Network and Cellular Effects of the Mu Opioid Receptor in Cortical Interneurons

Home , Opioidergic, Opioid receptor, Petersson, Petraitis

University of Vermont ScholarWorks @ UVM

Graduate College Dissertations and Theses Dissertations and Theses

2021

Adrian Dutkiewicz University of Vermont

Follow this and additional works at: https://scholarworks.uvm.edu/graddis

Part of the Neuroscience and Neurobiology Commons, and the Pharmacology Commons

Recommended Citation Dutkiewicz, Adrian, "Network and Cellular Effects of the Mu Opioid Receptor in Cortical Interneurons" (2021). Graduate College Dissertations and Theses. 1349. https://scholarworks.uvm.edu/graddis/1349

This Dissertation is brought to you for free and open access by the Dissertations and Theses at ScholarWorks @ UVM. It has been accepted for inclusion in Graduate College Dissertations and Theses by an authorized administrator of ScholarWorks @ UVM. For more information, please contact [email protected]. NETWORK AND CELLULAR EFFECTS OF THE MU OPIOID RECEPTOR IN CORTICAL INTERNEURONS

A Dissertation Presented

Adrian P. Dutkiewicz

The Faculty of the Graduate College

The University of Vermont

In Partial Fulfillment of the Requirements For the Degree of Doctor of Philosophy Specializing in Neuroscience

May, 2021

Defense Date: Feb 26th, 2021 Dissertation Examination Committee:

Anthony Morielli, PhD., Advisor Matthew Weston, PhD., Chairperson Donna Toufexis, PhD. Benedek Erdos, PhD. Cynthia J. Forehand, PhD., Dean of the Graduate College ABSTRACT

The µ opioid receptor (µOR) exerts a powerful excitatory effect in cortical circuits and cultured neurons by promoting glutamatergic activity after binding endogenous or exogenous opioids. While most research indicates that the receptor does this by decreasing activity or output of GABAergic interneurons that inhibit glutamate-releasing Pyramidal Neurons, other experiments suggest that the µOR directly upregulates excitatory Pyramidal Neurons instead. Thus, the cellular target of cortical opioid agonists remains unclear, and the µOR’s net excitatory mechanisms are not fully understood. Consequently, utilizing electrophysiology to detect µOR responses to the specific agonist [D-Ala2, N-Me-Phe4, Gly5-ol]-Enkephalin (DAMGO) has yielded incomplete information on its effects. This receptor has been shown to modulate both αDTX sensitive and insensitive currents in thalamic neurons to exert its effects, but this has not been investigated in cortical neurons. Here, we utilized a combination of calcium- imaging and patch-clamp electrophysiology in cultured rat neocortical neurons to investigate the network and cellular effects of the µOR to understand its mechanisms and their consequences. With our experiments testing its effects on spontaneous calcium oscillations, we found that the µOR exerts its net excitatory effects through inhibition of GABAergic interneurons. Our separate set of studies using patch-clamp electrophysiology reveals that the µOR has multiple inhibitory effects on firing frequency and action potential kinetics, including αDTX-sensitive and insensitive ones. Thus, opioids suppress GABAergic interneurons to promote net excitation in cortical circuits. Collectively, these findings promote our understanding of effects of endogenous or exogenous opioids on cortical networks, as well as provide robust analyses of the electrophysiological effects of the µOR which could provide insight into further studies of this receptor on cortical neurons.

ACKNOWLEDGEMENTS

First off, I would like to acknowledge my family for their interest in my research: Thank you Deborah Dutkiewicz, Leonard Dutkiewicz, Cassie Dutkiewicz, Nicholas Dutkiewicz, and Sariah Dutkiewicz. I am humbled by their unwavering support.

I would like to thank my mentor, Dr. Tony Morielli, for his guidance and teaching over the years. His refreshingly positive attitude and optimism has also been a comforting presence during these intense years. I am particularly thankful for the faith he had in me when undertaking some ambitious experiments and his appreciation for original ideas.

I would like to thank the Pharmacology department and its Chairperson Dr. Mark

Nelson for his leadership over these years. I am particularly thankful for their support for the lab space and resources that facilitated these experiments.

I would like to thank my committee, chaired by Dr. Matthew Weston and committee members Dr. Donna Toufexis and Dr. Benedek Erdos, who have provided insight and suggestions over the years. They gave me more ideas for extra analyses in

Chapter 3 and reinforced its methodological and analytical thoroughness. I would additionally like to thank Dr. Weston for referring us to the AAV that made Chapter 3 possible.

I would like to thank the former graduate students in the Morielli lab: Dr. Sharath

Madasu, Dr. Kutibh Chihabi, and Dr. Eugene Cilento. They taught me a lot, especially during the initial stages of my project, as I learned various techniques in the neuroscience lab.

