MOTHERHOOD, STRESS, AND SEROTONIN RECEPTORS – INFLUENCE ON POSTPARTUM SOCIAL AND AFFECTIVE BEHAVIORS IN FEMALE LABORATORY RATS By Erika Vitale A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Psychology---Doctor of Philosophy 2020 MOTHERHOOD, STRESS, AND SEROTONIN RECEPTORS – INFLUENCE ON POSTPARTUM SOCIAL AND AFFECTIVE BEHAVIORS IN FEMALE LABORATORY ABSTRACT RATS By Erika Vitale Mammalian mothers show a unique suite of behavioral responses beginning around the time of parturition that are necessary for successful rearing of young. These include caring for offspring, high levels of aggression, and low anxiety. These behaviors emerge in response to the unique neurochemical milieu resulting from pregnancy and parturition. Studies in this dissertation test the hypothesis that changes in receptors for the neurotransmitter serotonin (5- HT) are part of this neurochemistry in female laboratory rats. Experiments in chapter one found that there are reproductive state-dependent changes in expression of central 5-HT receptors that may be responsible for peripartum behavioral responses. Specifically, females examined at parturition and early lactation showed less serotonin 2C receptor (5-HT2C) mRNA expression in the midbrain dorsal raphe (DR), more serotonin 2A receptor (5-HT2A) mRNA in the medial preoptic area (mPOA), and more serotonin 1A (5-HT1A) mRNA in the shell subregion of the nucleus accumbens (NAcSh) compared to nulliparous females. Receptor autoradiography confirmed that binding density of 5-HT2A was higher in the mPOA of recently parturient females and that binding density of 5-HT1A in the NAcSh was higher in lactating females at particular rostrocaudal levels. Such differences in 5-HT receptor expression were not found in maternally acting virgin females, suggesting that pregnancy and parturition are necessary for these changes in central 5-HT receptors to occur. Because pregnancy stress derails most behavioral adaptations of motherhood, follow up experiments then explored whether the normative changes in 5-HT receptor expression across reproduction were prevented by daily application of mild-to-moderate stress beginning one week after mating. Stressed females showed lower maternal care and higher depression-like behaviors, which were correlated with 5- HT receptor mRNA in the mPOA, NAcSh and DR. Autoradiographic binding density of mPOA 5-HT2A receptors was not affected by pregnancy stress, although the stress reduced 5-HT1A binding in the NAcSh. Because the NAcSh is involved in motivation and reward processing, the last experiment directly tested whether 5-HT1A receptors in the NAcSh contribute to maternal caregiving and emotional behaviors. Long-term knock down of 5-HT1A in the NAcSh was established using an adeno-associated virus promoting shRNA against 5-HT1A mRNA. The 5- HT1A shRNA vector or a scrambled control vector was infused into the NAcSh during early pregnancy and mothers’ later postpartum social and affective behaviors (i.e. caregiving, maternal motivation, aggression, anxiety- and depression-like behaviors) were observed. 5-HT1A knock down resulted in higher frequencies of self-grooming and sleeping away from the nest, delayed retrieval of displaced pups back to the nest, and increased anxiety-like behavior. Overall, I found that female reproduction is associated with changes in serotonin receptor expression in numerous brain sites involved in postpartum behavior. Of particular interest, the normative change in 5-HT1A expression in the nucleus accumbens shell is altered in response to stress during pregnancy, and disrupting its expression reduces maternal motivation and increases postpartum anxiety-like behavior. Together, the results from this dissertation provide new insights into how the serotonergic system contributes to postpartum social and affective behaviors and offer a potential mechanism via the brain’s reward system through which pharmacological treatments that affect the serotonin system (e.g., SSRIs) may work to alleviate postpartum affective disorders in women. Copyright by ERIKA VITALE 2020 ACKNOWLEDGEMENTS I would first like to thank my advisor, Dr. Joe Lonstein, who has been incredibly supportive during my six years at Michigan State University. He was always excited to hear about my new data, and to listen to my new ideas even when they had nothing to do with what would be my dissertation projects. He knew when to be hands on and when to leave me on my own to figure things out, and he never doubted my ability to solve any problem I came across. I truly would not be the scientist, or the person, I am today without him. I would also like to thank my committee members – Drs. Cheryl Sisk, Lily Yan, and Alexa Veenema. Cheryl and Lily have been on several of my committees for various milestones throughout my time in the BNS program, and they have been very supportive throughout my journey, and so excited for me every time I passed a new milestone. Alexa came to the program just in time to be part of my dissertation committee, and I’m extremely glad for her kind words, support, and feedback, and for putting up with me lying on the floor in her office with her dog. I would also like to thank both past and current lab mates – Dr. Zach Grieb, Taryn Meinhardt, Dr. Christina Ragan, and Dr. Allie Holschbach. Zach and I overlapped for four and a half years in the lab, and he has become such a great friend over the course of us singing in lab together, laughing at YouTube videos, and arguing over lab equipment. I also had two rock star undergraduate research