Cortical Influences on Cognitive and Respiratory

CORTICAL INFLUENCES ON COGNITIVE AND

RESPIRATORY DYSFUNCTION IN A MOUSE MODEL OF

RETT SYNDROME

CODY JAMES HOWELL

Submitted in partial fulfillment of the requirements

for the degree of Doctor of Philosophy

Dissertation Advisor: David M. Katz Ph.D.

Department of Neurosciences

CASE WESTERN RESERVE UNIVERSITY

May 2019 Case Western Reserve University

School of Graduate Studies

We hereby approve the dissertation of

Cody James Howell

candidate for the degree of Doctor of Philosophy

Committee Chair...... Heather T. Broihier, Ph.D.

Committee Member...... David M. Katz, Ph.D.

Committee Member...... Evan S. Deneris, Ph.D.

Committee Member...... Thomas E. Dick, Ph.D.

February 12th 2019

*We also certify that written approval has been obtained for any proprietary

material contained therein

Contents

List of Figures ……………...... 7

List of Tables ...... 10

Abstract...... 11

Preface…………………………....…………………………………………………….13

Chapter 1: Introduction...... 16

Rett Syndrome: Clinical Presentation...... 16

Methyl-CpG-binding protein 2 (MECP2) and Rett Syndrome...... 19

Mouse Models of Rett Syndrome...... 23

Activity Mapping in the Mecp2 Mutant Brain...... 31

The Medial Prefrontal Cortex...... 39

Chapter 2: Chronic Intermittent Treatment with Low-dose Ketamine

Reverses Dendritic Spine Deficits in a Mouse Model of Rett Syndrome…..55

Abstract...... 56

Introduction...... 57

Methods...... 59

Results...... 62

Development of dendritic spine defects in the mPFC in Mecp2 mutant mice…..62

Ketamine rapidly stimulates phosphorylation of rpS6...... 63

Ketamine rapidly stimulates dendritic spine growth in symptomatic Mecp2 mutant mice...... 64

Ketamine stimulates dendritic spine growth following repeated dosing in symptomatic Mecp2 mutant mice...... 65

Discussion...... 66

Chapter 3: Activation of the Medial Prefrontal Cortex Reverses

Cognitive and Respiratory Symptoms in a Mouse Model of Rett

Syndrome...... 77

Abstract...... 78

Significance Statement...... 79

Introduction...... 79

Methods...... 81

Results...... 86

DREADD expression in mPFC pyramidal neurons...... 86

DREADD-positive fibers project to brain regions important for respiratory control and cue-dependent fear memory...... 88

Increasing mPFC pyramidal neuron activity restores normal respiration in

Mecp2 mutants...... 89

DREADD activation of mPFC pyramidal neurons impacts downstream function in respiratory-related neurons...... 90

DREADD activation of mPFC pyramidal neurons restores long-term

Retrieval of auditory conditioned fear...... 92

Discussion...... 93

Chapter 4: Specific Activation of the mPFC Projection Circuitry targeting the Dorsal Midline Thalamus and Its Impact on

Cue-Dependent Fear Conditioning Deficits in a Mouse Model of

Rett Syndrome……………………………………………………………...……....107

Introduction……………………………………………………………..………....….108

Methods………………………………………………………….………....…………110

Results……………………….………....………………………………………..……112

DREADD expression in mPFC pyramidal neurons projecting to the dMT……..112

DREADD activation of mPFC pyramidal neurons projecting to the dMT enhances expression of LTM retrieval in Wt but not Mecp2 mutant mice……..112

Discussion..……………………….………....……………………………………….114

Chapter 5: Discussion and Future Directions………...……………………….119

Low-dose ketamine treatment in mouse models of Rett syndrome………....….120

Activation of the mPFC in mouse models of Rett syndrome………....………....128

Translational considerations………....………………………………………..……135

Bibliography…………………………………………………………………………140

List of Figures

Chapter 1

Figure 1. Schematic of the human MECP2 gene including the most common point mutations in associated with Rett syndrome...... 51

Figure 2. The Mecp2 mutant brain displays altered neuronal activity compared to wild-type………………………………………………...... 52

Figure 3. Model circuitry demonstrating the role of mPFC in respiration...... 53

Figure 4. Model circuitry underlying cue-dependent fear conditioning……………54

Chapter 2

Figure 1. Dendritic spine density and maturity is reduced on apical oblique dendrites in the mPFC Mecp2 Null mice compared to Wt at

6 weeks of age……………………………………………………………………….…70

Figure 2. Dendritic spine maturity is reduced on apical oblique dendrites in the mPFC Mecp2 Null mice compared to Wt at 3 weeks of age…………….…71

Figure 3. The ratio of phospho-S6 (pS6) to total S6 is markedly reduced in the Mecp2 Null mPFC compared to Wt and is rescued by acute treatment with ketamine…………………………………………………………….…72

Figure 4. Mushroom spine deficits on oblique dendrites in the Mecp2 Null mPFC are rescued by either single acute treatment or chronic intermittent treatment with ketamine 24 hours after dosing……………………….73

Supplemental Figure 1. Same data set as shown in Figure 4, including segments removed as statistical outliers……………………….………………...…74

Chapter 3

Figure 1. DREADD-Gq expression in the mPFC…………………..………………98

Figure 2. Activation of mPFC pyramidal neurons by DREADD-Gq increases neuronal activity in vivo…………………………………………….……100

Figure 3. Photomicrographs showing mCherry-positive fibers in coronal sections through brain regions important for respiratory control and cue- dependent fear memory following DREADD-Gq infection of the mPFC...……101

Figure 4. DREADD-Gq activation of pyramidal neurons in the mPFC eliminates the apneic breathing phenotype in Mecp2 mutants..………………..102

Figure 5. Activation of pyramidal neurons in the motor cortex by

DREADD-Gq does not alter respiration……………………………..……………..103

Figure 6. DREADD-Gq activation of pyramidal neurons in the mPFC reduces respiratory frequency variability to Wt levels without impacting

8 the average frequency of respiration.…………………….………………………..104

Figure 7. Activation of mPFC pyramidal neurons by DREADD-Gq impacts neuronal function in respiratory subnuclei of the nTS……………………………105

Figure 8. Activation of mPFC pyramidal neurons by DREADD-Gq rescues long-term expression of cue-dependent fear memory……………………………106

Chapter 4

Figure 1. Representative images from the mPFC and dMT demonstrating

CRE-dependent DREADD-Gq and CRE expression respectively…………..….116

Figure 2. Preliminary data suggests that activation of mPFC projection neurons that target the dMT increases LTM expression in Wt but not in

Mecp2 mutant mice………………………………..…………………………………117

Figure 3. Mecp2 Hets display increased CRE-dependent DREADD-Gq expression in the mPFC but reduced LTM retrieval response to stimulation of mPFC to dMT projection neurons………………………………..……………..118

List of Tables

Table 1. Raw values for dendritic spine density listed for each spine type in either 6-week old or 3-week old Wt or Null mPFC………………………….……75

Table 2. Raw values for dendritic spine density listed for each spine type in 6-week old Wt and Null mPFC treated with either saline or one of three ketamine paradigms…..……………………...………………………………………..76

Cortical Influences on Cognitive and Respiratory

Dysfunction in a Mouse Model of Rett Syndrome

Abstract

CODY JAMES HOWELL

A fundamental goal of neuroscience is to understand how behavior is controlled by specific neural circuitry, and how abnormalities in such circuity can lead to neuropsychiatric disorders. Autism Spectrum Disorders are of increasing global concern, and while much is known about the behavioral and cognitive deficits in these disorders, little is known about the underlying brain circuitry abnormalities.

This is due, in part, to the lack of good models used for basic science. Fortunately, mouse models with high face and construct validity do exist for several syndromic forms of Autism and can be used to provide insight into Autism Spectrum Disorders as a whole. To this end, my thesis has used a mouse model of Rett syndrome to ask fundamental questions about specific cortical circuits and their role in the abnormal behavior observed in this disorder.

My first study produced a map of activity abnormalities in the Mecp2 mutant which helped bring my focus to the medial prefrontal cortex (mPFC), a hub region that receives inputs from many regions and projects widely throughout the brain

11 allowing it to play a role in diverse behavioral domains. Subsequently, we demonstrated that the mutant mPFC displays reduced excitatory drive onto excitatory neurons and fewer dendritic spines on excitatory pyramidal neurons.

Therefore, my first study sought to further define deficits in pyramidal neuron dendritic spine density and maturity as a surrogate for excitatory synaptic connectivity. We found that the mutant mPFC displayed reduced spine density and abnormal spine morphology maturity that were reversed by the NMDAR antagonist ketamine. The second study in my thesis sought to determine whether or not hypofunctional pyramidal neurons of the mPFC contribute to behavioral deficits in core symptom domains of Rett syndrome. To answer this question, I used a virally mediated chemogenetic approach to increase activity specifically in excitatory neurons of the mPFC and used both cognitive and respiratory outcome measures to determine if activation of these cells could eliminate deficits in these core domains of Rett syndrome symptomology. Supporting the hypothesis that hypofunction in mPFC projection circuitry contributes to behavioral deficits, activation of these cells led to the elimination of respiratory apneas and restoration of long-term memory after cue-dependent fear conditioning.

Preface

Disorders of the central nervous system are a tremendous burden on not only the affected individual, but also on the families and society as a whole. The UN

World Health Organization reports that nearly 1 in 6 people suffer from some form of neurological disorder across the globe[1]. Of these, autism spectrum and autism related disorders impact 1 in every 59 births, and that rate continues to climb as better awareness and diagnostic criteria develop[1]. Autism Spectrum

Disorders, or ASDs, are neurodevelopmental disorders that are diagnosed early in life but are generally persistent through adulthood. At least 3.5 million

Americans live with an ASD, and autism is currently the fastest-growing developmental disability[1].

In Germany, 1966, Andreas Rett described a neurodevelopmental disorder that predominantly occurred in females and led to apparently normal babies developing an extremely debilitating lifelong condition[2]. 20 years later, Hagberg and colleagues gave the first English language presentation of Rett syndrome[3].

Rett syndrome, an autism related disorder, has come to be diagnosed around the world and includes all ethnic and racial groups. Rett syndrome is rare, but is one of the most severe neurological disorders in females. Rett syndrome also has a unique complexity: on one hand it appears as an ASD, but also has features of

Parkinsonism and metabolic disorders which make is incredibly difficult to treat[4].

In fact, no treatment currently exists for patients with Rett syndrome.

However, there has been a recent flurry of translational and clinical activity in the field. In fact, my own drive to understand the disorder truly began in earnest when I first met both the patients with Rett syndrome, along with their families.

We had gathered for the announcement of a new clinical trial beginning for patients with Rett syndrome, a trial that my work had helped motivate. These families were looking for hope; they knew that their daughters were trapped inside of a body controlled by Rett syndrome, but hoped that somehow, they would eventually be free. This belief by the parents of Rett children, that the symptoms of the disorder can be reversed, has also been supported by data from mouse models of Rett syndrome.

The solution to neurological disease, like any disorder, lies in understanding the causal mechanisms. With neurological disorders, treatment often develops from the understanding of which cells and circuits are involved in the pathophysiology of the disorder, discovery of the mechanisms of the dysfunction, and the development of tools to combat that dysfunction. For example, researchers studying Parkinson’s disease found that deep brain stimulation of the basal ganglia ameliorated some of the associated motor symptoms (reviewed in [5]). In this case, finding a locus of behavioral dysfunction, and understanding the circuitry involved, led to a treatment which gives symptom relief to many. The goal of this thesis is much the same, to understand mechanisms that underlie circuitry abnormalities related to behavioral deficits observed in Rett syndrome, and to search for a brain locus that underlies multiple behavioral deficits and may become a therapeutic target. It is the hope of the author that this information will

14 eventually be used to benefit patients with Rett syndrome, and to increase the quality of life for them and their families.

Chapter 1

Rett Syndrome: Clinical Presentation

Rett syndrome affects 1 in 10,000 females worldwide[1]. On the whole, the burden of care for families is enormous, and there are currently no treatments for children with Rett syndrome. Generally, de novo mutations arise in the father’s germ line resulting in the vast majority of Rett syndrome patients being female heterozygotes mosaic for normal and mutant MECP2[6, 7]. The majority of male

Rett syndrome patients die in utero or soon after birth (reviewed in [8]). Initial developmental milestones are reached until a regression beings between 1.5 and

3 years when girls lose progress with speech, motor capabilities, and social skills with the loss of spoken language skills being common and on the whole resembling autism[9, 10]. As regression continues, girls begin to develop issue with anxiety, respiratory control, seizures, and repetitive behaviors such as hand wringing. Rett syndrome generally reaches a plateau after the regression stage, and some acquired skills that were lost can begin to return. For example, some autism-like features may also begin to decrease in some girls, and patients often develop an intense eye gaze that they may use to communicate with their parents[11, 12]. Dysregulation of breathing and autonomic homeostasis is common in girls with Rett syndrome. Respiratory dysfunction includes periods of hyperventilation as well as long breath holds that can cause both cyanosis and loss of consciousness all of which is worse during periods of anxiety or overstimulation but improves during sleep[13-18]. This suggests that behavioral and emotional state plays a role in symptom severity. Abnormalities in autonomic

16 function are a critical component of Rett syndrome pathology given that around

25% of Rett syndrome patient deaths are sudden and may result from complications with cardiorespiratory dysfunction. Other autonomic abnormalities include decreased heart rate variability and prolongation of QT interval, abnormal sweating, and vasodilation problems often associated with cold extremities[19-21].

Motor deficits commonly include abnormal gait, dystonia, tremors, myoclonus and repetitive motor behaviors such as hand wringing, as well as teeth grinding and grimacing (reviewed in [4]). Rett syndrome patients also suffer from more global deficits in growth, nutrition and gastrointestinal function[22]. Rett syndrome patients may display altered bone growth including scoliosis, which often requires surgical intervention[23], and reduced head size resulting in microcephaly[3].

Height and weight are often diminished as well, though a subset of patients struggle with obesity[24, 25]. These phenotypes are also associated with eating and gastrointestinal problems which can begin with deficits in chewing and swallowing, difficulty keeping food down and issues with reflux, as well as common and severe constipation all of which may lead to abnormal nutritional intake and discomfort.

Because patients with Rett syndrome have deficits in both communication and motor skills, determining the degree to which intellectual ability is impacted is difficult or inaccurate[9, 26]. Even so, there are numerous reports of severe cognitive disability in Rett patients that seems to be arrested at the onset of regression around 1 year of age (reviewed in [27]). This cognitive disability is also associated with reduced communicative ability. Interestingly, several studies

17 have suggested increased social interaction with family members and caregivers that increase over time while object permanence, a measure of cognitive ability, is often severely impacted in patients with Rett syndrome[27]. This may suggest that specific types of memory are impacted by the loss of MeCP2 while others are spared. Social interaction and communication in Rett syndrome patients in the later stages of regression often takes the form of an intense eye gaze[28, 29].

Using this eye gaze-based communication, recent studies have found that while patients with Rett syndrome are capable of following instructions to recognize and categorize simple pictures, visual processing and visual memory were significantly lower in patients compared to controls (discussed in [30]). In addition to cognitive and memory deficits, patients with Rett syndrome also display abnormal emotional control and periods of heightened anxiety (reviewed in [31]).

As discussed above, shifts in emotional state can influence the severity of other symptoms such as respiration and repetitive behaviors. Given the importance of cognitive and emotional symptoms to the disorder, several experiments discussed in Chapters 3 and 4 were designed to model cognitive dysfunction in

Rett syndrome in a task with both emotional and memory components

(discussed further below).

Although Rett syndrome is associated with reduced head size, patients do not show evidence of neural degeneration, reductions in myelination, or a reduction in the number of neurons or glia[32, 33]. However, reduced total brain volume, smaller cell size, and reduced complexity of dendritic branching has been observed in regions such as the cortex and hippocampus[34]. Additionally, brain

18 derived neurotrophic factor (BDNF), a growth factor that plays critical roles in cell growth and synapse formation, is also reduced in the brains of Rett patients and is likely associated with the smaller size and complexity of dendrites[35].

Rett syndrome was classified as one of 5 autism-related conditions, along with

Asperger syndrome, in the DSM-4. However, since the discovery that Rett syndrome is caused by mutations in a single gene, the DSM-5 no longer includes

Rett syndrome as an ASD. This has caused much controversy and concern. On one hand, early symptoms of Rett syndrome look much like autism, but in many patients these autistic features wane as the children age. On the other hand, there is concern, particularly among parents who have children with Rett syndrome, that less support will be available now that Rett syndrome is no longer on the spectrum. Given that there was also controversy in 1994 when Rett syndrome was included in the DSM-4 as an autism related disorder, this debate over the classification of Rett syndrome is likely far from over[36, 37]. If nothing else, this speaks to the numerous phenotypic domains impacted, and the diverse nature of the Rett syndrome clinical presentation.

Methyl-CpG-binding protein 2 (MECP2) and Rett Syndrome

Much progress has been made recently on understanding the mechanisms underlying the pathophysiology of Rett syndrome. This progress began in earnest after early studies had proposed a genetic basis for Rett syndrome which was later confirmed with the identification of mutations in the gene which

19 encodes MeCP2[38]. MECP2, located on chromosome Xq28, encodes methyl-

CpG-binding protein-2, a transcription factor which is ubiquitously expressed in mammalian tissue[39, 40]. MeCP2 is a member of a family of proteins that bind

DNA at methylated CpG islands and modify transcriptional expression of their targets[41]. MeCP2 is comprised of at least three critical functional domains: the methyl-binding domain (MBD), the transcriptional repressor domain (TRD), and the NCoR/SMRT interaction domain (NID) and mutations in any one of these domains can lead to Rett syndrome[42-44]. MeCP2 is generally categorized as a transcriptional repressor due to its ability to bind methylated CpG islands in DNA through its MBD, then recruit transcriptional repressors such as Sin3a and histone deacetylase complexes that alter the structure of DNA using its TRD[42]; or it may recruit TBL1/TBLR1, components of the NCoR/SMRT corepressor complex, to the NID[43]. However, the view of MeCP2 as strictly a global repressor of transcription has been called into question multiple groups have found that the expression of many genes is reduced in the absence of MeCP2, which has led some to suggest that MeCP2 may be capable of also acting as a transcriptional activator[45, 46]. This hypothesis gains some support from the finding that MeCP2 can bind with the transcriptional activator CREB1, and that overexpression of MeCP2 leads to global gene transcription being increased compared to Wt[45]. A separate study found that the type of methylation at each gene influenced the likelihood that the loss of MeCP2 would lead to either ran up or down regulation of their mRNA[46]. Specifically, the most active genes had more 5-hydroxymethylcytosine methylation, as opposed to 5-methylcytosine

20 residues, and were more likely to be downregulated with the loss of MeCP2 even though MeCP2 binds both methylation marks[46]. Other studies have also suggested that post-translational modifications, specifically differential phosphorylation which affected the binding affinity of MeCP2, could also play a role in the diverse functions of MeCP2 as either repressor or activator[47].

Although the indirect evidence suggests that MeCP2 may have roles as both a repressor and an activator of transcription, no conclusive transcriptional assay has demonstrated activation by MeCP2 binding as of yet.

Rett syndrome is caused by loss of function mutations in MECP2, and more than

250 have been identified thus far; however, there are several common point mutations (R106W, R133C, T158M, R168X, R255X, R270X, R294X, R306C) that account for more than half of Rett syndrome cases, while large deletions of exons or the 3’ end account for around 15% (IRSF MECP2 Variation Database).

Before the widespread use of genetic testing, patients with mutations in other genes, such as FOXG1 and CDKL5, were often misdiagnosed with Rett syndrome[48]. Likewise, genetic testing has also led to more a clearly defined diagnosis for Rett syndrome versus other disorders with similar symptomology such as Angelman syndrome, general ASDs, and Phenylketonuria. However, the majority of individuals with MECP2 mutations do meet the criteria for a diagnosis of Rett syndrome which includes: a period of regression followed by recovery or stabilization, partial or complete loss of purposeful hand skills and spoken language, gait abnormalities or absence of ability, and stereotypic movement especially of the hands. Rett syndrome is diagnosed by meeting at least two of

21 these main criteria and 5 supporting criteria such as breathing disturbances, bruxism, abnormal muscle tone, scoliosis, and intense eye communication as is common among these patients. Beyond the diversity of mutations that cause Rett syndrome, variability in clinical presentation could also be caused by differences in patterns of X chromosome inactivation whereby a normal process in all female cells allows for the expression of genes from only one X chromosome. As such, girls with Rett syndrome would each display differences in the patterning of cells with normal or mutated MECP2, leading to the possibility that each patient may display a different degree of impairment in each phenotypic domain based on this patterning.

It is also important to note here that a small, and possibly under-represented, population of surviving males with MECP2 mutations does exist. Often, males with MECP2 dysfunction display overexpression of the gene[49-53]. However,

MECP2 loss of function mutations also exist in surviving males who may display non-specific cognitive delay, progressive motor problems, and autistic features[54]. These male patients often display Rett syndrome in combination with

Klinefelter syndrome (XXY) because, similar to female patients with Rett syndrome, two different X chromosomes exist and lead to somatic mosaicism[8].

Mouse Models of Rett Syndrome

Since Rett syndrome was found to be caused by mutations within a single gene,

MECP2, a variety of mouse models of the disorder have been developed with each reproducing the behavioral and molecular components of the human condition to varying degrees, which will be discussed here. Mouse Mecp2 alleles can generally be categorized as follows: 1. Null mutations, which produce no

MeCP2, are similar to the approximately 10% of Rett syndrome patients with large deletions (Mecp2tm1.1Bird, Mecp2tm1.1Jae, and Mecp2tm1Tam)[55-57]; 2. truncations and single-nucleotide changes that reproduce or at least approximate mutations that are more commonly found in Rett syndrome patients (Mecp2 308 truncation, R168X, T158A, R255X, etc.)[58-62]; and 3. mutations designed to probe

MeCP2 function more specifically such as hypomorphs for Mecp2 which express approximately half the wild-type (Wt) level of MeCP2 and have more modest impairments[63, 64], or Mecp2 overexpression mice which have both overlapping

(social impairment, altered learning and memory, seizures, and behavioral inflexibility) and non-overlapping phenotypes (increased dendritic spine stability and maturity rather than reduced) with the models for Rett Syndrome[51, 65].

Because Mecp2 Null alleles are the most commonly studied in mouse models of

Rett syndrome, and the Mecp2tm1.1Jae Null allele is the model used for the studies described within this thesis, this section will focus primarily on those phenotypes which have been reported in models with Null alleles in either male hemizygous

(Nulls) or female heterozygous (Hets) mice, but will touch on the other models where necessary. Most of these models display a wide range of phenotypes that

23 are similar to those seen in Rett syndrome patients and are worse in male Nulls than in female Hets including reduced lifespan motor and sensory impairments, cognitive abnormalities, respiratory and autonomic dysfunction, and cellular and synaptic deficits (reviewed in [4, 66]). The choice of using male hemizygous mice or female heterozygous mice must be considered in any study on mouse models of Rett syndrome. On one hand the male Null expresses no MeCP2 leading to no variability from cell to cell unlike in female mouse models with somatic mosaicism. On the other hand, female heterozygous mice better model the more common female heterozygous human patients. For the studies described in this thesis, both male and female Mecp2 mutants were used; male mice were needed to study synaptic structure in genetically define cells in Chapter 2, but female mice were used for all other experiments described in Chapters 3 and 4.

Mecp2 mutant males with Null alleles appear to develop normally until around the

4th week of age when Rett syndrome-like phenotypes start to become apparent.

These male Nulls exhibit shortened lifespan, with the Mecp2Jae strain often dying around 4-6 weeks of age, whereas female Hets often live full lives but do experience a higher rate of sudden death (discussed in [4]). Between the 4th and

6th weeks of life, male Nulls become progressively worse in motor, cognitive, and respiratory domains. Phenotypes appear and worsen with age including abnormal gate and posture[67, 68], splayed hind limbs as well as hind limb clasping when suspended[55-57, 59, 61, 62, 69, 70], hypoactivity[56, 57, 62, 71], tremors and seizures[55-59, 61, 62, 70, 72], and grooming deficits leading to disheveled fur

(reviewed in [4, 66]). These phenotypes are also present in female Hets, but

24 generally do not develop until at least 10 weeks of age, and then, exhibit greater variability likely due to somatic mosaicism[70]. Many of these phenotypes mimic human Rett syndrome, but some, such as hind limb clasping, are less obviously

Rett syndrome-like phenotypes. Although it is yet to be determined, hind limb clasping and abnormal grooming behaviors could be similar to repetitive motor behaviors observed in Rett syndrome patients.

Respiratory dysfunction is common in humans with Rett syndrome and mouse models of the disorder alike. Phenotypes in this domain include highly variable respiratory frequency with variation in respiratory cycle length as well as breath holds and periods of tachypnea (reviewed in [4]). Respiratory pauses occur in

Nulls by 5 weeks of age in both Mecp2tm1.1Bird and Mecp2tm1.1Jae lines and as early as 10 weeks of age in female Hets from the Mecp2tm1.1Jae strain, but as late as 16-56 weeks in females from the Mecp2tm1.1Bird strain (reviewed in [4]). Both male and female mutants also display exaggerated hypoxic ventilatory responses and vagal respiratory pauses[73-77]. These phenotypes are consistently observed on the various background and allele types (reviewed in [4]).

Of particular interest to the hypotheses tested in this thesis is the fact that patients with Rett syndrome[14, 78, 79] as well as Mecp2 mutant mice[80, 81] display respiratory abnormalities that worsen during behavioral arousal. Like patients, mouse models of Rett syndrome display apneas are respirator frequency abnormalities that are better during sleep and worsen when awake or during periods of stress[79, 81]. In addition, our lab previously described an exaggerated respiratory response to the novelty induced orienting response in Mecp2

25 mutants[80], a behavioral paradigm which uses mild novel tones and is known to transiently increase respiration in Wt controls[82]. The fact that Mecp2 mutants display abnormal respiration that worsens during states of arousal suggests an abnormal modulation of respiratory control that will be further discussed below, and again in Chapter 3.

Cognition disability is also a hallmark of Rett syndrome in human patients, but modeling cognitive abilities in mouse models poses obvious challenges. That being said, there are several cognitive phenotypes in mouse models that are reliable, consistently observed between diverse Mecp2 alleles and genetic backgrounds, and may provide insight into mechanisms underlying cognitive deficits in Rett syndrome. Cognitive tests that are of particular interest to this thesis work are those which are influenced by well-defined brain circuitry and may provide insight into aspects of the cognitive domain that are abnormal in

Rett syndrome. A good example of such a cognitive task are fear conditioning paradigms which test both cognition and memory for emotionally salient stimuli.

One form of fear conditioning, contextual fear conditioning, has been found to be abnormal in multiple mouse models of Rett syndrome including male Nulls by about 6 weeks of age[70]. Likewise, mouse models of human mutations, such as male Mecp2tm1.1Joez and Mecp2tm1Hzo mice, show contextual fear conditioning deficits by 10 and 20 weeks of age, respectively[58, 61, 83]. A similar paradigm, cue- dependent fear conditioning (CDFC), has also been found to be abnormal in

Mecp2 mutants. This paradigm pairs a previously innocuous stimulus with an inherently aversive stimulus. Our lab and others have demonstrated that Mecp2

Nulls display reduced long term memory in this paradigm[84, 85]. This model of cognition will be used in the experiments discussed in Chapter 3 and will be described in greater detail below.

Mecp2 Null mice, including both Mecp2tm1.1Jae and Mecp2tm1.1Bird alleles, also phenocopy the morphologic abnormalities observed in brains of patients with

Rett syndrome. Such phenotypes include reduced cortical thickness, smaller neurons which are more densely packed, and reduced dendritic arborization and complexity of excitatory pyramidal neurons of the cortex and hippocampus

(reviewed in [4, 66]). In addition to having less dendritic complexity, the dendrites that are present exhibit reduced dendritic spine density in regions such as the somatosensory and motor cortices in Mecp2tm1.1Bird mice[86, 87], the mPFC in

Mecp2tm1.1Jae mice[88], and the hippocampus in both strains (reviewed in [4]).

Although reduced dendritic spine density has been observed in multiple regions, backgrounds, and allele types, the specific dendritic abnormality seems to vary some by region. This will be discussed further in Chapter 2.

Studies of network excitability have, likewise, found regional specificity to the phenotypes observed such as decreased excitatory synaptic drive in cortical circuits[88-90] and increased excitability in the hippocampus[66, 91-94] and multiple brainstem regions involved in autonomic control (reviewed in [95, 96]; see also

[93] discussed in more detail below). Inhibitory influences may also play a role in the network excitability changes, and these also likely have regional specificity.

