Reducing central serotonin in adulthood promotes hippocampal neurogenesis

Chronic administration of selective serotonin reuptake inhibitors (SSRIs), which up-regulates central serotonin (5-HT) system function, enhances adult hippocampal neurogenesis. However, the relationship between central 5-HT system and adult neurogenesis has not fully been understood. Here, we report that lowering 5-HT level in adulthood is also able to enhance adult hippocampal neurogenesis. We used tamoxifen (TM)-induced Cre in Pet1-CreERT2 mice to either deplete central serotonergic (5-HTergic) neurons or inactivate 5-HT synthesis in adulthood and explore the role of central 5-HT in adult hippocampal neurogenesis. A dramatic increase in hippocampal neurogenesis is present in these two central 5-HT-deficient mice and it is largely prevented by administration of agonist for 5-HTR2c receptor. In addition, the survival of new-born neurons in the hippocampus is enhanced. Furthermore, the adult 5-HT-deficient mice showed reduced depression-like behaviors but enhanced contextual fear memory. These findings demonstrate that lowering central 5-HT function in adulthood can also enhance adult hippocampal neurogenesis, thus revealing a new aspect of central 5-HT in regulating adult neurogenesis.


Depletion of central 5-HTergic neurons in adulthood.
To specifically ablate 5-HTergic neurons in the adult brain, we crossed Pet1-CreER T2 mice, in which inducible Cre recombinase is selectively activated in 5-HTergic neurons 17 , with Rosa26-DTA mice 18 to generate Pet1-CreER T2 ; Rosa26-DTA (referred to as DTA iPet1 ) offspring. TM administration led to the selective depletion of 5-HTergic neurons in the brainstem of DTA iPet1 mice, due to the toxic effect of DTA. One month after the final TM treatment, mice were sacrificed for analysis. When a lower dose of TM (80 mg/kg body weight, once daily for 3 consecutive days) was used, approximately 50% of 5-HTergic neurons were depleted, as illustrated by the striking decrease in the number of Tph2 + neurons in the dorsal raphe nucleus (Fig. 1a,b); these mice were referred to as lTM-DTA iPet1 mice. On the other hand, when a higher dose of TM (250 mg/kg body weight, once daily for 4 consecutive days) was used, more than 95% of 5-HTergic neurons were depleted (Fig. 1a,b); these mice were referred to as hTM-DTA iPet1 mice. Correspondingly, only a few 5-HTergic fibers were observed in the hippocampus, as shown by 5-HTT immunostaining (Fig. 1c). HPLC showed that 5-HT and its metabolite 5-HIAA were dramatically decreased in hTM-DTA iPet1 mice (Fig. 1d), which indicated that central 5-HT was dramatically reduced as a result of brainstem 5-HTergic neuronal loss. To determine whether other monoamines were affected or not after loss of 5-HTergic neurons, we examined the expression of tyrosine hydroxylase (TH), which labels dopaminergic neurons in the midbrain (Fig. 1e) and norepinephrinergic neurons in the hindbrain (Fig. 1f), respectively. We found that TH expression in hTM-DTA iPet1 mice was comparable to that in control mice. In consistent with these immunostaining data, HPLC showed that both the concentration of norepinephrine and homovanillic acid, the metabolite of dopamine, was not altered in hTM-DTA iPet1 mice relative to that of control mice (Fig. 1g). Together, 5-HTergic neurons in the brain can be selectively depleted in TM-DTA iPet1 mice in adulthood.