assistants – Maggie Ahern and Emma Ford – who helped me tremendously on my projects. I’m so proud of their work and wish them both well in Medical School. Katrina Linning was also instrumental in training me in research and lab maintenance, and has also been a great friend, mentor, and someone I look up to on a daily basis. I have also made so many lifelong friends in the Psychology department and the Neuroscience program. v While I am sad to be parting ways with them, I look forward to maintaining these close friendships and working with them as colleagues as we become the next generation of scientists. I would like to give a special thanks to my partner, Rachael Goodman-Williams, for her incredible support not only throughout my graduate school experience, but also in life. Graduate school has been the first of many adventures that we will share together, and I am so happy you will be by my side forever. Finally, I would like to thank my mother, sister, and grandmother for their continued love and support. They have been my loudest cheerleaders and my biggest fans (well, probably tied with Rachael), and I could not have achieved this without them. vi TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES CHAPTER 1: INTRODUCTION Overview of Midbrain Raphe Nuclei and Serotonin 1A, 2A, and 2C Receptors Role of Serotonin in Social Behaviors Laboratory Rodents In humans Serotonin and Steroid Hormones Serotonin and Postpartum Behaviors Serotonin and Affective Behaviors In non-human mammals In humans Serotonin, Stress, and Postpartum Affective Disorders Overview of Dissertation and Experiments CHAPTER 2: EXPRESSION OF FOREBRAIN AND MIDBRAIN SEROTONIN 1A, 2A, AND 2C RECEPTORS ACROSS FEMALE REPRODUCTION Experiment 2a – Effects of Female Reproductive State on Serotonin 1A, 2A, and 2C mRNA Expression Methods Subjects Real Time RT-PCR Data Analysis Results Experiment 2b – Effects of Maternal Sensitization on Serotonin 1A, 2A, and 2C Receptor mRNA Expression Methods Subjects Ovariectomy Maternal Sensitization Sacrifice and Tissue Processing Real-Time PCR Data Analysis Results Experiment 2c – Effects of Female Reproductive State on Serotonin 1A and 2A Receptor Autoradiograph Binding Density Methods Subjects Receptor Autoradiography vii x xi 1 21 2 3 4 7 10 12 15 15 16 18 20 26 26 26 26 28 28 33 33 33 33 34 35 35 36 36 38 38 38 38 50 Data Analysis Results Discussion Reproductive State Influences 5-HT Receptors in the DR Reproductive State Influences 5-HT Receptors in the mPOA Reproductive State Influences 5-HT Receptors in the NAcSh Conclusion CHAPTER 3: EFFECTS OF PREGNANCY STRESS ON POSTPARTUM SOCIAL AND AFFECTIVE BEHAVIORS AND BRAIN SEROTONIN RECEPTOR BINDING DENSITY Experiment 3a – Effects of Repeated Variable Stress During Pregnancy on Postpartum Caregiving and Affective Behaviors Methods Subjects Repeated Variable Stress (RVS) Maternal Caregiving Behavior and Retrieval Testing Anxiety-like behaviors Maternal Aggression Saccharin Preference Test Forced Swim Test (FST) Sacrifice and Blood Collection Data Analyses Results Maternal Caregiving Behavior Pup Retrieval Anxiety-like behavior Maternal aggression Depression-like behavior in the saccharin preference and forced swim tests Chapter 3b – Effects of Pregnancy Stress on Postpartum Serotonin 1A and 2A Receptor Binding Methods Subjects Repeated Variable Stress (RVS) Tissue Processing and Receptor Autoradiography Data Analyses Results Discussion CHAPTER 4: EFFECTS OF 5-HT1A RECEPTOR KNOCKDOWN IN THE NUCLEUS ACCUMBENS SHELL ON POSTPARTUM SOCIAL AND AFFECTIVE BEHAVIORS 82 Methods Subjects Surgery and Viral Vector Injection Maternal Behavior and Retrieval Testing Anxiety-like behaviors 39 40 42 42 43 47 49 54 54 54 54 56 57 58 58 59 59 60 60 60 61 61 62 62 70 70 70 70 70 71 72 74 85 85 86 86 87 viii Maternal Aggression Saccharin Preference Test Forced Swim Test (FST) Sacrifice and Tissue Collection Immunohistochemistry for Green-Fluorescent Protein In vivo determination of 5-HT1A receptor knockdown Results NAcSh infusions and degree of knockdown Maternal and Litter Health Maternal Behavior Anxiety-Like Behavior Maternal Aggression Depression-like Behavior Discussion CHAPTER 5: GENERAL DISCUSSION Overall summary of findings Future directions and relevant considerations Hormonal contributions to 5-HT receptor expression Characterization of 5-HT1A receptor-expressing neurons in the NAcSh 5-HT1A receptors in the approach/avoidance model of maternal behavior – implications for postpartum affective behaviors REFERENCES 88 89 89 90 90 91 93 93 93 94 95 95 95 105 112 114 114 117 119 112 122 ix LIST OF TABLES Table 1: 5-HT receptor mRNA in various brain sites across reproductive stages. Table 2: Incubation conditions for the radioligands used in Experiment 2c. Table 3: Sample schedule of RVS stressors. Table 4: Effects of RVS during pregnancy on dam and litter health across lactation. Table 5: Frequency (Mean ± SEM) of maternal behaviors displayed by non- stressed and stressed dams during two 30 min observations each day on postpartum days (PPD) 1-8. Table 6: Anxiety-like behaviors, maternal aggression behaviors, and depression- like behaviors (Means ± SEMs) in control and stressed dams. Table 7: Effects of 5-HT1A knockdown during pregnancy on dam and litter health across lactation. Table 8: Frequency (Mean ± SEM) of maternal behaviors displayed by 5-HT1A KD dams and scramble injected dams during two 30 min observations each day on postpartum days (PPD) 1-12. Table 9: Anxiety-like behaviors and depression-like behaviors (Means ± SEMs) in 5-HT1A knockdown and control injected dams. Table 10: Maternal aggression related behaviors (Means ± SEMs) in 5-HT1A knockdown and control injected dams. 30 32 63 64 66 68 97 99 102 103 x LIST OF FIGURES Figure 1: 5-HT receptor mRNA in brain sites across female reproductive states. Figure 2. 