I would like to thank our program coordinators Haley Olszewski, Carrie Perkins, and Darby Mayville for the logistical and personal support that they have demonstrated over the years. I would also like to thank the students of the NGP, with a special shout- out to Yu Han. It was great GTAing and working with you all.

I would like to thank the staff of Core Facilities at UVM, namely Thomm

Butolph, Todd Clason, Sheryl White, and the Dean of the Graduate College Dr. Cindy

Forehand. Their equipment and efforts facilitated these experiments and gave me the opportunity to collect a large volume of data.

Finally, I would like to thank all my undergraduates of NSCI 112 and the course directors Dr. Allison Anacker and Dr. Jaeda Coutinho-Budd for all of the pleasant distractions and for getting me out of the lab once in a while. Those mornings and afternoons are the things I will miss the most about graduate school.

iii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii LIST OF TABLES ...... vii LIST OF FIGURES ...... viii CHAPTER 1: COMPREHENSIVE LITERATURE REVIEW ...... 1 1.1 Opioids and neuropharmacology ...... 1 1.2 Overview of the cerebral cortex ...... 4 1.3 Cortical neuroanatomy of the endogenous opioid system ...... 6 1.4 Subcortical neuroanatomy of the endogenous opioid system ...... 11 1.5 Opioids: decisions and rewards ...... 13 1.6 Opioids: analgesia and pain-perception ...... 16 1.7 Introduction to neocortical neurons...... 19 1.8 Excitatory cortical neurons...... 20 1.9 Inhibitory cortical neurons ...... 21 1.10 µOR-expressing cortical interneurons ...... 24 1.11 Roles and features of µOR+ interneurons: PV+ and VIP+ ...... 29 1.12 Roles and features of µOR+ interneurons: NGFCs...... 30 1.13 µOR disinhibition ...... 33 1.14 Excitation/inhibition balance...... 34 1.15 Cognitive function and band oscillations ...... 35 1.16 Calcium transients and excitability ...... 38 1.17 Opioid receptors ...... 42 1.18 µ opioid receptor ...... 44 1.19 Endogenous opioid ligands ...... 47 1.20 DAMGO ...... 50 1.21 Opioid receptor intracellular cascades ...... 51 1.22 Opioid desensitization ...... 55 1.23 µOR endocytic cycling ...... 56 1.24 Basics of neurophysiology ...... 58 1.25 Hyperpolarization by the µOR ...... 62 iv

1.26 Voltage-sensitive K channels ...... 64 1.27 µOR regulation of voltage-sensitive channels ...... 68 1.28 Measures of excitability in individual neurons ...... 70 1.29 Cortical neurodevelopment and Dlx genes ...... 72 1.30 Primary neuron culture ...... 75 CHAPTER 2: NEOCORTICAL µ OPIOD RECEPTORS ENHANCE SPONTANEOUS CALCIUM OSCILLATIONS THROUGH DOWNREGULATION OF GABAA AND GABAB RECEPTORS ...... 76 2.1 Introduction ...... 76 2.2 Materials and methods ...... 80 2.2.1 Cell cultures ...... 80 2.2.2 Calcium imaging...... 80 2.2.3 Calcium-imaging analyses ...... 83 2.2.4 Validation for auto_mask ...... 85 2.2.5 Whole-cell recordings ...... 86 2.2.6 Sample sizes and statistics ...... 86 2.3 Results ...... 88 2.3.1 Measuring synchronized spontaneous calcium oscillations ...... 88 2.3.2 SCOs require neuronal activity and glutamatergic signaling ...... 88 2.3.3 First dataset experimental design ...... 92 2.3.4 µOR agonism enhances spontaneous Ca2+ oscillations ...... 93 2.3.5 The effect of DAMGO is mediated through both GABA receptors...... 95 2.3.6 Interneurons and Pyramidal Neurons ...... 100 2.3.7 Second dataset experimental design ...... 102 2.3.8 Changes similar in interneurons and excitatory neurons ...... 103

2.3.9 Effect of DAMGO in picrotoxin similar to GABAB blockade ...... 106 2.4 Discussion ...... 108 CHAPTER 3: µ OPIOID RECEPTORS MODULATE ACTION POTENTIAL KINETICS AND FIRING FREQUENCY IN NEOCORTICAL INTERNEURONS ... 114 3.1 Introduction ...... 114 3.2 Materials and methods ...... 119 3.3 Results ...... 126 v