A drastic increase of immature neurons in hTM-DTA iPet1 mice. Adult neurogenesis involves proliferation of neural stem cells and differentiation of new-born neurons, two processes that require coordinated expression of a number of different genes 3 . Firstly, to detect the change of immature neurons after loss of 5-HTergic neurons, DCX immunostaining was performed to label the neuroblasts and immature neurons in the SGZ. We found that the number of DCX-labeled cells in lTM-DTA iPet1 mice was comparable to that in control mice (Fig. 2a,c), indicating that overall neurogenesis level in the hippocampus was unaffected in lTM-DTA iPet1 mice. In contrast to lTM-DTA iPet1 mice, the number of DCX + cells in the SGZ was dramatically increased in hTM-DTA iPet1 mice (Fig. 2b,c). In addition to DCX + cells, the number of NeuroD + neuroblasts was also significantly increased in hTM-DTA iPet1 mice (Fig. 2d,e). Furthermore, the number of CR + , which is expressed in the immature neurons following the initiation of DCX expression 1 , were increased also in hTM-DTA iPet1 mice (Fig. 2f). There was no change of CR + fibers in the molecular layer (arrows, Fig. 2f), which are originated from hilar mossy neurons. Finally, the enhanced hippocampal neurogenesis was first detected two weeks after the last TM injection (Fig. 2g) and also observed two months after the last TM injection (Fig. 2h,i), demonstrating that this effect was not a transient one. Together, these results indicate that removing central 5-HTergic neurons during adulthood leads to a significant enhancement in the hippocampal neurogenesis.

Proliferation of neural stem cells is enhanced in hTM-DTA iPet1 mice.
To explore the cause of increase of neuroblasts and immature neurons after depleting 5-HTergic neurons, we checked the proliferation of adult stem cells in the SGZ. First, we examined neural stem cell proliferation by means of BrdU incorporation method, and found that the number of BrdU + cells was dramatically increased in hTM-DTA iPet1 mice compared with controls (Fig. 3a,d). Besides, we also examined the expression of two other proteins, MCM2 and Ki67, which are expressed in cell cycles of neural stem cells 19 . Similar to BrdU + cells, both MCM2 + and Ki67 + cells in hTM-DTA iPet1 mice were significantly increased ( Fig. 3b-d). These results indicate that 5-HTergic neuron depletion leads to increased neural stem cell proliferation in the SGZ.
SSRIs enhance adult hippocampal neurogenesis by increasing the number of amplifying neural progenitor cells (ANPs) rather than quiescent neural progenitor cells (QNPs) 20 . To determine whether QNPs or ANPs were affected in hTM-DTA iPet1 mice, we performed GFAP and Sox2 co-immunostaining 1 and found that the number of GFAP + /Sox2 + QNPs was significantly elevated in the SGZ of hTM-DTA iPet1 mice (Fig. 3e,f). A similar increase was also observed in GFAP -/Sox2 + ANPs (Fig. 3e,f). We therefore conclude that the enhanced adult hippocampal neurogenesis in hTM-DTA iPet1 mice is due to an increase in both QNPs and ANPs in the SGZ.
Increased survival of new-born neurons in hTM-DTA iPet1 mice. Chronic SSRI administration enhances the survival of adult-born neurons 11 , and we next set out to explore whether the survival of adult-born neurons was affected or not after adult depletion of 5-HTergic neurons. Firstly, to detect the survival of new-born cells generated before depleting 5-HTergic neurons, BrdU was injected prior to TM administration (Fig. 4a).
