Neurons in the central nervous system are produced from neural stem/progenitor cells (NSPCs) at embryonic or early postnatal stages in a process called neurogenesis. In the adult brain, this new-neuron production machinery is limited to two regions: the subgranular zone (SGZ) of the hippocampus, which generates glutamatergic granule cells of dentate gyrus (DG), and the subventricular zone (SVZ) lining the lateral ventricles, which produces GABAergic/dopaminergic cells committed to the olfactory bulb (Vadodaria and Gage, 2014). To date, substantial progress has been made to elucidate the mechanisms underlying NSPC activity and subsequent neuronal differentiation and integration into the existing neural network. Key effectors in these different developmental steps include cell intrinsic and extrinsic factors, such as transcriptional factors, morphogens, growth factors, neurotransmitters, and network activity.

Among neurotransmitters, serotonin (5-HT) has attracted the most interest because of the enhancement of adult hippocampal neurogenesis induced by selective serotonin reuptake inhibitors (SSRIs), which might contribute to their antidepressant effects (Sahay and Hen, 2007; Vaidya et al., 2007). SSRIs primarily target the 5-HT system rather than other neurotransmitter systems. Both acute and chronic administration of SSRIs increases the synaptic (extracellular) serotonin concentration by several folds. In the majority of studies of SSRI function in adult hippocampal neurogenesis in rodents, chronic rather than acute administration of SSRIs enhances the proliferation of NSPCs, increases the survival of adult-born neurons, and accelerates the maturation of immature neurons (Malberg et al., 2000; Santarelli et al., 2003; Wang et al., 2008).

Although the higher extracellular level of 5-HT induced by chronic SSRI administration enhances adult neurogenesis, some different conclusions have been drawn from individual genetic mouse models with dysfunctional 5-HT systems. Here, we focus on conclusions from (1) mouse models with altered central 5-HT levels from embryonic stages, (2) mouse models with deletion of 5-HT receptors from embryonic stages, and (3) mouse models with an altered central 5-HT system exclusively in adulthood. We have summarized recent findings to better understand the complicated roles of the central 5-HT system in adult hippocampal neurogenesis.

5-HTergic neurons are differentiated from progenitor cells at embryonic day (E) 10.5–11.5, and serotonin is synthesized around E12.5 in mouse brain (Gaspar et al., 2003; Suri et al., 2015). Thus, conventional or “non-inducible” conditional genetic mouse models targeting the genes expressed in 5-HTergic neurons would lead to changes in 5-HT levels from embryonic stages.

5-HT is released to synaptic cleft and functions after binding to its receptors. Thus far, 14 types of 5-HT receptors have been identified (Barnes and Sharp, 1999). The molecular and pharmacological features of 5-HT receptors have been reviewed elsewhere (Hoyer et al., 2002) and are not the major point of this review. How these receptors regulate adult neurogenesis is studied widely by acute or chronic administration of their agonists and antagonists. A recent review has summarized the effects of drugs on adult neurogenesis (Alenina and Klempin, 2015). Here, we focus on the data obtained from genetic mouse models of 5-HT receptors. Most of the genetic mouse models are conventional KO mice with total rather than specific deletion of 5-HT receptors.

5-HTR1A acts as both an autoreceptor and a heteroreceptor in mediating 5-HT function. 5-HTR1A inhibits firing of 5-HTergic neurons as an autoreceptor. It is highly expressed in many brain regions, including dorsal raphe nucleus and hippocampus. Most studies showed that the agonists of 5-HTR1A promote proliferation of NSPCs or survival of newborn neurons in adult hippocampal neurogenesis (Alenina and Klempin, 2015). Baseline level of adult neurogenesis is not affected in conventional 5-HTR1A KO mice, although increased proliferation of NSPCs in hippocampus induced by fluoxetine is blocked in these mice (Santarelli et al., 2003). A recent study found that higher proliferation of hippocampal NSPCs in mice housed in enriched environment is blocked in 5-HTR1A KO mice (Rogers et al., 2016). Different region-specific and time-controlled 5-HTR1A-deficient mice have been generated to investigate the function of 5-HTR1A in behaviors associated with mood and so on (Donaldson et al., 2013). More information and a precise conclusion about the function of 5-HTR1A in adult neurogenesis could be obtained from these genetic mouse models. Like 5-HTR1A, 5-HTR1B also acts as both an autoreceptor and a heteroreceptor. To date, no research has been performed to address adult neurogenesis in 5-HTR1B single KO mice. Double KOs of 5-HTR1A and 1B receptors, however, lead to decreased survival of adult-born neurons rather than proliferation of NSPCs in DG (Xia et al., 2012).