3.3.1 Identifying responders ...... 126 3.3.2 DAMGO effects on action potential activity and kinetics ...... 135 3.3.3 DAMGO alters AP waveform and pattern of evoked APs in H-responders .. 136 3.4 Discussion ...... 168 CHAPTER 4: COMPREHENSIVE DISCUSSION ...... 177 4.1 Overview ...... 177 4.2 αDTX-sensitive/insensitive effects ...... 179 4.3 Predicting a DAMGO response...... 180 4.4 Identifying responders by subtype ...... 182 4.5 Origin of spontaneous APs ...... 183 4.6 Drawbacks of whole-cell recording ...... 184 4.7 DAMGO and GABARs ...... 185 4.8 DAMGO inhibition and GABA release ...... 186 4.9 True role of µOR+ interneurons in cortical networks ...... 187

4.10 GABAA receptors counteract hypersynchrony...... 190 4.11 Unknown physiology of SCOs ...... 192 4.12 SCOs: glutamatergic input or neuronal output? ...... 194 4.13 SCOs: AMPARs, NMDRs or mGluRs? ...... 196 4.14 Relationship between SCO halfwidth and interneuronal function ...... 197 4.15 Factors other than GABARs can modulate SCO halfwidth ...... 198 4.16 Sub-threshold calcium oscillations...... 200 4.17 Additional peakfinder measures ...... 201 4.18 SCOs for drug discovery and assays of glutamatergic activity ...... 202 4.19 Between SCOs and interneuronal suppression...... 204 4.20 Summary ...... 205 BIBLIOGRAPHY ...... 206

LIST OF TABLES

Table 1. Validation for auto_mask ROI generator...... 86 Table 2. DAMGO enhances amplitude and halfwidth of SCOs...... 95 Table 3. Manual measurements positively correlate with script-derived values...... 121 Table 4. Control and drug groups for electrophysiology...... 129 Table 5. Responder groups demonstrate different proportions of spike patterns...... 135 Table 6. Summary of DAMGO effects in H-responders...... 141 Table 7. Few H-responders shift spiking patterns after DAMGO...... 142 Table 8. αDTX-sensitive channels reverse some DAMGO effects...... 147 Table 9. DAMGO Effect in S-responders...... 155 Table 10. Summary of effects for H-responders vs S-responders...... 156 Table 11. Comparisons for the effect of polarity shift...... 159 Table 12. Posthoc summary of electrophysiological changes...... 167

vii

LIST OF FIGURES

Figure 1. Representative example of calcium peaks in one neuron...... 83 Figure 2. Neurons in image field have synchronous spontaneous calcium oscillations. .. 85 Figure 3. Neuronal activity and glutamatergic signaling required for SCOs...... 90 Figure 4. SCOs correspond to bursts of activity...... 91 Figure 5. Timecourse of experiment...... 93 Figure 6. DAMGO enhances calcium oscillation duration and amplitude...... 94 Figure 7. Blockading both GABARs prevents DAMGO-enhancement of SCOs...... 99 Figure 8. Representative calcium peaks for various drug conditions...... 99 Figure 9. Cultured neurons transformed with AAV-mDlx-NLS-mRuby2...... 101 Figure 10. Second dataset experimental design...... 102 Figure 11. DAMGO enhancement of SCOs observed in both neuron types...... 105 Figure 12. Effect of GABAB inhibition...... 106 Figure 13. Sample AP diagram...... 123 Figure 14. Cultured neurons with AAV-mDlx-NLS-mRuby2...... 127 Figure 15. Time course for patch-clamp recordings...... 128 Figure 16. H-responders hyperpolarize...... 131 Figure 17. Spontaneous APs after DAMGO or saline...... 132 Figure 18. Representative Examples of Firing Patterns...... 134 Figure 19. Spike frequency altered in H-responders...... 138 Figure 20. AP kinetics are altered in H-responders...... 140 Figure 21. αDTX reverses DAMGO effects on ISI and AP number...... 145 Figure 22. αDTX partially reverses some effects of DAMGO in H-responders...... 150 Figure 23. Example of αDTX reversal of DAMGO...... 151 Figure 24. S-responders also hyperpolarize...... 152 Figure 25. S-responders have changes in AP waveform...... 153 Figure 26. Polarity shift does not account for other changes...... 158 Figure 27. Responder categories captured most large changes...... 162 Figure 28. Nicotinic modulation of SCOs...... 203

viii

CHAPTER 1: COMPREHENSIVE LITERATURE REVIEW

1.1 Opioids and neuropharmacology

Our understanding of neuroscience has been profoundly shaped by the pharmacological tools we have available. This is particularly true for the development and research in opioids, which have been used for thousands of years for both recreational and medicinal purposes (Santella & Triggle, 2007). Opium’s active component, morphine, was successfully isolated from the plant product by the mid-19th century (Blakemore &

White, 2002). Since then, this chemical has been successfully modified to enhance its properties; in some instances, to enhance its pain-killing properties, and in other cases for nefarious purposes by leveraging its addictive and euphoric qualities. Codeine, papaverine, buprenorphine, fentanyl, methadone, and diamorphine (heroin) are all opioids – either natural or synthetic that have been developed over the generations

(Brownstein, 1993; Pathan & Williams, 2012). These drugs are all exogenous opioids that are ingested, injected, or inhaled by the subject and they are not normally found in humans or other animals.