Scientific RepoRts | 6:20338 | DOI: 10.1038/srep20338 Because depleting 5-HTergic neurons only occurs after TM administration, BrdU could label similar numbers of proliferating cells in both hTM-DTA iPet1 and control mice, and BrdU + cells in the hippocampus 30 days after TM administration could be used to reflect changes in the survival of adult-born neurons in hTM-DTA iPet1 mice. We found that BrdU + cells in the SGZ of hTM-DTA iPet1 mice were increased to about 1.6 fold of that in control mice (Fig. 4b,d,e). Secondly, to determine the survival of new-born neurons generated after depleting 5-HTergic neurons, BrdU were injected 30 days after the TM treatment. BrdU + cells in the SGZ of hTM-DTA iPet1 mice were increased to about 2.6 fold of that in control mice (Fig. 4c-e). To confirm whether BrdU-incorporated cells eventually become neurons or not, we performed BrdU/NeuN double immunostaining. We found that most of BrdU + cells were NeuN + in both control and hTM-DTA iPet1 mouse brains without obvious differences ( Fig. 4f-g). It is concluded that the survival of new-born neurons in the SGZ is enhanced in hTM-DTA iPet1 mice. (d) HPLC assay shows hTM-DTA iPet1 mice have extremely low levels of 5-HT and its metabolite 5-HIAA comparing to control mice. Data of 5-HT level were compared using Mann-Whitney rank sum test (p = 0.008), and data of 5-HIAA using Student's t-test (p = 4.5 × 10 −6 ). (e,f) Immunostaining of TH shows similar expression in midbrain dopaminergic neurons (e) and hindbrain norepinephrinergic neurons (f) between control and hTM-DTA iPet1 mice. (g) HPLC analysis shows comparable levels of HVA, a metabolite of dopamine, and NE in control and hTM-DTA iPet1 mice. Data were compared using Student's t-test.CA1, field CA1 of hippocampus; DG, dentate gyrus; DR, dorsal raphe; 5-HIAA, 5-Hydroxyindoleacetic acid; HVA, homovanillic acid; LC, locus coeruleus; NE, norepinephrine; SNc, substantia nigra, compact part; VTA, ventral tegmental area. All data are presented as mean ± s.e.m. *p < 0.05; **p < 0.01; ***p < 0. Enhanced hippocampal neurogenesis is also present in adult Tph2 iPet1 mice. In hTM-DTA iPet1 mice, 5-HTergic neurons are depleted due to toxicity of DTA. To ascertain whether enhanced adult hippocampal neurogenesis is caused by the loss of central 5-HTergic neurons or loss of 5-HT itself, Pet1-CreER T2 mice were crossed with Tph2 flox/flox mice 21 , to create Pet1-CreER T2 ; Tph2 flox/flox (referred to as Tph2 iPet1 ) mice in which 5-HTergic neurons are intact but 5-HT synthesis is inactivated upon TM injection. Nissl staining showed that intensely-stained "5-HTergic-like" neurons were present in the dorsal raphe nucleus of Tph2 iPet1 mice (arrowheads, Fig. 5b), but absent from hTM-DTA iPet1 mice (Fig. 5a). The number of Tph2 + neurons remaining in Tph2 iPet1 mice was reduced to approximately 15% of controls (Fig. 5e,f), and HPLC data also revealed a drastic reduction of 5-HT and 5-HIAA in Tph2 iPet1 mice (Fig. 5d). The maintenance of 5-HTergic neurons in Tph2 iPet1 Data were compared using Student's t-test (p = 8.8 × 10 -5 ). (f) Immunohistochemical analysis reveals that CR + neurons are greatly increased in the hippocampal SGZ of hTM-DTA iPet1 mice relative to controls. No changes of CR + hilar mossy neuron projections in the molecular layer of DG (arrows). (g) Two weeks after the last TM injection, the number of DCX + cells is increased in the SGZ of hTM-DTA iPet1 mice. (h,i) Two months after the last TM injection, the number of DCX + cells remains significantly increased in the SGZ of hTM-DTA iPet1 mice. Data were compared using Student's t-test (p = 0.04). All data are presented as mean ± s.e.m. *p < 0.05; **p < 0.01; ***p < 0.001. Scale bars represent 150 μm (a,d,f) and 100 μm (b,g,h). mice was confirmed by in situ hybridization of aromatic L-amino acid decarboxylase (AADC; Fig. 5c), a marker for 5-HTergic neurons. Thus, 5-HTergic neurons remained in Tph2 iPet1 brain, despite Tph2 inactivation in most central 5-HTergic neurons. Importantly, BrdU + , NeuroD + and DCX + cell numbers were also increased in the SGZ of Tph2 iPet1 mice relative to controls ( Fig. 5g-l), although to a lesser extent than that observed in hTM-DTA iPet1 mice. Taken together, these findings indicate that reducing central 5-HT levels in adulthood by inactivation of Tph2 is also capable of increasing adult hippocampal neurogenesis.