Similar to 5-HTR1A, baseline level of proliferation of NSPCs and survival of adult-born neurons is not affected in 5-HTR2B KO mice. However, the increased proliferation and survival induced by chronic fluoxetine administration is blocked in 5-HTR2B KO mice (Diaz et al., 2012). 5-HTR3 is the only ionotropic receptor in the 5-HT receptor family. The baseline level of proliferation of NSPCs and survival of adult-born neurons are not affected in 5-HTR3 KO mice; however, elevated adult neurogenesis and antidepressant-like behaviors induced by exercise is blocked in such mice (Kondo et al., 2015).

For 5-HTR4, its agonist has rapid anxiolytic–antidepressant effects (Mendez-David et al., 2014) and stimulates adult neurogenesis, but enhanced hippocampal neurogenesis is not required for these effects. In 5-HTR4 conventional KO mice, the progenitors and immature neurons are not affected at the baseline level, although the increase of these two types of cells induced by fluoxetine is blocked (Imoto et al., 2015). In this study (Imoto et al., 2015), the authors showed that 5-HTR4 is expressed in mature granule cells rather than in immature neurons in the hippocampus, which indicates an indirect involvement of 5-HTR4 in adult hippocampal neurogenesis. For 5-HTR7 receptor, the proliferation of NSPCs in the hippocampus is not changed in KO mice (Sarkisyan and Hedlund, 2009). Until now, adult neurogenesis has not been addressed in some 5-HT receptor KO mice, such as 5-HTR1D, 1E, 1F, and so on. Table 2 summarizes the adult neurogenesis in 5-HT receptor conventional KO mice.

5-HT receptors may have synergic effects to modulate adult neurogenesis. Additionally, in conventional KO mice with the defined deletion of a 5-HT receptor, the activity of other 5-HT receptors may compensate. Therefore, further study may need to investigate the adult hippocampal neurogenesis in multiple-gene KO mice with deletion of several receptors simultaneously.

Different receptors may play different roles in regulating adult hippocampal neurogenesis: some receptors contribute to the baseline level of adult hippocampal neurogenesis, whereas others are required for excise- or enriched environment-induced hippocampal neurogenesis.

In conventional or non-inducible KO mice, a change in the 5-HT level or 5-HT receptors begins in the embryonic stages, and as mentioned, developmental defects or possible compensations present in these mice interfere with conclusions obtained from these mice. Therefore, time-controlled inducible conditional KO mice with normally developed brains are required to address the roles of 5-HT system in adult hippocampal neurogenesis. In our recent work, time-controlled inducible genetic mouse models are generated with central 5-HT deficiency exclusively from adulthood. Three mouse models with time-controlled 5-HT deficiency have been used: (1) Pet1-Cre; Rosa26-DTR (diphtheria toxin receptor) mice obtained by crossing Pet1-Cre with Rosa26-DTR mice in which 5-HTergic neurons are depleted after diphtheria toxin (DT) is injected (Jia et al., 2014); (2) Pet1-CreER^T2; Rosa26-DTA mice generated by crossing Pet1-CreER^T2 with Rosa26-DTA mice in which 5-HTergic neurons would be deleted after tamoxifen is injected; and (3) Pet1-CreER^T2; Tph2^flox/flox CKO mice are generated by crossing Pet1-CreER^T2 with Tph2^flox/flox mice in which central 5-HT synthesis are blocked after tamoxifen is injected (Song et al., 2016a).