5-HT receptor mRNA in brain sites in maternally sensitized and non- sensitized females. Figure 3. 5-HT receptor binding in the NAc and mPOA across female reproductive states. Figure 4: Schematic representation of experimental timeline used to determine the effects of repeated variable stress during pregnancy on postpartum caregiving and affective behaviors. Figure 5: Effects of repeated-variable stress on maternal caregiving behaviors. Figure 6: Effects of RVS during pregnancy on anxiety-like behavior, maternal aggression, and depression-like behavior. Figure 7: Nucleus accumbens 5-HT1A and medial preoptic area 5-HT2A receptor binding density in non-stressed and stressed dams. Figure 8: Schematic representation of experimental timeline used to determine the effects of 5-HT1A knockdown in the nucleus accumbens shell on postpartum caregiving and affective behaviors. Figure 9: 5-HT1A mRNA and GFP immunoreactivity in the nucleus accumbens shell. Figure 10: Effects of 5-HT1A knockdown in the nucleus accumbens shell on maternal caregiving and maternal motivation. Figure 11: Effects of 5-HT1A knockdown on anxiety-like behavior and maternal aggression. Figure 12: Effects of 5-HT1A knockdown on postpartum depression-like behaviors. 31 37 40 62 65 67 72 95 96 98 100 101 xi CHAPTER 1: INTRODUCTION Mammalian mothers show a unique suite of behavioral responses beginning around the time of parturition that are critical for successful reproduction. These include high maternal caregiving of their young, maternal aggression against intruders to the nest, and a dampening of anxiety-related behaviors that prevent over-reactivity to threats (Bridges, 2015; Fleming & Luebke, 1981; Lonstein et al., 2014). Under most conditions, these behavioral changes associated with motherhood initially depend on the hormonal fluctuations associated with pregnancy. In rats and a number of other mammals, these include a gradual rise in circulating estradiol across pregnancy that peaks at parturition, high levels of progesterone that peak at the end of pregnancy followed by a steep drop during parturition, and sharp bursts of oxytocin and prolactin during parturition (Bridges, 2015; Lonstein et al., 2014). Interfering with this hormonal profile severely disrupts females’ ability to display caregiving behaviors (Bridges et al., 1978), indicating that this hormone combination is necessary for maternal caregiving to occur. After the typical onset of caregiving, peripheral estradiol and progesterone appear to instead facilitate the waning of caregiving behavior as the postpartum period progresses, as postpartum removal of the ovaries increases licking and upright-crouched (i.e., kyphotic) nursing at the end of lactation when they should be declining before weaning (Grieb et al., 2017). In contrast, postpartum oxytocin (OT) continues to facilitate the display of postpartum caregiving behaviors such as retrieving pups to the nest (Okabe et al., 2017), whereas blocking OT signaling in various brain regions significantly disrupts this and other maternal behaviors (Bosch et al., 2005; Bosch et al., 2012; Grieb and Lonstein, under review; Pedersen et al., 1994; Van Leengoed et al., 1987). While ovarian and peptide hormones along with interactions with offspring also modulate a host of neurotransmitters - including dopamine, norepinephrine, and GABA - to 1 maintain mothers’ postpartum social and emotional behaviors (Lonstein, 2007; Pederson et al., 1994; Numan and Stolzenberg, 2009; Rosenblatt et al., 1988), little research has examined any role for serotonin (5-HT), a neurotransmitter that is well known to influence social and affective behaviors in non-parous animals (reviewed in Graeff et al., 1996). Overview of Midbrain Raphe Nuclei and Serotonin 1A, 2A, and 2C Receptors In the mammalian brain, serotonin is synthesized from the amino acid, tryptophan (TPH), in cell bodies located in the midbrain and hindbrain raphe nuclei. Five nuclei residing in the hindbrain comprise the descending serotonergic pathway that mainly goes to the spinal cord, and these include the nucleus raphe obscurus (NRO; B2), nucleus raphe pallidus (NRPa; B1 and B4), nucleus raphe magnus (NRM; B3), neurons in the ventrolateral medulla (B1 and B3), and the area postrema. Four raphe nuclei, mainly residing in the midbrain, send ascending serotonergic projections to the forebrain, and these include the caudal linear nucleus (CLN; B8), median raphe nucleus (MRN, B8 and B5), a group of neurons just dorsal to the medial lemniscus, and the dorsal raphe nucleus (DR; B7 and B6). The DR is the largest collection of forebrain-projecting 5- HT neurons, with the DR of the rat brain containing 11,500 neurons that synthesize serotonin (Jacobs and Azmitia, 1992). The behavioral effects of central serotonin depend on the receptor subtype activated, and at least three (5-HT1A, 2A, and 2C) of the 15 different 5-HT receptors known to date have been implicated in social and affective behaviors in rats and in humans. 5-HT1A receptors are inhibitory, leading to activation of Gi protein and, therefore, a reduction in neuronal activity when activated (Raymond et al., 1999). 