Administration of 5-HTR2c agonist prevents the increased neurogenesis in hTM-DTA iPet1 mice. We speculate that the depletion of 5-HTergic neurons or lowering 5-HT level leads to impairments of activation of some 5-HTRs, which in turn results in the enhanced hippocampal neurogenesis. We therefore set out to explore whether administration of agonists of 5-HTRs in the hTM-DTA iPet1 mice could prevent the increased adult neurogenesis or not. As shown in Fig. 6a, both control and hTM-DTA iPet1 mice were injected with 8-OH DPAT, α -methyl-5-hydroxytryptamine maleate, or WAY161503, which are the agonists for 5-HTR1a, 5-HTR2 family and 5-HTR2c, respectively, for consecutive 24 days and were sacrificed for immunostaining of DCX. To screen out which receptor was involved, as the first step we used DCX increase index (normalizing the numbers of DCX + cells in agonist-treated hTM-DTA iPet1 mice to that of agonist-treated controls) for comparison in each set of experiment. After administration of HTR1a and HTR2 family agonists, the DCX increase index in hTM-DTA iPet1 mice was not changed relative to vehicle-treated hTM-DTA iPet1 mice (Fig. 6b,c). By contrast, the index was significantly reduced in hTM-DTA iPet1 mice treated with WAY161503, agonist of 5-HTR2c (Fig. 6b,c), suggesting that 5-HTR2c is involved in the enhanced neurogenesis in hTM-DTA iPet1 mice. To know whether the effect was achieved by reducing neural stem cell proliferation, we next examined BrdU-labeled cells in both control and hTM-DTA iPet1 mice after 3-week treatment with WAY161503. BrdU was administrated 4 times with a 2-hour interval before the mice were sacrificed. We found that the number of BrdU + cells reduced dramatically in number in hTM-DTA iPet1 mice with no significant differences compared with controls (Fig. 6e,f). DCX + cells were reduced in number as well although it was still higher in hTM-DTA iPet1 mice relative to control (Fig. 6g). These results suggest that impaired activation of 5-HTR2c contributes to the enhanced adult neurogenesis in hTM-DTA iPet1 mice. Altered anxiety-and depression-like behaviors, and fear memory in hTM-DTA iPet1 mice. The enhanced adult hippocampal neurogenesis is required for fluoxetine to exert its antidepressant effect 22 . To date, there have been no published reports describing the behavioral characteristics associated with a genetic mouse model lacking the vast majority of central 5-HTergic neurons from adulthood, but having enhanced hippocampal neurogenesis. We thus set out to characterize the behavioral phenotype of hTM-DTA iPet1 mice, starting with anxiety-like behaviors. In the light/dark choice test, we found that hTM-DTA iPet1 mice spent significantly more time in the lit compartment compared to control mice (Fig. 7a). Consistently, hTM-DTA iPet1 mice spent more time in the open arms of the elevated plus maze than control mice did, although this difference did not reach statistical significance (Fig. 7b; p = 0.086). Basal depression-like behaviors were also assessed with two different tests. No difference was observed in the forced swim test, but immobility time in the tail-suspension test was significantly reduced in hTM-DTA iPet1 mice compared with controls (Fig. 7c,d). These results suggest that the basal anxiety-and depression-like behaviors are altered in hTM-DTA iPet1 mice.
Our previous study showed that Pet1-Cre;Lmx1b flox/flox mice, which essentially lack all central 5-HTergic neurons from late embryonic stage, displayed enhanced contextual fear memory 23 . It has also been reported that mice with decreased adult hippocampal neurogenesis exhibit reduced fear memory 24 . We were promoted to examine whether contextual fear memory is altered in our hTM-DTA iPet1 mice. On the first day (Day 0), mice were given five foot shocks to acquire fear memory. Both hTM-DTA iPet1 and control mice showed similar increases in freezing time after each foot shock (Fig. 7e). One day (Day 1) and 30 days (Day 30) later, mice were tested to retrieve recent and remote fear memory, respectively. We found that freezing times remained high in hTM-DTA iPet1 mice at both time points, but decreased in control mice (Fig. 7f,g), thus demonstrating that contextual fear memory is enhanced in hTM-DTA iPet1 mice.