For example, the cerebellum of Mecp2tm1.1Jae mice has increased GABA positive terminals onto Purkinje cells, while the density of GABAergic terminals in the

27 somatosensory cortex and mPFC is unaffected[88, 97]. Intrinsic electrical properties are also known to be abnormal in several regions of the Mecp2 mutant brain, such as the locus coeruleus and the substantia nigra[98, 99] but normal in others such as the mPFC[88], and transmitter release has been shown to be abnormal in cultured hippocampal neurons and chromaffin cells[100, 101].

Studies of both long-term potentiation (LTD) and long-term depression (LTD), two forms of synaptic plasticity thought to underlie memory processes, have found abnormalities in both when analyzing population responses in brain slices from each of the Mecp2 Null models (reviewed in [102]). In particular, the hippocampus displays reductions in both LTP, observed at the CA3 to CA1 synapse[103], and reductions in LTD in the CA1 region[83, 103]. In addition, reduced

LTP has been observed in the somatosensory cortex layers II/III when studied at the population level[97]. However, other studies using monosynaptic approaches have found that the somatosensory cortex, layer V specifically, has intact LTP in

Mecp2Jae Null mice[90]. Therefore, further work is required to determine if the observed deficits in LTP and LTP are due to differences in the region and layer studied, or if these deficits reflect reductions in either weak or sparse synaptic connectivity as has been observed in regions such as the mPFC and somatosensory cortex (discussed in [88]).

Cell signaling pathways underlying synapse development, maintenance, and circuit activity are of growing interest in Rett syndrome. In particular, diverse mouse models reproduce the reduced levels of BDNF and biogenic amines found in the brains of patients with Rett syndrome, as well as some of the

28 associated signaling pathways such as the mammalian target of rapamycin

(mTOR)[35, 101, 104-107]. Importantly, BDNF and mTOR, a downstream target of

BDNF signaling, are known to play critical roles in synapse formation, development and excitability[108, 109]. Indeed, the severity of phenotypes in Mecp2 mutant mice is affected by the level of BDNF expression. Deletion of BDNF from

Mecp2 mutants leads to an earlier onset of Rett syndrome-like symptoms, and deletion of BDNF from otherwise Wt animals leads to several Rett syndrome-like phenotypes including reduced brain size and hind-limb clasping[110]. Moreover,

BDNF is known to modulate excitability directly at synapses such as the primary afferent synapse in the nTS where BDNF inhibits responses to glutamatergic excitation[111]. Deficits in BDNF are therefore likely to underlie abnormalities in network excitability and synaptic connectivity in at least some regions of the

Mecp2 mutant brain. Further, signaling pathways downstream of BDNF that regulate synapse development, such as mTOR, are commonly dysregulated in other synaptopathies such as Fragile X, Angelman’s syndrome, and depression[112-117] (discussed more below), and may be a central mechanism by which synapse development is dysregulated in Mecp2 mutants.

Mouse models of Rett syndrome have not only been used to better understand the mechanisms of dysfunction after the loss of MeCP2, but in the past decade have also been used to ask critical questions about the reversibility of this dysfunction. This topic is of the utmost importance to the patients and their families, as treatment success may depend on the degree of reversibility in patients who are diagnosed after abnormalities become apparent. Key studies

29 using genetic re-expression of MeCP2 in symptomatic mice have now demonstrated that phenotypic reversal is possible even in symptomatic animals[69, 71, 72, 118-122]. These data demonstrate that the phenotypes associated with Rett syndrome are caused by functional disturbances of neural cells and circuits rather than irreversible brain abnormalities.

Since these findings, numerous therapeutic approaches have been tested in mouse models; however, most have failed to translate in human clinical trials.

However, from the increasing knowledge surrounding MeCP2 biology, several new approaches to reversing the phenotypes of Rett syndrome have been developed. One such example is the pharmacological targeting of BDNF signaling. MeCP2 is known to control transcription of BDNF, a peptide that is reduced in the brains of patients with Rett syndrome and mouse models alike, in an activity dependent manner[123]. In addition, levels of BDNF are directly linked to the severity of Rett syndrome-like phenotypes, with overexpression of BDNF in

Mecp2 mutants leading to an increase in lifespan, ameliorated motor phenotypes, and reversal of spontaneous firing frequency deficits in the cortex[110]. However, BDNF itself has very poor drug-like qualities, so researchers have developed other means to increase BDNF signaling in Mecp2 mutants. The resulting studies have demonstrated that the BDNF/TrkB pathway may be a promising target of therapeutic intervention. An early approach to increasing

BDNF signaling used overexpression of BDNF in the hippocampus of Mecp2 mutants to demonstrate the reversibility of deficits in dendritic morphology[124].

Since then, small molecule agonists of the BDNF receptor, TrkB, have been used

30 to ameliorate both the signaling deficits in TrkB receptor phosphorylation as well as functional deficits associated with respiratory deficits in Mecp2 mutants[80, 125,

126]. Another study found that fingolimod, a sphingosine-1 phosphate receptor modulator was capable of increasing BDNF levels and significantly increased both the survival and motor skills of Mecp2 mutants[127]. Together, studies such as these have demonstrated that Rett syndrome-like phenotypes are reversible even without MeCP2 re-expression, raising the potential for clinical trials to be based on pharmacological approaches rather than being restricted to gene therapy.

Activity Mapping in the Mecp2 Mutant Brain

The brains of Mecp2 mutants are characterized by reductions in cortical thickness, cell body size and dendritic complexity[32-34]. However, there is no evidence to suggest that cell death or reduced cell number leads to reductions in cortical thickness. Furthermore, the brain abnormalities that have been described tend to be region specific[96]. The loss of MeCP2 is known to cause a wide range of behavioral phenotypes which suggest the involvement of multiple brain regions. For this reason, we hypothesized that Null animals would display regional differences in activity that could be used to better understand global circuit dysfunction in the Null brain[93]. To test this hypothesis, we set out to map activity throughout the brain based on the knowledge that behavioral phenotypes could arise from abnormalities in the activity of neurons and the way they communicated within behaviorally relevant circuits. Further, generating an activity

31 map for the Mecp2 mutant brain could be used to direct future studies toward circuits with abnormal patterns of activity that may play a role in specific behaviors.

Fos immunostaining has been widely used as a surrogate marker of neuronal depolarization to map circuits and pathways in the normal brain that are activated by specific types and patterns of neural stimulation[128]. To determine whether or not Mecp2 mutants displayed regional differences in Fos expression, we initially surveyed Fos expression in serial sections throughout the rostro-caudal extent of the brain, from the olfactory bulbs to the spinomedullary junction, in Wt and Null mice at 3 and 6 weeks of age, i.e., before and after the appearance of overt Rett syndrome-like symptoms. We found no obvious effect of Mecp2 genotype on the number of Fos-positive neurons in 3-week-old animals in any brain region examined. However, there were marked and reproducible differences between

Null and Wt animals at 6 weeks of age across the neuraxis (summarized in Fig.

2).

Some of the most dramatic effects of Mecp2 genotype on Fos expression were observed in cortical and subcortical limbic structures, including the prelimbic and infralimbic cortices, retrosplenial cortex, cingulate cortex, the nucleus accumbens

(nAC, both core and shell), as well as the piriform cortex, the motor cortex and lateral septal nuclei, all of which showed significantly less Fos labeling in Null animals compared with Wt. Quantitative analysis revealed that the number of

Fos-positive cells in Nulls was reduced by up to 80% compared with Wt. In the brainstem, the periaqueductal gray (PAG) and the nucleus of the solitary tract

(nTS), both of which are major cell groups involved in modulation of autonomic homeostasis[129-132], exhibited the strongest genotype-dependent differences in

Fos expression levels. Specifically, the Null PAG had significantly fewer Fos- positive cells compared with Wt. In contrast, Null mice exhibited significantly more Fos-positive cells throughout the rostro-caudal extent of the three major cardiorespiratory subnuclei in nTS [medial (mnTS), commissural (nComm) and lateral nTS, including interstitial, lateral and ventrolateral subnuclei (lnTS)] compared with Wt.

Reduced expression of Fos in forebrain cortices is consistent with reports of hypoconnectivity in layer V cortical circuits in Mecp2 mutants[88, 96]. However, a particularly striking feature of the Fos map in Null mice is the marked reduction in labeling throughout the midline limbic network, including the medial prefrontal

(mPFC), cingulate and retrosplenial cortices compared with Wt. This pattern of hypoactivity is significant because 1, these cortices are key nodes in the default mode network (DMN), a forebrain circuit that also exhibits hypoactivity and/or reduced connectivity in human autism (for review, see [133]), and 2, the midline limbic cortices play a critical role in behavioral state regulation of autonomic homeostasis, which is abnormal in Rett syndrome.

We also found reduced Fos activity in sensory cortices which, in humans, interact with the default mode network through a frontoparietal control system[134] that includes the dorsal PFC and cingulate cortex. The default mode network is considered critical for self-referential cognition and theory of mind[135-137], planning, remembering, and monitoring the external environment[136, 138]. Social

33 deficits in ASDs have been correlated with weaker connections between the cingulate and frontal/temporal cortices, while stereotyped and repetitive behaviors are correlated with weaker connections between the cingulate cortex and the mPFC[139]. The fact that Mecp2 mutants exhibit reduced Fos expression in these structures suggests that hypoactivity in the default mode network may be a common feature of Rett syndrome as well as non-syndromic autism and may underlie overlapping features of behavioral dysfunction in these disorders.

Significant decreases in Fos labeling were also observed in the Null piriform, visual, somatosensory, auditory and motor cortices, as well as the nucleus accumbens, septal nuclei, and the caudate/putamen. Although further work is needed to fully understand the relationship between reduced activity in these structures and specific functional abnormalities, hypoactivity in motor cortex and the basal ganglia correlates well with the hypokinesis characteristic of young Rett syndrome patients and mouse models[8]. On the other hand, we saw no significant effect of Mecp2 genotype on Fos expression in the hippocampus, a region which is prone to network hyperexcitability in vitro[66, 83, 91, 140, 141]. It is possible, however, that genotype effects on Fos expression in the Null hippocampus would be detectable in the context of stimuli that trigger seizure activity, given that, in some studies, hyperexcitability in hippocampal networks is only apparent after challenge with excitatory stimuli[91]. It is also possible that this apparent discrepancy reflects differences in network excitability between intact animals and in vitro preparations as observed previously[91, 142].

Although Fos has been widely validated as a marker of neural activity in normal animals, this has not previously been analyzed in animals lacking MeCP2.

Therefore, to determine whether or not Mecp2 genotype effects on Fos labeling indeed reflect differences in neural activity, patch clamp electrophysiological recordings were used to compare synaptic and neuronal excitability in Wt and

Null mice at 5–7 weeks of age, using the nTS as a model system. The nTS is ideally suited for such analyses because of a clear anatomic segregation between presynaptic inputs in the solitary tract (TS) and second-order neurons within the various nTS subnuclei. Indeed, EPSC amplitudes evoked by TS stimulation were significantly larger in Nulls compared with Wt in both the lnTS and nComm. Spontaneous excitatory currents at primary afferent synapses in the

Null nTS were also enhanced compared with Wt. On the other hand, basic membrane properties and action-potential properties were similar between the genotypes[93].

It is possible that reduced Fos expression in the cortex of Mecp2 Null mice is due to an inability to express Fos at normal levels in the absence of MeCP2. To determine whether or not this was the case, we turned to the literature to find a pharmacological method to increase activity in the cortex and evaluate the effect of such a treatment on Fos expression in the Mecp2 mutant. Ketamine, an N- methyl-D-aspartate receptor (NMDAR) channel blocker, was previously shown to increase Fos in the rodent forebrain by disinhibiting cortical pyramidal neurons

(PNs)[143]. Therefore, we treated both Mecp2 mutant and Wt mice with a single low dose of ketamine, 8mg/kg i.p., and dissected for Fos immunostaining 90

35 minutes later. We found that in all midline limbic cortices, Fos was restored to Wt levels after this treatment[93]. Our finding that forebrain deficits in Fos expression in Nulls can be rescued by acute treatment with ketamine, even at subpsychotomimetic doses, supports the hypothesis that reduced Fos labeling reflects a reversible deficit in network activity, rather than an intrinsic inability to express Fos. Thus, the recovery of Fos expression in ketamine-treated Nulls likely reflects a restoration of excitatory drive in circuits that are otherwise hypoactive in the absence of MeCP2.

We next sought to determine if increased network activity in the frontal cortex of

Mecp2 mutants was associated with functional changes in a behavior that relies on these circuits. Therefore, we performed pre-pulse inhibition (PPI) of the acoustic startle reflex, a measure of sensorimotor gating widely used as an index of cognitive function in neuropsychiatric disorders, including ASDs[144-147]. PPI measures the ability of a weak sensory input to modulate behavioral responses to a subsequent strong sensory stimulus and thereby reflects the function of inhibitory circuitry thought to be critical for normal cognition. Because the circuitry underlying PPI includes structures that exhibit reduced Fos staining in Nulls, such as the mPFC and nAC[148, 149], and because ketamine treatment of Nulls rescues Fos expression in these regions, we decided to use PPI as an index of forebrain circuit function in the absence and presence of ketamine. We found that while Hets normally exhibit a significant increase in PPI amplitude compared to

Wt, acute treatment with ketamine restored PPI in Hets to Wt levels[93].

These data indicate that ketamine treatment is also effective at restoring circuit function in MeCP2 deficient mice. Although the specific site(s) at which ketamine acts to restore Wt PPI in Hets remain to be defined, there is overlap between structures involved in cortical modulation of PPI and those exhibiting Fos rescue in ketamine-treated animals, including the mPFC and nAC[148, 149]. It is also possible that the rescue of PPI is related to the ability of ketamine to acutely increase translation of brain derived neurotrophic factor (BDNF)[150], as BDNF levels are reduced in Mecp2 mutants compared with Wt controls[126] and PPI amplitude has previously been shown to be inversely related to BDNF availability[151, 152]. Regardless of mechanism, these data highlight the potential of low-dose ketamine treatment to improve behavioral dysfunction in Rett syndrome, as in other neuropsychiatric disorders[150]. The therapeutic potential of ketamine will be discussed more below and further studied in the experiments described within Chapter 2.

Together, these findings demonstrated marked effects of Mecp2 genotype on expression of the activity-dependent, immediate-early gene product Fos within specific forebrain and hindbrain networks, including many previously unrecognized sites of circuit dysfunction within the Mecp2 mutant brain. In view of the close spatial and temporal association between genotype effects on neural activity and Fos expression, our data indicate that loss of MeCP2 results in a stereotyped pattern of activity alterations within a defined subset of functionally interrelated brain circuits that emerges during late postnatal development, coincident with the appearance of overt symptoms.

Because Rett syndrome patients exhibit complex disturbances of breathing and heart rate control[153], attention has focused on nTS as a possible locus at which loss of MeCP2 deranges cardiorespiratory homeostasis. Indeed, our electrophysiological and Fos mapping data from this previous study[93] indicate a shift toward a more excited default state in cardiorespiratory subnuclei in the mutant nTS, which would be expected to result in a loss of sensory gating within nTS and destabilization of cardiorespiratory homeostasis. In fact, Mecp2 mutant mice exhibit an exaggerated hypoxic ventilatory reflex[73, 75, 77, 154], a homeostatic increase in ventilation triggered by primary chemoafferent inputs to the nComm[155, 156]. Similarly, hyperexcitability at the first synapse in lnTS, where pulmonary stretch receptors terminate[131, 157], would be expected to lower the activation threshold for the Hering-Breuer reflex pathway, which inhibits inspiration and promotes expiration. Indeed, Mecp2 mutants exhibit prolonged respiratory pauses in response to vagal stimulation[76], indicative of exaggerated

Hering-Breuer reflex activation. Our data therefore highlight the lnTS as a key site at which synaptic hyperexcitability likely contributes to the frequent respiratory pauses that characterize breathing in Mecp2 mutants and Rett syndrome patients.

However, sensory gating deficits at the level of nTS cannot explain all of the cardiorespiratory phenotypes associated with loss of MeCP2 function. In particular, cardiorespiratory dysfunction in Rett syndrome patients is also dependent on behavioral state, with symptoms improving during sleep[14] and worsening in the context of emotional stress[78]. Thus, it is particularly noteworthy

38 that Nulls exhibit reduced Fos levels in interconnected forebrain and midbrain structures critical for behavioral modulation of respiration, including the mPFC[158-

160] and the PAG[161]. Our data, indicating hypoactivity in the mPFC, which projects to both the PAG and to brainstem respiratory targets such as the nTS, and the fact that similar brainstem respiratory regions are also targets of the

PAG[129, 162-164], suggest that excitatory/inhibitory imbalance across the forebrain- midbrain-hindbrain neuraxis may be a substrate for abnormal behavioral state regulation of respiratory and autonomic homeostasis in Rett syndrome. This hypothesis will be further tested in Chapter 3.

The Medial Prefrontal Cortex

The data discussed in this thesis will use the mPFC as a model region to study how hypofunction in particular circuits underlies or contributes to critical behavioral domains that are abnormal in both Rett syndrome patients and Mecp2 mutant models. Hypoactivity in the pyramidal neurons of the mPFC, the rostral most portion of the midline limbic cortex, is of particular interest in Rett syndrome given the role of this region in multiple behaviors and functional domains, from autonomic and respiratory control to cognition, that are abnormal in Rett syndrome. The mPFC projects heavily throughout the brain including to both cortical and subcortical targets that allow it to exert influence over visceral, limbic, and cognitive functions[158-160]. For example, the ventral mPFC is also known as the ‘visceral cortex’ as it is responsible for regulating behavioral state-dependent changes in respiration and autonomic homeostasis such as during periods of

39 stress or in response to conditioned learning[165-168]. Structures in the ventral mPFC, including the prelimbic (PL), infralimbic (IL), and dorsal peduncular cortex

(dPC) give rise to extensive direct projections to cardiorespiratory cell groups in the pons and medulla, as well as indirect projections to subcortical forebrain cell groups that project to the brainstem, including the hypothalamus and amygdala

(Fig. 3)[169]. Although some of these brainstem targets, such as the preBotzinger complex, have internal pacemaker rhythms, the mPFC has the connectivity to be capable of adjusting respiratory and autonomic rhythms during situations with emotional valence. However, we hypothesize that when these projections from the mPFC have reduced excitability, as in mouse models of Rett syndrome, their ability to influence either baseline or stimulated respiratory or autonomic responses may be abnormal.

In addition to the projections that allow the mPFC to influence ‘visceral’ functions, the mPFC is more commonly known for its role in cognition. Within the cognitive domain, the mPFC is known to have a role in decision making and executive control[170-175], theory of mind and social cognition, as well as retrieval and expression of both short- and long-term memories (reviewed in [176]). Influence over these diverse cognitive functions is provided both by the broad pattern of connectivity produced by mPFC projection neurons, and by the large number of inputs received back to the mPFC. In addition, the mPFC has reciprocal projections to other limbic cortical regions such as the cingulate and retrosplenial cortices which, together with the mPFC, make up the DMN.

In addition to its connections within the DMN, the mPFC has reciprocal connections with other limbic regions such as the amygdala which is necessary for memory of emotionally charged stimuli[169, 177]. Not only is the amygdala involved in emotionally salient events, but the mPFC itself, likely due to connections with both the amygdala and sensory cortices, responds strongly to both positive and negative events (reviewed in [176]; see also [171]), as well as to reward expectancy or negative outcomes[178-183]. Together with the amygdala, projections from the mPFC to subcortical regions such as the periaqueductal gray and brainstem autonomic targets allow the individual to produce emotionally relevant behavioral action appropriate for the event, whether this be freezing in response to a predator or painful stimulus, altered respiration and heart rate, etc.

[169, 184-186]. Indeed, it is in the context of this connectivity that the ventral mPFC became known as the visceral cortex as described above[168]. On the other hand, the dorsal mPFC has weaker connectivity with emotional and visceral regions, but stronger connectivity with motor and pre-motor regions[158, 169, 184]. The dorsal mPFC also projects to the spinal cord in rodents[169]. Together, the mPFC gains both sensory and emotionally salient stimuli and has control over output behaviors by way of autonomic and skeletal muscle activity.

There has been some debate on the degree to which the rodent mPFC is homologous to the mPFC of the human and other primates[187-191]. One side to this argument claims that the rodent mPFC acts as an undifferentiated region with functional aspects of both medial and lateral mPFC in primates[187, 191]. This is supported by the fact that the rodent mPFC plays a role in behaviors linked to

41 orbital cortex-like functions, such as socioaffective behaviors, dorsallateral cortex-like functions such as working memory, or anterior cingulate-like functions such as visceromotor behavior in the primate frontal cortex[187]. However, the mPFC of both rodent and primate follow similar organizational schemes, with anatomical evidence, projection mapping, and lesion studies suggesting that the medial aspect of the PFC is more similar between the two, even if less differentiated than the primate frontal cortical subdivisions[171, 189, 192-194].

Important to my thesis experiments, projection targets and anatomical makeup of the mPFC overlaps significantly between primate and murine models. This allows for the study of mPFC function generally, as well as the function of specific projection neurons, in a translationally relevant manner.

The principle projection neuron in the mPFC is the cortical PN which fit into two broad classes. The first set of PNs have cell bodies that reside in shallow layers of the mPFC, layers II and III, and primarily have short-range projections that terminate in cortical targets such as the DMN, interconnectivity with deeper layers within the mPFC itself, and some long range projections to regions such as the amygdala (discussed in [160]; see also [195]). The short-range layer II/III

PNs allow for communication within the DMN both projecting to and receiving inputs from DMN regions including the contralateral mPFC[160]. The other class of mPFC PNs have cell bodies that reside in layers V and VI and are responsible for long range connectivity with subcortical and brainstem targets, as well as other cortical targets via the corpus collosum[160]. Dendrites of both classes of PN are comprised of distinct dendritic domains, including the basal dendrites which

42 emanate within the layer of the cell bodies, a long apical dendrite which projects toward the pial surface and is topped by the apical tuft which radiates through layer I, and the apical oblique dendritic domain which branches from the main apical dendrite and extends parallel to the pial surface within layers II/III and V.

Both classes of PN also receive inputs from other cortical regions and limbic regions such as the lateral amygdala[158, 169, 196].

It should be noted that, unlike the primary sensory cortices the mPFC is an agranular cortex, and lacks a layer 4 thalamic recipient zone (discussed in [160]; see also [187]). However, the mPFC does have reciprocal connections with the dorsal medial thalamus from layer 6 of the mPFC (discussed in [197]; see also

[169]). These projections, in addition to projections from the mPFC to both the amygdala and hippocampus, play distinct roles in memory formation and retrieval in the context of cue-dependent fear conditioning (CDFC). CDFC involves the exposure of the animal to a conditioned stimulus (CS, such as a tone) which co- terminates with an aversive event (the unconditioned stimulus or US, usually an electric shock). This association can be retrieved throughout the lifetime of the animal and is stored in circuits involving the basolateral amygdala (BLA) and the central nucleus of the amygdala (CeA)[198-203]. The prelimbic cortex of the mPFC

(PL) sends projections to the amygdala through both a direct pathway to the

BLA, and through an indirect pathway in which the PL projects to the dorsal midline thalamus (dMT, which included the periventricular thalamus) which then, in turn, projects to the CeA (Fig. 4)[197, 204, 205]. During recall of the cue-shock pairing either at short-term (usually 4-12 hours after conditioning) or long-term

43 time points (24-72 hours after conditioning) activation of the mPFC is required for normal cue-evoked expression of fear memory, as inactivation of the mPFC at either time point reduces freezing to the cue[206]. Interestingly, the critical pathway for expression of this fear memory changes between the short- and long-term time points. In the short-term, fear expression requires activation of the PL to

BLA pathway and inhibition of these projections significantly reduces fear memory expression; however, by the long-term time point the critical PL projection shifts to those cells that project to the dMT and the PL to BLA projection becomes unnecessary for expression[197].

These circuits are of particular interest given that Mecp2 mutants are impaired in the ability to retain long-term memory of auditory conditioned fear but not short- term memory[84, 85, 207, 208]. This suggests that a subset of mPFC projection circuitry may be more strongly affected by the loss of MeCP2, i.e. layer VI projections to the dMT which are required for long-term fear memory expression.

Indeed, mPFC hypofunction and reduced long-term fear memory expression in

Mecp2 mutants are likely linked given studies of Wt murine models that found inhibition of either the mPFC[206] or the dMT[209, 210] during long-term memory recall leads to reduced fear memory expression. In particular, this suggests that reduced activity in the mPFC projections to the dMT may be the primary mechanism of reduced fear memory retrieval in Mecp2 mutants; however, projections to the amygdala, which are necessary for short-term consolidation of learning[209], cannot be ruled out.

Understanding mechanisms underlying mPFC hypofunction becomes critical to answering these questions, and for developing approaches to reversing Rett syndrome phenotypes associated with mPFC hypofunction. To begin elucidating mechanisms underlying mPFC hypofunction in Rett syndrome, our lab previously examined circuit dysfunction in the mPFC using a variety of techniques to study

E/I synaptic balance, NMDA and AMPA responses and synaptic protein expression, neuronal population excitability, and dendritic spine density on excitatory PNs[88]. This study found that the synaptic E/I balance was shifted toward reduced excitability in the Mecp2 Null with no change to inhibition within the same cell. Because this excitatory response is made up of both AMPA and

NMDA components, understanding whether or not the ratio of these components was normal may lead to a better understanding of why E/I balance was shifted.

Indeed, it was found that Mecp2 Null neurons had reduced NMDA current but normal AMPA current. NMDA currents contribute to total synaptic drive as well as action potential timing and synchrony[211-214]; therefore, reduced NMDA current likely contributes to overall hypofunction of the mPFC, and to signal propagation in particular. To test the latter possibility, slices of the mPFC were stimulated at the corpus callosum and activity patterns emanating from the electrode toward the pia matter where observed and quantified using a calcium sensitive dye. In this way, it was found that Mecp2 Null mPFC displayed reduced responsiveness in deeper layer V as well as the shallow layers II and III, consistent with decreased excitatory synaptic connectivity. Finally, to determine if reduced responsiveness to stimulation was associated with reduced density of dendritic

45 spines, a structural read out of synapse number, DiOlistic neuronal labeling was used to compare the morphology of apical dendrites on layer V PNs. Indeed, spine density was significantly reduced in the Mecp2 Null[88] . Together, these data suggest the Mecp2 mutant mPFC is hypofunctional due at least in part to reduced excitatory drive and reduced excitatory post-synaptic structural connectivity.

Dysfunction of the frontal cortex is also common among several other neurological disorders in addition to depression including Fragile X, Angelman syndrome, and non-syndromic ASDs. The common thread in mouse models of these disorders is abnormal regulation of synaptic protein expression and excitatory synaptic connectivity. Angelman syndrome, which is categorized by intellectual disability, a pervasively happy demeanor, seizures and motor dysfunction is cause by loss of function mutations to the gene UBE3A which encodes an ubiquitin ligase responsible for degradation of synaptic proteins; while on the other hand, over expression of the same gene is associated with autism[215]. Both the under- and over-expression of the UBE3A gene leads to abnormal regulation of synaptic structure in frontal cortical circuits. Likewise,

Fragile X syndrome is caused by mutations in the gene encoding the fragile X mental retardation protein (FMRP) which regulates protein synthesis at synapses, generally in a repressive manner, and results in intellectual disability, repetitive behaviors, and abnormal social interaction (reviewed in [216]). As with mouse models of Rett syndrome, Fragile X mouse models display frontal cortical hypofunction in part due to reduced synaptic maturity and synaptic protein

46 expression, which unlike Rett syndrome presents as an overall increase in immature dendritic spines rather than an overall reduction in spines[112, 216, 217].

One target of FMRP is the SHANK3 gene, which is responsible for organizing the post synaptic density in nearly all excitatory glutamatergic synapses[218-222].