Chemists eventually discovered drugs (such as naloxone and naltrexone) that can be co-administered to reverse the effects of morphine. But the mechanisms that all these opioid-derived drugs used to induce their effects on living beings remained a mystery for several generations. Some of earliest studies in modern neuropharmacology were dedicated to determining how these drugs physically interact with our physiology to produce their characteristic pain relief and euphoria.

Over 100 years later, in the 1970s, researchers identified the means that opioids use to exert their chemical effects using radio-labelled ligands; these drugs interacted with various receptor proteins located on cell membranes and were distributed unevenly in the brain and body (Goldstein, Lowney, & Pal, 1971; Kuhar, Pert, & Snyder, 1973;

Pert & Snyder, 1973; E. J. Simon, Hiller, & Edelman, 1973; Terenius, 1973). The µ

Opioid Receptor (µOR) is presently known to be the main target of these drugs and is the principal opioid receptor (OR) that mediates pain analgesia and euphoric effects of opioid drugs (Valentino & Volkow, 2018).

Since it seemed unlikely that animals produce the µOR specifically for exogenous opiates, scientists predicted that these receptors were evolved for unidentified natural, neurotransmitters that may fulfill a pain-killing role in the brain (Brownstein, 1993).

These hypothesized neurotransmitters were called endogenous opioids because they were believed to occur naturally within animals and bind to the same opioid receptors that the exogenous opioids were already known to target. Their existence was shown in experiments in the 1970s; it was discovered naloxone, which can reverse the effects of morphine, could also reverse the analgesia seen after repeated, painful shocks (Akil,

Mayer, & Liebeskind, 1976). Thus, it appeared that animals secrete endogenous opioids which can be reversed the same way that an administered opioid drug could also be reversed.

Several different endogenous opioids, and a few receptors for those endogenous opioids, were subsequently discovered over the decades. Fascinatingly, consuming opioid drugs are not the only means that the opioid neurotransmitter system can be disrupted;

2 several neuropsychiatric conditions are shown to feature a disrupted endogenous opioid system, which can lead then to aberrations in pain-sensitivity, and deficient motivation and poor mood (Ashok, Myers, Reis Marques, Rabiner, & Howes, 2019; Der-Avakian &

Markou, 2012). The endogenous opioid neurotransmitter system is now a target for treatment in mood disorders exhibiting these symptoms (Butelman & Kreek, 2015; Lutz

& Kieffer, 2013; Valentino & Volkow, 2018). The endogenous opioid system is also implicated in binge-eating disorders, which arise from defective reward-valuation processes by drive satiated animals to consume food (Giuliano & Cottone, 2015). Much of this research has centered on the neocortex, the center of higher-function and decision- making.

Yet despite the long history of medical opioids, the usage of opioids is perilous.

Specifically, the addictiveness and rapid accumulation of tolerance to the drug has birthed the “opioid epidemic” of abuse. Some opioids, like heroin, originally were developed for medical applications but rapidly became popular as a drug of abuse

(Hosztafi, 2001). Heroin is one of the most addictive substances in human history. In addition to the addictiveness and tolerance issues, opiates can induce respiratory depression through the µOR and, ultimately, to death in high doses (Valentino &

Volkow, 2018).

Around 750,000 Americans have died since 1999 from opioid overdose; about

2/3rds of all drug overdoses are associated with opioids. This problem is also attributed partly to abuse of prescription opioids, which accounted for about 1/3rd of opioid-related deaths in 2018 (Services, 2020). Currently, there has been an interest in developing new

3 opioidergic drugs that can treat depression, impulsivity, and paradoxically, to also prevent addictions (Shippenberg, 2009; Sullivan, 2018). There is a critical need to understand how the µOR and opioidergic drugs function to exert their effects on brains.

1.2 Overview of the cerebral cortex

The human cortex constitutes almost half of the human brain and is associated with higher-order functions, such as perception, cognition, and voluntary movements (Molnár

& Pollen, 2014; Rakic, 2009). This region of the brain has expanded immensely over mammalian evolution leading to greater brain complexity and is crucial for the intellectual capability and flexibility (Lui, Hansen, & Kriegstein, 2011). However, this greater complexity has been accompanied by greater vulnerability; this vast brain region is implicated in schizophrenia (Hashimoto et al., 2008; David A. Lewis, Curley, Glausier,

& Volk, 2012), autism (Amaral, Schumann, & Nordahl, 2008; Carper & Courchesne,

2005; Courchesne & Pierce, 2005; Zilbovicius et al., 1995), major depressive disorder

(Bremner et al., 2002; Drevets, 2007; George et al., 2000; Lemogne, Delaveau, Freton,

Guionnet, & Fossati, 2012), and attention-deficit hyperactivity disorder (Rubia et al.,

2010; Wolf et al., 2009).