Discussion
In the present study, we employed two approaches to genetically lower central 5-HT levels starting from adulthood. Both mouse models with lower adult central 5-HT took advantage of the inducible Cre system, in order to avoid the potential developmental defects associated with depleting central 5-HT during the development.   (g,h) The number of BrdU + cell is significantly increased in Tph2 iPet1 mice relative to controls. Data were compared using Student's t-test (p = 0.013). (i,j) The number of NeuroD + neuroblasts is also significantly increased in Tph2 iPet1 mice relative to controls. Data were compared using Student's t-test (p = 0.048). (k,l) The number of DCX + cells is significantly increased in Tph2 iPet1 mice compared with controls. Data were compared using Student's t-test (p = 0.032). All data are presented as mean ± s.e.m. *p < 0.05; ***p < 0.001. Scale bars represent 200 μm (a-c,e,g) and 30 μm (i,k). We found that removing 5-HTergic neurons in hTM-DTA iPet1 mice and deleting Tph2 in 5-HTergic neurons in Tph2 iPet1 mice both led to significant enhancement of adult neurogenesis in the hippocampus.
Previous studies have examined hippocampal adult neurogenesis in knockout mice constitutively lacking Tph2 or 5-HTT or 5-HT receptors. Normal hippocampal neurogenesis has been reported in adult Tph2 −/− and 5-HTT −/− mice 9,10 . 5-HTR1a antagonists can cause a decrease in adult neurogenesis 3 , while 5-HTR1a knockout mice appear normal in this regard 22 . Central 5-HTRs are known to be functional from embryonic stages onwards, and central 5-HT plays important roles in neuronal morphogenesis and neural circuitry formation  . (f,g) Quantification of BrdU + cells and DCX + cells. There is no significant differences between control and hTM-DTA iPet1 mice in number of BrdU + cells while more DCX + cells are present in hTM-DTA iPet1 mice. Data were compared using Student's t-test (p = 0.041 in g). All data are presented as mean ± s.e.m. *p < 0.05. Scale bars represent 100 μ m (b,e).
during embryonic and early postnatal development. Therefore, it is likely that these conventional knockout lines would exhibit developmental defects, and in fact growth retardation has been reported in Tph2 −/− mice 25,26 . Due to known and potential unknown secondary developmental defects, these knockouts may be suboptimal for conclusively characterizing the role of central 5-HT in adult neurogenesis.
It is well known that up-regulating 5-HT system by SSRIs enhances adult hippocampal neurogenesis 8 . In this study, we revealed that lowering 5-HT system function in adult brain by depleting 5-HTergic neurons or inactivating Tph2 expression could also enhance adult hippocampal neurogenesis. BrdU-labeled, Ki67 + and MCM2 + cells in the SGZ were all increased, and GFAP + /Sox2 + QNPs and GFAP -/Sox2 + ANPs were increased as well. In addition to increase of neural stem cells, we also found that the survival of new-born neurons generated before and after depleting 5-HTergic neurons was also enhanced in hTM-DTA iPet1 mice, which is consistent with the previous findings that the survival of adult-born cells is enhanced in the SGZ of mice with central 5-HT deficiency from embryonic stages including Pet1 −/− mice, 5-HTT cre/+ ; VMAT2 f/f and Tph2 KI mice 12,13 . Based on the data mentioned above, we think that both enhanced new-born survival and increased neural stem cell proliferation contribute to the enhancement of hippocampal neurogenesis in our mouse models.
It has been shown that 5-HTR2c antagonist increases adult hippocampal neurogenesis 27 . We revealed that chronic administraion of 5-HTR2c agonist lowered the increase of BrdU-labled cells in hTM-DTA iPet1 mice, suggesing that the increased adult neurogenesis in our 5-HT-deficient mice may be if not all at least partially caused by impaired activitaion of 5-HTR2c. However, althrough DCX + cells were lowered in number after adminstration of 5-HTR2c agonist, it was still higher than control. These results suggest that other 5-HT receptors may be involved, although we did not detect obvouis changes in DCX + cells after administration of agonist for 5-HTR1a and 5-HTR2 family. Among 5-HT receptors distributed in the dentate gyrus and hippocampal CA1-3 regions, relative expression level of 5-HTR2c is low 28 . As 5-HT is believed to be an extrinscic factor in regualting adult neurogenesis 1-3 , we speculate that adult 5-HT deficiency may lead to unknown alterations in activities of some neural networks which in turn result in the enhanced adult neurogenesis. Further studies are needed to examine the mechanisms underlying the enhanced hippocampal neurogensis with central 5-HT deficiency exclusively in adulthood.