Interestingly, mutations leading to either loss or gain of function in SHANK3 leads to autism spectrum disorders (reviewed in [223]). Mouse models of Shank3 loss of function display reduced social interaction and repetitive behaviors and also display reductions in spine number and size in frontal cortical regions (reviewed in [223]) very similar to the phenotypes observed in Fragile X syndrome[112, 116],

Angelman syndrome[113, 117, 224], and Rett syndrome[88, 225-227] mPFC hypofunction and reduced connectivity is also shared with mouse models of depression[114, 115]. In fact, murine models of depression display deficits in excitatory neuron structure and function very similar to those found in the mPFC of Mecp2 mutants. Most notably, chronic stress paradigms that are used to induce depressive-like features in murine models also induce reductions in cortical volume, atrophy of pyramidal neurons, reduction of dendritic spine density, and reduced excitatory transmission in the mPFC among other regions

(reviewed in [115]). Another striking link between Rett syndrome and depression models are the abnormalities in signaling pathways that are related to the maintenance, maturation, and stability of dendritic spines. Specifically, rodent models of depression have reduced expression of BDNF in the mPFC and other regions, as well as reduced activity in the mTOR pathway, a signaling pathway downstream of BDNF that regulates activity dependent protein synthesis and is

47 required for protein translation related to synaptic plasticity[228, 229]. Likewise, expression of BDNF and activity in the mTOR pathway are known to be reduced in the cortex of Mecp2 mutants[35, 107]. Together, these data suggest that similar mechanisms may underlie mPFC hypofunction and reduced synaptic connectivity in both depression and Rett syndrome.

Clinical studies have demonstrated that ketamine produces a rapid antidepressant response even in patients who are resistant to traditional antidepressants. Knowing this, researchers have since discovered that one critical mechanism underlying the rapid antidepressant effects of ketamine (such as reduced anxiety, and increased time to failure in the forced swim test) is the rapidly induced translation of BDNF and the stimulation of mTOR dependent signaling; specifically, blockade of either BDNF/TrkB interaction or mTOR signaling eliminates the antidepressant effect of ketamine treatment in mouse models of depression (reviewed in [230]). Moreover, ketamine treatment reverses dendritic spine deficits in the mPFC in mouse models of depression, a rescue that is associated with the upregulation a host of synaptic proteins such as activity regulated cytoskeletal-associated protein (Arc), the structural protein

PSD95, the AMPA receptor subunit GluR1, and synapsin[231]. Indeed, these protein and structural changes were also associated with increased excitatory drive in the mPFC, all of which was blocked by the mTOR inhibitor rapamycin

(reviewed in [115, 232]; see also [231]).

Abnormal modulation of the mTOR signaling pathway is common feature of frontal cortex dysfunction in all these models, as are abnormalities in dendritic

48 morphology and excitatory synaptic connectivity[88, 109, 114, 115, 216, 233, 234]. This raises the possibility that disorders such as these result in circuit deficits with common underlying mechanisms which could act as therapeutic targets, such as pyramidal cell activity and/or synaptic protein regulation and mTOR signaling within these cells. Therefore, gaining a better understanding of frontal cortex circuit dysfunction in Rett syndrome, with a particular focus on mPFC pyramidal neurons, may provide insight into similar cellular and behavioral phenotypes observed in these other neurological disorders.

If the mechanisms underlying mPFC hypofunction in depression are similar to those in Rett syndrome, as the evidence suggests they may be, then this raises the possibility that ketamine may be capable of reversing deficits in structural connectivity, mPFC hypoactivity, and ameliorate behavioral phenotypes that depend on mPFC function. Our previous study demonstrated that the activity dependent protein cFos could be upregulated by ketamine treatment even in the complete absence of MeCP2[93]. This suggests that cellular responses to activity remain intact, and that cellular hypofunction is due to a lack of proper synaptic communication rather than an inability to respond to activity in the absence of

MeCP2. This hypothesis will be tested first in Chapter 2 by determining whether or not ketamine treatment can restore structural connectivity in the mPFC of

Mecp2 Null mice, and second, in Chapter 3 by directly increasing activity in excitatory neurons of the mPFC to determine if restoring activity in circuits underlying specific behaviors can reverse behavioral phenotypes in Mecp2 mutants.

Rett syndrome research often focuses on individual circuits, the mechanisms of dysfunction within them, and methods of restoring function of a particular circuit as a means to ameliorating Rett syndrome phenotypes. What has been found is great variation from region to region, with some regions being hyperactive while others are hypoactive; therefore, a successful therapeutic approach targeting a mechanism in one region may be completely unsuccessful in another (reviewed in [96]; see also [93]). For this reason, this thesis work aims to restore function in a highly connected hub region naturally capable of influencing many other regions throughout the brain. Specifically, the experiments in this thesis will test the hypothesis that restoring function to the mPFC, a hub region that projects throughout the brain and modulates activity in a wide variety of circuits, may have a beneficial impact on multiple behavioral domains impacted by the loss of

MeCP2 function. This region has been sparsely studied in the context of Rett syndrome, but restoring function to the mPFC could have wide ranging impact on disease phenotypes. Because the Mecp2 mutant brain is characterized by activity imbalance between forebrain and hindbrain, one aspect of these studies is to determine how the restoration of function in the mPFC can influence activity in distant targets such as the brainstem, and determine whether or not this is associated with beneficial behavioral outcomes. The understanding of how specific circuit dysfunction relates to specific behavioral abnormalities in freely behaving animals is a knowledge gap within the study of Rett syndrome that this thesis seeks to address.

Figure 1. Schematic of the human MECP2 gene including the most common point mutations in associated with Rett syndrome. MBD, methyl-binding domain; TRD, transcriptional repression domain; NLS, nuclear localization sequence; NID, NCoR/SMRT interaction domain.

Figure 2. The Mecp2 mutant brain displays altered neuronal activity shifted towards reduced (in blue) or increased (in red) excitation compared to wild-type. This model was produced using data from our lab and others which included electrophysiological recordings as well as mapping of Fos expression. BLA; basolateral amygdala; HIP, hippocampal formation; IL, infralimbic cortex; LC; locus coeruleus; LSN, lateral septal nucleus; NaS, nucleus accumbens shell; nTS, nucleus of the solitary tract; PAG, periaqueductal gray; Pir; piriform cortex; PrL, prelimbic cortex; VLM, ventrolateral medulla.

Model created using the Allan Brain Atlas Composer application.

Figure 3. Model circuitry demonstrating the role of mPFC in respiration. mPFC, medial prefrontal cortex; PAG, periaqueductal gray; LC, locus coeruleus; nTS, nucleus of the solitary tract; VLM; ventral lateral medulla; green arrows indicate direct projections from the mPFC; red arrows indicate secondary projections by which the mPFC may influence respiration.

Figure 4. Model circuitry underlying cue-dependent fear conditioning. PL, prelimbic cortex;

L5, layer 5; L6, layer 6; PVT, paraventricular thalamic nucleus; CE, central nucleus of the amygdala; BLA, basolateral amygdala; red arrows indicate mPFC projections required for long-term fear memory retrieval (FMR); blue arrows indicate mPFC projections required for acquisition and short-term retrieval of auditory fear conditioning; dashed lines indicate projections from regions other than the mPFC that are involved in cue-dependent fear conditioning and FMR.

Chapter 2

Chronic Intermittent Treatment with Low-dose Ketamine Reverses

Dendritic Spine Deficits in a Mouse Model of Rett Syndrome

C. James Howell*, Saloni U. Lad*, Lina Ghosh, Anmol Gupta, Samichhya Aryal,

Karthik Ravichandran, David M. Katz

*These authors contributed equally to this work

Abstract

Rett syndrome (RTT), a severe neurodevelopmental disorder caused by loss-of- function mutations in the MECP2 gene, is characterized by neurological regression leading to deficits in motor, respiratory and autonomic control and cognition. Similar to other autism spectrum disorders, abnormal synaptic connectivity is a hallmark of RTT, including reduced density and maturity of cortical dendritic spines. In the medial prefrontal cortex (mPFC), spine deficits on layer 5/6 pyramidal neurons are associated with decreased excitatory synaptic drive and reduced neuronal activity[88, 93] which, in turn, appear to contribute to respiratory and cognitive dysfunction[84]. These findings suggest that restoration of excitatory connectivity in the mPFC could be of therapeutic benefit in RTT. To approach this issue, we examined the reversibility of deficits in spine density in the mPFC in Mecp2 Null (Null) mice in response to ketamine, an NMDA antagonist that is known to promote dendritic spine growth in other disease models[235]. Specifically, we compared dendritic spine density in the mPFC of wildtype (Wt) and Null mice 24 hours after single or repeated intermittent dosing

(q72 hours x 4 weeks) with a sub-anesthetic dose of ketamine (10 mg/kg, i.p.).

To visualize dendritic spines, we used Thy1-GFPm Mecp2 Null or Wt mice in which GFP is expressed in a subset of layer 5/6 pyramidal neurons. 24 hours after single or repeated ketamine dosing, deficits in the density of mushroom spines were reversed in Null mice without a change in total spine density. Spine recovery in the mPFC was associated with increased expression of the phosphorylated ribosomal protein S6, which has previously been linked to

56 ketamine effects on synaptic plasticity in other systems. Our findings support the therapeutic potential of repeated intermittent treatment with low-dose ketamine for reversing structural synaptic deficits caused by MeCP2 deficiency, as in Rett syndrome.

Introduction

Rett syndrome (RTT) is a severe neurodevelopmental disorder resulting from loss-of-function mutations in the gene encoding methyl-CpG-binding protein 2

(MeCP2)[38, 236], a transcriptional regulatory protein[237]. RTT patients exhibit a diverse spectrum of symptoms including severe deficits in motor control and cognitive function, dysregulation of respiratory and autonomic homeostasis, autistic behaviors and increased risk of seizures[3, 8, 13, 14, 45, 236, 238, 239]. Although reduced MeCP2 signaling does not cause neuronal cell death or degeneration, the brains of RTT patients and Mecp2 mutant mice exhibit microcircuit abnormalities, including decreased dendritic complexity[86, 87, 123, 226, 240-243], as in other disorders on the autism spectrum[109]. In particular, cortical pyramidal neurons exhibit deficits in the density and maturity of dendritic spines[87, 88, 109, 123,

226, 244-247] that are believed to contribute to the pathophysiology of RTT by disrupting normal excitatory synaptic drive (reviewed in [96]). Interventions that promote dendritic spine growth in Mecp2 mutants, such as genetic reversal of

MeCP2 deficiency[122] are associated with improved neurological function, suggesting that treatments aimed at restoring normal dendritic structure may provide therapeutic benefit to RTT patients.

In other disease models, including rodent models of depression, the non- selective NMDA receptor antagonist ketamine has proven highly effective at acutely enhancing cortical dendritic spine growth and improving performance in cortically-mediated behavioral tasks[115, 231, 232, 235, 248]. These effects of ketamine appear to be mediated by activity-dependent mechanisms including stimulation of glutamate and BDNF release and activation of mTOR/S6 kinase-dependent translation[115, 232, 235, 249]. Moreover, these effects are quite rapid and appear within hours of ketamine treatment (reviewed in [115]).

Ketamine is also effective at improving neurological function in mouse models of

RTT, including abnormalities in sensorimotor gating, respiration and visual cortical processing[93, 250]. In addition, a recent case report provides evidence for clinical efficacy of ketamine treatment in a patient with RTT[251]. However, given the importance of MeCP2 for activity-dependent processes, including BDNF expression[252] and release[105], we asked whether or not ketamine can reverse structural synaptic defects in MeCP2 deficient neurons. This is an important issue because it bears on the question of whether or not ketamine treatment can durably impact circuit dysfunction in RTT by improving structural connectivity.

Therefore, the present study was undertaken to define the acute effects of low- dose ketamine treatment on the growth and maturation of dendritic spines on cortical pyramidal neurons in Mecp2 mutant mice. We focused in particular on layer 5/6 pyramidal neurons in the medial prefrontal cortex (mPFC), a region critical for behaviors that are disrupted in RTT, including learning and memory and behavioral state-dependent control of respiration and autonomic

58 homeostasis[176]. Previous studies in our laboratory demonstrated that reduced spine density on Layer 5/6 pyramidal neurons in the medial prefrontal cortex

(mPFC) of Mecp2 mutant mice is associated with reduced excitatory synaptic drive, neuronal hypoactivity and reduced population activity in response to white matter stimulation[88] . The functional impact of reduced excitatory connectivity in the mPFC is underscored by our recent finding that increasing the activity of mPFC pyramidal neurons reverses disease phenotypes in heterozygous (Het) female Mecp2 mutants[84]. The present studies demonstrate that, 24 hours after a single sub-anesthetic dose of ketamine, or 24 hours after the last of repeated intermittent doses delivered over 4 weeks, dendritic spine deficits on Layer 5/6 pyramidal neurons in Mecp2 Null mice are reversed. These data indicate that low-dose ketamine treatment has the potential to reverse structural microcircuit abnormalities that contribute to the pathophysiology of RTT caused by MeCP2 deficiency, after single and repeated dosing.

Materials and Methods

Mice: Experiments were performed on male (Wt) or male Null Mecp2tm1.1Jae littermates maintained on a 129S/BalbC/B6 background. These mice were also heterozygous for a Thy1-EGFPmJrs/J transgene. All animals were genotyped before and after each experiment to confirm genetic identity.

Tissue preparation for spine analysis and immunocytochemistry: Mice were deeply anesthetized with isoflurane and perfused transcardially with phosphate-

59 buffered saline (PBS) followed by 4% paraformaldehyde in 0.1M phosphate buffer (pH 7.4). Brains were post-fixed in 4% paraformaldehyde in 0.1M phosphate buffer (pH 7.4) for 1 hour, then cryoprotected in 25% sucrose overnight. The brains were frozen in 2-methylbutane at -45°C and stored at -

80°C until sectioning. Coronal sections were cut at 50µm on a cryostat microtome (Jung Frigocut 2800N) and sections containing the prelimbic, infralimbic, and anterior cingulate cortices were mounted for confocal imaging or processed for immunostaining as described below.

Dendritic Spine Analysis: Z-stack images of GFP-labeled layer 5/6 pyramidal neurons in the mPFC of Mecp2 Wt and Null mice were obtained by an investigator blinded to genotype and treatment group, using a Leica SP8 confocal microscope. Imaging was restricted to dendritic segments in cortical layers 2/3 and 4 from the prelimbic, infralimbic, and anterior cingulate cortices between +2.00mm and +1.20mm from Bregma. Neurolucida 360 software (MBF

Biosciences) was used to quantify dendritic spine number and morphology using software preset definitions for spine types (i.e., mushroom, thin and stubby). All dendritic spine analysis was performed by at least two investigators blinded to genotype and treatment group. For the text figures, segment values for each spine type were removed if they fell outside of the range of the group average ±

3x standard deviation. This outlier test was applied to all data sets but only led to removal of segments within the data set containing dendritic spines from animals treated with ketamine. 3 segments were removed in this manner: 1 from the Wt

60 saline treated group and 2 from the Null saline treated group. The complete data set including outliers, is shown in Supplemental Figure 1.

Western blots: Mice received injections of either saline or ketamine (10 mg/kg, i.p.) one hour prior to being anesthetized with isoflurane overdose followed by decapitation and dissection of the mPFC. Tissue was lysed by sonication in RIPA buffer containing Halt protease and phosphatase inhibitor cocktail (Thermo, catalog #1861281). Total protein concentration was determined using the Pierce

BCA protein assay (Thermo Scientific) and 20µg of protein per lane was loaded into Biorad precast gels. The following antibodies were used: anti-phospho-S6 and anti-S6 (Cell Signaling; Cat#D68F8 and #5G10, respectively; 1:1000), and goat anti-rabbit secondary antibody from Jackson (Cat#ABIN964977; 1:5000).

Blots were developed using Supersignal West Pico Plus Chemiluminescent substrate from Thermo Scientific and imaged in a FluorChem M (Protein Simple).

Immunohistochemistry: Alternate 40 µm frozen sections were processed for S6 and pS6 immunostaining by first blocking in 10% goat serum in dilution buffer

(PBS, 2% BSA, 0.3% Triton X-100) for 90 minutes and then incubating in the same solution overnight with primary antibody (Cell Signaling antibodies used above for Western blotting at 1:1000 dilution). After rinsing in dilution buffer, sections were incubated in biotinylated goat anti-rabbit IgG secondary antibody

(1:400, Vector Labs) for 1 hour. Sections were rinsed in dilution buffer and PBS and then incubated in avidin and biotinylated horseradish peroxidase complex

(1:150, Vector Labs) followed by chromogen development using Sigma Fast diaminobenzidine kit.

Statistics: Group sizes were determined by a priori power analysis using

G*Power3 software [253]. All data are presented as mean ± SEM. Genotype comparisons were analyzed by unpaired two-tailed Student’s T-test. Multiple group data sets were analyzed by one-way ANOVA with post hoc Tukey tests to identify the sources of any observed differences. Results were considered significant at p < 0.05.

Results

Development of dendritic spine defects in the mPFC in Mecp2 mutant mice

Prior to examining potential effects of ketamine treatment on spine density, and to help define an optimal treatment window, we first sought to characterize the development of spine phenotypes in the mPFC of Wt and Mecp2 Null mice. Initial studies focused on the impact of MeCP2 loss on the density and morphology of dendritic spines on L5/6 pyramidal neurons in the mPFC at 5-6 weeks of age, a time at which Null mice exhibit RTT-like impairments across multiple symptom domains [4]. Our analyses included the apical shaft, apical oblique and apical tuft dendritic compartments. Other than an increase in the density of apical shaft thin spines in Null animals compared to Wt, there were no other effects of Mecp2 genotype on spine densities in the apical shaft or apical tuft compartments (Fig. 1 and Table 1). However, multiple changes in spine density were observed between Null and Wt animals in the apical oblique compartment. Specifically,

Null animals exhibited significant reductions in total and mushroom spine density

(Fig. 1 and Table 1). To determine whether or not MeCP2 deficiency results in mPFC spine abnormalities prior to the onset of overt behavioral symptoms, we also analyzed dendritic spines in the apical oblique compartment in 3 week-old

Wt and Mecp2 Null mice. In contrast to 6 week-old animals, 3 week-old Nulls exhibited a reduction in mushroom spine density only, without significant changes in total, stubby or thin spine density compared to Wt (Fig. 2 and Table

1). These data demonstrate that Mecp2 Null mice begin to develop dendritic spine deficits in the mPFC prior to overt behavioral symptoms and that these spine abnormalities worsen through adolescence. Given this finding we asked whether or not early treatment could prevent the development of dendritic spine deficits, and this helped guide the timing of chronic intermittent ketamine paradigm described below.

Ketamine rapidly stimulates phosphorylation of rpS6

In other rodent models, ketamine stimulation of dendritic spine growth requires phosphorylation of S6 kinase, a downstream effector of mTOR and other signaling pathways involved in synapse plasticity (reviewed in [115]). To determine whether or not ketamine treatment acts similarly in the absence of

MeCP2, we examined expression of the phosphorylated form of the ribosomal protein S6 (rpS6) in the mPFC of Wt and Null mice treated with either saline or ketamine (10 mg/kg). rpS6 is a substrate of S6 kinase and an effector of the mTORC1 signaling pathway[254, 255] that is required for ketamine-induced spine growth in rodent models of depression[235, 249] . Using Western blots of the mPFC,

63 we found that expression of total rpS6 was not different between saline-treated

Wt and Null mice. However, Null mice expressed markedly reduced levels of p- rpS6 compared to Wt controls (Fig 3; p-rpS6; Wt Saline; 100.00 ± 12.63 vs Null

Saline; 49.00 ± 10.68. p = 0.004), consistent with previous results[107] . Within one hour after ketamine treatment (10 mg/kg, i.p.), Null mice showed an increase in p-rpS6 levels compared to saline-treated Null animals (Fig 4; p-rpS6; Null

Ketamine; 85.94 ± 7.23, p = 0.033). These data were confirmed by immunohistochemistry for total rpS6 and p-rpS6 (Fig 3).

Ketamine rapidly stimulates dendritic spine growth in symptomatic Mecp2 mutant mice

To determine whether or not low-dose ketamine can reverse spine deficits in the absence of MeCP2, the density of apical oblique spines was compared among 6 week-old Mecp2 Null mice treated with ketamine or saline, as well as saline treated Wt littermates, 24 hours after treatment. Consistent with our initial studies in naïve animals, Mecp2 Nulls exhibited significant reductions in total and mushroom spine densities on apical oblique dendrites (Fig. 4 and Table 2) compared to Wt littermates. These cohorts also exhibited a modest but significant decrease in apical oblique thin spines (Fig. 4).

Treatment with a single sub-anesthetic dose of ketamine (10 mg/kg I.P.) significantly increased total spine density in Mecp2 Nulls (Table 2) and completely rescued mushroom spine density (Fig. 4 and Table 2). Ketamine also

64 significantly increased the density of stubby spines (Table 2), but had no effect on thin spine density (Table 2). The fact that total spine density was increased by ketamine treatment indicates that the increase in mushroom and stubby spine density was associated with the growth of new spines.

Ketamine stimulates dendritic spine growth following repeated dosing in symptomatic Mecp2 mutant mice

In light of the potential for tachyphylaxis in response to repeated ketamine exposure (discussed in [256]), we next examined whether or not the spine- promoting effect of ketamine treatment in Mecp2 mutants persists following chronic intermittent dosing. To address this issue, we compared spine density and morphology in mPFC pyramidal neurons among Wt animals treated with saline and Null animals treated with either saline or ketamine (10 mg/kg), once every 3 days, for a total of 4 weeks; animals were sacrificed 24 hours after the last treatment. As with single dosing, mutant animals treated for 4 weeks exhibited a significant increase, to Wt levels, in the density of mushroom spines compared to saline-treated mutant controls (Fig 4 and Table 3). However, this effect was relatively transient in that animals examined 72 hours after completion of the chronic intermittent dosing paradigm showed no difference in mushroom spine density between saline- and ketamine-treated Nulls (Fig. 4 and Table 3).

Discussion

Perturbations in cortical synaptic connectivity are a hallmark of RTT and other disorders on the autism spectrum[225, 257]. In the case of RTT, loss of function mutations in Mecp2 are associated with reduced excitatory connectivity onto pyramidal neurons in multiple cortical regions, including the mPFC[87, 88, 226, 246,

247], and impairments in spine density appear to play a critical role in this hypoconnectivity phenotype[88]. Moreover, genetic reversal experiments have shown that cortical spine deficits in mice carrying a conditional Mecp2 Null allele are reversible upon restoration of Mecp2 signaling, in association with improvements in neurological function[69, 71, 72, 118-122]. Together, these findings indicate that treatments that enhance cortical excitatory connectivity in the context of reduced MeCP2 activity may be of therapeutic benefit in RTT.

The present study provides the first detailed analysis of pyramidal neuron spine phenotypes during postnatal development in the mPFC of Mecp2 mutants. Our findings demonstrate that spine deficits appear within three weeks after birth and primarily impact the density of mature, mushroom-type spines in the oblique dendritic compartment, consistent with previous results from our laboratory[2].

Reductions in dendritic spine density have also been reported in the hippocampus [227, 258], somatosensory cortex[87], motor cortex[226, 242] and visual cortex[259] in Mecp2 mutants. As in the mPFC, spine deficits in other cortical regions have also been shown to be compartment-specific. For example, in motor cortex, the apical oblique and apical tuft compartments show reductions in total spine density, but the primary apical and basal compartments are

66 unaffected[242]. Studies in motor cortex also demonstrated that the loss of MeCP2 can have both cell autonomous and non-autonomous effects on spine density; specifically, that both MeCP2 positive and MeCP2 negative cells in Het mice exhibit reductions in the density and maturity of dendritic spines compared to Wt control animals[226].

We previously found that acute treatment of Mecp2 Null or Het mice with a single, sub-anesthetic dose of ketamine rapidly increases cortical expression of the activity-dependent immediate early gene product Fos, which is deficient in

Mecp2 mutants, and reverses abnormal pre-pulse inhibition of acoustic startle[93].

These rapid effects, which occur within 90 minutes of ketamine administration, most likely reflect acute changes in network excitability leading to increased pyramidal cell activity[93]. Potential mechanisms include selective inhibition of inhibitory cortical interneurons or homeostatic adjustments in pyramidal cell firing, as described in other systems[143, 260-262]. However, in addition to acute changes in synaptic function, ketamine can stimulate growth of cortical dendritic spines through activity-dependent mechanisms, including glutamate and BDNF release and activation of mTORC1 signaling[263], leading to durable changes in dendritic morphology[264]. Given the importance of MeCP2 for activity-dependent neuronal plasticity[123, 265, 266], we therefore asked whether or not ketamine could also reverse deficits in structural connectivity in neurons lacking MeCP2. Our findings demonstrate that, indeed, low-dose ketamine treatment stimulates dendritic spine growth in Mecp2 Null mice, sufficient to reverse mushroom spine deficits in the mPFC within 24 hours after treatment. Our finding that dendritic spine deficits are

67 reversible even in the absence of MeCP2 is consistent with previous studies demonstrating that treatments involving choline supplementation[267] or IGF signaling[246, 259, 268] can effectively reverse deficits in dendritic spines.

Given the importance of mushroom spines as sites of strong excitatory synaptic connectivity[269-271], these data suggest that the reversal of spine deficits that we observed following low-dose ketamine treatment may promote enhanced cortical synaptic function in Mecp2 mutants. However, further work is needed to test this hypothesis directly. The ability of low-dose ketamine to promote mushroom spine growth in Mecp2 Nulls, and the fact that this reversal persisted even after 4 weeks of dosing, indicates that the intermittent dosing paradigm we used did not result in tolerance to ketamine with respect to these specific treatment endpoints.

There were, however, subtle differences in the impact of single vs repeated dosing on spine profiles overall. In particular, whereas single dosing increased total spine density, repeated dosing had no such effect (Table 2). This difference may indicate that the initial effect of ketamine is to stimulate both a burst of new spine growth and the maturation of existing spines, whereas, over time, the net effect of repeated dosing is to promote maturation at the expense of ongoing generation of new spines. This possibility is supported by the fact that single, but not repeated dosing significantly increased the density of stubby spines, which are thought to be relatively transient structures associated with spine immaturity[269, 270] (Table 2).

Acute treatment with ketamine also led to enhanced phosphorylation of rp-S6 in

Mecp2 Nulls, one hour after the injection of ketamine, as is typical of ketamine

68 responses in Wt cortical neurons[264]. In other disease models that exhibit reduced mTOR activity and deficits in dendritic spine density in the mPFC, such as rodent models of depression, ketamine acts rapidly to reverse structural connectivity and signaling deficits and acts as an anti-depressant[115]. In these models, increased activity in the mTOR pathway following ketamine treatment, as evidenced by increased phosphorylation of the mTOR target rp-S6, is associated with the upregulation of multiple synaptic proteins such as activity regulated cytoskeletal-associated protein (Arc), PSD95, the AMPA receptor subunit GluR1, and synapsin[231]. In fact, the reversal of structural synaptic connectivity by ketamine treatment in the depression model depends on activation of the mTOR pathway, and these antidepressant effects of ketamine can be blocked by mTOR inhibitors such as rapamycin[115, 235]. It remains to be determined if the reversal of dendritic spine deficits we have observed in Mecp2

Nulls is mediated by similar mechanisms and, in addition, whether or not spine recovery underlies or contributes to recovery of mPFC dependent behaviors.

Nonetheless, our data indicate that ketamine treatment can stimulate the growth and maturation of dendritic spines on pyramidal neurons in Mecp2 Null mice and thereby reverse structural neuronal phenotypes associated with MeCP2 deficiency. Furthermore, our data demonstrate that low-dose ketamine is well tolerated and remains efficacious even after repeated dosing in mice, providing further preclinical support for the therapeutic potential of NMDA receptor antagonists for the treatment of Rett syndrome.

Figure 1. Dendritic spines on apical oblique dendrites in the mPFC were analyzed in Wt and Mecp2 Null mice at 6 weeks of age. Representative images of dendritic segments used for analysis from both Wt and Null animals are shown at the top. There is a significant reduction in both the total density and in the density of mushroom spines on oblique dendrites in the Mecp2 Null mPFC compared to Wt. *p<0.05; Student’s T-test. n=30 segments for Wt, n=20 segments for Null. Data from each segment are shown as open circles within each treatment group.

Figure 2. Dendritic spines on apical oblique dendrites in the mPFC were analyzed in Wt and Mecp2 Null mice at 3 weeks of age, prior to the appearance of overt behavioral symptoms. Representative images of dendritic segments used for analysis from both Wt and Null animals are shown at the top. At this age there is a significant reduction in the density of mushroom spines and trends towards reduced densities of thin and total spines. *p<0.05; Student’s t.test. Total; Wt n=20 segments, Null n=23, Thin; Wt n= 20,

Null n=23, Stubby; Wt n=20, Null=22, Mushroom; Wt n=19, Null n=23. Data from each segment are shown as open circles within each treatment group.