The human cerebral cortex encompasses the “wrinkly” superficial surface of the brain and is heavily gyrated (Lui et al., 2011). This region includes about 50 Brodmann’s

Areas that often have differences in their cytoarchitecture, depending on their function and connections (Johns, 2014; Rakic, 2009). All cortical areas are constituted of several layers of neurons, and the various types of neurons can be distributed unevenly through the layers and region of cortex (Lui et al., 2011).

The cerebral cortex can be divided into 3 major components; firstly, the neocortex

(also called isocortex); secondly, the allocortex, constituted of the paleocortex (primary olfactory cortex) and the archicortex (the hippocampus); and thirdly, the mesocortex

(cingulate cortex and parahippocampal gyrus) (Johns, 2014). The neocortex usually has 6 layers of stratified cortical neurons while the allocortex has 3 layers. The mesocortex, which is found between the allocortex and neocortex, has between 3-6 layers (Johns,

2014). This dissertation project deals exclusively with the neocortex (specifically the frontal cortex), but the hippocampus expresses similar cell types, and they are sometimes used as models for each other. However, some differences between the neurons of the hippocampus and neocortex will be noted in this text. Whenever possible, I have cited experiments done in neocortical neurons, but I cite hippocampal studies in lieu of neocortical studies because of the sometimes-disproportionate volume of relevant research done on hippocampal neurons.

All branches of science have benefited greatly from animal models. Animal models for neuroscience is particularly important because only animals possess brains, and brain research is sometimes a dangerous endeavor for the subject. While modern brain scans (PET, fMRI) are virtually harmless in moderation, much of neuroscience research requires “breaking” certain pathways and genes, or by introducing toxic reagents to understand how neurons work and interact with each other. Otherwise, all of neuroscience would strictly be correlational. Indeed, most every branch of science therefore requires models to serve as a stand-in for human participants, so that researchers can conduct science as ethically as possible. These findings can eventually be

5 translated into human medical science and improve the lives of humans (and nonhuman animals). However, neuroscience has a unique problem when compared to other branches of science; the brains of humans are incredibly complex when compared to other animals.

Thus, while the genes of a fruit fly and a human may be sufficiently similar with genetics research - but the structure of their brains are not similar enough for most brain research.

This problem is especially true for the neocortex of the human brain, which is uniquely complex and therefore difficult to compare with other animals.

Our fellow primates (for example, chimpanzees) have well-developed neocortices, but they are difficult to handle and carry serious ethical ramifications.

Rodents are a sensible starting point for studying neocortex because they are a model that are frequently used, and they can be handled relatively easily. Therefore, when scientists notice things about the brain that are initially observed in the brain scans of human subjects, other scientists can follow up on this research with studies done in rodents.

1.3 Cortical neuroanatomy of the endogenous opioid system

Much of the discussion in the literature about the roles of the endogenous opioid system revolves around several regions in the frontal cortex of humans. This region, fortunately, has homologs that can be studied in rats, and so I will delineate the relevant components.

This includes the prefrontal cortex (PFC), the orbitofrontal cortex (OFC), and the anterior cingulate cortex (ACC). In humans, the PFC is the dorsolateral prefrontal cortex (dlPFC), ventrolateral prefrontal cortex (vlPFC), the medial prefrontal cortex (mPFC) and the ventral prefrontal cortex (vPFC) (Striedter, 2005).

But do rats have a region homologous to the human PFC? At first, the consensus was that rats do not, because their own medial PFC was agranular; it lacked a layer of small (Pyramidal) neurons that populated the middle cortical layers that was present in humans. But, with the evolution of techniques that allowed scientists to trace afferent and efferent fiber pathways in the brain, more attention was devoted to their connections than their histology. The functions of the human prefrontal cortex are present in rats as the mPFC in a simplified form. Currently the use of rat mPFC as a model for human PFC is justified for several reasons, including being targeted by fibers from the mediodorsal nucleus of the thalamus, which is a thalamic region associated with memory and planning

– thus suggesting that rat mPFC was involved in these “higher functions” (Funahashi,

2013; Rose & Woolsey, 1948). Similarities are also expressed on the neurotransmitter receptors; both human prefrontal cortex and rat mPFC have reciprocal connections of cholinergic fibers from basal forebrain nuclei, and as well reciprocal dopaminergic fibers to/from the ventral tegmental area (the VTA is part of the opioidergic reward system)