5-HT1A receptors are highly expressed in limbic regions including the hippocampus, amygdala, and cingulate cortex (Chalmers and Watson, 1991). These 2 receptors are also highly abundant in the DR, where they are found on terminals and somata of 5- HT neurons and regulate the release of serotonin in the forebrain (Hamon et al., 1991). The 5- HT2 family of receptors is excitatory, and activation of these receptors leads to Gq protein activation (Chang et al., 2000). 5-HT2A and 2C receptors are not as highly expressed in the DR, but can be found with particularly high density in the CA3 region of the hippocampus, frontal cortex, the amygdala, and the striatum (Pompeiano et al., 1993). 5-HT2A receptors are also found in the hypothalamus in greater amounts than are 5-HT2C receptors (Pompeiano et al., 1993). Based on the high abundance of these three receptors in cortical and limbic regions, it is unsurprising that they contribute to the display of various social and emotional behaviors in non- parous mammals (to be reviewed below). Role of Serotonin in Social Behaviors Serotonin has been widely implicated in numerous social behaviors in humans and non- human animals (Graef et al., 1996). Importantly, serotonergic neurons in the DR project to forebrain regions that are key regulators of prosocial and affective behaviors, including the medial prefrontal cortex (mPFC), the nucleus accumbens (NAc), the medial preoptic area (mPOA), the bed nucleus of the stria terminalis (BST), and the amygdala (Berk & Finklestein, 1981; Halberstadt & Balaban, 2006; Van Bockstaele et al., 1993). For instance, the mPFC regulates numerous reproductive behaviors in female rats, evidenced by the fact that excitotoxic lesions of it disrupt pup retrieval, pup licking, and proceptive sexual behaviors (Afonso et al., 2007). The mPOA is also critical for maternal caregiving, with lesions disrupting pup retrieval and nest building (Jacobson et al., 1980) and also reducing bar pressing in rats to gain access to pups (Lee et al., 2000). Regions of the limbic system, such as the dorsal BST and central and 3 basolateral amygdala, are highly involved in reproduction and also play a role in social investigation in juvenile male mice (Maaswinkle et al., 1996) and adult male and female rats (Dumais et al., 2016), social affiliation in male prairie voles (Lei et al., 2017), and social vigilance in female California mice (Duque-Wilckens et al., 2018). I will first discuss the specific role of serotonin on these behaviors in non-human mammals, and then turn to what is known about serotonin’s role in these processes in humans. Laboratory Rodents Given that most if not all brain sites regulating social behaviors receive serotonergic innervation, it is unsurprising that manipulating this neurotransmitter affects a large suite of social behaviors, including sexual behavior, social play behavior, and aggression, and this has (not surprisingly) been most studied in males. With regards to sexual behavior, serotonin is generally considered to inhibit copulation. For example, serotonin release in the lateral hypothalamus reduces sexual motivation and increases sexual satiety during the post-ejaculatory interval in male laboratory rats (Lorrain et al., 1999). Selective serotonin reuptake inhibitors (SSRIs), which increase central serotoninergic signaling after long-term use, also reduce sexual motivation in male rats (Matusczyk et al., 1998). Infusing serotonin directly into the mPOA, a brain site that regulates sexual behavior, also impairs male copulation (Verma et al., 1989). Conversely, reducing serotonergic activity by inhibiting synthesis of serotonin facilitates male sexual behavior by increasing ejaculatory activity (Kondo & Yamanouchi, 1997). Although these studies provide evidence for an inhibitory role of serotonin on copulatory behaviors themselves, serotonin release during sexual behavior may promote the rewarding properties of reproduction. In support, depleting serotonin reduces the rewarding properties of sexual experience in males, 4 as demonstrated by a reduction in conditioned place preference to a context paired with a sexual stimulus (Straiko et al., 2007). Additionally, serotonin in the nucleus accumbens, an important region for reward processing, is increased during male sexual activity (Ahlenius et al., 1987). Different serotonin receptor subtypes may mediate these seemingly contrasting effects of serotonin on the components of sexual activity. For example, activating inhibitory 5-HT1A receptors stimulates sexual behavior and decrease ejaculation latency in male rats, while also increasing mounting behavior in female rats that were administered testosterone (Haensel et al., 1991). In contrast, activating 5-HT1A receptors inhibits sexual behavior in female rats by reducing proceptive and receptive behaviors (Mendelson & Gorzalka, 1986). Serotonin acting on excitatory 5-HT2 receptors instead facilitates sexual behavior in female rats, since blocking these receptors inhibits sexual receptivity and lordosis (Mendelson and Gorzalka, 1985). In males, administering a non-specific 5-HT2 agonist, DOI, inhibits copulation by increasing ejaculation latency (Foreman et al., 1989; Klint & Larsson, 1995). Thus, serotonergic influence on reproductive behavior may be opposite in males and females, with the inhibitory 1A receptor promoting copulation males but the excitatory 2 receptors promoting it in females. Serotonin regulates non-sexual social behaviors as well, with 5-HT1B receptor antagonism in the NAc abolishing conditioned place preference for a socially paired context in mice (Dolen et al., 2013). Juvenile play (another highly motivating social behavior) is greatly