Here we found that depleting over 95% of central 5-HTergic neurons in hTM-DTA iPet1 mice enhances adult hippocampal neurogenesis, but removing approximately 50% of them has no such effect. In our Tph2 iPet1 mice, Tph2 is inactivated in approximately 85% of central 5-HTergic neurons, and increased adult neurogenesis is also observed, although the increasing level is less than that in hTM-DTA iPet1 mice. Based on these data, we can hypothesize that a threshold of central 5-HT level exists, below which hippocampal neurogenesis becomes enhanced. Different 5-HTRs show different roles in regulation of adult neurogenesis 29 . Different 5-HT receptors are expressed at varying levels within the hippocampus 28 (http://mouse.brain-map.org), and their specific affinity Data were compared using Student's t-test (p = 0.0363 in d). (e-g) Enhanced fear memory in hTM-DTA iPet1 mice. Normal fear freezing is observed in hTM-DTA iPet1 mice on Day 0 (e). Data were compared using Two-way Repeated-Measures ANOVA. On Day 1 (f) and Day 30 (g), however, hTM-DTA iPet1 mice exhibit significantly increased contextual freezing. Data were compared using Student's t-test (p = 0.017 in (f) and p = 8.2 × 10 −4 in g). All data are presented as mean ± s.e.m. *p < 0.05; ***p < 0.001. to 5-HT is also variable 30 (http://www.iuphar-db.org). It is possible that different 5-HT receptor combinations might be activated depending on whether 5-HT levels are high or low, thus affecting adult neurogenesis differentially. Enhancing 5-HT function by SSRIs is also capable of incresesing adult hippocampal neurogenesis 22 . Thus, the hippocampal neurogenesis is enhanced in the two opposite conditions by 5-HT, in which central 5-HT levels are extremely low (the present study), as well as in SSRIs-treated mice, in which 5-HT levels are high 22 through activation different 5-HT receptors in normally-developed brain.
Our hTM-DTA iPet1 mice display reduced anxiety-like behaviors but enhanced contextual fear memory. Our previous results indicated that Pet1-Cre;Lmx1b flox/flox mice lacking central 5-HTergic neurons at embryonic stage show similar alterations in these types of behaviors 23 but have normal adult hippocampal neurogenesis. These results suggest that reduced anxiety-like behaviors and enhanced contextual fear memory in hTM-DTA iPet1 mice may be due at least in part to low levels of 5-HT itself. Running wheel training enhances the adult hippocampal neurogenesis in rodents, and physical exercises display the therapeutic effects in major depression and other behavioral deficit [31][32][33] . Lower hippocampal volume has been reported in patients suffering from depression 34 , and SSRIs-induced anti-depressant effect requires the enhanced adult hippocampal neurogenesis in mice 8,22 . Importantly, recent studies from Pet1-Cre;Lmx1b flox/flox and Tph2 KO mice have shown that central 5-HT deficiency is not sufficient to induce depression-like behaviors in mouse 35,36 . Considering the important role of adult hippocampal neurogenesis in depression, it is likely that the enhanced hippocampal neurogenesis may be a key factor for lowered basal depression-like behaviors in our hTM-DTA iPet1 mice.
In summary, we designed and utilized two novel mouse models, in which the brain develops normally but central 5-HT is depleted in adulthood, in order to investigate the regulation of adult hippocampal neurogenesis. These models uncovered an unexpected new role for central 5-HT, demonstrating that lowering it to a certain level leads to enhanced adult hippocampal neurogenesis and this effect may be partially achieved by inactivation of 5-HTR2c.