(Carr & Sesack, 2000). The mPFC is also well-connected with the periaqueductal gray, a pain-killing neuroanatomical region that I will be discussing later (An, Bandler, Öngür, &

Price, 1998; Hardy & Leichnetz, 1981; Ong, Stohler, & Herr, 2019; Y.-f. Xie, Huo, &

Tang, 2009). Lesion studies also indicate that the human dlPFC and rat mPFC both mediate executive functions, such as decision-making, directing attention, and working memory (Hoover & Vertes, 2007). Therefore, the complex “human” brain functions that are distributed throughout a relatively large area, but can be found in a small, simplified form in the rat mPFC where it fulfills a similar function to humans.

Whether rats have a PFC has been a contentious topic, but the mPFC in rats is also intimately related to decision-making and the endogenous opioid system. This region expresses opioid receptors and opioid peptides (C. Jiang et al., 2019). This region is involved in suppressing impulsivity, binge-eating, mood disorders, reactivity to fear

(Blasio, Steardo, Sabino, & Cottone, 2014; Selleck et al., 2015; Tejeda et al., 2015). In rats, the ORs in this region can instigate overeating with a preference for carbohydrate- enriched foods(Mena, Sadeghian, & Baldo, 2011; Mena, Selleck, & Baldo, 2013).

Like the mPFC, the orbitofrontal cortex (OFC) in rats is also connected with the mediodorsal nucleus and subcortical structures associated with the endogenous opioid system in humans and rats, which will be discussed later in this review (Heilbronner,

Rodriguez-Romaguera, Quirk, Groenewegen, & Haber, 2016; Rose & Woolsey, 1948).

This includes afferents from the ventral tegmental area, a region associated with the reward system (M. J. M. Murphy & Deutch, 2018). The OFC is implicated in addiction, decision-making, and depression (Bremner et al., 2002; Rolls & Grabenhorst, 2008;

Volkow & Fowler, 2000).

The ACC is located along the medial surface of the brain and consists of a ribbon of tissue that encloses the anterior portion of the corpus callosum. The OFC and ACC also play an important role in decision-making, attention, and behavior. They show prominent expression of opioid receptors and endogenous opioids in both humans and rodents (Lau, Ambrose, Thomas, Qiao, & Borgland, 2020; Mitchell et al., 2012). Rat neurons cultured from the ACC also express these receptors in vitro (E. Tanaka & North,

1994). The ACC is implicated in behavior, affective disorders, schizophrenia and effort-

8 related decisions (Benes, McSparren, Bird, SanGiovanni, & Vincent, 1991; Devinsky,

Morrell, & Vogt, 1995; Elston, Croy, & Bilkey, 2019; Schweimer, Saft, & Hauber,

2005). This region also interacts with the reward and antinociceptive systems(Narita et al., 2010; Navratilova et al., 2015).

Although the hippocampal formation is not a neocortical region, this literature review will frequently reference their µORs. Some hippocampal neurons express µORs and endogenous opioids that bind to the receptor (Drake & Milner, 2006; H. K. Lee,

Dunwiddie, & Hoffer, 1980; Madison & Nicoll, 1988; Przewlocki et al., 1999; Stumm,

Zhou, Schulz, & Höllt, 2004; M. Whittington, Traub, Faulkner, Jefferys, & Chettiar,

1998). Interfering with the endogenous opioid system there, by direct infusion of µOR antagonists into the CA3, interferes with both the encoding and the retrieval of spatial memory (Meilandt, Barea-Rodriguez, Harvey, & Martinez, 2004). These receptors may also interfere with memory retrieval during stress (Shi et al., 2020). Chronic administration of opioids is believed to reduce neurogenesis in this region, which may then contribute to some of the long-lasting cognitive effects of opioid addiction (Eisch,

Barrot, Schad, Self, & Nestler, 2000; Persson et al., 2003). Therefore, the µORs and the endogenous opioids of the hippocampal formation are an important feature of memory encoding and retrieval.

The endogenous opioid system of the neocortex collectively modulates antinociception, reward valuation and reward-seeking behaviors. Disruption of this neurotransmitter system is believed to contribute to pathological and compulsive behaviors, such as behaviors eating disorders, pathological gambling, and drug-seeking

(B. Bencherif et al., 2005; Badreddine Bencherif et al., 2005; Joutsa et al., 2018; Mick et al., 2016; Mitchell et al., 2012; Qu, Huo, Huang, & Tang, 2015; Zubieta et al., 1996).