reduced in rats with a genetic knockdown of the serotonin transporter (Homberg et al., 2007), and is reduced following acute SSRI treatment (Paksepp et al., 1987; Knutson et al., 1996). Central lesions of serotoninergic fibers using 5,7-DHT increases play in dominant rats but decreases play in non-dominant rats (Knutson & Paksepp, 1997). Mice lacking the serotonin transporter (SERT) gene also show disruptions in social behavior. Specifically, SERT knockout 5 mice are unable to distinguish between a novel and familiar mouse in a social recognition task, and basal firing rate of DR serotonin neurons is also significantly reduced (Veenstra- VanderWeele et al., 2012). Taken together, these studies provide support for the regulation of a variety of social behaviors by serotonin. Arguably the largest research effort has been given to studying the role of serotonin in male aggression, with the overall consensus indicating an inverse relationship between the two (Duke et al., 2013). This “serotonin deficiency” hypothesis of aggression is supported by negative correlations between trait-like impulsive aggression and/or violence with cerebrospinal fluid (CSF) concentrations of the serotonin metabolite 5-HIAA in humans and non-human primates (for reviews see Berman et al., 1997; Krakowski et al., 2003). Studies of non-human primates have shown that serotonin influences social behaviors along two axes – dominant/submissive and agnostic/affiliative. Experimental manipulations that lower serotonergic function are most often found to increase dominance, while those that increase serotonergic function lower dominance and increase affiliative behavior (Chamberlain et al., 1987; Raleigh et al., 1991). Additionally, dominant monkeys have higher blood platelet serotonin, a possible peripheral indication of neuronal 5-HT function, and platelet serotonin levels fall as dominance is lost (Raleigh & McGuire, 1991). Many studies in rodents have also demonstrated that reducing serotonin using neurotoxins for 5-HT neurons, or heightening brain serotonin by increasing tryptophan hydroxylase, the precursor to 5-HT, respectively increases or decreases aggressive behaviors in males (see Miczek et al., 2002 for review). Additionally, systemically administering 5-HT1A and 1B (inhibitory receptors) agonists or 5-HT2A and 2C (excitatory receptors) antagonists reduces aggression in male rodents (Blanchard et al., 1988; Lindgren & Kantak, 1987; Sijbesma et al., 6 1991; for additional review see Olivier, 2004). While the majority of studies support this conclusion, particularly those that incorporate brain-wide manipulations of serotonin, many studies fail to tease apart the differential pre- and postsynaptic effects of 5-HT receptor pharmacological manipulations. For example, systemically activating 5-HT1A receptors reduces male aggression, but peripherally injecting S-15535 (an agonist to somatodendritic 5-HT1A receptors but a competitive antagonist at 5-HT1A heteroreceptors), reduces serotonin release and reduces male aggression (de Boer and Koolhaas, 2005). Studies in hamsters also support an inhibitory role for serotonin on aggression (Ferris et al., 1999), although recent findings discovered a sex difference in the serotonergic control of aggression, with systemic injection of fluoxetine increasing the duration of aggression in females while significantly reducing aggression in males (Terranova et al., 2018). Additionally, activating 5-HT1A receptors in the anterior hypothalamus almost completely abolishes aggression in males, but significantly increases it in females (Terranova et al., 2018). Of note, no studies have explored a sex difference in serotonergic modulation of aggression in other laboratory rodents, even though sex differences in aspects of the serotonergic system have been known to exist since the 1980s (Carlsson & Carlsson, 1988; as discussed in more detail below) In humans Serotonin also influences social behaviors including social affiliation, sexual behavior, and aggression in humans, and seems to do so in a similar fashion as in non-human mammals. For example, volunteers treated with the SSRI, paroxetine, for one week showed increased social cooperation during a partner puzzle task compared to placebo administered volunteers (Knutson et al., 1998). Although social cooperation did not significantly differ between the SSRI and 7 placebo groups after 4 weeks of administration, plasma levels of paroxetine were positively correlated with social cooperation during this time point. Acute administration of tryptophan facilitates the recognition of happy facial expressions and reduces processing of disgusted faces (Murphy et al., 2006). Interestingly, these effects were only present in women, while tryptophan administration had no effect on facial recognition in men. Women given tryptophan supplementation also showed reduced startle reactivity to neutral, pleasant, and unpleasant pictures. Conversely, reducing serotonin levels by tryptophan depletion attenuates the attractiveness of positive faces, while also attenuating the negative intensity of threatening faces (Beacher et al., 2011). This suggests that reducing serotonin impairs the ability to perceive and interpret social stimuli. Acutely administering an SSRI also reduces steady-state visually evoked potentials, which is a measure of cortical processing, in the frontal and occipital cortices following visualization of unpleasant images (Kemp et al., 2004). Because acute SSRI administration causes short-term attenuations in serotonin release due to