe an increase in the luciferase signal. Treatment of the various cultures with positive controls, i.e., the combination of valproic acid and retinoic acid, robustly increased the luciferase signal, thus confirming the potency of the cultured cells to undergo neuronal differentiation (Figure 5b).

C57BL/6N and C57BL/6J Sub-Strains Differ in Behavioral and Cognitive Tasks
From the results of the various behavioral tests, we observed differences between WT mice on the C57BL/6J and C57BL/6N genetic backgrounds. We retrospectively compared

Discussion
In the present study, we submitted young and aged GAL2-KO and GAL3-KO mice to a spatial learning task and two anxiety-related behavioral tests to examine if the GAL sys-

Discussion
In the present study, we submitted young and aged GAL 2 -KO and GAL 3 -KO mice to a spatial learning task and two anxiety-related behavioral tests to examine if the GAL system is involved in these cognitive functions and affected by aging. Furthermore, we scrutinized for potential correlations between behavior and the levels of cell proliferation and survival within the neurogenic niche of the hippocampus as a function of age. We observed that young GAL 3 -KO mice show reduced spatial learning in the MWM test compared to WT, whereas older GAL 3 -KO mice develop an anxiety-like behavior, according to the EPM test. This GAL 3 -KO anxiety-like phenotype is in line with previous findings from our laboratory, although the anxiety-like phenotype was also detected in young GAL 3 -KO mice on the C57BL/6N background and in the OF test in our previous study [31]. A possible explanation could be the occurrence of genetic drift in this mouse line in the time since the previous study, although the lines were backcrossed to the C57BL/6N background every seventh generation. Additionally, we confirmed genetic uniformity to the C57BL/6N genetic background by genetic homogeneity analysis performed by the mouse distributor (data not shown). The behavior of mice on the C57BL/6J background was similar, independent of age or GAL 2 -KO genotype (except for the EPM test, where young GAL 2 -KO mice spent more time in the closed arms). Previous findings from Bailey and colleagues reported a more robust anxiety-like behavioral response for GAL 2 -KO mice in the EPM. Importantly, their mice were on the 129S1Sv/ImJ genetic background [32]. Further background-specific phenotypes were observed in GAL-KO mice on the 129/OlaHsd background, which had a severe lactation deficit [32,33]. Intriguingly, Bailey et al. also reported EPM-but not OF-related anxiety-like behavior in GAL 1 -KO mice [32,39]. Holmes and colleagues gave a possible explanation for this anxiety-like behavior restricted to the EPM. They reported that the EPM test leads to especially high secretion of stress hormones as compared to other behavioral tests such as the OF. These circulating hormones, such as adrenocorticotropic hormone (ACTH) and corticosterone, may lead to a different stress-induced GAL response during the EPM as compared to other behavioral challenges [39]. Stress response circuits are highly conserved between different mammalian species and were shown to be already present in the earliest vertebrates [42]. Therefore, we speculate that, given the evolutionary importance of stress responses, genetic compensation still guarantees adequate responses in KO mice, as reported previously for other genes, where phenotypes in knockdowns and KO differed significantly [43]. Therefore, we hypothesize that the effects of a full germ-line KO of GAL 2 -R or GAL 3 -R might lead to partial compensation due to the latter mentioned genetic robustness of the system, which might decline during aging, resulting in the observed increased anxiety-like behavior of GAL 3 -KO mice.
Interestingly, the deficit in learning and the anxiety-like behavior detected in GAL 3 -KO mice were not reflected by changes in proliferation and survival of cells in the neurogenic niche of the dDG. In line with this, no change in the activity of the DCX promoter, which correlates with neurogenic activity, was observed following application of GAL to MEF cell cultures. Mennicken et al. [28] demonstrated that GAL 3 -R appears to be restricted to the hypothalamus. The hypothalamus is, beside the hippocampus, part of the limbic system, which is strongly associated with memory formation and emotion [44]. Therefore, we think that the therapeutic enhancement of GAL 3 -signaling could improve age-related comorbidities like loss of memory function or anxiety. However, a direct correlation to the neurogenic niche seems to be unlikely. In contrast, we detected improved survival of newly generated cells in the dDG of 3-month-old GAL 2 -KO mice. Hence, no direct correlation was observed between the proliferation and survival of newly generated cells in the dDG of mice with GAL 2 -R or GAL 3 -R KO and their performance in the behavioral tests. However, in accordance with the literature [5], we observed drastic reductions of proliferating and integrating cells in the older mice of both lines. Interestingly, we detected differences in learning capabilities and behavior between C57BL/6N and C57BL/6J mice which adds knowledge to previously reported differences between C57BL/6N and C57BL/6J mice ranging from chronobiology and metabolic responses [45] to general behavior [46,47]. We speculate that the lower learning capabilities of C57BL/6J mice and their overall anxious behavior upon aging compared to C57BL/6N mice might result from the overall increased proliferation of cells in the dDG, which was shown to induce "forgetting" of unpleasant events [48].