Some neuropsychiatric disorders, such as schizophrenia are associated with abnormalities in the cortical opioid system, which may produce some of the cognitive symptoms

(impairments in executive control) featured in that disorder (Volk, Radchenkova, Walker,

Sengupta, & Lewis, 2012). People with schizophrenia also have reduced affective responses to pain (Ashok et al., 2019; de la Fuente-Sandoval, Favila, Gómez-Martín,

León-Ortiz, & Graff-Guerrero, 2012; Linnman, Coombs, Goff, & Holt, 2013; Stubbs et al., 2015), which can also be found in animal models for the disease (Kekesi et al., 2011;

Szűcs, Büki, Kékesi, Horváth, & Benyhe, 2016). In a broader context, poor inhibitory control and motivational drive is featured in many psychiatric conditions, such as depression, attention deficit hyperactivity disorder, and drug addiction (American

Psychiatric Association; Jentsch & Pennington, 2014; Kaye et al., 2013; Merriam, Thase,

Haas, Keshavan, & Sweeney, 1999; Puumala & Sirviö, 1998; Smith, Mattick, Jamadar, &

Iredale, 2014; Wise, Murray, & Gerfen, 1996). Although the pain-modulation component appears related to cortical connections to the PAG, these latter functions are likely mediated through cortical projections to the hypothalamus and nucleus accumbens

(Baldo, 2016).

Taken together, the anterior cingulate, orbitofrontal and mPFC of rodents express

ORs, opioid neurotransmitters, and are interconnected with other neuroanatomical regions that mediate the notable functions of opioids. These cortical regions interact with a vast network of subcortical regions to compose the endogenous opioid system.

1.4 Subcortical neuroanatomy of the endogenous opioid system

Although the primary focus of these experiments is the ORs of the neocortex, both the cortical and subcortical opioidergic systems interact to exert their effects on animals. This is particularly true because opioids that are taken orally or injected intravenously can often bind to ORs in both areas. In addition, neuropsychiatric conditions that afflict the endogenous opioid neurotransmitter system may also include both cortical and subcortical regions, and extensive research on subcortical µORs has been performed for many years. These experiments can be used to make predictions about their mechanisms in the neocortex. I will therefore describe these subcortical components to explain how they collaborate with cortical regions.

The periaqueductal gray (PAG) of the midbrain (in the brainstem) is perhaps the best-known pain-modulating center of the brain (Devinsky et al., 1995; Hardy &

Leichnetz, 1981; Royce, 1983). The PAG manages descending pain inhibition mostly through its input on the rostroventral medulla (RVM). The RVM sends serotonergic fibers (interacting regions may send noradrenergic fibers) down the spinal cord and activate neurons in the dorsal horn of the spinal cord (Ossipov, Morimura, & Porreca,

2014). The effect of serotonin on their spinal targets is predicted by the particular receptor subtype, but opioids administered to directly to the PAG results in recruitment of

RVM neurons, which then leads to descending inhibition of pain (for review, see

(Ossipov et al., 2014)). Many parts of the brain modulate activity of the PAG because, by extension, this allows many parts of the brain to modulate pain transmission.

The ventral tegmental area (VTA) is a part of the brainstem that has strong control over motivated/rewarding behaviors (Fields & Margolis, 2015). The VTA is rich the dopamine-secreting efferent fibers (fibers from principal cells of the VTA) which project to several parts of the brain, including the prefrontal cortex. Application of opioids into the VTA increases firing rates of the principal neurons there, and leads to greater dopamine secretion (Carr & Sesack, 2000). It serves as a critical link between opioids and reward/addictive behaviors; ORs in the VTA can modulate the secretion of dopamine to the nucleus accumbens to reinforce drug-seeking, while also projecting to cortical regions to influence decisions being made there (Langlois & Nugent, 2017). For instance, the

VTA’s bidirectional connections with the anterior cingulate cortex facilitates the consideration of cost/benefit decisions in animals (rats, in this experiment) (Elston et al.,

2019). The prefrontal cortex also provides excitatory input back to the VTA and is therefore able to influence activity there as well by synapsing on the neurons of the VTA

(Carr & Sesack, 2000).

The amygdala is involved in fear and anxiety responses of animals, but it also plays a role in decision-making which will be discussed in greater detail later. But its function extends beyond its typical role of the “fear sensor” of the brain, and this region likewise contributes to reward-valuation in animals (Wassum, Ostlund, Maidment, &

Balleine, 2009). Sensory cortices and the sensory portion of the thalamus project here to provide this region with their input (Janak & Tye, 2015). The central nucleus of the amygdala (CeA) projects to the PAG to modulate pain sensation. Meanwhile, the basolateral amygdala (BLA) projects to the prefrontal cortex to contribute to decision-

12 making with rewards or aversive consequences (Barbas, 2000; Bechara, Damasio, Tranel,

& Damasio, 1997; Morrison & Salzman, 2010). Opioids in this region can modulate excitability in the central nucleus and basal lateral amygdala (Winters et al., 2017).