activation of raphe 5- HT1A autoreceptors (Czachura et al., 2000), these results suggest that reductions in central serotonin impair the cortical activation that accompanies the processing of social stimuli. Together, these results suggest that serotonin regulates the processing of socially salient stimuli across various contexts. Regarding sexual behavior, serotonin appears to have the same inhibitory effect in humans as it does in rodents. For instance, there have been many reports of low sexual motivation and arousal in men and women taking SSRIs (Aldrich et al., 1996; Montejo-Gonzalez et al., 1997). Several case studies have also reported erectile dysfunction, genital anesthesia, and ejaculatory anhedonia following chronic (4 months to 2 years) treatment with a variety of SSRIs (Reisman et al., 2017). Additional evidence for an inhibitory effect of serotonin on sexual 8 behavior in humans is that daily treatment with SSRIs such as sertraline and fluoxetine delays ejaculation in men who experience premature ejaculation (Waldinger & Olivier, 2004). SSRI influence on male ejaculation may be regulated by serotonergic mechanisms that differ from those that regulate depression, given that the antidepressant effects of SSRIs only appear after 4 to 6 weeks of chronic administration while the effects of SSRIs on male ejaculation can be seen at shorter treatment time points (Waldinger & Olivier, 2004). Overall, the inhibitory effects of serotonin on human sexual behavior appear to be similar to those found in laboratory rodents. As stated in the previous section, serotonin has historically been negatively correlated with aggression based on many studies showing that highly aggressive human subjects have low CSF levels of 5-HIAA, the metabolite of serotonin (Brown et al., 1979). Low serotonin is also associated with violent, impulsive, and suicidal behavior, as well as personality measures of hostility and aggression (Cleare & Bond, 1997; Cremniter et al., 1989). In addition, 5-HT1A receptor binding potential in the dorsal raphe nucleus is significantly negatively correlated with lifetime aggression score, and this finding was true for both men and women (Parsey et al., 2002). Since inhibitory 5-HT1A receptors are presynaptic in the DR, this suggests that higher inhibitory control over serotonin release has anti-aggressive effects, in contrast to the majority of studies indicating an inverse relationship between serotonin and aggression. It is important to note that in this study, though, all individuals were healthy and did not meet criteria for any psychiatric disorders and were not on any medications. Therefore, individual differences in serotonin release via 5-HT1A receptor expression may influence individual variation in aggressive tendencies differently than in individuals with psychopathological forms of aggression. 9 Serotonin and Steroid Hormones Although the review above indicates an important role for serotonin in social behaviors mostly studied in male mammals, serotonin may differentially contribute to these behaviors during female reproduction. The serotonergic system is heavily influenced by ovarian hormones, which are critical for the onset of maternal caregiving. As described earlier in this chapter, steroid hormones follow a particular pattern of release during the peripartum period in rats and numerous other non-primate mammals, involving a gradual rise in estradiol during the middle and end of pregnancy and a decline following parturition. Progesterone, on the other hand, stays high throughout almost all of pregnancy and rapidly declines right around parturition (Bridges, 2015). Many midbrain serotonergic neurons express estrogen and progesterone receptors (Alves et al., 1998; Nomura et al., 2005; VanderHorst et al., 2005), and these hormones are known to influence the serotonergic system of non-parous mammals; therefore, the hormonal changes associated with pregnancy, parturition, and lactation very likely influence the maternal serotonergic system. Indeed, estradiol injections increase TPH2 mRNA in the DR of ovariectomized virgin rats (Lu et al., 1999), and increase DR serotonin synthesis (Hiroi et al., 2011). Estradiol also increases the expression of 5-HT2A receptors in the NAc, the anterior frontal cortex, anterior cingulate cortex, and olfactory cortex of virgin female rats (Sumner and Fink, 1995), while it decreases 5-HT1A receptor expression in the amygdala and hippocampus (Osterlund et al., 2000). Chronic progesterone increases serotonergic activity within the brain of female macaques and guinea pigs by increasing TPH2 protein and decreasing 5-HT1A autoreceptor mRNA within the DR (Lu et al., 1999; Pecins-Thompson & Bethea, 1999; Lu & Bethea, 2002; Hiroi & Neumaier, 2009). Thus, elevations in estradiol and progesterone during 10 late pregnancy may produce behaviorally relevant increases in mother’s central serotonergic activity. Administration of either estrogen or progesterone also reduces serotonin transporter mRNA (Pecins-Thompson & Bethea, 1998), and reduces the serotonin degradation enzyme monoamine oxidase A in the brain of rhesus macaques and rodents (Gundlah et al., 2002; Smith et al., 2004), suggesting that once serotonin is released when these hormones are elevated, it remains in the synapse for a longer amount of time. Spontaneous firing rate of serotonin cells in the DR is also increased during the peripartum period, and this is due to the progesterone metabolite allopregnanolone (Robichaud & Debonnel, 2005). Interestingly, co-administration of estradiol and