In these three mouse models, central 5-HT levels are reduced specifically after administration of DT or tamoxifen in adulthood, and these levels also can be controlled by the drug dosage (Song et al., 2016a). We found that adult hippocampal neurogenesis is enhanced as shown by increased proliferation of NSPCs and survival of newborn neurons in the DG of DT-injected Pet1-Cre; Rosa26-DTR mice (PC/DTR mice) and high-dose tamoxifen-administrated adult Pet1-CreER^T2; Rosa26-DTA mice (hTM-DTA^iPet1 mice) with a loss of about 95% of central 5-HTergic neurons (Jia et al., 2014; Song et al., 2016a). In contrast, the loss of half of the central 5-HTergic neurons in adult Pet1-CreER^T2; Rosa26-DTA mice (lTM-DTA^iPet1 mice) with a low dose of tamoxifen administration does not lead to significant alterations of adult hippocampal neurogenesis (Song et al., 2016a). In Pet1-CreER^T2; Tph2^flox/flox CKO mice with tamoxifen administration in adulthood, however, 5-HT level are estimated in 20% of controls, and the progenitors and immature neurons in the DG also are increased but to a lesser extent than in hTM-DTA^iPet1 mice (Song et al., 2016a). Thus, lowering central 5-HT to an extremely low level also can enhance adult hippocampal neurogenesis in mice with normally developed brains. This result, in part, is consistent with a previous study demonstrating that survival rather than proliferation is enhanced in mice chronically treated with para-chlorophenylalanine (PCPA), an inhibitor of 5-HT synthesis (Diaz et al., 2013). In addition, we also found that the dendritic length of adult-born neurons is increased in DT-treated adult PC/DTR mice (Song et al., 2016b). Combined with data from administration of SSRIs, we conclude that in a normally developed brain, hippocampal neurogenesis is enhanced in two opposite conditions of central 5-HT levels, extremely low or high. It is possible that an extremely low or high central 5-HT level activates different 5-HT receptor combinations, which results in similar phenotypes of adult hippocampal neurogenesis.

Chronic administration of SSRIs, which inhibits the function of Sert and therefore augments extracellular 5-HT level, promotes adult hippocampal neurogenesis. Until now, however, there is no inducible Sert KO mouse available with Sert-deficiency exclusively in adulthood. In mice with downregulation of Sert by RNA interference in adulthood, the proliferation of NSPCs and the number of immature neurons are increased in the DG (Ferres-Coy et al., 2013). Because of the limitations of in vivo RNAi, time-controlled Sert conditional KO mice by crossing Sert-floxed mice (Chen et al., 2015) with Pet1-CreER^T2 would be helpful to address the effect of higher extracellular 5-HT on adult hippocampal neurogenesis. Table 1 provides a summary of the mouse models discussed here.

Collectively, central 5-HT regulates adult neurogenesis in different ways, as shown by data from mice with dysfunctional 5-HT from embryonic stages or adulthood. The function of SSRIs on depression- and anxiety-like behaviors show age dependence (Ansorge et al., 2008; Homberg et al., 2010; Olivier et al., 2011), which may be caused by the broad and transient expression of Sert in the developing brain (Narboux-Neme et al., 2008; Homberg et al., 2010). Increased neurogenesis by SSRIs, however, is reduced or blocked in aged mouse models (Couillard-Despres et al., 2009), suggesting that aging is a key factor affecting adult hippocampal neurogenesis induced by experimental manipulations.

In Tph2 conventional KO mice, the proliferation of NSPCs in the DG is normal at young and adult age, but it is increased at aged stage (Klempin et al., 2013). Similar phenotypes are also present in Tph2 and Lmx1b conditional KO mice in which these genes have been deleted in the embryonic stage (Song et al., 2016b). Additionally, an increase in the dendritic length of newborn neurons is present in aged rather than young adult Tph2 and Lmx1b CKO mice (Song et al., 2016b). As mentioned, this late-onset phenotype of enhanced proliferation of NSPCs also is observed in Sert KO mice (Schmitt et al., 2007). It is still unclear why these mouse lines with altered 5-HT systems from the embryonic stage have normal hippocampal neurogenesis at young and adult age but show enhanced neurogenesis when aged.

Up- or downregulation of 5-HT system function does not simply cause opposite effects on adult hippocampal neurogenesis. 5-HT may function in different ways in regulating adult neurogenesis with altered 5-HT system from embryonic stages or from adulthood. Future studies are needed to explore the multifaceted roles of central 5-HT on adult hippocampal neurogenesis.

N-NS, YH, XY, BL, Y-QD, and LZ wrote the paper.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.