Figure 3. The ratio of phospho-S6 (pS6) to total S6 is markedly reduced in the Mecp2

Null mPFC compared to Wt and is rescued by acute treatment with ketamine. Acute treatment with low-dose ketamine (10 mg/kg, i.p.) significantly increased the pS6/S6 ratio to Wt levels one hour following injection, example images of pS6 and S6 immunostained sections from the mPFC are shown on the left. Representative blots are shown on the right. * p<0.05 ANOVA with LSD posthoc test. n=6 animals for the Wt group, n=6 animals for the Null saline treated group, n=5 animals for the Null Ketamine treated group. Individual data from each animal are shown as open circles within each treatment group.

Figure 4. Mushroom spine deficits on oblique dendrites in the Mecp2 Null mPFC are rescued by either single acute treatment (Null acute) or chronic intermittent treatment

(Null chronic) with ketamine (10 mg/kg, i.p.), 24 hours after dosing. However, spine density in Null mice returns to baseline within 72 hours after dosing (Null Chronic 72 hrs). Representative images of dendritic segments used for analysis from each group are shown on the right. * p<0.05, ***p<0.001 ANOVA with Tukey posthoc test. Group sizes (n) =143 segments for Wt saline, 131 segments for Null saline, 37 segments for

Null Acute, 51 segments for Null Chronic, 43 segments for Null Chronic 72 hrs.

Supplemental Figure 1. Same data set as shown in Figure 4, including segments removed as statistical outliers. * p<0.05, ***p<0.001 ANOVA with Tukey posthoc test.

Group sizes (n) =149 segments for Wt saline, 138 segments for Null saline, 39 segments for Null Acute, 52 segments for Null Chronic, 43 segments for Null Chronic 72 hrs.

Table 1. Raw values for dendritic spine density listed for each spine type in both 6-week old or 3-week old Wt and Null mPFC. Data are displayed as group average ± SEM. P values were determined using Student’s T-test.

Table 2. Raw values for dendritic spine density listed for each spine type in 6-week old Wt and Null mPFC treated with either saline or one of three ketamine paradigms. Data are displayed as group average ± SEM. P values were determined using one-way ANOVA with post hoc Tukey test.

Chapter 3

Activation of the Medial Prefrontal Cortex Reverses Cognitive and

Respiratory Symptoms in a Mouse Model of Rett Syndrome

C. James Howell, Michael P. Sceniak, Min Lang, Wenceslas Krakowiecki,

Fatimah E. Abouelsoud, Saloni U. Lad, Heping Yu, and David M. Katz

Abstract

Rett syndrome (RTT) is a severe neurodevelopmental disorder caused by loss- of-function mutations in the gene encoding methyl-CpG-binding protein 2

(MeCP2)[38], a transcriptional regulatory protein[272]. Mouse models of RTT

(Mecp2 mutants) exhibit excitatory hypoconnectivity in the medial prefrontal cortex (mPFC)[88], a region critical for functions that are abnormal in RTT patients, ranging from learning and memory to regulation of visceral homeostasis[160]. The present study was designed to test the hypothesis that increasing the activity of mPFC pyramidal neurons in heterozygous female

Mecp2 mutants (Hets) would ameliorate RTT-like symptoms, including deficits in respiratory control and long-term retrieval of auditory conditioned fear. Selective activation of mPFC pyramidal neurons in adult animals was achieved by bilateral infection with an AAV8 vector expressing excitatory hm3D(Gq) DREADD

(Designer Receptors Exclusively Activated by Designer Drugs)[273] under the control of the CamKIIa promoter. DREADD activation in Mecp2 Hets completely restored long-term retrieval of auditory conditioned fear, eliminated respiratory apneas, and reduced respiratory frequency variability to wild-type (Wt) levels.

Reversal of respiratory symptoms following mPFC activation was associated with normalization of Fos protein levels, a marker of neuronal activity, in a subset of brainstem respiratory neurons. Thus, despite reduced levels of MeCP2 and severe neurological deficits, mPFC circuits in Het mice are sufficiently intact to generate normal behavioral output when pyramidal cell activity is increased.

These findings highlight the contribution of mPFC hypofunction to the

78 pathophysiology of RTT and raise the possibility that selective activation of cortical regions such as the mPFC could provide therapeutic benefit to RTT patients.

Significance Statement

Rett syndrome (RTT) is a devastating disorder for which no treatments are currently available beyond supportive care. Although significant progress has been made in understanding RTT pathophysiology at the molecular and cellular levels, the relationship between dysfunction in specific brain circuits and the expression of specific RTT symptoms remains unclear. The present study provides the first direct evidence of a link between hypoactivity in the mPFC and cognitive and respiratory symptoms in Mecp2 mutants by demonstrating that activation of the mPFC restores wild-type (Wt) function in these domains. Thus, in addition to highlighting the contribution of mPFC dysfunction to the pathophysiology of RTT, these findings raise the possibility that targeted activation of specific cortical regions could provide therapeutic benefit to RTT patients.

Introduction

Rett syndrome (RTT) is caused by loss-of-function mutations in the gene encoding methyl-CpG-binding protein 2 (MeCP2) and is one of the most physically debilitating disorders on the autism spectrum. RTT patients exhibit a complex constellation of symptoms ranging from deficits in motor function and cognition to dysregulation of breathing and autonomic control (Amir et al., 1999).

Studies in RTT mouse models, which recapitulate the symptomatology of human

RTT, as well as human postmortem studies have revealed that loss of MeCP2 does not result in neuronal degeneration or cell loss[274] but rather in abnormalities in the structure and function of brain microcircuits[96]. These changes include marked alterations in synaptic strength and connectivity[275] which differ among brain regions and appear to be reversible[69, 122]. One of the most striking effects of MeCP2 loss on brain circuit function is a decrease in excitatory synaptic connectivity in the motor, somatosensory, visual, and midline limbic cortices, including the medial prefrontal cortex[275]. Cortical hypoconnectivity is associated with multiple factors, including reduced density and maturity of dendritic spines on pyramidal neurons[88, 226, 227, 242, 276], a shift in the balance of excitatory and inhibitory synaptic signaling molecules toward decreased excitation[88, 277] and, in some cases, increased inhibitory connectivity[277]. As a result, many cortical regions in the Mecp2 mutant brain are hypoactive at rest compared to wild-type (Wt) controls[93].

Hypoactivity of pyramidal neurons in the mPFC in Mecp2 mutants is of particular interest given the role of the mPFC in multiple brain functions that are abnormal in RTT, ranging from learning and memory to respiratory and autonomic homeostasis. Despite this, the role of mPFC dysfunction in the pathophysiology of RTT has been little explored. For example, the ventral mPFC, or “visceral cortex”[167, 168], is responsible for regulating behavioral state-dependent changes in respiratory and autonomic homeostasis, as during stress or in response to conditioned learning[165, 166]. Structures in the ventral mPFC, including the prelimbic (PL), infralimbic (IL), and dorsal peduncular cortex (dPC) give rise to

80 extensive direct projections to cardiorespiratory cell groups in the pons and medulla, as well as indirect projections to subcortical forebrain cell groups that project to the brainstem, including the hypothalamus and amygdala[169]. On the basis of these observations, we hypothesize that hypofunction of the mPFC may influence cardiorespiratory control in RTT. This possibility is supported by the fact that abnormal breathing in RTT worsens with stress and can be nearly normal during sleep, arguing strongly for a significant behavioral component[14, 81].The ventral mPFC has also been shown to be required for cognitive tasks, such as fear memory consolidation and retrieval[207, 208], that are disrupted in RTT. For example, Mecp2 mutants are impaired in their ability to retain memory of auditory conditioned fear 24 h after training[85], a task that requires ongoing activity in the

PL[206] and its projections to the dorsal midline thalamus (dMT)[209]. Therefore, the present study was designed to determine whether or not increasing activity in mPFC pyramidal neurons would ameliorate symptoms of cognitive impairment as well as abnormal breathing in Mecp2 mutant mice.

Materials and Methods

Animals

All experiments were performed on female Mecp2tm1.1Jae Wt and Het mice on a

129S/BalbC/Bl6 background, and each animal was genotyped before and after every experiment to confirm genetic identity.

Surgery and viral constructs

Animals (8-10 weeks of age) were sedated with isoflurane (5% for induction, 1.5-

0.8% for maintenance), fixed in a stereotaxic frame (Stoelting) and then

81 subjected to bilateral craniotomy at bregma +1.75 mm, ±0.25 mm from the midline. Infusions of 500-nl viral construct (AAV8- CamKIIa-hm3D-mCherry; UNC vector core) were made at a depth of 2.0 mm using a Hamilton NeuroS syringe.

For injections in the motor cortex, infusions were targeted at bregma 1.9 mm, 1.5 mm from the midline, and 0.1 mm below the dorsal surface of the brain. To reduce backflow of the injectate along the needle path, infusions were performed at a rate of 100 nl per 2 min with an additional 2 min following the final 100 nl.

After surgery, animals were returned to their cages for two weeks to allow for viral expression.

Designer Receptors Exclusively Activated by Designer Drugs (DREADD)-Gq

Expression Mapping

To verify the distribution of DREADD-Gq labeling following injection (Fig. 1), low- magnification images of mCherry expression from a representative cohort of 14 injected animals (seven Wt and seven Het) were taken using a Zeiss Axiophot.

These images were montaged using Adobe Photoshop, then overlaid onto coronal atlas images[278]. We then used Corel Draw tracing and lenses to create heat maps of DREADD-Gq expression showing the extent of overlap among all

14 animals. Additionally, these images were used to quantify the number of infected cells in the mPFC in Mecp2 Wt and Het mice. ImageJ was used to select and quantify the number of cells in a 200-m2 circular area surrounding each injection site. Data were collected from three tissue sections spanning the rostro- caudal extent of the injection sites from each of five animals per genotype.

Electrophysiology

In vivo extracellular recordings were made in a subset of mPFC-DREADD-Gq animals, two weeks after infection. Animals were anesthetized with urethane

(1000 mg/kg) and placed in a stereotaxic apparatus (Stoelting). Craniotomies were made as described above and tungsten sharp electrodes (AM-Systems) were lowered into the mPFC at the same coordinates used for virus injections.

Filtered extracellular recordings (1 kHz high-pass) were amplified (AM-Systems) and digitally stored on a computer running custom software (Matlab) through an

A/D converter (National Instruments). Extracellular recordings of spontaneous activity within the mPFC were collected under urethane anesthesia while simultaneously monitoring respiration. Recordings of spontaneous activity were collected during the baseline condition for 10 min, and then 45 min after administration of clozapine-N-oxide (CNO; 0.03 mg/kg, i.p.) to determine the effects of mPFC-DREADD-Gq activation on spiking activity. All animals were perfused after recordings to perform a histologic reconstruction of the recording site and to confirm the expression of the mCherry tag associated with the AAV8- hm3-DREADD construct.

Plethysmography

Breathing analysis was performed on unrestrained mice using whole-body plethysmographs (EMMS) in which constant air flow was provided at 1 l/min.

Analysis was restricted to periods of quite breathing, defined as times when the animal was not involved in ambulation, grooming, or investigatory sniffing and all four paws were resting on the chamber floor. Animals were treated with either

83 saline or CNO (0.03 mg/kg, i.p.) and then allowed to acclimate to the chamber for

1 h. After acclimation, respiratory data were gathered over the next 2–3 h. eDaq software (EMMS) was used to analyze the respiratory data for frequency, coefficient of frequency variation (CV), and apneas/min (where an apnea was defined as a respiratory pause lasting at least twice the duration of the average expiratory time, Te). Each animal served as its own control, being first injected with saline before plethysmography, then with CNO 24 – 48 h later, again followed by plethysmography. In addition, in one cohort of animals, respiration was measured a third time, 24 – 48 h after CNO injection and 1 h following saline treatment.

Cue-dependent fear conditioning

Animals were habituated in a fear conditioning chamber (Med Associates) for 90 s before being exposed to four sequential pairings of a 5 kHz, 80 dB tone for 30 s

(conditioned stimulus; CS) which co-terminated with a 0.5-mA foot shock

(unconditioned stimulus; US); each pairing was separated by 30-s inter-stimulus intervals and the time spent immobile during the CS presentation is reported as

% freezing. Animals that did not learn the CS-US pairing, i.e., that did not exhibit increased freezing during conditioning, were excluded from further analysis.

To assess short-term memory (STM) retrieval (4 h after CS-US pairing), animals were exposed to the CS alone, repeated twice with a 30-s interval, and the % freezing during the two CS exposures was averaged. These data are displayed as a percentage of the freezing to CS4 during conditioning ([FreezingTone(LTM or

STM) * 100]/FreezingCS4). Long-term memory retrieval was tested 24 h after fear

84 conditioning using the same protocol as described for STM retrieval testing. To avoid environmental context as a confounding variable, both STM and long-term memory (LTM) testing were performed in a novel environment produced by modifying the fear conditioning chamber with a white floor, black a-frame insert, black and white striped background, and a vanilla scent. Some of the animals used for fear conditioning were heterozygous carriers of Thy1-EGFPMJrs/J that were generated by crossing homozygous male Thy1-EGFPMJrs/J mice on a Bl6 background with heterozygous female Mecp2tm1.1Jae mice from our main colony.

The presence of the transgene was incidental to the present study; these animals were included to increase group sizes during a period of limited animal availability and did not differ in performance from the main colony stock on either fear memory acquisition or short- and long-term fear memory retrieval (p=0. 49, p= 0.51, p= 0.92, respectively by Student’s t test, comparing animals with and without the transgene).

Immunohistochemistry

After behavioral testing, mice were deeply anesthetized with isoflurane and then perfused with 0.9% saline, followed by 4% paraformaldehyde (PFA) by cardiac perfusion. PFA fixed brains were sectioned at 40 µm on a cryostat microtome

(Jung Frigocut 2800N). Free floating sections were exposed to a primary antibody directed against either mCherry (Life technologies M11217; 1:500) or

Fos protein (SySy 226003; 1:3000) overnight at room temperature. The next day, the tissue was incubated sequentially with a biotinylated secondary antibody

(Jackson #112-065-003; 1:400 or Vector BA1000), ABC kit (Vector), and

85 developed with the Sigmafast diaminobenzidine kit. Mounted sections were imaged using a light microscope (Zeiss Axiophot) and photographed with a CCD camera (Q-Imaging).

Blinding and randomization

Animals were selected for all experiments, and assigned to experimental groups without prior knowledge of phenotypic severity. For fear conditioning experiments and Fos-labeling studies, the investigators were blind to genotype and treatment group at all stages including data analysis. Blinding for treatment group was not necessary for plethysmography and electrophysiology experiments because each animal received both saline and CNO and thereby served as its own control.

Statistics

Minimum group sizes were determined using G*Power 3 power analysis software[253]. Each data set was tested for normality using the Shapiro–Wilk test in the Statistical Package for Social Sciences (SPSS), and parametric tests were performed either on the raw data for normally distributed data sets or on log transformed data for non-normally distributed data sets.

Results

DREADD expression in mPFC pyramidal neurons

To enable selective activation of mPFC pyramidal neurons, we injected an AAV8 viral construct expressing the excitatory DREADD AAV8-CamKIIa-hm3DmCherry

(DREADD-Gq; UNC Vector Core) bilaterally into the mPFC of female Wt and

Mecp2tm1.1Jae/+ (Het) littermates (bregma +1.75 mm, ±0.25 mm from the midline, -

2.00 mm depth). Following injection of the viral construct, two weeks were allowed for the expression of DREADD-Gq in the infected neurons before beginning subsequent experimental protocols. Reconstruction of the injection sites from 14 animals revealed that our injection coordinates generated bilateral infection of neurons in the pre- and infra-limbic cortices (PL and IL, respectively) with some spread into the dPC cortex ventrally, and the cingulate (CG) and motor cortices (MCtxs) dorsally (Fig. 1 ; see Methods for details). Animals that displayed either unilateral expression or expression surrounding the lateral ventricles were eliminated from further analysis. To compare the efficiency of viral transduction in Wt and Het animals, we estimated the number of DREADD- mCherry positive cells in the injection site in a subset of animals (see Materials and Methods for details) and found no difference between genotypes (Wt, 106.13

± 8.59; Het, 106.233 ± 11.13). Consistent with previous evidence that AAV8 is not transported either transynaptically or in the retrograde direction[279], we found no cell bodies labeled with DREADD-Gq outside of the regions mentioned above.

To verify that CNO injection altered neuronal activity in the mPFC of mPFC-

DREADD-Gq-infected Het mice, we analyzed in vivo extracellular recordings of spiking activity before and after CNO administration (0.03 mg/kg, i.p.) in anesthetized animals. These experiments revealed that the firing rate of neurons in the mPFC increased 2-fold after CNO (average firing rate 19.67 ± 3.43 pre-

CNO vs 47.60 ± 5.20 post-CNO; Wilcoxon rank sum test; Fig. 2), consistent with previous studies[273, 280-289].

DREADD-positive fibers project to brain regions important for respiratory control and cue-dependent fear memory

To begin elucidating how activation of the mPFC might impact Mecp2 mutant respiratory and cognitive phenotypes, we asked whether or not mPFC-DREADD-

Gq infected neurons project to regions known to be important for either respiratory control or cue-dependent fear memory. To approach this issue, we took advantage of the fact that the mCherry tag in the DREADD-Gq construct can be visualized throughout the infected neurons, including their axonal projections, and amplified by immunocytochemical staining. Analysis of brain sections from mPFC-DREADDGq-infected animals revealed extensive projections of mCherry-labeled axons within numerous components of the ponto- medullary respiratory network, including the parabrachial nucleus and locus coeruleus in the pons, and the caudal raphe, nucleus of the solitary tract (nTS) and ventrolateral reticular formation in the medulla (Fig. 3). Corticobulbar projections to the medulla could be clearly seen exiting the pyramidal tract before ramifying dorsally in the caudal raphe and nTS and laterally in the ventrolateral medulla (VLM; Fig. 3). Fibers were also present in mid- and forebrain structures that influence breathing and project to the ponto-medullary network, including the periaqueductal gray (PAG), hypothalamus, and the basolateral nucleus of the amygdala (BLA), a structure which is also critically important for cue-dependent fear memory consolidation.

Increasing mPFC pyramidal neuron activity restores normal respiration in Mecp2 mutants

To determine whether or not changes in the activity of mPFC pyramidal neurons would alter Mecp2 mutant respiratory phenotypes, we compared breathing in mPFC DREADD-Gq-infected Wt and Het animals 1 h after injection of either saline (control) or the DREADD ligand CNO, using whole-body plethysmography

(Fig. 4, experimental timeline). In a subset of animals, saline was also given 24-

48 h after CNO, and animals were again tested by plethysmography to determine the reversibility of any observed effects of the DREADD ligand. To address the possibility that CNO may have nonspecific effects on respiration[290, 291], we also analyzed the respiratory response to CNO in Hets that had not been infected with

DREADD-Gq (naïve Hets).

As previously described for Mecp2 mutants[275], naïve Hets treated with CNO, as well as mPFC-DREADD-Gq-infected Hets treated with saline exhibited significantly more apneas than Wt animals (ANOVA with Bonferroni post hoc test; Fig. 4). However, in mPFC-DREADD-Gq-infected Hets treated with CNO, the apnea score was reduced to Wt levels (apneas/min; Wt Gq saline, 0.33 ±

0.16; Het Gq saline, 0.87 ± 0.15; Het Gq CNO, 0.29 ± 0.06; Fig. 4). This effect was completely reversible as mPFC-DREADD-Gq-infected Het mice injected with saline 24-48 h after receiving CNO exhibited the same level of apnea as they did before CNO (Het Gq saline #2, 0.91 ± 0.20; Fig. 4). To determine whether or not

CNO activation of neurons infected by spread of the virus into the MCtx contributed to the apnea rescue in mPFC-DREADD-Gq-infected Hets, a parallel

89 series of experiments was performed in animals in which DREADD-Gq infection was restricted to the MCtx. CNO injection had no effect on the number of apneas in these animals (Student’s t test; Het MCtx saline, 1.04 ± 0.34; Het MCtx CNO,

1.24 ± 0.52; Fig. 5). Together, these data demonstrate that activation of excitatory neurons in the mPFC can eliminate the apnea phenotype in Mecp2

Hets.

We next sought to determine whether CNO treatment of animals expressing

DREADD-Gq in the mPFC impacted other respiratory phenotypes previously observed in Mecp2 mutant mice[275]. Therefore, we analyzed both instantaneous respiratory frequency and the CV. Although we found no effect of either genotype or treatment on respiratory frequency (Fig. 6), CV was significantly increased in

DREADD-Gq Hets treated with saline compared to DREADD-Gq Wt controls (Wt

Gq saline, 100.00 ± 4.77, Het Gq saline, 117.41 ± 6.71 CV; Fig. 6). In contrast,

CNO treatment of Hets expressing DREADD-Gq in the mPFC restored CV to Wt levels (Het Gq CNO, 96.85 ± 7.60 CV; ANOVA with LSD post hoc test; Fig. 6).

The increase in CV in Het mice, and the subsequent CV rescue following CNO treatment could be attributed to the effects of genotype and treatment, respectively, on the number apneic pauses, as these effects disappeared when

CV was calculated with apneas removed (data not shown).

DREADD activation of mPFC pyramidal neurons impacts downstream function in respiratory-related neurons

To determine whether activation of the mPFC in Mecp2 mutants altered neuronal

90 activity in downstream respiratory targets, we used immunocytochemical staining to compare expression of Fos protein in response to CNO treatment in mPFC-

DREADD-Gq-infected Wt and Het animals. We previously showed that Fos, the protein product of the activity-dependent immediate early gene, cFos, is a sensitive and accurate marker of differences in neural circuit activity between Wt and Mecp2 mutant mice[93]. We focused in particular on the respiratory subnuclei of the nTS, which exhibit synaptic hyperexcitability and markedly increased Fos expression in Mecp2 mutants compared to Wt[93]. Synaptic hyperexcitability in respiratory subnuclei of nTS, including those involved in regulation of the inspiratory off-switch through the Hering-Breuer reflex pathway, is thought to be a substrate for respiratory hyperreflexia in Mecp2 mutants and may thereby contribute to apneic breathing[93, 292]. In the present study, CNO treatment of mPFC-DREADD-Gq-infected Het mice reduced Fos expression in nTS subnuclei to Wt levels (average Fos expression Wt Gq saline, 19.56 ± 5.76, Het Gq saline,

45.15 ± 15.06, Het Gq CNO, 14.59 ± 2.68; ANOVA with LSD post hoc test; Fig.

7), indicating that activation of mPFC pyramidal neurons can impact neuronal function in downstream respiratory targets. Similar changes were not seen in other brainstem respiratory cell groups, including the VLM and PAG [dorsal PAG

(dPAG); Wt Gq saline, 48.23 8.11, Het Gq saline, 50.38 ± 9.83; Het Gq CNO,

79.32 ± 12.58; lateral PAG (lPAG); Wt Gq saline, 152.58 ± 14.19, Het Gq saline,

169.51 ± 27.71; Het Gq CNO, 209.24 ± 16.34; VLM; Wt Gq saline, 70.19 ± 12.32,

Het Gq saline, 86.99 ± 13.60; Het Gq CNO, 72.43 ± 14.67; Fig. 7], although there was a trend toward increased expression in the dPAG in the Het Gq CNO group.

DREADD activation of mPFC pyramidal neurons restores long-term retrieval of auditory conditioned fear

In addition to cortical modulation of visceral homeostasis, the mPFC plays critical roles in cognition that are mediated through projections to diverse cortical and subcortical cell groups[164, 169]. In particular, direct projections from mPFC pyramidal neurons to the BLA mediate the consolidation of auditory conditioned fear memory following conditioned fear learning, and the short-term retrieval of this memory while mPFC projections to the dMT mediate long-term retrieval[197].

The following series of experiments was designed to determine whether or not hypoactivity in the mPFC was associated with a deficit in retrieval of auditory conditioned fear memory and, if so, whether the deficit is reversible by DREADD activation of mPFC pyramidal neurons. To approach these issues, we compared acquisition and retrieval of auditory conditioned fear memory among five groups of animals; non-infected Het control animals given either saline or CNO (0.03 mg/kg, i.p.), mPFC-DREADD-Gq-infected Het animals treated with either CNO or saline, and mPFC-DREADD-Gq-infected Wt animals treated with saline; all groups received saline or CNO,1 h before training. Animals were trained with exposure to pairings of CS and US across four trials (CS = 5 kHz, 80 dB tone;

US = 0.5 mA shock co-terminating with the last second of the CS). No significant differences in learning were observed between groups across the four conditioning trials (CS4-CS1% freezing; Wt Gq saline, 44.25 ± 4.66; Het Gq saline, 38.51 ± 5.40; Het naïve saline, 38.56 ± 6.18; Het naïve CNO, 34.05 ±

4.37; Het Gq CNO, 33.25 ± 3.80; ANOVA with LSD post hoc test; Fig. 8). To

92 measure fear memory retrieval, animals were subsequently exposed to the CS alone at 4 h after training (testing STM) and 24 h after training (testing LTM;

Schafe and LeDoux, 2000). No significant differences were observed among the groups 4 h after training [freezing (%CS4); Wt Gq saline, 102.62 ± 8.00; Het Gq saline, 78.73 ± 5.06; Het naïve saline, 93.06 ± 15.80; Het naïve CNO, 82.96 ±

8.02; Het Gq CNO, 78.14 ± 8.27; ANOVA with LSD post hoc test; Fig. 8].

However, 24 h after training, mPFC-DREADD-Gq-infected Het mice treated with saline exhibited a significant deficit in their freezing response, as did non-infected

Het controls treated with either saline or CNO, compared to mPFC-DREADD-Gq- infected Wt mice. In contrast, the freezing response in mPFC-DREADD-Gq- infected Het mice treated with CNO was indistinguishable from mPFC-DREADD-

Gq-infected Wt mice treated with saline [freezing (%CS4); Wt Gq saline, 125.53

± 11.42; Het Gq saline, 70.67 ± 8.42; Het naïve saline, 81.98 ± 12.23; Het naïve

CNO, 87.05 ± 6.19; Het Gq CNO, 123.16 ± 16.87; ANOVA with LSD post hoc test; Fig. 8]. These data demonstrate that activation of the mPFC in Mecp2 mutants can restore long-term retrieval of auditory conditioned fear memory to

Wt levels, 24 h after training. Given that the duration of CNO action is 6h[293], these data indicate that transient activation of the mPFC in Mecp2 mutants leads to a durable change in mechanisms underlying fear memory retrieval.

Discussion

The present findings demonstrate the importance of mPFC hypofunction in RTT by showing that activation of mPFC pyramidal neurons can reverse abnormalities

93 in breathing and long-term retrieval of conditioned fear learning in Mecp2 Het mice, a model of RTT that recapitulates the genetic mosaicism and many phenotypic characteristics of the human disorder. It seems likely that distinct mechanisms downstream of enhanced pyramidal neuron activity underlie the reversal of these respiratory and cognitive abnormalities, respectively. This is underscored by the fact that the normalization of respiratory apneas and respiratory variability was transient and undetectable 24 h after CNO treatment, the same time point at which fear memory retrieval was rescued in mPFC-

DREADD mutants. These differences could be explained by the fact that different populations of mPFC neurons innervate brainstem versus cortical and subcortical fore brain targets[169].

In general, loss of Mecp2 is thought to disrupt resting breathing by shifting synaptic excitatory-inhibitory balance toward increased excitability at multiple loci throughout the brainstem respiratory network[76, 79, 153]. We hypothesize, therefore, that activation of the mPFC normalizes resting breathing in Het mice by acutely modulating hyperexcitability of the brainstem network. Moreover, our data suggest that loss of cortical modulation resulting from hypoactivity of pyramidal neurons in the mPFC may underlie or contribute to behavioral dysregulation of breathing in RTT[88]. Based on the projections of mPFC-

DREADD-Gq-infected neurons, the modulation of respiratory output that we observed could have been mediated by direct corticobulbar inputs to ponto- medullary respiratory cell groups, including the nTS, VLM, and LC, and/or by projections to structures that indirectly modulate breathing, such as the

94 hypothalamus, amygdala, and PAG. The fact that mPFC activation can normalize mutant levels of Fos expression in respiratory subnuclei of nTS suggests that cortical inputs may influence the breathing pattern in RTT, at least in part by modulating hyperexcitability of brainstem neurons involved in reflex regulation of respiratory motor output. Such a mechanism would be consistent with previous studies demonstrating that the mPFC regulates gain and sensitivity of cardiovascular reflexes mediated by the nTS[167, 294, 295]. However, given the complexity of the brainstem respiratory network, we cannot rule out the possibility that stabilization of the breathing pattern in DREADD-activated mutants involves alterations in excitatory/inhibitory balance in structures other than the nTS as well[296].