progesterone for 14 days significantly increases serotonin release within the hypothalamus, while administering either hormone alone does not have this effect (Lu et al., 1999). Therefore, the changes in estradiol and progesterone throughout pregnancy and parturition may together alter serotonergic activity in peripartum females. Either E alone or E+P also reduce expression of genes related to cell death in the DR, suggesting a neuroprotective effect of reproductive hormones on the serotonergic system, potentially during the peripartum period (Bethea et al., 2009). In fact, our lab recently found that cell death in the DR is relatively low when cells are born during late pregnancy, when both E and P are relatively high, compared to those that are born during the first week of lactation when E and P are much lower (Holschbach & Lonstein, 2018). Thus, ovarian and other hormones released across pregnancy, parturition, and lactation may profoundly influence serotonin release, serotonin receptor expression, and DR function in ways that help initiate and/or maintain the display of postpartum caregiving and affective behaviors. 11 Serotonin and Postpartum Behaviors Postpartum females show a unique collection of behavioral responses, including maternal care towards offspring, high maternal aggression, and low emotionality, and serotonin release in the larger maternal behavior neural network - including the NAc, mPOA, and BST (Lonstein et al., 2014; Numan and Insel 2003) - may play a role. During the first week of lactation, when mothering and aggression are high and anxiety is low, there is higher serotonin turnover in the mPOA and both the dorsal and ventral subregions of the BST compared to non-parous female rates (Lonstein et al., 2003; Smith et al., 2013). This pattern suggests that serotonin release in these regions may contribute to postpartum behaviors. Indeed, mice with life-long knockdown of tryptophan hydroxylase 2 (TPH2, rate-limiting enzyme for serotonin synthesis) show little to no postpartum maternal behavior or nest building, and offspring survival is greatly reduced (Angoa- Perez et al, 2014). Serotonin coming specifically from the DR appears to be important for lactation, because both pre- and postpartum lesions of the serotonergic cells in the DR reduce pup survival and decrease prolactin secretion in rats (Barofsky et al., 1983a; Barofsky et al., 1983b). Although serotonin is clearly necessary for caregiving, postpartum emotional behaviors, such as maternal aggression and anxiety-like behaviors, were not tested in these studies. A more recent study from our lab selectively lesioned the serotonin cells in the medial DR two days after parturition and found that lesions altered the temporal patterning of kyphosis (i.e., arched- back/upright crouched nursing), reduced pup licking, and prominently reduced maternal aggression. The same study found that DR serotonin lesions during pregnancy reduced maternal aggression to an even greater extent (Holschbach et al., 2018). This effect may be mediated by a reduction in serotonin fiber innervation of the hypothalamus, as the lesions reduced fiber length 12 in the anterior hypothalamus (Holschbach et al. 2018), where serotonin is known to influence aggression in both males and females (Terronova et al., 2018). Although it is clear that an intact midbrain serotonin system is necessary for normal mothering, there has yet to be an in-depth characterization across female reproduction of serotonin receptor expression in the DR or in its major projections sites. Such an exploration could help clarify what receptors are involved in the onset and maintenance of maternal caregiving and affective behaviors. Broadly, studies using peripheral drug injections have shown that central 5-HT1A, 2A, and 2C receptors contribute to postpartum caregiving behaviors in female rats. 5-HT2A and 2C receptors generally have opposing effects on mothering, with systemic blockade of 5-HT2A receptors during early lactation disrupting retrieval (Chen et al., 2014) while systemic activation of 5-HT2C similarly disrupts the behavior (Zhao et al., 2009). 5- HT2C activation also reduces arched back nursing (kyphosis). Of note, pup retrieval and kyphosis are not affected when a 5-HT2C receptor agonist is injected directly into the mPFC, NAc, or mPOA (Wu et al., 2016), suggesting that other brain sites are responsible for the behavioral effects of 5-HT2C activation, or that a collection of sites act together for 5-HT2C receptors’ effects. Supporting this possibility, systemic activation of 5-HT2C receptors during the first week of lactation decreases basal Fos expression in the DR (Wu et al., 2016), so 5-HT2C receptor activation might influence overall neuronal activity and, therefore, serotonin release from the DR. Because 5-HT2C receptors in the DR directly affect central serotonin release (Queree et al., 2009), it is possible that activating 5-HT2C receptors in the DR disrupts postpartum caregiving by affecting serotonin output to the forebrain. This is plausible because 5- HT2C receptors in the DR are found almost exclusively on GABAergic neurons (Serrats et al., 13 2004), and injecting a 5-HT2C agonist into the DR decreases 5-HT neuronal firing and increases Fos expression selectively in DR GABA neurons (Boothman et al., 2006). With regards to maternal aggression, serotonin appears to be facilitatory, which is in contrast to the majority of work in male aggression suggesting an inverse relationship between serotonin and aggression. Reducing overall serotonin output