Methods
Experimental animals. To deplete 5-HTergic neurons in adulthood, Pet1-CreER T2 mice 17 were crossed with Rosa26-DTA mice 18 , and DTA iPet1 (i.e. Pet1-CreER T2 ; Rosa26-DTA) were obtained from the progeny. To block 5-HT synthesis in the adult brain, Pet1-CreER T2 mice were crossed with Tph2 floxed mice and Tph2 iPet1 (i.e. Pet1-CreER T2 ; Tph2 flox/flox ) mice were obtained. In DTA iPet1 mice, tamoxifen (TM) dissolved in corn oil was administered by oral gavage for a total of three times (Day 1, 2, 3) or four times (Day 1, 2, 4, 5) beginning 2.5-3.0 months after birth. Littermates of other genotypes (wild type, Rosa26-DTA or Pet1-CreER T2 ) received the same TM regimen and were used as controls in each set of experiments. In Tph2 iPet1 mice, TM was administered for a total of six times (Day 1, 2, 5, 6, 9, 10) beginning 1. For BrdU pulse labeling experiment to analyze cell proliferation, mice received 4 injections of BrdU at 50 mg/kg body weight at a 2-hour interval, and were sacrificed 2 hours after the last injection. For analysis of new-born cell survival, BrdU was injected before or after TM administration. In BrdU injection before depleting 5-HTergic neurons, BrdU was injected once daily for 3 consecutive days, followed by 4 times of TM administration; mice were sacrificed 30 days after the last TM administration. In BrdU injection after depleting 5-HTergic neurons, BrdU was injected 30 days after the last TM treatment for 3 consecutive days in the same way, and mice were allowed to survive for further 30 days after the BrdU injection. Brain sections were immersed in 0.01 M citrate buffer at 95 °C for 5 min, in 2 N HCl at 37 °C for 20 min and in 0.1 M sodium borate for 10 min, and then washed in PBS. Treated sections were immunostained with anti-BrdU antibody as described above.
AADC in situ hybridization was performed as described previously . Images were captured with an epifluorescence (Eclipse 80i, Nikon) or confocal (TCS SP5, Leica) microscope.
High performance liquid chromatography (HPLC). Samples containing the cerebral cortex and hippocampus were collected and HPLC were performed as described previously 23 .
Behavioral tests. Behavioral observation was performed in adult male mice 1 month after the final TM injection. For forced swim test, mice were placed into glass cylinders (height: 24 cm; diameter: 16 cm) containing 18 cm of 23-25 °C water for 6 min. A mouse was considered to be immobile when it floated in the water, and immobility time was recorded during the last 4 min of the 6-min testing period, after 2 min of habituation.
Tail suspension test. Mice were suspended by the tail with adhesive tape (distance from tip of tail: 2 cm). Their behavior was video recorded for the duration of the 6-min testing period, with immobility time being measured during the last 4 min.
Elevated plus maze test. Mice were initially placed at the far end of one of the closed arms and allowed to freely explore the maze for 5 min. The percentage of time spent in the two open arms was then recorded.
Light/dark choice test. The light/dark choice test apparatus consisted of a small dark chamber (30 cm × 20 cm × 30 cm) connected by an opening (5 cm × 7 cm) to a larger lit chamber (30 cm × 30 cm × 30 cm). A single mouse was initially placed in one corner of the dark chamber and the percentage of time spent in the lit chamber was measured over 5 min.
Contextual fear conditioning. Contextual fear conditioning was performed as previously described 23 . FreezeFrame and FreezeView software systems were used to record and analyze freezing behaviors. On the first day (Day 0), mice were given five foot shocks (0.7 mA, 2s) at 2 min intervals during which the mice were able to move freely. The percentage of freezing time was measured during each inter-shock interval. On the second day (Day 1) and one month later (Day 30), mice were placed back in the box for 11 min without receiving any foot shocks, and freezing time was measured to test contextual fear memory.
Statistical analysis. All samples passed both the Shapiro-Wilk normality test and equal variancetest except for 5-HT level in Fig. 1d, BrdU + cell numbers in Fig. 3d and BrdU + cells in BrdU after TM groups in Fig. 4d-e. Comparisons were performed using the Student's t-test, one-way ANOVA with post hoc Tukey test or Mann-Whitney rank sum test (for samples which didn't pass normality or equal variance test). All data are presented as mean ± s.e.m. P values of less than 0.05 were considered statistically significant.