Further studies will be required to determine which subregion(s) of the mPFC are responsible for the rescue of Het breathing phenotypes described here, as the

PL, IL, and dPC all have direct and/or indirect connections with brainstem nuclei involved in cardiorespiratory control[169], as well as the impact of DREADD activation of mPFC pyramidal neurons on the strength of these connections.

More generally, there is growing appreciation for the fact that hypofrontality[297], and abnormalities in the corticofugal projection network in particular contribute to the pathophysiology of diverse neuropsychiatric diseases[298]. However, despite the long-standing recognition that respiratory dysfunction in RTT exhibits features of limbic cortical dysfunction[299], a role for prefrontal corticofugal influences on breathing abnormalities in RTT has not previously been described. By highlighting the ability of corticofugal pathways to modulate respiratory output in

Mecp2 mutants, these findings raise the possibility that cortical influences on breathing could be harnessed to improve respiratory control in RTT.

Ongoing activity of mPFC pyramidal neurons is also critical for diverse cognitive tasks, including consolidation and retrieval of cue-dependent fear conditioning, which requires activity of neurons projecting from the mPFC to the BLA and the dMT for short- and long-term retrieval, respectively[197, 206, 209]. In fact, the LTM deficit that we and others have observed in Mecp2 mutants, 24h after training[70,

85], is reminiscent of similar deficits described in normal mice following pharmacologic blockade of activity in the PL after training and before testing for fear memory retrieval[206]. Interestingly, our data indicate that the PL to BLA projection circuitry involved in short-term retrieval of conditioned fear memory is functional in Mecp2 mutants whereas the PL to dMT projection circuitry required for retrieval at 24 h is impaired. Given that mPFC activation during fear conditioning restores long-term retrieval to Wt levels, our data suggest that increasing the activity or excitability of mPFC pyramidal neurons during training enables the shift in dependence from PL-BLA to PL-dMT projections that occurs by 24 h after training[197].

The fact that both respiratory apneas and expression of cue-dependent conditioned fear memory can be restored to Wt levels in Mecp2 mutants indicates that the circuitry required for these behaviors is sufficiently intact to respond appropriately to activation of pyramidal neurons in the mPFC. This finding is consistent with the fact that loss of MeCP2 results in reversible changes in neural circuit function rather than neuronal cell loss or

96 degeneration[34, 69, 122, 300-303]. Given that loss of Mecp2 results in functional hypoconnectivity in multiple cortical regions in addition to the mPFC[95], we predict that the approach described here is likely to be effective at ameliorating other symptoms in RTT mice as well. Moreover, these data raise the possibility that treatment strategies aimed at increasing activity in specific cortical networks may provide therapeutic benefit to RTT patients (see also [301]).

Figure 1. DREADD-Gq expression in the mPFC. The distribution of DREADD-Gq-labeled neurons was plotted from low-magnification micrographs of mCherry labeling in serial coronal sections through the forebrain in 14 animals; the plots were then overlaid to produce heat maps of the injection sites in which the color coding represents the extent of overlap in the distribution of labeled neurons in all 14 animals. These heat maps are shown on the right side of each figure, with representative hemi-brain sections shown on the opposite side (note that all animals received bilateral injections). The faint labeling seen lateral to the mPFC at 1.77 and 1.41 mm rostral to bregma is mCherry expression in the axons of infected neurons. Bottom right, Representative photomicrograph of mCherry- labeled neurons in a coronal section through the mPFC of an infected animal. Anatomic

98 labels for this and all subsequent figures as follows: AC, anterior cingulate; AID, dorsal agranular insular cortex; AIV, ventral agranular insular cortex; CPu, caudate putamen; DI, dysgranular insular cortex; DP, dorsal peduncular cortex; DTT, dorsal tenia tecta; fmi, forceps minor of the corpus callosum; Fr3, frontal cortex area 3; IL, infralimbic cortex; LO, lateral orbital cortex; M1, primary motor cortex; M2, secondary motor cortex; MO, medial orbital cortex; PrL, prelimbic cortex; S1, primary somatosensory cortex; S1DZ, somatosensory cortex dysgranular zone; S1FL, forelimb region of the somatosensory cortex; S1J, jaw region of the somatosensory cortex; VO, ventral orbital cortex. Inset shows a high-magnification view of an infected neuron.

Figure 2. Activation of mPFC pyramidal neurons by DREADD-Gq increases neuronal activity in vivo. Left, Representative extracellular traces of spontaneous activity in the mPFC of mPFC-DREADD-Gq Het animals before and after CNO treatment. Right,

Summary data from two animals; the total number of recorded cells is shown within each bar. **p < 0.01 by Wilcoxon rank sum test.

100

Figure 3. Photomicrographs showing mCherry-positive fibers in coronal sections through brain regions important for respiratory control and cue-dependent fear memory following

DREADD-Gq infection of the mPFC. Numbers in the top right of each micrograph correspond to the distance from bregma (mm). 4V, fourth ventricle; Aq, aqueduct of

Sylvius; BA, basal amygdala; CC, central canal; CE, central amygdala; cnTS, commissural nTS; LC, locus coeruleus; LA, lateral amygdala; mnTS, medial nTS; PB, parabrachial nucleus; py, pyramidal tract; vlnTS, ventrolateral nTS (abbreviations from Franklin and

Paxinos, 2012).

101

Figure 4. DREADD-Gq activation of pyramidal neurons in the mPFC eliminates the apneic breathing phenotype in Mecp2 mutants. The timeline of DREADD injection, CNO

(or saline) treatment and plethysmographic recordings is shown at the top of the figure.

The bar graph shows summary data illustrating the apneas/min for each experimental group with group sizes included within each bar. * p < 0.05 by ANOVA with Bonferroni post hoc test. Bottom right, Representative respiratory traces; ▼indicates typical respiratory pauses scored as apneas.

102

Figure 5. Activation of pyramidal neurons in the motor cortex by DREADD-Gq does not alter respiration. Left, the distribution of DREADD-Gq-labeled neurons was plotted from low-magnification micrographs of mCherry labeling in coronal sections through the forebrain in all animals tested; the plots were then overlaid to produce a heat map of the injection sites in which the color coding represents the extent of overlap in the distribution of labeled neurons in all seven animals. Note that all animals received bilateral injections.

Right, Animals expressing DREADD-Gq in the motor cortex were subjected to two consecutive days of plethysmographic recording of respiration, 1 h following either saline treatment (day 1) or CNO treatment (day 2). The bar graph shows summary data illustrating the apneas/min on day 1 (saline) and day 2 (CNO), with group sizes included within each bar.

103

Figure 6. DREADD-Gq activation of pyramidal neurons in the mPFC reduces respiratory frequency variability to Wt levels without impacting the average frequency of respiration.

Left, Summary data showing the mean instantaneous respiratory frequency for each experimental group. Right, Summary data showing the coefficient of variation (CV) for instantaneous respiratory frequency for each experimental group. Group sizes are shown within each bar. * p < 0.05 by ANOVA with LSD post hoc test.

104

Figure 7. Activation of mPFC pyramidal neurons by DREADD-Gq impacts neuronal function in respiratory subnuclei of the nTS. Animals were sacrificed 90 min after either saline CNO injection and processed for Fos immunostaining. Photomicrographs show representative sections through the caudal nTS from 3 different animals, each of which received bilateral injections of DREADD-Gq in the mPFC. CNO treatment restored Fos levels in the nTS of mPFC-DREADD-Gq Het mice to Wt levels. * p < 0.05, ** p < 0.01,

ANOVA with LSD post hoc test. No significant differences were observed between treatment groups in the dPAG, lPAG or VLM.

105

Figure 8. Activation of mPFC pyramidal neurons by DREADD-Gq rescues long-term expression of cue-dependent fear memory. Left, Conditioning to CS-US pairings across four trials demonstrates similar responses in all groups using % freezing (immobility) during the CS presentations as an index of fear learning. Right, Short-term (4 h after conditioning) and long-term (24 h after conditioning) memory retrieval were tested by exposure to the CS alone, with the animal in a test chamber that was distinguished from the conditioning chamber by novel environmental features (see Materials and Methods).

Data are presented as a percentage of the freezing to CS4 during conditioning

([FreezingTone(LTM or STM) * 100]/FreezingCS4). * p < 0.05 by ANOVA with LSD post hoc test.

106

Chapter 4

Specific Activation of the mPFC Projection Circuitry targeting the

Dorsal Midline Thalamus and Its Impact on Cue-Dependent Fear

Conditioning Deficits in a Mouse Model of Rett Syndrome

107

Introduction

As the data from Chapter 3 demonstrate, activation of the mPFC and its projection targets is sufficient to restore expression of conditioned cue-dependent fear memory and eliminate respiratory apneas in heterozygous female Mecp2 mutant mice. Activation of the mPFC may influence these behaviors through several direct or indirect projections. As it relates to cue-dependent fear conditioning (CDFC), activation of the PNs in the mPFC would be expected to exert influence both directly onto the amygdala, and indirectly through its projection to the dorsal midline thalamus (dMT) [197, 204, 205]. Direct projections to the amygdala from the mPFC have little interaction with the central amygdala

(CeA), the output region of the amygdala that projects to motor and autonomic regions which allow for the expression of fear[304], but rather, target the BLA, whose principle neurons subsequently project to the CeA (discussed in [305]).

Projections from the mPFC to the BLA are heterogeneous, with the prelimbic cortex (PL) enhancing CS-evoked freezing when stimulated and reducing freezing when lesioned[206, 208, 306] while, on the other hand, the infralimbic cortex

(IL) is thought to drive fear inhibition and is critical for fear memory extinction[208,

306-314]. This heterogeneity of mPFC projections to the BLA is thought to result from differences in the degree to which projections from each sub-region interact with inhibitory intercalated cells (ITCs) within the BLA, an inhibitory cell group that projects to the CeA to inhibit fear related behaviors such as freezing to the

CS[315-319]. Specifically, the IL terminates more heavily near ITCs than does the

PL[304, 320], and stimulation of the IL leads to activation of the ITC population[321].

108

So how then would activation of both the PL and IL, as in Chapter 3, lead to increased freezing responses to CS presentation 24 hours after CDFC given that

PL and IL likely have opposing influence on BLA output? A likely explanation for this apparent conundrum comes from studies demonstrating that the mPFC to

BLA projection is only necessary for short-term consolidation and expression of fear memory, but is dispensable at long-term time points[197]. Specifically, optogenetic inhibition of PL projections to the BLA significantly reduce freezing to the CS at 6 hours after conditioning but not at 7 days after conditioning. On the other hand, specific optogenetic inhibition of PL projections to the dMT reduces freezing to the CS at 7 days after conditioning but not at 6 hours[197]. These data, along with the data from Chapter 3, suggest that PL projections related to long- term memory in CDFC, such as those to the dMT, may be specifically impacted by the loss of MeCP2 while other PL projections, such as those to the BLA, may be spared. To test this hypothesis, the following preliminary experiments have been carried out to determine whether or not activation of the specific projections from the mPFC to the dMT is sufficient to restore expression of long-term fear conditioned memory in the CDFC paradigm.

109

Methods

Surgery and viral constructs

Female mice at 15-16 weeks of age were sedated with isoflurane (5% for induction, 1.5-0.8% for maintenance), fixed in a stereotaxic frame (Stoelting) and then subjected to bilateral craniotomy at bregma +1.75 mm, ±0.25 mm from the midline and a single midline craniotomy at bregma -1.00. An infusions of 500-nl of the CRE-dependent viral construct (AAV2-hSyn-DIO-hm3D-mCherry;

Addgene) was made bilaterally into the mPFC of Mecp2 Hets and Wt mice at a depth of -2.00 mm, similar to the experiments in Chapter 3. In addition, a 130 nl infusion of a retrogradely transported rAAV2 viral vector expressing CRE (rAAV2-

CAG-CRE; UNC) was made into the dMT along the midline at Bregma -0.75 mm and a depth of 3.30 mm. Control animals were infused with the CRE-dependent

DREADD virus into the mPFC without the corresponding dMT infusions to confirm that DREADD-Gq expression was tightly controlled in the absence of

CRE. All infusions were made using the Nanoject II system (Drummond) at a rate of 13.5 nl per minute for the dMT and 46.0 nl per minute for the mPFC. After surgery, animals were returned to their cages for two weeks to allow for viral expression.

Cue-dependent fear conditioning

The animals were treated with either saline or CNO (0.03 mg/kg, i.p.) 1 hour prior to conditioning. Animals were habituated in a fear conditioning chamber (Med

Associates) for 90 s before being exposed to four sequential pairings of a 5 kHz,

110

80 dB tone for 30 s (conditioned stimulus; CS) which co-terminated with a 0.5- mA foot shock (unconditioned stimulus; US); each pairing was separated by 30-s inter-stimulus intervals and the time spent immobile during the CS presentation is reported as % freezing. Short-term and long-term memory were assessed by exposure to the CS alone as described in Chapter 3. Data are displayed as a percentage of the freezing to CS4 during conditioning ([FreezingTone(LTM or STM) *

100]/FreezingCS4).

Quantification of CRE-Dependent DREADD Expression

(Jung Frigocut 2800N). Mounted sections were imaged using a light microscope

(Zeiss Axiophot) and photographed with a CCD camera (Q-Imaging). To verify the distribution of CRE-dependent DREADD-Gq labeling following injection, we imaged serial sections through the mPFC of each animal and quantified the number of cell bodies labeled with mCherry.

Immunohistochemistry

40 µm free floating sections throughout the dMT were exposed to a primary antibody directed against CRE (Cell Signaling D7L7L; 1:150) overnight at room temperature then incubated with a 488 conjugated secondary antibody (Life

Technologies A11034; 1:400). Mounted sections were imaged using a light microscope (Zeiss Axiophot) and photographed with a CCD camera (Q-Imaging).

111

Results

DREADD expression in mPFC pyramidal neurons projecting to the dMT

To achieve selective activation of neurons which project from the mPFC to the dMT I injected a retrogradely transported rAAV2 CRE-expressing virus into the dMT and a CRE-dependent virus expressing DREADD-Gq-mCherry bilaterally into the mPFC. This allowed for DREADD-Gq-mCherry expression in a subset of

Layer V/VI PNs in the mPFC (Fig. 1). The number of DREADD-Gq-mCherry PNs was quantified in serial sections from each animal and the average number of cells per section was created from the combined counts for each genotype. This analysis revealed a significant increase in the number of cells expressing mCherry in the Het animals compared to Wt (Wt = 5.62 ± 1.94; Het = 17.97 ±

5.46; n = 4 for Wt, n = 7 for Het. p < 0.05, Student’s T.test; Fig. 3).

DREADD activation of mPFC pyramidal neurons projecting to the dMT enhances expression of LTM retrieval in Wt but not Mecp2 mutant mice.

The results from Chapter 3 demonstrated that global DREADD-Gq activation in mPFC PNs was sufficient to reverse deficits in LTM memory expression after

CDFC in Mecp2 Hets. This finding led us to wonder if hypofunction in a specific projection of the mPFC could be responsible for the deficit in LTM expression.

Recent data from the Quirk lab has demonstrated that while activity in the mPFC is necessary for expression of fear memory at any time point, the projections from the mPFC to the dMT in particular are necessary for expression of LTM[197].

112

This data, combined with our finding that LTM, but not STM, is impacted by the loss of MeCP2 led us to hypothesize that activation of the projection neurons from the mPFC to dMT would be sufficient to reverse deficits in LTM memory expression after CDFC in Mecp2 Hets.

To test this hypothesis, animals were treated with either saline or CNO (0.03 mg/kg, i.p.) 1 hour prior to training. Animals were trained with exposure to pairings of CS and US across four trials (CS = 5 kHz, 80 dB tone; US = 0.5 mA shock co-terminating with the last second of the CS). Fear memory retrieval was then tested using exposure to the CS alone at either 4 or 24 hours after conditioning as in Chapter 3. No significant differences were observed among the groups 4 hours after training [freezing (%CS4); Wt Gq saline, 103.50 ± 16.04;

Het Gq saline, 95.43 ± 13.71; Wt Gq CNO, 121.10 ± 49.74; Het Gq CNO, 107.15

± 15.47; ANOVA with LSD post hoc test; Fig. 2]. However, 24 hours after training, Wt DREADD-Gq expressing animals treated with CNO displayed significantly increased freezing to the CS compared to all other groups [freezing

(%CS4); Wt Gq saline, 73.87 ± 7.63; Het Gq saline, 103.99 ± 15.48; Wt Gq CNO,

207.28 ± 121.15; Het Gq CNO, 90.94 ± 16.89; ANOVA with LSD post hoc test;

Fig. 2]. These data, while very preliminary, suggest that while activation of the projections from mPFC to dMT is sufficient to increase expression of LTM in the

Wt, it is not sufficient in the Mecp2 Het.

113

Discussion

These preliminary data suggest that specific activation of the projection from the mPFC to the dMT is sufficient to increase freezing to the CS at long-term time points in Wt animals, but is not sufficient to alter freezing in Hets. If these findings are supported with larger group sizes it may suggest that while activation of all excitatory PNs in the mPFC is sufficient to increase LTM fear memory in Mecp2

Hets, specific activation of the subset of PNs that project to the dMT is not sufficient to overcome deficits in this paradigm. Future studies will be required to determine whether or not activation of this single projection from the mPFC is insufficient, and subsequent studies could use similar approaches to determine whether or not activation of mPFC projections to other targets could restore LTM in Mecp2 Hets. For example, activation of mPFC projections to the BLA may be required either instead of, or in addition to, the mPFC projections to the dMT to mimic the rescue described in Chapter 3.

Additionally, we used the mCherry tag to determine the pattern of CRE- dependent DREADD-Gq expression in the mPFC for each animal. This revealed that Wt animals expressed DREADD-Gq in fewer mPFC neurons than did the

Hets (Fig. 3) suggesting the possibility of a wiring abnormality in the Het mPFC.

Therefore, future studies could determine whether or not an increased number of projections from the mPFC to the dMT is the result of improper pruning, for example. In addition, we compared freezing responses during LTM testing to the number of DREADD-Gq positive cells in the mPFC of Hets versus Wt animals.

This revealed a higher freezing response in the Wt animals compared to Hets

114

(Fig. 3) despite the fact that Mecp2 Hets display a greater number of DREADD-

Gq positive cells in the mPFC. Together, these data suggest that the deficit in

LTM observed in Mecp2 Het may be the result of abnormalities in multiple pathways that involve direct and/or indirect projections from the mPFC. Further experiments will be required to determine how global activation of mPFC PNs differs from this selective activation of mPFC projections to the dMT to better understand the mechanisms underlying abnormal expression of LTM.

115

Figure 1. Representative images from the mPFC and dMT demonstrating CRE- dependent DREADD-Gq and CRE expression respectively. Left, representative image of

CRE-dependent DREADD-Gq-mCherry-labeled neurons in a coronal section from the mPFC from an infected animal. Note that all animals received bilateral injection. Right, representative image of CRE expression in a coronal section from the dMT from an infected animal. Anatomic map and labels are from Paxinos and Franklin’s Mouse Brain

Atlas 4th edition and abbreviations are as follows: CM, central medial thalamic nucleus;

D3V, dorsal 3rd ventricle; MD, medial dorsal thalamic nucleus; PT, paratenial thalamic nucleus; PVA, paraventricular thalamic nucleus anterior.

116

Figure 2. Preliminary data suggests that activation of mPFC projection neurons that target the dMT increases LTM expression in Wt but not in Mecp2 mutant mice. A) Conditioning to CS-US pairings across four trials demonstrates similar responses in all groups using % freezing (immobility) during the CS presentations as an index of fear learning. B) Short- term (4 h after conditioning) and C) long-term (24 h after conditioning) memory retrieval were tested by exposure to the CS alone, with the animal in a test chamber that was distinguished from the conditioning chamber by novel environmental features. Data are presented as a percentage of the freezing to CS4 during conditioning ([FreezingTone(LTM or STM) * 100]/FreezingCS4). For A-C, n = 4 for Wt saline and Wt CNO treated groups, n

= 7 for Het saline and CNO treated groups.

117

Figure 3. Mecp2 Hets display increased CRE-dependent DREADD-Gq expression in the mPFC (Left) but reduced LTM retrieval response to stimulation of mPFC to dMT projection neurons (Right). The number of cells were determined by quantifying mCherry+ cell bodies in every other section thorughout the mPFC (Left). The graph on the right is produced by dividing %Freezing during LTM testing by the number of cells quantified on the left. n = 4 for Wt and n = 7 for Het. Only animals treated with CNO are shown. * = p < 0.05, ** = p <

0.01 by Student’s T.test.

118

Chapter 5

The studies described in this thesis were designed to better understand how loss of MeCP2 leads to dysfunction in brain circuitry underlying specific symptoms at the core of Rett syndrome, and to determine if a particular brain region could become a potential target for future therapeutic approaches. Our earlier study provided a map[93] which made it clear that the Mecp2 mutant brain displayed a distinct pattern of activity imbalance; namely, that excitatory neurons in the cortex and several other forebrain regions were hypoexcitable while neurons in brainstem regions were more likely hyperexcitable. The pattern of Fos expression we discovered led to the hypothesis that activity imbalance in regions across the neuraxis is likely connected, and that a hub region capable of influencing many regions throughout the brain could be a particularly powerful therapeutic target.

I chose to focus my studies on the mPFC due in part to its extensive connectivity to both cortical and subcortical targets throughout the brain which allow it to influence visceral, limbic, and cognitive functions[158-160], all of which are abnormal in Rett syndrome. This connectivity to many critical brain regions across the neuraxis, along with our previous data demonstrating that excitatory pyramidal neurons (PNs) of the mPFC display reductions in Fos expression, excitatory drive, and dendritic spine density[88, 93] led to me choosing to focus on the mPFC for these studies. Using the mPFC as a model region, these experiments were designed to test the hypothesis that excitatory hypoconnectivity could be corrected, and that reduced activity in projection

119 neurons could be reversed as a means to restore function in numerous behavioral domains impacted by the loss of MeCP2. Moreover, I sought to determine whether or not hypofunction in specific brain circuitry underlies or contributes to specific behavioral dysfunction in a mouse model of Rett syndrome.

Our lab previously demonstrated that excitatory PNs of the Mecp2 mutant mPFC display reduced excitatory drive and reduced Fos expression, and found that these were likely due, at least in part, to reduced excitatory synaptic connectivity onto PNs[88, 93]. The experiments described in this thesis took two separate approaches to understanding the role and reversibility of mPFC PN hypofunction in Mecp2 mutants: first, to determine whether or not the deficits in excitatory connectivity were reversible even in the absence of MeCP2 (discussed in

Chapter 2), and second, to bypass deficits in excitatory synaptic drive in the mPFC by directly increasing the excitability of excitatory PNs using viral expression of the DREADD-Gq receptor, then determine the influence of increasing activity in mPFC projection circuitry on behavioral outcomes dependent on mPFC function (discussed in Chapter 3).

Low-dose ketamine treatment in mouse models of Rett syndrome

To address the first question regarding the reversibility of deficits in excitatory connectivity in the absence of MeCP2, I turned again to ketamine. In our previous study[93], we found that Mecp2 mutants responded to low-dose ketamine

120 treatment by increasing activity, as measured by Fos expression, in cortical PNs, similar to their Wt counterparts. This acute increase in activity was also associated with a rescue of sensory-motor gating in pre-pulse inhibition of acoustic startle, a behavior that is regulated by forebrain regions such as the limbic cortices where reduced Fos expression was rescued by ketamine[93]. This suggests that reduced excitability in PNs in these regions may underlie or contribute to cognitive phenotypes in the context of reduced MeCP2 expression.

It is also suggestive that increasing activity in Mecp2 mutant limbic cortical circuits was sufficient to restore processing in a cognitive task.

Interest in ketamine and other N-methyl-D-aspartate receptor (NMDAR) antagonists as potential therapeutics for Rett syndrome comes from two initial preclinical studies (reviewed in [95]). First, our findings in 2012 demonstrated that acute low-dose ketamine treatment can increase Fos expression in the frontal cortex to Wt levels and reverse abnormalities in sensorimotor gating[93]. The second line of evidence is a series of findings demonstrating that abnormalities in the expression of NMDAR subunits leads to increased inhibitory connectivity in visual circuits and is associated with abnormal visual processing[277, 322, 323]. More specifically, NMDARs in young animals contain GluN2B, but this subunit is later downregulated in favor of the GluN2A subunit as circuits mature[322]. This developmental shift is reversed in Mecp2 mutants[323], and interestingly, genetic reduction of the GluN2A subunit can prevent deficits in visual processing in

Mecp2 mutants[277]. This link between abnormal NMDAR signaling and inhibitory hyperconnectivity was further supported in a recent study which demonstrated

121 that daily low-dose ketamine treatment is effective at reversing inhibitory hyperconnectivity and increases activity in PNs of the visual cortex[250]. In addition to the evidence supporting the therapeutic potential of NMDAR antagonism in mouse models of Rett syndrome, early safety trials for ketamine in human patients with Rett syndrome is also showing potential efficacy (reviewed in [275]).

However, ketamine is known not only for its ability to acutely increase the excitability of principle cortical neurons, but also for the ability to stimulate mechanisms of synaptic plasticity leading to long-term changes in cellular morphology and synaptic connectivity[115, 250]. Therefore, in the studies described in Chapter 2, I used dendritic spine density and maturity as endpoints to determine whether or not doses of ketamine shown to be effective at reversing dendritic spine deficits in other models could also reverse these deficits in a mouse lacking MeCP2[231]. It was unknown whether or not the necessary mechanisms could be properly engaged in the absence of MeCP2, namely the increase in mTOR signaling necessary for translation of synaptic proteins, given the importance of MeCP2 in activity-dependent neuronal plasticity. Therefore, I designed the studies in Chapter 2 to provide a detailed analysis of spine deficits on PNs in the mPFC of Mecp2 mutants, and determine whether or not these could be reversed by ketamine treatment.

Dendritic spines are sites of excitatory synaptic connectivity[324] and have been used as a measure of connectivity in other models[325]. These protrusions are known to grow from sites on the dendrite that are in close proximity to glutamate

122 release[269]. Further, it has been found that spines are closely juxtaposed to presynaptic terminals; in young circuits, multiple presynaptic terminals may be associated with a single spine, but as circuits mature, spines generally correspond to presynaptic terminals in a 1:1 ratio[324]. Studies of dendritic spine morphology have also described mechanisms by which stimulated spines develop into mature, electrically active structures responsive to glutamate[269]. In their thin, immature state, dendritic spines express NMDA receptors but not

AMPA receptors and correspond to ‘silent synapses’[270]. During the maturation of a spine, structural proteins such as actin and PSD-95 are localized to the postsynaptic density and enable the synaptic localization and stabilization of AMPA receptors[270]. This, in turn, enlarges the spine into the more mature stubby or mushroom structure[324]. Given this understanding of dendritic spines as the postsynaptic portion of an excitatory synapse, the study in Chapter 2 used dendritic spine density as a measure of excitatory synaptic connectivity; however, it is worth considering the limitation of such a proxy for connectivity given the possibility that the processes underlying synaptogenesis are abnormal in the context of reduced MeCP2.

The findings in Chapter 2 demonstrate that deficits in dendritic spine density and maturity begin to appear early in postnatal development, at or before 3 weeks of age. These deficits primarily impact the apical oblique compartment of mPFC

PNs, and reductions in total spine density are a result of a specific reduction in the density of spines with the mature mushroom morphology. This phenotype persists and worsens with age, with 6-week-old mutants displaying more severe

123 reductions in oblique dendritic spine density and maturity. Dendritic spines have been studied in several other regions in mouse models of Rett syndrome, and the data have demonstrated some similarities and some differences with our data from the mPFC (reviewed in [247]). Similarities included deficits in dendritic spine density in other regions of the cortex including somatosensory[87], motor[86, 242], and visual cortices[259] as well as the mPFC in our previous study[88], and the hippocampus both in vivo and in vitro[227, 258]. Unlike the study described in

Chapter 2, the majority of these studies did not differentiate between dendritic compartments or analyze spine morphology. However, in the motor cortex, deficits in dendritic spine density are present on apical oblique and apical tuft domains while the primary apical and basal domains are spared[242]. This differs from our findings for the mPFC in that only the apical oblique dendrite displays reduced density, while the other domains are mostly spared. Knowing which dendritic compartments are impacted by the loss of MeCP2, and which are spared, may provide insight into which inputs to these cells are impacted in Rett syndrome. For example, layer V PNs in the mPFC, such as those studied in

Chapter 2, receive direct input to the apical and apical oblique domains from midline thalamic nuclei[326] and input to their basal dendrites from the BLA[177, 327].