with serotonin-specific lesions to the DR significantly reduces maternal aggression towards a male intruder male (Holschbach et al., 2018). Pharmacological manipulations have implicated 5-HT1A, 2A, and 2C receptors in maternal aggression, with intracerebroventricular (i.c.v.) injection of 8-OH-DPAT, a 5-HT1A receptor antagonist, decreasing maternal aggression in rats. Furthermore, activating postsynaptic 5-HT1A receptors in the amygdala, dorsal periaqueductal gray, or median raphe also reduces maternal aggression (de Almeida & Lucion, 1994; de Almeida & Lucion, 1997). In contrast, 5- HT2A and 2C receptors can either increase or reduce maternal aggression depending on the site of injection- infusing a mixed 5-HT2A/2C agonist into the dorsal PAG reduces aggression in postpartum females, yet the same drug appears to increase maternal aggression when injected into the central amygdala (de Amleida et al., 2005; de Almeida et al., 2006). Thus, it is likely that these site-specific effects of 5-HT2A/2C agonists may depend on whether 2A or 2C receptors are predominantly expressed in particular brain regions. Reproductive-state changes in 5-HT receptor expression in particular brain regions may also explain the site-specific effects of these pharmacological manipulations; however, this has yet to be determined. 14 Serotonin and Affective Behaviors In non-human mammals While 5-HT and its receptors clearly influence postpartum caregiving and maternal aggression, they might do so indirectly by regulating postpartum affective behaviors. Serotonin depletion studies indicate that reducing central serotonin increases anxiety-like behaviors in laboratory rodents (Mosienko et al., 2012). Furthermore, genetic manipulations that knockdown expression of serotonin-related genes, such as those for the serotonin transporter (Holmes et al., 2003a; Holmes et al., 2003b) or 5-HT1A receptor (Ramboz et al., 1998; Toth, 2003; Zhuang et al., 1999) significantly alter serotonin release and affect anxiety-like behavior. In fact, 5-HT receptors have been common pharmaceutical targets for treating anxiety and depression, and downregulating 5-HT1A receptors is proposed to underlie the alterations in 5-HT output seen in individuals with major depression (reviewed in Drevets, 2007). In rats, activating 5-HT1A receptors in the DR reduces stress-induced anxiety-like behaviors (Kennett et al., 1987) and rats with a genetic “resistance” to depression-like symptoms (“Flinders resilient rats”) show higher 5- HT1A receptor density in many brain regions compared to controls (Nishi et al., 2008). 5-HT1A receptor binding in the cortex, CA1, and DR is also reduced by early life stress, a paradigm that induces later depression-like behaviors in rodents, and these alterations are reversed if offspring are given fluoxetine in adulthood (Leventopoulos et al., 2009; Sillaber et al., 2008; Czeh et al., 2006). The 1A receptor is not the only one relevant for affective behaviors, as 5-HT2C antagonists have potent, and immediate, antidepressant effects in mice by reducing immobility in the forced swim test, with a faster onset (five days) than that of SSRIs (several weeks) (Opal et al., 2014). 5-HT2C antagonists also increase brain derived neurotrophic factor (BDNF) in the 15 medial prefrontal cortex within a few days, which is also associated with chronic SSRI antidepressant action (Opal et al., 2014). This same study found that 5-HT2C blockade specifically in the ventral tegmental area (VTA) was sufficient to induce the same antidepressant effects and increased BDNF in the medial prefrontal cortex. In addition, combining selective serotonin reuptake inhibitor (SSRI) treatment with a 5-HT2C antagonist reduces the amount of time that SSRIs require for symptom relief in human patients suffering from anxiety or depression (Cremers et al., 2007). While forebrain 5-HT2C receptors have been implicated in the potent antidepressant effects of 5-HT2C antagonists, these receptors specifically in the DR also modulate affective behaviors. For instance, 5-HT2C receptor blockade reduces anxiety under some conditions by increasing serotonergic activity in the DR (Craige et al., 2015). Postsynaptic 5-HT2C receptors are highly localized on GABA neurons in the DR, and activating them increases GABA release. Importantly, injecting GABA directly into the DR reduces forebrain serotonin release and increases anxiety-like behaviors in mice, thus DR 5-HT2C receptors likely regulate anxiety-like behaviors by influencing GABA release in the DR (Xiao et al., 2017). Together, these results not surprisingly suggest that 5-HT receptors play an important role in maintaining 5-HT output, and alterations in 5-HT activity within the DR can dysregulate affective behaviors. In humans One of the first examples of serotonin’s involvement in human affective state was the lowering of mood after acute tryptophan depletion (ATD) (Young et al., 1985). Since then, approximately half of all published studies of ATD in healthy volunteers show a similar subclinical reduction in mood, characterized by self-reported boredom and reduced interest in 16 rewards (reviewed in Young & Leyton, 2002). Women report lowered mood following ATD more often than men (Murphey et al., 2006), which is consistent with the higher rate of mood disorders found in women compared to men (Bebbington, 1998; Kessler et al., 1993; Wolk and Weissman, 1995). Others have found sex differences in serotonin synthesis in humans (M>F) (Nishizawa et al., 1997; Sakai et al., 2006), central 5-HT1A receptor binding potentials (M