Animals
Mice were bred and housed in the animal facility of the Paracelsus Medical University. GAL 3 -KO mice were initially obtained from the European Mouse Mutant Archive and were backcrossed to the C57BL/6N background and phenotypically characterized in our lab [31]. GAL 2 -KO on the C57BL/6J background [49] were kindly provided by Marina Picciotto's lab (Yale University). KO and WT mice were bred as homozygotes. C57BL/6N and C57BL/6J mice (Charles River, Sulzfeld, Germany) were used to backcross GAL 2 -KO and GAL 3 -KO mice every 7th generation to limit genetic drift. All animals were analyzed with respect to their GAL 2 -R and GAL 3 -R genotype before and after the experiments [31,49]. Male mice were used exclusively in the study to avoid effects of the estrous cycle on the results. All mice were kept in groups of a maximum of five, with a 12 h light/dark cycle and access to water and food ad libitum. All experiments were approved by the Austrian federal ethical commission (BMBWF-66.019/0031-V/3b/2018) and were conducted in accordance with the guidelines of "Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes".

BrdU Injection
To study survival and integration of newly generated cells in the dDG, the DNA of dividing cells was labeled through daily intraperitoneal injections of 50 mg/kg bodyweight BrdU (Sigma-Aldrich, Inc., Darmstadt, Germany) from day 1 to 4. Following perfusion on day 28, BrdU + cells were visualized by immunohistochemical techniques.

Elevated Plus Maze
Elevated Plus Maze (EPM) (Ugo Basile, Gemonio, Italy) trials were performed on day 19 after the start of the BrdU injections, as previously described [50]. Mice were placed in the central region of the maze and their position was tracked for 5 min using an Ethovision XT system (Noldus, Wageningen, The Netherlands). The maze was cleaned between every trial to eliminate olfactory cues. The time spent in the open arms and central region was designated as exploratory behavior, whereas time spent in the closed arms was defined as anxious-like behavior. For statistical analyses, total time spent in the open/center zone vs. total time spent in closed arms was analyzed.

Open Field
The open-field (OF) test, which assesses spontaneous locomotor and exploratory activities, was performed on day 20 after the start of the BrdU injections. Mice were placed in the middle of a circular arena 1 m in diameter and allowed to move freely for 5 min. Parameters of movement and position were recorded using Ethovision XT software (Noldus, Wageningen, The Netherlands). The arena was cleaned between every trial to eliminate olfactory cues.

Morris Water Maze
The Morris Water Maze (MWM) test was performed to assess spatial learning and memory from day 23 to 28 after the start of the BrdU injections, as previously described [50]. The time taken to find the hidden platform was recorded using Ethovision XT tracking software (Noldus, Wageningen, The Netherlands). For the evaluation of memory on day 28, the trial was performed in the absence of the hidden platform. Time spent in the platform quadrant, as well as mean distance to the platform and numbers of platform crossings were analyzed.

Statistical Analysis
All data were analyzed using GraphPad Prism 7 (Graphpad Inc., San Diego, CA, USA). Unpaired Student's t-test and one-way analysis of variance (ANOVA) were used to evaluate differences between mice. Outliers were identified using the ROUT test [54] and were excluded for further analyses. A p-value ≤ 0.05 was considered significant. All values are represented as mean ± standard error of the mean (SEM).

Conclusions
We showed that neuropeptide GAL signaling through GAL 3 -R is relevant for learning and anxiety in an age-dependent manner. In addition, we observed that proliferation of cells in the neurogenic niche can be modulated by GAL signaling; however, there was no correlation with the behavioral outcome. Furthermore, we want to emphasize that the genetic background of the animal models used can significantly impact behavioral outcomes, which underscores the high complexity of analyzing neuropeptide functions in mice. Finally, we conclude that the GAL neuropeptide system might be a promising target for the development of therapeutics which target age-related comorbidities such as anxiety or reduced memory function either as an antagonist (GAL 2 -R) or agonist (GAL 3 -R). However, a direct correlation of GAL 3 -signaling and the neurogenic niche seems to be unlikely.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.
Sample Availability: Remaining tissue samples of the study are available from the authors.