This may suggest that inputs from the BLA to the mPFC are spared while those from the dorsal midline thalamus (dMT) may display a reduced number of synapses onto PNs of the mPFC, a hypothesis that gains further support from the findings in Chapter 3 discussed below.

124

In this study, we took two different approaches to determine whether or not deficits in dendritic spine density and maturity were reversible in the absence of

MeCP2. First, Mecp2 mutant mice were treated with a single low dose of ketamine (10mg/kg) at 5-6 weeks of age to determine whether or not established deficits in dendritic spine density and maturity are reversible in response to acute treatment, as in other rodent models[232]. Indeed, a single dose of ketamine stimulated a modest, but significant, increase in total spine density on oblique dendrites and a complete rescue of spine maturity, as measured by the density of mushroom spines, to Wt levels (Chapter 2, Fig. 4). My second approach asked two questions: 1, would early treatment prevent the deficits in dendritic spine density and maturity, and 2, would chronic intermittent treatment with low-dose ketamine, designed to mimic a translatable treatment paradigm, restore dendritic spine density after one month of treatment. This second question was of particular importance given the possibility of reduced efficacy, or tachyphylaxis, after repeated exposure[256]. Pertaining to both questions, we found that chronic intermittent treatment led to a complete reversal of mushroom spine density with a partial reversal of total spine density. This demonstrates that ketamine treatment is capable of at least partially reversing deficits in mPFC PN dendritic spine morphology at either early or late stages in disease progression, and repeated treatment is well-tolerated with no reduction in efficacy.

Low-doses of ketamine such those used in Chapter 2 have been shown to have rapid antidepressant-like effects in murine models of depression as measured by improvements in forced swim tests, learned helplessness, and other

125 assessments of depression in rodents (reviewed in [115]). This antidepressant effect depends on at least 2 connected mechanisms of action of ketamine: first, low-dose ketamine treatment rapidly induces BDNF translation which is known to be required for the anti-depressant effects of ketamine[328]; and second, activity in the mTOR signaling pathway, which is stimulated by BDNF interaction with the

TrkB receptor, is increased and leads to the translation of a host of synaptic proteins involved in both remodeling synaptic structure and increasing receptor densities, including AMPA receptors which act to increases excitability of the synapse (reviewed in [114, 115]; also see [231]). Blockade of either of these processes is sufficient to block the antidepressant effects of ketamine and to block the remodeling of dendritic spines on excitatory PNs[231, 329]. This raises the question of whether or not all of these mechanisms are intact in the absence of

MeCP2. For example, increased translation and insertion of AMPA receptors into synapses is required for their stabilization and the functional integration of dendritic spines into circuits after ketamine treatment, and in mouse models of depression, these mature spines persist for a week or more[231, 235]. On the one hand, our data suggest that mTOR dependent translation of synaptic proteins may be intact in Mecp2 Null mice given that reduced activity in the mTOR pathway as measured by phospho-rpS6 levels is restored by acute ketamine treatment (Chapter 2). On the other hand, improvements to dendritic spine maturity in the mPFC of Mecp2 mutants after chronic intermittent ketamine treatment persist to 24 hours after the last treatment, but deficits in spine maturity return by 72 hours after treatment unlike in depression models[231, 235] (Chapter

126

2). This may suggest that while PNs in the Mecp2 mutant mPFC are capable of undergoing structural synaptic changes after ketamine treatment, these processes involved in the long-term stabilization of these spines require MeCP2.

Future studies will be required to determine the degree to which dendritic spine growth and maturation induced by ketamine is associated with an increase in functional synapse formation in the mPFC of Mecp2 mutant mice. Current studies in our lab are asking whether or not ketamine treatment is capable of upregulating BDNF expression as well as the expression of synaptic structural proteins, such as PSD95, and functional synaptic components, such as AMPA receptor subunits, in the mPFC of Mecp2 mutants and, specifically, in dendritic spines on PNs. Studies such as this will be required to better understand if the increased density of mature dendritic spines described in Chapter 2 is associated with an increase in both structural and functional synaptic proteins as seen in other models. In combination with the findings in Chapter 2, these future studies will continue to elucidate mechanisms of PN hypofunction in the mPFC of Rett syndrome mouse models and help determine whether or not synaptic proteins are upregulated and localized into synaptic compartments as is observed in Wt animals treated with ketamine. Other studies will be required to determine whether or not upregulated synaptogenesis after ketamine treatment is associated with increased excitatory drive onto PNs of the mPFC as observed in mouse models of depression[330].

127

Activation of the mPFC in mouse models of Rett syndrome

The findings in Chapter 2 further elucidated deficits in excitatory dendritic morphology that likely contribute to reduced excitatory drive onto PNs in the

Mecp2 Null mPFC[93]. Moreover, this study demonstrated the reversibility of the deficit in mature spine morphology even in the complete absence of MeCP2. To further determine the role that reduced excitability of mPFC PNs may play in Rett syndrome associated phenotypes, the experiments in Chapter 3 were designed to determine whether or not RTT phenotypes were reversible upon activation of mPFC PNs (or words to that effect). Therefore, I used the DREADD-Gq receptor under the control of CaMKII delivered via AAV vector which allowed me to specifically manipulate the activity of excitatory neurons in the mPFC[331]. The population of cells studied in Chapter 3 has significant overlap with the cell population studied in Chapter 2; however, the CaMKII promotor is not restricted to any specific cortical layer and would lead to expression in nearly all PNs in the infected region unlike the sparse number of layer V/VI PNs[332] studied in Chapter

In Chapter 3, I analyzed both a cognitive and a respiratory behavior for two primary reasons: first, these are both behavioral domains that are significantly impacted by the loss of MeCP2 in both humans and mice; and second, the mPFC is known to project strongly to subcortical regions responsible for cue- dependent fear conditioning[169, 177, 333], a cognitive behavioral readout that is abnormal in Mecp2 mutant mice[84, 85], and to numerous brainstem respiratory targets[165-169] that are known to have abnormal Fos expression in Mecp2

128 mutants[93]. Initially, it was critical to verify that the DREADD-Gq receptor used in this study was targeting appropriate regions of the mPFC that were known to be important for the behavioral outcomes being studied. To determine if this was the case, we took advantage of the mCherry tag that was expressed by the

DREADD-Gq construct to trace axonal projections from the mPFC throughout the brain. DREADD-mCherry+ projections traced in this manner heavily innervated the BLA, a critical region for the consolidation of cue-dependent fear conditioning

(discussed in [334]). Projections targeting the BLA arise from mPFC PNs that reside within both layers II and V[169]. In addition, dense projection fiber bundles were observed entering into the medulla from the pyramidal tract and terminating in the VLM and more heavily in the ventrolateral nTS. The cell bodies of these neurons are known to reside almost exclusively in layer V of the mPFC[169].

(CeA), the output region of the amygdala that projects to motor and autonomic

129 regions which allow for the expression of fear[304], but rather, target the BLA, whose principle neurons subsequently project to the CeA (discussed in [305]).

(IL) is thought to drive fear inhibition and is critical for fear memory extinction[208,

CS[315-319]. Specifically, the IL terminates more heavily near ITCs than does the

PL[304, 320], and stimulation of the IL leads to activation of the ITC population[321].

On the other hand, the principle target of PL projections to the BLA appear to be excitatory cells themselves, or feed-forward inhibition circuitry, which together increase output of the CeA (discussed in [305]). This may suggest that activation of both the PL and IL results in a balance between their opposing effects on fear memory retrieval during STM testing. However, the data presented in Chapter 3 demonstrate a rescue of fear memory expression 24 hours after conditioning, long after CNO has left the system and PN activity would have returned to baseline[293]. Interestingly, there is also a role for dMT projections to the CeA in

CDFC as early as 3 hours after conditioning when these projections increase their activity and release BDNF into the lateral nucleus of the central amygdala

(CeL). This BDNF release was shown to be necessary for enhancement of

130 excitatory input into this region after learning in the CDFC paradigm[335, 336]. In the

Mecp2 mutant, activation of the PL may lead to increasing activity in this dMT to

CeA pathway thus increasing BDNF signaling and strengthening connections that will be used for memory retrieval hours later. Future studies could test whether or not BDNF release is reduced in the CeA after CDFC in Mecp2 mutants, and if this can be increased by stimulation of either the mPFC and/or the dMT to CeA projection neurons.

Projections from the mPFC that target the dMT and subsequently influence BLA activity, are known to play a role specifically in LTM retrieval after CDFC[197].

This, along with the results from Chapter 3, raises the possibility that mPFC projections to the dMT may be specifically impacted by the loss of MeCP2 while other mPFC projections, such as those to the BLA, may be spared given that

Mecp2 mutants display a deficit in LTM but not STM. To test this hypothesis, preliminary experiments described in Chapter 4 sought to determine whether or not activation of the specific projections from the mPFC to the dMT is sufficient to restore expression of long-term fear conditioned memory in the CDFC paradigm.

Those experiments revealed that while stimulation of this pathway is sufficient to increase LTM expression in Wt animals, Mecp2 Hets remain unaffected by the same stimulation. This was in spite of the fact that Hets expressed significantly more DREADD-Gq-mCherry labeled cells in the mPFC compared to Wt animals.

This may suggest that the global expression of DREADD-Gq in mPFC PNs is activating another target which is required for consolidation of the memory after

CDFC. For example, Hets may lack proper activation of the BLA during

131 conditioning and consolidation, and this may be required for expression of that memory at the later time; or, it could simply make that memory more stable.

Another possible explanation for why selective activation of the mPFC to dMT projection cells is not sufficient to increase expression of LTM in Hets may be that, unlike Wt animals, the temporal shift in the importance of mPFC projections from those going to the BLA to those targeting the dMT does not happen in Hets.

Likewise, abnormal wiring in the pathways underling CDFC could also potentially explain differences in the responsiveness of Wt versus Hets to specific stimulation of the mPFC to dMT projections. Evidence for such a wiring deficit comes from our finding in Chapter 4 demonstrating that Hets display significantly more DREADD-mCherry positive cells after the combinatorial DREADD expression paradigm than do the Wt animals. While the evidence presented in

Chapter 4 is very preliminary, questions such as these have been raised by the initial findings and provide a rationale for future studies seeking to understand mechanisms of dysfunction in Mecp2 mutants after CDFC.

As is the case for mPFC projections responsible for CDFC, the mPFC projections that exert influence over respiratory circuits are known to take both direct and indirect pathways to the brainstem. Direct projection targets of the mPFC that may influence respiration included the nTS, ventrolateral medulla (VLM), and locus coeruleus (LC)[167, 169]. Indirect pathways could include projections to subcortical targets such as the hypothalamus, amygdala, and PAG which themselves then modulate ponto-medullary respiratory cell groups[169]. Activation of these mPFC projection cells, which primarily reside in layer 5 of the mPFC[167,

132

169], led to Fos expression in the nTS of Mecp2 mutants being reduced to Wt levels[84]. This suggests that one possible mechanism underlying the elimination of respiratory apneas in Mecp2 mutants is modulation of activity within this critical brainstem respiratory target; however, the role of mPFC modulation of other respiratory targets in the elimination of apneas cannot be ruled out by this study.

Studies into the mechanisms underlying abnormal respiratory control in mouse models of Rett syndrome have focused almost entirely on dysfunction in structures within the ponto-medullary respiratory network (reviewed in [79]).

Interestingly, respiratory phenotypes are present even in reduced preparations, such as the working heart brainstem preparation which does not include forebrain influence[292]. In addition, re-expression of MeCP2 in just the medullary respiratory network of a Mecp2 mutant is sufficient to prevent respiratory apnea[337]. However, both human Rett patients, as well as mouse models of Rett syndrome, display breathing disturbances that are behavioral state dependent, worsening during arousal and periods of anxiety while improving during sleep[81].

This strongly suggests that respiratory dysfunction in organisms lacking MeCP2 is modulated by cortical influence. The data described in Chapter 3 support this hypothesis by demonstrating not only that activation of the mPFC completely abolishes respiratory apneas, but it also reduces hyperexcitability in brainstem respiratory targets such as the nTS consistent with previous findings that activity in the mPFC can modulate the gain of cardo-respiratory reflex output in the nTS[167, 294, 295]. Taken together, reduced excitability in excitatory neurons of the mPFC may underlie or contribute to respiratory apneas in Mecp2 mutants; and,

133 at the very least, stimulation of these mPFC projection neurons is capable of overcoming abnormal function in ponto-medullary regions to restore normal respiration in a mouse model of Rett syndrome. However, it remains to be determined whether or not mPFC hypofunction contributes to the development of hyperactive ponto-medullary respiratory circuitry in Mecp2 mutants, or if brainstem hyperfunction and mPFC hypofunction develop independently.

Using the data described in this thesis, it is possible to hypothesize that dendritic spine deficits in the mPFC of Mecp2 mutants may be specific to a subset of projection neurons rather than a global feature of mPFC PNs. For example, the findings from Chapter 3 demonstrating that Mecp2 Hets express normal STM, but have deficits in LTM expression that are rescued by mPFC PN activation, provides some evidence that not all mPFC projection cells are equally impacted by the loss of MeCP2. This hypothesis is further supported by the recent findings that short- and long-term memory retrieval each depend on different mPFC projection circuity with STM being dependent on mPFC to BLA projection while

LTM is dependent on mPFC to dMT projections[305]. These two projection subsets are further differentiated by the layer in which their cell bodies reside, with the

BLA-projecting PNs being distributed between layers 2 and 5 while the dMT projecting neurons are primarily restricted to layer 6 of the mPFC[169].

Additionally, mPFC projections to both the brainstem and the PAG have cell bodies primarily in layer 5 of the mPFC[169]. Therefore, future studies could use retrograde tracers from these downstream targets of the mPFC to determine whether or not specific subsets of mPFC PNs, either based on layer or projection

134 target, are more or less strongly impacted by the loss of MeCP2 in terms of their dendritic density and morphology. The data contained in this thesis support a hypothesis that layer 5/6 PNs may be have greater dendritic abnormalities than would PNs in layer 2 given that STM is spared while behaviors in which layer 5/6 neurons participate, such as LTM and respiration, are abnormal in Mecp2 mutants.

Translational considerations

The findings contained in this thesis support the hypothesis that the circuitry of the Mecp2 mutant brain remains sufficiently intact that, if properly activated, normal behavior can be restored. Further, synaptic structural connectivity remains plastic even in the complete absence of MeCP2. A major goal of this thesis project was to determine whether or not mPFC hypofunction was a mechanism that underlies or contributes to behavioral abnormalities in Rett syndrome. These data support this hypothesis, and suggest that the mPFC may be a hub of dysfunction that could be targeted therapeutically to reverse behavioral deficits in diverse behavioral domains.

The data described in Chapter 2, along with data from other labs using chronic low-dose ketamine treatment in Mecp2 mutants[250], demonstrates that ketamine can be well tolerated even after weeks of dosing. Further, the efficacy of repeated ketamine treatment is not diminished with repeat exposure suggesting that clinical trials using similar ketamine treatment paradigms in patients with Rett

135 syndrome may prove successful. Further hope may be gathered from our data demonstrating that deficits in dendritic spines on PNs can be reversed even at late stages of the disorder suggesting that it is not too late for effective treatment to be found for older patients. However, our data also suggests that the synaptic plasticity-inducing effects of ketamine may be short lived after treatment ends, given that the restoration of dendritic spine density and maturity is lost somewhere between 24 and 72 hours after the last dose. Future translational studies or clinical trials should take this into account when determining dosing paradigms and time points for studying outcome measures after treatment.

The studies described in Chapter 3 demonstrate that increasing activity in excitatory PNs of the mPFC can reverse deficits in both cognition and respiration.

This was determined using AAV viral expression of the DREADD-Gq receptor; however, future studies could determine whether or not less invasive approaches to stimulating the mPFC could be effective in mouse models of Rett syndrome and eventually human patients. In addition to pharmacologic approaches, transcranial magnetic stimulation (TMS), or non-invasive electrical stimulation, is also a potential method of increasing mPFC activity[338]. Repetitive TMS (rTMS) is known to exert lasting effects on activity and brain function after stimulation is ended [339] and can have either inhibitory effects at low frequency[340] or excitatory effects at high frequency (>5 Hz)[341]. Interestingly, rTMS has been shown to have anti-depressant like effects when used on the prefrontal cortex[342] and is capable of increasing BDNF expression[343], which may suggest a potential for efficacy in easing the severity of mPFC dependent behavioral deficits in Rett

136 syndrome. Although this has not been tested in patients with Rett syndrome,

TMS has been used to study motor cortex function in patients with Rett syndrome without consequence[344].

Gaining a better understanding of deficits in mPFC PN function and connectivity in mouse models of Rett syndrome may additionally provide insight into frontal cortical dysfunction and related behavioral outcomes in other ASDs. Much is known about behavioral abnormalities in non-syndromic ASDs but relatively little is known about the circuitry underlying these behavioral abnormalities, due in part to a lack of good mouse models. On the other hand, autism related disorders and syndromic ASDs, such as…. do have mouse models, and insofar as there is overlap between disorders on the spectrum, deficits in one disorder may result from mechanisms better studied in another. One major area of overlap among disorders on the spectrum that has been gaining attention is dendritic spine dysgenesis on PNs of the frontal cortex and synaptopathology more generally

(reviewed in [109]). As we observed in the Mecp2 Null, mouse models of

Angelman syndrome, Fragile X syndrome, and the Shank3 loss of function model of non-syndromic autism all have reduced dendritic spine density or maturity on

PNs in various regions of the frontal cortex including the mPFC[109, 345].

Additionally, the Angelman and Shank3 models exhibit reduced activity in the mTOR pathway, while in the Fragile X model mTOR activity is increased[109]. This suggests that ketamine treatment, which increases spine density and maturity, and increases signaling in the mTOR pathway may be beneficial in both

Angelman and Shank3 models. On the other hand, given that mTOR activity is

137 increased in the Fragile X model, ketamine treatment may not be beneficial, and could actually cause overgrowth of dendritic spines on PNs of the frontal cortex.

Such an overgrowth of dendritic spines associated with hyperactivity in the mTOR signaling pathway has been observed in mouse models of PTEN

Autism[346, 347]; and notably, inhibition of the mTOR pathway has proven beneficial to frontal cortex dysfunction and dendritic spine overgrowth phenotypes in PTEN

Autism[348]. Together, these findings highlight the importance of the mTOR signaling pathway and dendritic spine dysgenesis in ASDs (discussed in [109]).

Future studies could be designed to test whether or not dendritic spine deficits in the mPFC of Angelman syndrome and Shank3 mutants could be reversed using similar approaches to those described in Chapter 2, and determine the degree to which dendritic spine deficits on PNs of the mPFC in each of these models contributes to overlapping phenotypes in ASDs more generally.

Together the studies in this thesis demonstrate that reduced excitability of excitatory PNs in the mPFC, resulting in part from reduced dendritic spine density and maturity on these cells, is a mechanism underlying behavioral abnormalities in numerous phenotypic domains in mouse models of Rett syndrome. The findings described within this thesis will help form a foundation for future studies into the mechanisms of mPFC hypofunction that contribute to Rett syndrome symptomology and potentially lead to therapeutic approaches. Further, the data from Chapter 3, combined with preliminary data described in this chapter, demonstrate that hypofunction in specific mPFC projection circuitry

138 underlies or contributes to specific behavioral abnormalities in mouse models of

Rett syndrome.

139

Bibliography

1. Baio, J., et al., Prevalence of Autism Spectrum Disorder Among Children Aged 8 Years - Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2014. MMWR Surveill Summ, 2018. 67(6): p. 1-23. 2. Rett, A., [On a unusual brain atrophy syndrome in hyperammonemia in childhood]. Wien Med Wochenschr, 1966. 116(37): p. 723-6. 3. Hagberg, B., et al., A progressive syndrome of autism, dementia, ataxia, and loss of purposeful hand use in girls: Rett's syndrome: report of 35 cases. Ann Neurol, 1983. 14(4): p. 471-9. 4. Katz, D.M., et al., Preclinical research in Rett syndrome: setting the foundation for translational success. Dis Model Mech, 2012. 5(6): p. 733-45. 5. Muthuraman, M., et al., Deep Brain Stimulation and L-DOPA Therapy: Concepts of Action and Clinical Applications in Parkinson's Disease. Front Neurol, 2018. 9: p. 711. 6. Wan, M., et al., Rett syndrome and beyond: recurrent spontaneous and familial MECP2 mutations at CpG hotspots. Am J Hum Genet, 1999. 65(6): p. 1520-9. 7. Trappe, R., et al., MECP2 mutations in sporadic cases of Rett syndrome are almost exclusively of paternal origin. Am J Hum Genet, 2001. 68(5): p. 1093-101. 8. Chahrour, M. and H.Y. Zoghbi, The story of Rett syndrome: from clinic to neurobiology. Neuron, 2007. 56(3): p. 422-37. 9. Neul, J.L., et al., Rett syndrome: revised diagnostic criteria and nomenclature. Ann Neurol, 2010. 68(6): p. 944-50. 10. Hagberg, B., Clinical manifestations and stages of Rett syndrome. Ment Retard Dev Disabil Res Rev, 2002. 8(2): p. 61-5. 11. Kaufmann, W.E., et al., Social impairments in Rett syndrome: characteristics and relationship with clinical severity. J Intellect Disabil Res, 2012. 56(3): p. 233-47. 12. Coenraads, M., Face to face with Rett syndrome. Epigenetics, 2007. 2(1): p. 2-4. 13. Weese-Mayer, D.E., et al., Autonomic nervous system dysregulation: breathing and heart rate perturbation during wakefulness in young girls with Rett syndrome. Pediatr Res, 2006. 60(4): p. 443-9. 14. Weese-Mayer, D.E., et al., Autonomic dysregulation in young girls with Rett Syndrome during nighttime in-home recordings. Pediatr Pulmonol, 2008. 43(11): p. 1045-1060. 15. Marcus, C.L., et al., Polysomnographic characteristics of patients with Rett syndrome. J Pediatr, 1994. 125(2): p. 218-24. 16. Julu, P.O. and I. Witt Engerstrom, Assessment of the maturity-related brainstem functions reveals the heterogeneous phenotypes and facilitates clinical management of Rett syndrome. Brain Dev, 2005. 27 Suppl 1: p. S43-S53. 17. Julu, P.O., et al., Characterisation of breathing and associated central autonomic dysfunction in the Rett disorder. Arch Dis Child, 2001. 85(1): p. 29-37. 18. Elian, M. and N.D. Rudolf, EEG and respiration in Rett syndrome. Acta Neurol Scand, 1991. 83(2): p. 123-8. 19. Sekul, E.A., et al., Electrocardiographic findings in Rett syndrome: an explanation for sudden death? J Pediatr, 1994. 125(1): p. 80-2. 20. McCauley, M.D., et al., Pathogenesis of lethal cardiac arrhythmias in Mecp2 mutant mice: implication for therapy in Rett syndrome. Sci Transl Med, 2011. 3(113): p. 113ra125.

140

21. Guideri, F., et al., Cardiac dysautonomia and serotonin plasma levels in Rett syndrome. Neuropediatrics, 2004. 35(1): p. 36-8. 22. Tarquinio, D.C., et al., Growth failure and outcome in Rett syndrome: specific growth references. Neurology, 2012. 79(16): p. 1653-61. 23. Percy, A.K., et al., Profiling scoliosis in Rett syndrome. Pediatr Res, 2010. 67(4): p. 435-9. 24. Schultz, R.J., et al., The pattern of growth failure in Rett syndrome. Am J Dis Child, 1993. 147(6): p. 633-7. 25. Renieri, A., et al., Diagnostic criteria for the Zappella variant of Rett syndrome (the preserved speech variant). Brain Dev, 2009. 31(3): p. 208-16. 26. Baptista, P.M., et al., Cognitive performance in Rett syndrome girls: a pilot study using eyetracking technology. J Intellect Disabil Res, 2006. 50(Pt 9): p. 662-6. 27. Berger-Sweeney, J., Cognitive deficits in Rett syndrome: what we know and what we need to know to treat them. Neurobiol Learn Mem, 2011. 96(4): p. 637-46. 28. Djukic, A. and M.V. McDermott, Social preferences in Rett syndrome. Pediatr Neurol, 2012. 46(4): p. 240-2. 29. Burford, B. and C. Trevarthen, Evoking communication in Rett syndrome: comparisons with conversations and games in mother-infant interaction. Eur Child Adolesc Psychiatry, 1997. 6 Suppl 1: p. 26-30. 30. Ahonniska-Assa, J., et al., Assessing cognitive functioning in females with Rett syndrome by eye-tracking methodology. Eur J Paediatr Neurol, 2018. 22(1): p. 39-45. 31. Barnes, K.V., et al., Anxiety-like behavior in Rett syndrome: characteristics and assessment by anxiety scales. J Neurodev Disord, 2015. 7(1): p. 30. 32. Reiss, A.L., et al., Neuroanatomy of Rett syndrome: a volumetric imaging study. Ann Neurol, 1993. 34(2): p. 227-34. 33. Jellinger, K., et al., Neuropathology of Rett syndrome. Acta Neuropathol, 1988. 76(2): p. 142-58. 34. Armstrong, D., et al., Selective dendritic alterations in the cortex of Rett syndrome. J Neuropathol Exp Neurol, 1995. 54(2): p. 195-201. 35. Deng, V., et al., FXYD1 is an MeCP2 target gene overexpressed in the brains of Rett syndrome patients and Mecp2-null mice. Hum Mol Genet, 2007. 16(6): p. 640-50. 36. Volkmar, F.R. and B. Reichow, Autism in DSM-5: progress and challenges. Mol Autism, 2013. 4(1): p. 13. 37. Rutter, M., Debate and argument: there are connections between brain and mind and it is important that Rett syndrome be classified somewhere. J Child Psychol Psychiatry, 1994. 35(2): p. 379-81. 38. Amir, R.E., et al., Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-binding protein 2. Nat Genet, 1999. 23(2): p. 185-8. 39. Meehan, R.R., J.D. Lewis, and A.P. Bird, Characterization of MeCP2, a vertebrate DNA binding protein with affinity for methylated DNA. Nucleic Acids Res, 1992. 20(19): p. 5085-92. 40. Lewis, J.D., et al., Purification, sequence, and cellular localization of a novel chromosomal protein that binds to methylated DNA. Cell, 1992. 69(6): p. 905-14. 41. Hendrich, B. and A. Bird, Identification and characterization of a family of mammalian methyl-CpG binding proteins. Mol Cell Biol, 1998. 18(11): p. 6538-47. 42. Nan, X., F.J. Campoy, and A. Bird, MeCP2 is a transcriptional repressor with abundant binding sites in genomic chromatin. Cell, 1997. 88(4): p. 471-81.

141

43. Kruusvee, V., et al., Structure of the MeCP2-TBLR1 complex reveals a molecular basis for Rett syndrome and related disorders. Proc Natl Acad Sci U S A, 2017. 114(16): p. E3243- E3250. 44. Nan, X., R.R. Meehan, and A. Bird, Dissection of the methyl-CpG binding domain from the chromosomal protein MeCP2. Nucleic Acids Res, 1993. 21(21): p. 4886-92. 45. Chahrour, M., et al., MeCP2, a key contributor to neurological disease, activates and represses transcription. Science, 2008. 320(5880): p. 1224-9. 46. Li, H. and Q. Chang, Regulation and function of stimulus-induced phosphorylation of MeCP2. Front Biol (Beijing), 2014. 9(5): p. 367-375. 47. Bellini, E., et al., MeCP2 post-translational modifications: a mechanism to control its involvement in synaptic plasticity and homeostasis? Front Cell Neurosci, 2014. 8: p. 236. 48. Srivastava, S., et al., Monogenic disorders that mimic the phenotype of Rett syndrome. Neurogenetics, 2018. 19(1): p. 41-47. 49. Ramocki, M.B., Y.J. Tavyev, and S.U. Peters, The MECP2 duplication syndrome. Am J Med Genet A, 2010. 152A(5): p. 1079-88. 50. Miguet, M., et al., Further delineation of the MECP2 duplication syndrome phenotype in 59 French male patients, with a particular focus on morphological and neurological features. J Med Genet, 2018. 55(6): p. 359-371. 51. Kanai, S., et al., Successful corpus callosotomy for post-encephalopathic refractory epilepsy in a patient with MECP2 duplication syndrome. Brain Dev, 2018. 52. Friez, M.J., et al., Recurrent infections, hypotonia, and mental retardation caused by duplication of MECP2 and adjacent region in Xq28. Pediatrics, 2006. 118(6): p. e1687-95. 53. Meins, M., et al., Submicroscopic duplication in Xq28 causes increased expression of the MECP2 gene in a boy with severe mental retardation and features of Rett syndrome. J Med Genet, 2005. 42(2): p. e12. 54. del Gaudio, D., et al., Increased MECP2 gene copy number as the result of genomic duplication in neurodevelopmentally delayed males. Genet Med, 2006. 8(12): p. 784-92. 55. Pelka, G.J., et al., Mecp2 deficiency is associated with learning and cognitive deficits and altered gene activity in the hippocampal region of mice. Brain, 2006. 129(Pt 4): p. 887- 98. 56. Guy, J., et al., A mouse Mecp2-null mutation causes neurological symptoms that mimic Rett syndrome. Nat Genet, 2001. 27(3): p. 322-6. 57. Chen, R.Z., et al., Deficiency of methyl-CpG binding protein-2 in CNS neurons results in a Rett-like phenotype in mice. Nat Genet, 2001. 27(3): p. 327-31. 58. Shahbazian, M., et al., Mice with truncated MeCP2 recapitulate many Rett syndrome features and display hyperacetylation of histone H3. Neuron, 2002. 35(2): p. 243-54. 59. Lawson-Yuen, A., et al., Ube3a mRNA and protein expression are not decreased in Mecp2R168X mutant mice. Brain Res, 2007. 1180: p. 1-6. 60. Jentarra, G.M., et al., Abnormalities of cell packing density and dendritic complexity in the MeCP2 A140V mouse model of Rett syndrome/X-linked mental retardation. BMC Neurosci, 2010. 11: p. 19. 61. Goffin, D. and Z.J. Zhou, The neural circuit basis of Rett syndrome. Front Biol (Beijing), 2012. 7(5): p. 428-435. 62. Brendel, C., et al., Readthrough of nonsense mutations in Rett syndrome: evaluation of novel aminoglycosides and generation of a new mouse model. J Mol Med (Berl), 2011. 89(4): p. 389-98.

142

63. Samaco, R.C., et al., A partial loss of function allele of methyl-CpG-binding protein 2 predicts a human neurodevelopmental syndrome. Hum Mol Genet, 2008. 17(12): p. 1718-27. 64. Kerr, B., et al., Defective body-weight regulation, motor control and abnormal social interactions in Mecp2 hypomorphic mice. Hum Mol Genet, 2008. 17(12): p. 1707-17. 65. Ash, R.T., et al., Increased Axonal Bouton Stability during Learning in the Mouse Model of MECP2 Duplication Syndrome. eNeuro, 2018. 5(3). 66. Calfa, G., J.J. Hablitz, and L. Pozzo-Miller, Network hyperexcitability in hippocampal slices from Mecp2 mutant mice revealed by voltage-sensitive dye imaging. J Neurophysiol, 2011. 105(4): p. 1768-84. 67. Santos, M., et al., Evidence for abnormal early development in a mouse model of Rett syndrome. Genes Brain Behav, 2007. 6(3): p. 277-86. 68. Picker, J.D., et al., An altered neonatal behavioral phenotype in Mecp2 mutant mice. Neuroreport, 2006. 17(5): p. 541-4. 69. Guy, J., et al., Reversal of neurological defects in a mouse model of Rett syndrome. Science, 2007. 315(5815): p. 1143-7. 70. Stearns, N.A., et al., Behavioral and anatomical abnormalities in Mecp2 mutant mice: a model for Rett syndrome. Neuroscience, 2007. 146(3): p. 907-21. 71. Jugloff, D.G., et al., Targeted delivery of an Mecp2 transgene to forebrain neurons improves the behavior of female Mecp2-deficient mice. Hum Mol Genet, 2008. 17(10): p. 1386-96. 72. Alvarez-Saavedra, M., et al., Cell-specific expression of wild-type MeCP2 in mouse models of Rett syndrome yields insight about pathogenesis. Hum Mol Genet, 2007. 16(19): p. 2315-25. 73. Bissonnette, J.M. and S.J. Knopp, Separate respiratory phenotypes in methyl-CpG- binding protein 2 (Mecp2) deficient mice. Pediatr Res, 2006. 59(4 Pt 1): p. 513-8. 74. Johnson, R.A., et al., 7,8-dihydroxyflavone exhibits therapeutic efficacy in a mouse model of Rett syndrome. J Appl Physiol (1985), 2012. 112(5): p. 704-10. 75. Roux, J.C., E. Dura, and L. Villard, Tyrosine hydroxylase deficit in the chemoafferent and the sympathoadrenergic pathways of the Mecp2 deficient mouse. Neurosci Lett, 2008. 447(1): p. 82-6. 76. Stettner, G.M., et al., Breathing dysfunctions associated with impaired control of postinspiratory activity in Mecp2-/y knockout mice. J Physiol, 2007. 579(Pt 3): p. 863-76. 77. Voituron, N., et al., Early breathing defects after moderate hypoxia or hypercapnia in a mouse model of Rett syndrome. Respir Physiol Neurobiol, 2009. 168(1-2): p. 109-18. 78. Kerr, A.M. and P.O. Julu, Recent insights into hyperventilation from the study of Rett syndrome. Arch Dis Child, 1999. 80(4): p. 384-7. 79. Ramirez, J.M., C.S. Ward, and J.L. Neul, Breathing challenges in Rett syndrome: lessons learned from humans and animal models. Respir Physiol Neurobiol, 2013. 189(2): p. 280- 7. 80. Kron, M., et al., A BDNF loop-domain mimetic acutely reverses spontaneous apneas and respiratory abnormalities during behavioral arousal in a mouse model of Rett syndrome. Dis Model Mech, 2014. 7(9): p. 1047-55. 81. Ren, J., et al., Anxiety-related mechanisms of respiratory dysfunction in a mouse model of Rett syndrome. J Neurosci, 2012. 32(48): p. 17230-40. 82. Nalivaiko, E., et al., Respiratory component of the orienting reflex: a novel sensitive index of sensory-induced arousal in rats. Front Physiol, 2011. 2: p. 114.

143

83. Moretti, P., et al., Learning and memory and synaptic plasticity are impaired in a mouse model of Rett syndrome. J Neurosci, 2006. 26(1): p. 319-27. 84. Howell, C.J., et al., Activation of the Medial Prefrontal Cortex Reverses Cognitive and Respiratory Symptoms in a Mouse Model of Rett Syndrome. eNeuro, 2017. 4(6). 85. Tai, D.J., et al., MeCP2 SUMOylation rescues Mecp2-mutant-induced behavioural deficits in a mouse model of Rett syndrome. Nat Commun, 2016. 7: p. 10552. 86. Belichenko, P.V., et al., Widespread changes in dendritic and axonal morphology in Mecp2-mutant mouse models of Rett syndrome: evidence for disruption of neuronal networks. J Comp Neurol, 2009. 514(3): p. 240-58. 87. Fukuda, T., et al., Delayed maturation of neuronal architecture and synaptogenesis in cerebral cortex of Mecp2-deficient mice. J Neuropathol Exp Neurol, 2005. 64(6): p. 537- 44. 88. Sceniak, M.P., et al., Mechanisms of Functional Hypoconnectivity in the Medial Prefrontal Cortex of Mecp2 Null Mice. Cereb Cortex, 2016. 26(5): p. 1938-1956. 89. Dani, V.S., et al., Reduced cortical activity due to a shift in the balance between excitation and inhibition in a mouse model of Rett syndrome. Proc Natl Acad Sci U S A, 2005. 102(35): p. 12560-5. 90. Dani, V.S. and S.B. Nelson, Intact long-term potentiation but reduced connectivity between neocortical layer 5 pyramidal neurons in a mouse model of Rett syndrome. J Neurosci, 2009. 29(36): p. 11263-70. 91. Zhang, L., et al., The MeCP2-null mouse hippocampus displays altered basal inhibitory rhythms and is prone to hyperexcitability. Hippocampus, 2008. 18(3): p. 294-309. 92. Medrihan, L., et al., Early defects of GABAergic synapses in the brain stem of a MeCP2 mouse model of Rett syndrome. J Neurophysiol, 2008. 99(1): p. 112-21. 93. Kron, M., et al., Brain activity mapping in Mecp2 mutant mice reveals functional deficits in forebrain circuits, including key nodes in the default mode network, that are reversed with ketamine treatment. J Neurosci, 2012. 32(40): p. 13860-72. 94. Kline, D.D., et al., Exogenous brain-derived neurotrophic factor rescues synaptic dysfunction in Mecp2-null mice. J Neurosci, 2010. 30(15): p. 5303-10. 95. Katz, D.M., F.S. Menniti, and R.J. Mather, N-Methyl-D-Aspartate Receptors, Ketamine, and Rett Syndrome: Something Special on the Road to Treatments? Biol Psychiatry, 2016. 79(9): p. 710-2. 96. Shepherd, G.M. and D.M. Katz, Synaptic microcircuit dysfunction in genetic models of neurodevelopmental disorders: focus on Mecp2 and Met. Curr Opin Neurobiol, 2011. 21(6): p. 827-33. 97. Lonetti, G., et al., Early environmental enrichment moderates the behavioral and synaptic phenotype of MeCP2 null mice. Biol Psychiatry, 2010. 67(7): p. 657-65. 98. Taneja, P., et al., Pathophysiology of locus ceruleus neurons in a mouse model of Rett syndrome. J Neurosci, 2009. 29(39): p. 12187-95. 99. Gantz, S.C., et al., Loss of Mecp2 in substantia nigra dopamine neurons compromises the nigrostriatal pathway. J Neurosci, 2011. 31(35): p. 12629-37. 100. Nelson, E.D., E.T. Kavalali, and L.M. Monteggia, MeCP2-dependent transcriptional repression regulates excitatory neurotransmission. Curr Biol, 2006. 16(7): p. 710-6. 101. Wang, H., et al., Dysregulation of brain-derived neurotrophic factor expression and neurosecretory function in Mecp2 null mice. J Neurosci, 2006. 26(42): p. 10911-5. 102. Boggio, E.M., et al., Synaptic determinants of rett syndrome. Front Synaptic Neurosci, 2010. 2: p. 28.

144

103. Asaka, Y., et al., Hippocampal synaptic plasticity is impaired in the Mecp2-null mouse model of Rett syndrome. Neurobiol Dis, 2006. 21(1): p. 217-27. 104. Katz, D.M., Brain-derived neurotrophic factor and Rett syndrome. Handb Exp Pharmacol, 2014. 220: p. 481-95. 105. Li, W., et al., Activity-dependent BDNF release and TRPC signaling is impaired in hippocampal neurons of Mecp2 mutant mice. Proc Natl Acad Sci U S A, 2012. 109(42): p. 17087-92. 106. Li, W. and L. Pozzo-Miller, Beyond Widespread Mecp2 Deletions to Model Rett Syndrome: Conditional Spatio-Temporal Knockout, Single-Point Mutations and Transgenic Rescue Mice. Autism Open Access, 2012. 2012(Suppl 1): p. 5. 107. Ricciardi, S., et al., Reduced AKT/mTOR signaling and protein synthesis dysregulation in a Rett syndrome animal model. Hum Mol Genet, 2011. 20(6): p. 1182-96. 108. Park, H. and M.M. Poo, Neurotrophin regulation of neural circuit development and function. Nat Rev Neurosci, 2013. 14(1): p. 7-23. 109. Phillips, M. and L. Pozzo-Miller, Dendritic spine dysgenesis in autism related disorders. Neurosci Lett, 2015. 601: p. 30-40. 110. Chang, Q., et al., The disease progression of Mecp2 mutant mice is affected by the level of BDNF expression. Neuron, 2006. 49(3): p. 341-8. 111. Balkowiec, A. and D.M. Katz, Cellular mechanisms regulating activity-dependent release of native brain-derived neurotrophic factor from hippocampal neurons. J Neurosci, 2002. 22(23): p. 10399-407. 112. Comery, T.A., et al., Abnormal dendritic spines in fragile X knockout mice: maturation and pruning deficits. Proc Natl Acad Sci U S A, 1997. 94(10): p. 5401-4. 113. Dindot, S.V., et al., The Angelman syndrome ubiquitin ligase localizes to the synapse and nucleus, and maternal deficiency results in abnormal dendritic spine morphology. Hum Mol Genet, 2008. 17(1): p. 111-8. 114. Duman, R.S. and G.K. Aghajanian, Synaptic dysfunction in depression: potential therapeutic targets. Science, 2012. 338(6103): p. 68-72. 115. Duman, R.S., et al., Synaptic plasticity and depression: new insights from stress and rapid-acting antidepressants. Nat Med, 2016. 22(3): p. 238-49. 116. Irwin, S.A., et al., Abnormal dendritic spine characteristics in the temporal and visual cortices of patients with fragile-X syndrome: a quantitative examination. Am J Med Genet, 2001. 98(2): p. 161-7. 117. Yashiro, K., et al., Ube3a is required for experience-dependent maturation of the neocortex. Nat Neurosci, 2009. 12(6): p. 777-83. 118. Giacometti, E., et al., Partial rescue of MeCP2 deficiency by postnatal activation of MeCP2. Proc Natl Acad Sci U S A, 2007. 104(6): p. 1931-6. 119. Luikenhuis, S., et al., Expression of MeCP2 in postmitotic neurons rescues Rett syndrome in mice. Proc Natl Acad Sci U S A, 2004. 101(16): p. 6033-8. 120. McGraw, C.M., R.C. Samaco, and H.Y. Zoghbi, Adult neural function requires MeCP2. Science, 2011. 333(6039): p. 186. 121. Nguyen, M.V., et al., MeCP2 is critical for maintaining mature neuronal networks and global brain anatomy during late stages of postnatal brain development and in the mature adult brain. J Neurosci, 2012. 32(29): p. 10021-34. 122. Robinson, L., et al., Morphological and functional reversal of phenotypes in a mouse model of Rett syndrome. Brain, 2012. 135(Pt 9): p. 2699-710.

145

123. Zhou, Z., et al., Brain-specific phosphorylation of MeCP2 regulates activity-dependent Bdnf transcription, dendritic growth, and spine maturation. Neuron, 2006. 52(2): p. 255- 69. 124. Larimore, J.L., et al., Bdnf overexpression in hippocampal neurons prevents dendritic atrophy caused by Rett-associated MECP2 mutations. Neurobiol Dis, 2009. 34(2): p. 199- 211. 125. Li, W., et al., A small-molecule TrkB ligand restores hippocampal synaptic plasticity and object location memory in Rett syndrome mice. Dis Model Mech, 2017. 10(7): p. 837- 845. 126. Schmid, D.A., et al., A TrkB small molecule partial agonist rescues TrkB phosphorylation deficits and improves respiratory function in a mouse model of Rett syndrome. J Neurosci, 2012. 32(5): p. 1803-10. 127. Deogracias, R., et al., Fingolimod, a sphingosine-1 phosphate receptor modulator, increases BDNF levels and improves symptoms of a mouse model of Rett syndrome. Proc Natl Acad Sci U S A, 2012. 109(35): p. 14230-5. 128. Dragunow, M. and R. Faull, The use of c-fos as a metabolic marker in neuronal pathway tracing. J Neurosci Methods, 1989. 29(3): p. 261-5. 129. Subramanian, H.H. and G. Holstege, Periaqueductal gray control of breathing. Adv Exp Med Biol, 2010. 669: p. 353-8. 130. Subramanian, H.H., R.J. Balnave, and G. Holstege, The midbrain periaqueductal gray control of respiration. J Neurosci, 2008. 28(47): p. 12274-83. 131. Kubin, L., et al., Central pathways of pulmonary and lower airway vagal afferents. J Appl Physiol (1985), 2006. 101(2): p. 618-27. 132. Andresen, M.C. and D.L. Kunze, Nucleus tractus solitarius--gateway to neural circulatory control. Annu Rev Physiol, 1994. 56: p. 93-116. 133. Minshew, N.J. and T.A. Keller, The nature of brain dysfunction in autism: functional brain imaging studies. Curr Opin Neurol, 2010. 23(2): p. 124-30. 134. Vincent, J.L., et al., Evidence for a frontoparietal control system revealed by intrinsic functional connectivity. J Neurophysiol, 2008. 100(6): p. 3328-42. 135. Raichle, M.E., et al., A default mode of brain function. Proc Natl Acad Sci U S A, 2001. 98(2): p. 676-82. 136. Gilbert, S.J., et al., Distinct regions of medial rostral prefrontal cortex supporting social and nonsocial functions. Soc Cogn Affect Neurosci, 2007. 2(3): p. 217-26. 137. Buckner, R.L., J.R. Andrews-Hanna, and D.L. Schacter, The brain's default network: anatomy, function, and relevance to disease. Ann N Y Acad Sci, 2008. 1124: p. 1-38. 138. Hahn, T.T., B. Sakmann, and M.R. Mehta, Differential responses of hippocampal subfields to cortical up-down states. Proc Natl Acad Sci U S A, 2007. 104(12): p. 5169-74. 139. Weng, S.J., et al., Neural activation to emotional faces in adolescents with autism spectrum disorders. J Child Psychol Psychiatry, 2011. 52(3): p. 296-305. 140. Nelson, E.D., et al., Selective impact of MeCP2 and associated histone deacetylases on the dynamics of evoked excitatory neurotransmission. J Neurophysiol, 2011. 106(1): p. 193-201. 141. Fischer, M., et al., Enhanced hypoxia susceptibility in hippocampal slices from a mouse model of rett syndrome. J Neurophysiol, 2009. 101(2): p. 1016-32. 142. D'Cruz, J.A., et al., Alterations of cortical and hippocampal EEG activity in MeCP2- deficient mice. Neurobiol Dis, 2010. 38(1): p. 8-16. 143. Behrens, M.M., et al., Ketamine-induced loss of phenotype of fast-spiking interneurons is mediated by NADPH-oxidase. Science, 2007. 318(5856): p. 1645-7.

146

144. Yuhas, J., et al., Brief report: Sensorimotor gating in idiopathic autism and autism associated with fragile X syndrome. J Autism Dev Disord, 2011. 41(2): p. 248-53. 145. Perry, W., et al., Sensorimotor gating deficits in adults with autism. Biol Psychiatry, 2007. 61(4): p. 482-6. 146. McAlonan, G.M., et al., Brain anatomy and sensorimotor gating in Asperger's syndrome. Brain, 2002. 125(Pt 7): p. 1594-606. 147. Frankland, P.W., et al., Sensorimotor gating abnormalities in young males with fragile X syndrome and Fmr1-knockout mice. Mol Psychiatry, 2004. 9(4): p. 417-25. 148. Alsene, K.M., et al., Discrete forebrain neuronal networks supporting noradrenergic regulation of sensorimotor gating. Neuropsychopharmacology, 2011. 36(5): p. 1003-14. 149. Swerdlow, N.R., M.A. Geyer, and D.L. Braff, Neural circuit regulation of prepulse inhibition of startle in the rat: current knowledge and future challenges. Psychopharmacology (Berl), 2001. 156(2-3): p. 194-215. 150. Autry, A.E., et al., NMDA receptor blockade at rest triggers rapid behavioural antidepressant responses. Nature, 2011. 475(7354): p. 91-5. 151. Takahashi, M., et al., Sustained brain-derived neurotrophic factor up-regulation and sensorimotor gating abnormality induced by postnatal exposure to phencyclidine: comparison with adult treatment. J Neurochem, 2006. 99(3): p. 770-80. 152. Papaleo, F., et al., Working memory deficits, increased anxiety-like traits, and seizure susceptibility in BDNF overexpressing mice. Learn Mem, 2011. 18(8): p. 534-44. 153. Katz, D.M., et al., Breathing disorders in Rett syndrome: progressive neurochemical dysfunction in the respiratory network after birth. Respir Physiol Neurobiol, 2009. 168(1- 2): p. 101-8. 154. Ward, C.S., et al., MeCP2 is critical within HoxB1-derived tissues of mice for normal lifespan. J Neurosci, 2011. 31(28): p. 10359-70. 155. Chitravanshi, V.C. and H.N. Sapru, Chemoreceptor-sensitive neurons in commissural subnucleus of nucleus tractus solitarius of the rat. Am J Physiol, 1995. 268(4 Pt 2): p. R851-8. 156. Finley, J.C. and D.M. Katz, The central organization of carotid body afferent projections to the brainstem of the rat. Brain Res, 1992. 572(1-2): p. 108-16. 157. Davies, R.O., L. Kubin, and A.I. Pack, Pulmonary stretch receptor relay neurones of the cat: location and contralateral medullary projections. J Physiol, 1987. 383: p. 571-85. 158. Hoover, W.B. and R.P. Vertes, Anatomical analysis of afferent projections to the medial prefrontal cortex in the rat. Brain Struct Funct, 2007. 212(2): p. 149-79. 159. Miller, E.K. and J.D. Cohen, An integrative theory of prefrontal cortex function. Annu Rev Neurosci, 2001. 24: p. 167-202. 160. Riga, D., et al., Optogenetic dissection of medial prefrontal cortex circuitry. Front Syst Neurosci, 2014. 8: p. 230. 161. Subramanian, H.H., R.J. Balnave, and C.M. Chow, Behavioural control of breathing in mammals: role of the midbrain periaqueductal gray. Adv Exp Med Biol, 2004. 551: p. 135-41. 162. Fisk, G.D. and J.M. Wyss, Descending projections of infralimbic cortex that mediate stimulation-evoked changes in arterial pressure. Brain Res, 2000. 859(1): p. 83-95. 163. Huang, Z.G., et al., Roles of periaqueductal gray and nucleus tractus solitarius in cardiorespiratory function in the rat brainstem. Respir Physiol, 2000. 120(3): p. 185-95. 164. Vertes, R.P., Differential projections of the infralimbic and prelimbic cortex in the rat. Synapse, 2004. 51(1): p. 32-58.

147

165. Alexandrov, V.G., T.G. Ivanova, and N.P. Alexandrova, Prefrontal control of respiration. J Physiol Pharmacol, 2007. 58 Suppl 5(Pt 1): p. 17-23. 166. Frysztak, R.J. and E.J. Neafsey, The effect of medial frontal cortex lesions on respiration, "freezing," and ultrasonic vocalizations during conditioned emotional responses in rats. Cereb Cortex, 1991. 1(5): p. 418-25. 167. Hassan, S.F., J.L. Cornish, and A.K. Goodchild, Respiratory, metabolic and cardiac functions are altered by disinhibition of subregions of the medial prefrontal cortex. J Physiol, 2013. 591(23): p. 6069-88. 168. Neafsey, E.J., Prefrontal cortical control of the autonomic nervous system: anatomical and physiological observations. Prog Brain Res, 1990. 85: p. 147-65; discussion 165-6. 169. Gabbott, P.L., et al., Prefrontal cortex in the rat: projections to subcortical autonomic, motor, and limbic centers. J Comp Neurol, 2005. 492(2): p. 145-77. 170. Bechara, A. and M. Van Der Linden, Decision-making and impulse control after frontal lobe injuries. Curr Opin Neurol, 2005. 18(6): p. 734-9. 171. Rushworth, M.F., et al., Frontal cortex and reward-guided learning and decision-making. Neuron, 2011. 70(6): p. 1054-69. 172. Ridderinkhof, K.R., et al., The role of the medial frontal cortex in cognitive control. Science, 2004. 306(5695): p. 443-7. 173. Posner, M.I., et al., The anterior cingulate gyrus and the mechanism of self-regulation. Cogn Affect Behav Neurosci, 2007. 7(4): p. 391-5. 174. Holroyd, C.B., M.G. Coles, and S. Nieuwenhuis, Medial prefrontal cortex and error potentials. Science, 2002. 296(5573): p. 1610-1 author reply 1610-1. 175. Botvinick, M.M., J.D. Cohen, and C.S. Carter, Conflict monitoring and anterior cingulate cortex: an update. Trends Cogn Sci, 2004. 8(12): p. 539-46. 176. Euston, D.R., A.J. Gruber, and B.L. McNaughton, The role of medial prefrontal cortex in memory and decision making. Neuron, 2012. 76(6): p. 1057-70. 177. Gabbott, P., et al., Amygdala afferents monosynaptically innervate corticospinal neurons in rat medial prefrontal cortex. J Comp Neurol, 2012. 520(11): p. 2440-58. 178. Quilodran, R., M. Rothe, and E. Procyk, Behavioral shifts and action valuation in the anterior cingulate cortex. Neuron, 2008. 57(2): p. 314-25. 179. Gilmartin, M.R. and M.D. McEchron, Single neurons in the medial prefrontal cortex of the rat exhibit tonic and phasic coding during trace fear conditioning. Behav Neurosci, 2005. 119(6): p. 1496-510. 180. Pratt, W.E. and S.J. Mizumori, Neurons in rat medial prefrontal cortex show anticipatory rate changes to predictable differential rewards in a spatial memory task. Behav Brain Res, 2001. 123(2): p. 165-83. 181. Baeg, E.H., et al., Fast spiking and regular spiking neural correlates of fear conditioning in the medial prefrontal cortex of the rat. Cereb Cortex, 2001. 11(5): p. 441-51. 182. Gruber, A.J., et al., More is less: a disinhibited prefrontal cortex impairs cognitive flexibility. J Neurosci, 2010. 30(50): p. 17102-10. 183. Burton, B.G., et al., Lesion of the ventral and intermediate hippocampus abolishes anticipatory activity in the medial prefrontal cortex of the rat. Behav Brain Res, 2009. 199(2): p. 222-34. 184. Heidbreder, C.A. and H.J. Groenewegen, The medial prefrontal cortex in the rat: evidence for a dorso-ventral distinction based upon functional and anatomical characteristics. Neurosci Biobehav Rev, 2003. 27(6): p. 555-79. 185. Nelson, E.E., et al., Brain systems underlying response flexibility in healthy and bipolar adolescents: an event-related fMRI study. Bipolar Disord, 2007. 9(8): p. 810-9.

148

186. Sewards, T.V. and M.A. Sewards, Fear and power-dominance drive motivation: neural representations and pathways mediating sensory and mnemonic inputs, and outputs to premotor structures. Neurosci Biobehav Rev, 2002. 26(5): p. 553-79. 187. Uylings, H.B., H.J. Groenewegen, and B. Kolb, Do rats have a prefrontal cortex? Behav Brain Res, 2003. 146(1-2): p. 3-17. 188. Seamans, J.K., C.C. Lapish, and D. Durstewitz, Comparing the prefrontal cortex of rats and primates: insights from electrophysiology. Neurotox Res, 2008. 14(2-3): p. 249-62. 189. Preuss, T.M., Do rats have prefrontal cortex? The rose-woolsey-akert program reconsidered. J Cogn Neurosci, 1995. 7(1): p. 1-24. 190. Chudasama, Y., Animal models of prefrontal-executive function. Behav Neurosci, 2011. 125(3): p. 327-43. 191. Brown, V.J. and E.M. Bowman, Rodent models of prefrontal cortical function. Trends Neurosci, 2002. 25(7): p. 340-3. 192. Wise, S.P., Forward frontal fields: phylogeny and fundamental function. Trends Neurosci, 2008. 31(12): p. 599-608. 193. Ostlund, S.B. and B.W. Balleine, Orbitofrontal cortex mediates outcome encoding in Pavlovian but not instrumental conditioning. J Neurosci, 2007. 27(18): p. 4819-25. 194. Camille, N., et al., Ventromedial frontal lobe damage disrupts value maximization in humans. J Neurosci, 2011. 31(20): p. 7527-32. 195. Douglas, R.J. and K.A. Martin, Neuronal circuits of the neocortex. Annu Rev Neurosci, 2004. 27: p. 419-51. 196. Sesack, S.R., et al., Topographical organization of the efferent projections of the medial prefrontal cortex in the rat: an anterograde tract-tracing study with Phaseolus vulgaris leucoagglutinin. J Comp Neurol, 1989. 290(2): p. 213-42. 197. Do-Monte, F.H., K. Quinones-Laracuente, and G.J. Quirk, A temporal shift in the circuits mediating retrieval of fear memory. Nature, 2015. 519(7544): p. 460-3. 198. Ciocchi, S., et al., Encoding of conditioned fear in central amygdala inhibitory circuits. Nature, 2010. 468(7321): p. 277-82. 199. Rogan, M.T., U.V. Staubli, and J.E. LeDoux, Fear conditioning induces associative long- term potentiation in the amygdala. Nature, 1997. 390(6660): p. 604-7. 200. Johansen, J.P., et al., Molecular mechanisms of fear learning and memory. Cell, 2011. 147(3): p. 509-24. 201. Gale, G.D., et al., Role of the basolateral amygdala in the storage of fear memories across the adult lifetime of rats. J Neurosci, 2004. 24(15): p. 3810-5. 202. Maren, S., Neurobiology of Pavlovian fear conditioning. Annu Rev Neurosci, 2001. 24: p. 897-931. 203. LeDoux, J.E., Emotion circuits in the brain. Annu Rev Neurosci, 2000. 23: p. 155-84. 204. Li, S., J.H. Kim, and R. Richardson, Differential involvement of the medial prefrontal cortex in the expression of learned fear across development. Behav Neurosci, 2012. 126(2): p. 217-25. 205. Vertes, R.P. and W.B. Hoover, Projections of the paraventricular and paratenial nuclei of the dorsal midline thalamus in the rat. J Comp Neurol, 2008. 508(2): p. 212-37. 206. Corcoran, K.A. and G.J. Quirk, Activity in prelimbic cortex is necessary for the expression of learned, but not innate, fears. J Neurosci, 2007. 27(4): p. 840-4. 207. Sierra-Mercado, D., Jr., et al., Inactivation of the ventromedial prefrontal cortex reduces expression of conditioned fear and impairs subsequent recall of extinction. Eur J Neurosci, 2006. 24(6): p. 1751-8.

149

208. Sierra-Mercado, D., N. Padilla-Coreano, and G.J. Quirk, Dissociable roles of prelimbic and infralimbic cortices, ventral hippocampus, and basolateral amygdala in the expression and extinction of conditioned fear. Neuropsychopharmacology, 2011. 36(2): p. 529-38. 209. Adhikari, A., et al., Basomedial amygdala mediates top-down control of anxiety and fear. Nature, 2015. 527(7577): p. 179-85. 210. Do-Monte, F.H., et al., Revisiting the role of infralimbic cortex in fear extinction with optogenetics. J Neurosci, 2015. 35(8): p. 3607-15. 211. Molina, L.A., I. Skelin, and A.J. Gruber, Acute NMDA receptor antagonism disrupts synchronization of action potential firing in rat prefrontal cortex. PLoS One, 2014. 9(1): p. e85842. 212. Jackson, M.E., H. Homayoun, and B. Moghaddam, NMDA receptor hypofunction produces concomitant firing rate potentiation and burst activity reduction in the prefrontal cortex. Proc Natl Acad Sci U S A, 2004. 101(22): p. 8467-72. 213. Anver, H., et al., NMDA receptor hypofunction phase couples independent gamma- oscillations in the rat visual cortex. Neuropsychopharmacology, 2011. 36(2): p. 519-28. 214. Flint, A.C. and B.W. Connors, Two types of network oscillations in neocortex mediated by distinct glutamate receptor subtypes and neuronal populations. J Neurophysiol, 1996. 75(2): p. 951-7. 215. Copping, N.A., et al., Neuronal overexpression of Ube3a isoform 2 causes behavioral impairments and neuroanatomical pathology relevant to 15q11.2-q13.3 duplication syndrome. Hum Mol Genet, 2017. 26(20): p. 3995-4010. 216. Bhakar, A.L., G. Dolen, and M.F. Bear, The pathophysiology of fragile X (and what it teaches us about synapses). Annu Rev Neurosci, 2012. 35: p. 417-43. 217. Krueger, D.D. and M.F. Bear, Toward fulfilling the promise of molecular medicine in fragile X syndrome. Annu Rev Med, 2011. 62: p. 411-29. 218. Sheng, L., et al., Prenatal polycyclic aromatic hydrocarbon exposure leads to behavioral deficits and downregulation of receptor tyrosine kinase, MET. Toxicol Sci, 2010. 118(2): p. 625-34. 219. Kreienkamp, H.J., Scaffolding proteins at the postsynaptic density: shank as the architectural framework. Handb Exp Pharmacol, 2008(186): p. 365-80. 220. Gundelfinger, E.D., et al., A role for zinc in postsynaptic density asSAMbly and plasticity? Trends Biochem Sci, 2006. 31(7): p. 366-73. 221. Grabrucker, A.M., et al., Postsynaptic ProSAP/Shank scaffolds in the cross-hair of synaptopathies. Trends Cell Biol, 2011. 21(10): p. 594-603. 222. Grabrucker, A.M., et al., Concerted action of zinc and ProSAP/Shank in synaptogenesis and synapse maturation. EMBO J, 2011. 30(3): p. 569-81. 223. Jiang, Y.H. and M.D. Ehlers, Modeling autism by SHANK gene mutations in mice. Neuron, 2013. 78(1): p. 8-27. 224. Sato, M. and M.P. Stryker, Genomic imprinting of experience-dependent cortical plasticity by the ubiquitin ligase gene Ube3a. Proc Natl Acad Sci U S A, 2010. 107(12): p. 5611-6. 225. Armstrong, D.D., Neuropathology of Rett syndrome. J Child Neurol, 2005. 20(9): p. 747- 53. 226. Belichenko, N.P., P.V. Belichenko, and W.C. Mobley, Evidence for both neuronal cell autonomous and nonautonomous effects of methyl-CpG-binding protein 2 in the cerebral cortex of female mice with Mecp2 mutation. Neurobiol Dis, 2009. 34(1): p. 71-7. 227. Chao, H.T., H.Y. Zoghbi, and C. Rosenmund, MeCP2 controls excitatory synaptic strength by regulating glutamatergic synapse number. Neuron, 2007. 56(1): p. 58-65.

150

228. Leal, G., D. Comprido, and C.B. Duarte, BDNF-induced local protein synthesis and synaptic plasticity. Neuropharmacology, 2014. 76 Pt C: p. 639-56. 229. Hoeffer, C.A. and E. Klann, mTOR signaling: at the crossroads of plasticity, memory and disease. Trends Neurosci, 2010. 33(2): p. 67-75. 230. Duman, R.S. and B. Voleti, Signaling pathways underlying the pathophysiology and treatment of depression: novel mechanisms for rapid-acting agents. Trends Neurosci, 2012. 35(1): p. 47-56. 231. Li, N., et al., Glutamate N-methyl-D-aspartate receptor antagonists rapidly reverse behavioral and synaptic deficits caused by chronic stress exposure. Biol Psychiatry, 2011. 69(8): p. 754-61. 232. Duman, R.S., et al., Signaling pathways underlying the rapid antidepressant actions of ketamine. Neuropharmacology, 2012. 62(1): p. 35-41. 233. Huber, K.M., et al., Dysregulation of Mammalian Target of Rapamycin Signaling in Mouse Models of Autism. J Neurosci, 2015. 35(41): p. 13836-42. 234. Lee, E., J. Lee, and E. Kim, Excitation/Inhibition Imbalance in Animal Models of Autism Spectrum Disorders. Biol Psychiatry, 2017. 81(10): p. 838-847. 235. Li, N., et al., mTOR-dependent synapse formation underlies the rapid antidepressant effects of NMDA antagonists. Science, 2010. 329(5994): p. 959-64. 236. Shahbazian, M.D. and H.Y. Zoghbi, Rett syndrome and MeCP2: linking epigenetics and neuronal function. Am J Hum Genet, 2002. 71(6): p. 1259-72. 237. Klose, R.J. and A.P. Bird, Genomic DNA methylation: the mark and its mediators. Trends Biochem Sci, 2006. 31(2): p. 89-97. 238. Ogier, M. and D.M. Katz, Breathing dysfunction in Rett syndrome: understanding epigenetic regulation of the respiratory network. Respir Physiol Neurobiol, 2008. 164(1- 2): p. 55-63. 239. Vorsanova, S.G., I.Y. Iourov, and Y.B. Yurov, Neurological, genetic and epigenetic features of Rett syndrome. J Pediatr Neurol, 2004. 2(4): p. 179-190. 240. Belichenko, N.P., et al., Comparative study of brain morphology in Mecp2 mutant mouse models of Rett syndrome. J Comp Neurol, 2008. 508(1): p. 184-95. 241. Belichenko, P.V., et al., Rett syndrome: 3-D confocal microscopy of cortical pyramidal dendrites and afferents. Neuroreport, 1994. 5(12): p. 1509-13. 242. Stuss, D.P., et al., MeCP2 mutation results in compartment-specific reductions in dendritic branching and spine density in layer 5 motor cortical neurons of YFP-H mice. PLoS One, 2012. 7(3): p. e31896. 243. Wang, I.T., A.R. Reyes, and Z. Zhou, Neuronal morphology in MeCP2 mouse models is intrinsically variable and depends on age, cell type, and Mecp2 mutation. Neurobiol Dis, 2013. 58: p. 3-12. 244. Chapleau, C.A., et al., Hippocampal CA1 pyramidal neurons of Mecp2 mutant mice show a dendritic spine phenotype only in the presymptomatic stage. Neural Plast, 2012. 2012: p. 976164. 245. Chapleau, C.A., et al., Dendritic spine pathologies in hippocampal pyramidal neurons from Rett syndrome brain and after expression of Rett-associated MECP2 mutations. Neurobiol Dis, 2009. 35(2): p. 219-33. 246. Landi, S., et al., The short-time structural plasticity of dendritic spines is altered in a model of Rett syndrome. Sci Rep, 2011. 1: p. 45. 247. Xu, X., E.C. Miller, and L. Pozzo-Miller, Dendritic spine dysgenesis in Rett syndrome. Front Neuroanat, 2014. 8: p. 97.

151

248. Duman, R.S. and N. Li, A neurotrophic hypothesis of depression: role of synaptogenesis in the actions of NMDA receptor antagonists. Philos Trans R Soc Lond B Biol Sci, 2012. 367(1601): p. 2475-84. 249. Dwyer, J.M., A.E. Lepack, and R.S. Duman, mTOR activation is required for the antidepressant effects of mGluR(2)/(3) blockade. Int J Neuropsychopharmacol, 2012. 15(4): p. 429-34. 250. Patrizi, A., et al., Chronic Administration of the N-Methyl-D-Aspartate Receptor Antagonist Ketamine Improves Rett Syndrome Phenotype. Biol Psychiatry, 2016. 79(9): p. 755-64. 251. Kimura, F., et al., [Case of Rett syndrome monitored with BIS and neuromuscular monitor during total intravenous anesthesia]. Masui, 2011. 60(6): p. 700-2. 252. Li, H., et al., Loss of activity-induced phosphorylation of MeCP2 enhances synaptogenesis, LTP and spatial memory. Nat Neurosci, 2011. 14(8): p. 1001-8. 253. Faul, F., et al., G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods, 2007. 39(2): p. 175-91. 254. Holz, M.K., et al., mTOR and S6K1 mediate assembly of the translation preinitiation complex through dynamic protein interchange and ordered phosphorylation events. Cell, 2005. 123(4): p. 569-80. 255. Holz, M.K. and J. Blenis, Identification of S6 kinase 1 as a novel mammalian target of rapamycin (mTOR)-phosphorylating kinase. J Biol Chem, 2005. 280(28): p. 26089-93. 256. Iadarola, N.D., et al., Ketamine and other N-methyl-D-aspartate receptor antagonists in the treatment of depression: a perspective review. Ther Adv Chronic Dis, 2015. 6(3): p. 97-114. 257. Bourgeron, T., From the genetic architecture to synaptic plasticity in autism spectrum disorder. Nat Rev Neurosci, 2015. 16(9): p. 551-63. 258. Baj, G., et al., Developmental and maintenance defects in Rett syndrome neurons identified by a new mouse staging system in vitro. Front Cell Neurosci, 2014. 8: p. 18. 259. Castro, J., et al., Functional recovery with recombinant human IGF1 treatment in a mouse model of Rett Syndrome. Proc Natl Acad Sci U S A, 2014. 111(27): p. 9941-6. 260. Kavalali, E.T. and L.M. Monteggia, Synaptic mechanisms underlying rapid antidepressant action of ketamine. Am J Psychiatry, 2012. 169(11): p. 1150-6. 261. Nakao, S., et al., Ketamine-induced c-Fos expression in the mouse posterior cingulate and retrosplenial cortices is mediated not only via NMDA receptors but also via sigma receptors. Brain Res, 2002. 926(1-2): p. 191-6. 262. Nakao, S., et al., Inhibitory effect of propofol on ketamine-induced c-Fos expression in the rat posterior cingulate and retrosplenial cortices is mediated by GABAA receptor activation. Acta Anaesthesiol Scand, 2003. 47(3): p. 284-90. 263. Lepack, A.E., et al., Fast-acting antidepressants rapidly stimulate ERK signaling and BDNF release in primary neuronal cultures. Neuropharmacology, 2016. 111: p. 242-252. 264. Phoumthipphavong, V., et al., Longitudinal Effects of Ketamine on Dendritic Architecture In Vivo in the Mouse Medial Frontal Cortex. eNeuro, 2016. 3(2). 265. Chen, W.G., et al., Derepression of BDNF transcription involves calcium-dependent phosphorylation of MeCP2. Science, 2003. 302(5646): p. 885-9. 266. Martinowich, K., et al., DNA methylation-related chromatin remodeling in activity- dependent BDNF gene regulation. Science, 2003. 302(5646): p. 890-3. 267. Chin, E.W.M., et al., Choline Rescues Behavioural Deficits in a Mouse Model of Rett Syndrome by Modulating Neuronal Plasticity. Mol Neurobiol, 2018.

152

268. Tropea, D., et al., Partial reversal of Rett Syndrome-like symptoms in MeCP2 mutant mice. Proc Natl Acad Sci U S A, 2009. 106(6): p. 2029-34. 269. Kasai, H., et al., Structural dynamics of dendritic spines in memory and cognition. Trends Neurosci, 2010. 33(3): p. 121-9. 270. Kasai, H., et al., Structure-stability-function relationships of dendritic spines. Trends Neurosci, 2003. 26(7): p. 360-8. 271. Yadav, A., et al., Morphologic evidence for spatially clustered spines in apical dendrites of monkey neocortical pyramidal cells. J Comp Neurol, 2012. 520(13): p. 2888-902. 272. Klose, R.J., et al., DNA binding selectivity of MeCP2 due to a requirement for A/T sequences adjacent to methyl-CpG. Mol Cell, 2005. 19(5): p. 667-78. 273. Armbruster, B.N., et al., Evolving the lock to fit the key to create a family of G protein- coupled receptors potently activated by an inert ligand. Proc Natl Acad Sci U S A, 2007. 104(12): p. 5163-8. 274. Akbarian, S., The neurobiology of Rett syndrome. Neuroscientist, 2003. 9(1): p. 57-63. 275. Katz, D.M., et al., Rett Syndrome: Crossing the Threshold to Clinical Translation. Trends Neurosci, 2016. 39(2): p. 100-113. 276. Kishi, N. and J.D. Macklis, MeCP2 functions largely cell-autonomously, but also non-cell- autonomously, in neuronal maturation and dendritic arborization of cortical pyramidal neurons. Exp Neurol, 2010. 222(1): p. 51-8. 277. Durand, S., et al., NMDA receptor regulation prevents regression of visual cortical function in the absence of Mecp2. Neuron, 2012. 76(6): p. 1078-90. 278. Franklin KBJ, P.G., Paxinos and Franklin's The mouse brain in stereotaxic coordinates. 2012. 279. Aschauer, D.F., S. Kreuz, and S. Rumpel, Analysis of transduction efficiency, tropism and axonal transport of AAV serotypes 1, 2, 5, 6, 8 and 9 in the mouse brain. PLoS One, 2013. 8(9): p. e76310. 280. Vardy, E., et al., A New DREADD Facilitates the Multiplexed Chemogenetic Interrogation of Behavior. Neuron, 2015. 86(4): p. 936-946. 281. Urban, D.J. and B.L. Roth, DREADDs (designer receptors exclusively activated by designer drugs): chemogenetic tools with therapeutic utility. Annu Rev Pharmacol Toxicol, 2015. 55: p. 399-417. 282. Scofield, M.D., et al., Gq-DREADD Selectively Initiates Glial Glutamate Release and Inhibits Cue-induced Cocaine Seeking. Biol Psychiatry, 2015. 78(7): p. 441-51. 283. Stachniak, T.J., A. Ghosh, and S.M. Sternson, Chemogenetic synaptic silencing of neural circuits localizes a hypothalamus-->midbrain pathway for feeding behavior. Neuron, 2014. 82(4): p. 797-808. 284. Krashes, M.J., et al., Rapid, reversible activation of AgRP neurons drives feeding behavior in mice. J Clin Invest, 2011. 121(4): p. 1424-8. 285. Kong, D., et al., GABAergic RIP-Cre neurons in the arcuate nucleus selectively regulate energy expenditure. Cell, 2012. 151(3): p. 645-57. 286. Ferguson, S.M., et al., Transient neuronal inhibition reveals opposing roles of indirect and direct pathways in sensitization. Nat Neurosci, 2011. 14(1): p. 22-4. 287. Chen, X., et al., The first structure-activity relationship studies for designer receptors exclusively activated by designer drugs. ACS Chem Neurosci, 2015. 6(3): p. 476-84. 288. Atasoy, D., et al., Deconstruction of a neural circuit for hunger. Nature, 2012. 488(7410): p. 172-7. 289. Alexander, G.M., et al., Remote control of neuronal activity in transgenic mice expressing evolved G protein-coupled receptors. Neuron, 2009. 63(1): p. 27-39.

153

290. Loffler, S., et al., Comment on "Impaired respiratory and body temperature control upon acute serotonergic neuron inhibition". Science, 2012. 337(6095): p. 646; author reply 646. 291. Gomez, J.L., et al., Chemogenetics revealed: DREADD occupancy and activation via converted clozapine. Science, 2017. 357(6350): p. 503-507. 292. Dhingra, R.R., et al., Decreased Hering-Breuer input-output entrainment in a mouse model of Rett syndrome. Front Neural Circuits, 2013. 7: p. 42. 293. Whissell, P.D., S. Tohyama, and L.J. Martin, The Use of DREADDs to Deconstruct Behavior. Front Genet, 2016. 7: p. 70. 294. Fassini, A., et al., NOP receptors in the prelimbic cortex have an inhibitory influence on cardiovascular responses induced by restraint stress. Neuropeptides, 2016. 57: p. 35-44. 295. Scopinho, A.A., R.F. Tavares, and F.M. Correa, The medial forebrain bundle mediates cardiovascular responses to electrical stimulation of the medial prefrontal cortex. Auton Neurosci, 2009. 147(1-2): p. 38-47. 296. Abdala, A.P., et al., Deficiency of GABAergic synaptic inhibition in the Kolliker-Fuse area underlies respiratory dysrhythmia in a mouse model of Rett syndrome. J Physiol, 2016. 594(1): p. 223-37. 297. Pratt, J.A., et al., Modelling prefrontal cortex deficits in schizophrenia: implications for treatment. Br J Pharmacol, 2008. 153 Suppl 1: p. S465-70. 298. Shepherd, G.M., Corticostriatal connectivity and its role in disease. Nat Rev Neurosci, 2013. 14(4): p. 278-91. 299. Woodyatt, G.C. and B.E. Murdoch, The effect of the presentation of visual and auditory stimuli on the breathing patterns of two girls with Rett syndrome. J Intellect Disabil Res, 1996. 40 ( Pt 3): p. 252-9. 300. Bauman, M.L., T.L. Kemper, and D.M. Arin, Microscopic observations of the brain in Rett syndrome. Neuropediatrics, 1995. 26(2): p. 105-8. 301. Hao, S., et al., Forniceal deep brain stimulation rescues hippocampal memory in Rett syndrome mice. Nature, 2015. 526(7573): p. 430-4. 302. Lang, M., et al., Selective preservation of MeCP2 in catecholaminergic cells is sufficient to improve the behavioral phenotype of male and female Mecp2-deficient mice. Hum Mol Genet, 2013. 22(2): p. 358-71. 303. Lu, H., et al., Loss and Gain of MeCP2 Cause Similar Hippocampal Circuit Dysfunction that Is Rescued by Deep Brain Stimulation in a Rett Syndrome Mouse Model. Neuron, 2016. 91(4): p. 739-747. 304. McDonald, A.J., F. Mascagni, and L. Guo, Projections of the medial and lateral prefrontal cortices to the amygdala: a Phaseolus vulgaris leucoagglutinin study in the rat. Neuroscience, 1996. 71(1): p. 55-75. 305. Arruda-Carvalho, M. and R.L. Clem, Prefrontal-amygdala fear networks come into focus. Front Syst Neurosci, 2015. 9: p. 145. 306. Vidal-Gonzalez, I., et al., Microstimulation reveals opposing influences of prelimbic and infralimbic cortex on the expression of conditioned fear. Learn Mem, 2006. 13(6): p. 728- 33. 307. Burgos-Robles, A., et al., Consolidation of fear extinction requires NMDA receptor- dependent bursting in the ventromedial prefrontal cortex. Neuron, 2007. 53(6): p. 871- 80. 308. Milad, M.R., I. Vidal-Gonzalez, and G.J. Quirk, Electrical stimulation of medial prefrontal cortex reduces conditioned fear in a temporally specific manner. Behav Neurosci, 2004. 118(2): p. 389-94.

154

309. Kim, J.H. and R. Richardson, New findings on extinction of conditioned fear early in development: theoretical and clinical implications. Biol Psychiatry, 2010. 67(4): p. 297- 303. 310. Milad, M.R. and G.J. Quirk, Neurons in medial prefrontal cortex signal memory for fear extinction. Nature, 2002. 420(6911): p. 70-4. 311. Santini, E., M. Sepulveda-Orengo, and J.T. Porter, Muscarinic receptors modulate the intrinsic excitability of infralimbic neurons and consolidation of fear extinction. Neuropsychopharmacology, 2012. 37(9): p. 2047-56. 312. Fontanez-Nuin, D.E., et al., Memory for fear extinction requires mGluR5-mediated activation of infralimbic neurons. Cereb Cortex, 2011. 21(3): p. 727-35. 313. Chang, C.H. and S. Maren, Strain difference in the effect of infralimbic cortex lesions on fear extinction in rats. Behav Neurosci, 2010. 124(3): p. 391-7. 314. Quirk, G.J., et al., The role of ventromedial prefrontal cortex in the recovery of extinguished fear. J Neurosci, 2000. 20(16): p. 6225-31. 315. Marek, R., et al., The amygdala and medial prefrontal cortex: partners in the fear circuit. J Physiol, 2013. 591(10): p. 2381-91. 316. Berretta, S., et al., Infralimbic cortex activation increases c-Fos expression in intercalated neurons of the amygdala. Neuroscience, 2005. 132(4): p. 943-53. 317. Sah, P. and M. Lopez De Armentia, Excitatory synaptic transmission in the lateral and central amygdala. Ann N Y Acad Sci, 2003. 985: p. 67-77. 318. Sah, P., et al., The amygdaloid complex: anatomy and physiology. Physiol Rev, 2003. 83(3): p. 803-34. 319. Millhouse, O.E., The intercalated cells of the amygdala. J Comp Neurol, 1986. 247(2): p. 246-71. 320. Pinard, C.R., F. Mascagni, and A.J. McDonald, Medial prefrontal cortical innervation of the intercalated nuclear region of the amygdala. Neuroscience, 2012. 205: p. 112-24. 321. Amir, A., T. Amano, and D. Pare, Physiological identification and infralimbic responsiveness of rat intercalated amygdala neurons. J Neurophysiol, 2011. 105(6): p. 3054-66. 322. Sheng, M., et al., Changing subunit composition of heteromeric NMDA receptors during development of rat cortex. Nature, 1994. 368(6467): p. 144-7. 323. Mierau, S.B., et al., Cell-Specific Regulation of N-Methyl-D-Aspartate Receptor Maturation by Mecp2 in Cortical Circuits. Biol Psychiatry, 2016. 79(9): p. 746-54. 324. Maiti, P., et al., Molecular regulation of dendritic spine dynamics and their potential impact on synaptic plasticity and neurological diseases. Neurosci Biobehav Rev, 2015. 59: p. 208-37. 325. Mancuso, J.J., et al., Methods of dendritic spine detection: from Golgi to high-resolution optical imaging. Neuroscience, 2013. 251: p. 129-40. 326. Lambe, E.K. and G.K. Aghajanian, Hypocretin (orexin) induces calcium transients in single spines postsynaptic to identified thalamocortical boutons in prefrontal slice. Neuron, 2003. 40(1): p. 139-50. 327. Bacon, S.J., et al., Amygdala input to medial prefrontal cortex (mPFC) in the rat: a light and electron microscope study. Brain Res, 1996. 720(1-2): p. 211-9. 328. Gideons, E.S., E.T. Kavalali, and L.M. Monteggia, Mechanisms underlying differential effectiveness of memantine and ketamine in rapid antidepressant responses. Proc Natl Acad Sci U S A, 2014. 111(23): p. 8649-54. 329. Bjorkholm, C. and L.M. Monteggia, BDNF - a key transducer of antidepressant effects. Neuropharmacology, 2016. 102: p. 72-9.

155

330. Liu, R.J., et al., Ketamine Strengthens CRF-Activated Amygdala Inputs to Basal Dendrites in mPFC Layer V Pyramidal Cells in the Prelimbic but not Infralimbic Subregion, A Key Suppressor of Stress Responses. Neuropsychopharmacology, 2015. 40(9): p. 2066-75. 331. Taft, C.E. and G.G. Turrigiano, PSD-95 promotes the stabilization of young synaptic contacts. Philos Trans R Soc Lond B Biol Sci, 2014. 369(1633): p. 20130134. 332. Feng, G., et al., Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron, 2000. 28(1): p. 41-51. 333. Sotres-Bayon, F., C.K. Cain, and J.E. LeDoux, Brain mechanisms of fear extinction: historical perspectives on the contribution of prefrontal cortex. Biol Psychiatry, 2006. 60(4): p. 329-36. 334. Ressler, R.L. and S. Maren, Synaptic encoding of fear memories in the amygdala. Curr Opin Neurobiol, 2018. 54: p. 54-59. 335. Penzo, M.A., et al., The paraventricular thalamus controls a central amygdala fear circuit. Nature, 2015. 519(7544): p. 455-9. 336. Li, Y., et al., Learning and reconsolidation implicate different synaptic mechanisms. Proc Natl Acad Sci U S A, 2013. 110(12): p. 4798-803. 337. Huang, T.W., et al., Progressive Changes in a Distributed Neural Circuit Underlie Breathing Abnormalities in Mice Lacking MeCP2. J Neurosci, 2016. 36(20): p. 5572-86. 338. Rossini, P.M., et al., Non-invasive electrical and magnetic stimulation of the brain, spinal cord, roots and peripheral nerves: Basic principles and procedures for routine clinical and research application. An updated report from an I.F.C.N. Committee. Clin Neurophysiol, 2015. 126(6): p. 1071-1107. 339. Terao, Y. and Y. Ugawa, Basic mechanisms of TMS. J Clin Neurophysiol, 2002. 19(4): p. 322-43. 340. Chen, R., et al., Depression of motor cortex excitability by low-frequency transcranial magnetic stimulation. Neurology, 1997. 48(5): p. 1398-403. 341. Pascual-Leone, A., et al., Responses to rapid-rate transcranial magnetic stimulation of the human motor cortex. Brain, 1994. 117 ( Pt 4): p. 847-58. 342. Lefaucheur, J.P., et al., Evidence-based guidelines on the therapeutic use of repetitive transcranial magnetic stimulation (rTMS). Clin Neurophysiol, 2014. 125(11): p. 2150- 2206. 343. Baek, A., et al., High-Frequency Repetitive Magnetic Stimulation Enhances the Expression of Brain-Derived Neurotrophic Factor Through Activation of Ca(2+)-Calmodulin- Dependent Protein Kinase II-cAMP-Response Element-Binding Protein Pathway. Front Neurol, 2018. 9: p. 285. 344. Krajnc, N. and J. Zidar, The role of transcranial magnetic stimulation in evaluation of motor cortex excitability in Rett syndrome. Eur J Paediatr Neurol, 2016. 20(4): p. 597- 603. 345. Rotaru, D.C., et al., Adult Ube3a Gene Reinstatement Restores the Electrophysiological Deficits of Prefrontal Cortex Layer 5 Neurons in a Mouse Model of Angelman Syndrome. J Neurosci, 2018. 38(37): p. 8011-8030. 346. Haws, M.E., et al., PTEN knockdown alters dendritic spine/protrusion morphology, not density. J Comp Neurol, 2014. 522(5): p. 1171-90. 347. Spinelli, L., et al., Functionally distinct groups of inherited PTEN mutations in autism and tumour syndromes. J Med Genet, 2015. 52(2): p. 128-34. 348. Zhou, J., et al., Pharmacological inhibition of mTORC1 suppresses anatomical, cellular, and behavioral abnormalities in neural-specific Pten knock-out mice. J Neurosci, 2009. 29(6): p. 1773-83.

156