The data I have cited here mostly come from data about focal or application of opioids to a confined neuroanatomical region through microinfusions. This procedure allows researchers to determine the exact contribution of µORs in a given region of interest, but systemically-administered opioids tend to cause many effects (e.g., euphoria and pain-relief) to come hand-in-hand because they act on virtually all the components of the opioidergic system. For instance, by causing pain relief and euphoria to coincide by acting on opioid receptors in the cortex, PAG, amygdala, and VTA all simultaneously.

Therefore, while the µORs in precise regions may fulfill narrow roles, all these effects tend to happen simultaneously when someone takes an opioidergic drug.

1.5 Opioids: decisions and rewards

One of the most well-known features of the opioid system is its relationship with reward valuation, addiction, and drug abuse. The relationship between opioids and rewards is partially due to the high levels of opioid receptors expressed in the VTA – a region rich in dopamine-expressing principal neurons which, in turn, then lead to rewarding/pleasurable feelings and could encourage certain behaviors in the future (for review, see (Fields & Margolis, 2015)). Decisions and reward-seeking behaviors is a closely related feature based on valuations, which is a function of the VTA and cortical areas. Binge-eating, for instance, is an aberration in the valuation process that leads an

13 animal to overconsume food even when they are presently satiated (Giuliano & Cottone,

2015; Mena et al., 2011; Mena et al., 2013; Selleck et al., 2015).

To explore the link between brain regions and the activity of endogenous opioids, researchers use radio-labeled µOR-agonists and brain scans to monitor the activity of those receptors when a subject is participating in a designated task. Drug use is correlated with the release of endogenous opioids in the human orbitofrontal and ACC (Colasanti et al., 2012; Mick et al., 2014; Mitchell et al., 2012). These scans also allowed researchers to explore the relationship between expected outcomes and activity in the neocortex as well; placebo effects enhance dopamine and opioid release in the cortex and VTA, while nocebo effects correlate with reduced dopaminergic and opioidergic activity there (Scott et al., 2008).

Beyond the brain scans and correlative data, a great deal of information about the cortical contribution to reward-valuation and decision-making is also derived from microinfusions of opioids into discrete regions of the brain. Infusion of the opioid agonists (specifically, the µOR agonist DAMGO) into rat mPFC drives feeding, carbohydrate intake, and locomotive behaviors in satiated rats and is accompanied by enhanced activity in the hypothalamus and nucleus accumbens shell (Mena et al., 2011;

Mena et al., 2013; Selleck et al., 2015). Blockade of the µOR from endogenous opioid neurotransmitter with naltrexone in the mPFC inhibits binge-eating in rats, which suggests a causal link between engaging in compulsive actions and release of endogenous opioids in the mPFC (Bartus et al., 2003; Blasio et al., 2014). Opioid antagonists (namely naltrexone) are also FDA-approved for treating poor impulse-control, presumably due to

14 their ability to unlink opioidergic and dopaminergic activity in the brain (Pettinati et al.,

2006; Soyka & Rosner, 2008).

The VTA’s efferent connections to other parts of the brain allow it to stimulate reward-seeking behaviors where effort is invested (Narita et al., 2010; Ramsey, Gerrits,

& Van Ree, 1999; Schweimer et al., 2005). In the VTA, like other areas of the brain, opioid receptors suppress local inhibitory neurons, which then produces an acute increase in activity of efferent principal neurons (Billy Chieng, Azriel, Mohammadi, & Christie,

2011; Johnson & North, 1992; Margolis, Toy, Himmels, Morales, & Fields, 2012). These dopaminergic connections then project to several locations, including the nucleus accumbens and hypothalamus (Narita et al., 2010; Pierce & Kumaresan, 2006; Zheng,

Patterson, & Berthoud, 2007). The nucleus accumbens is important to reinforce addictions/rewards, while the hypothalamus can drive feeding behavior (Carr & Sesack,

2000). Additionally, the VTA also projects to cortical regions, such as the mPFC, OFC,

ACC to facilitate the decision-making processes (Narita et al., 2010; Pierce &

Kumaresan, 2006; Zheng et al., 2007)

Both the mPFC and OFC play important roles in decision-making, but they do not have exactly the same roles. The OFC’s contribution to the reward system is encoding the hedonic value of the reward. For example, this region is associated with the devaluation of the reward during satiety or illness (Rudebeck & Rich, 2018). Consistent with this, infusion of opioid agonists in this region enhances the valuation of rewards (Castro &

Berridge, 2017). Meanwhile the mPFC mediates the execution and monitoring of the

15 selected action once the decision has been made (Phillips, MacPherson, & Della Sala,

2002).

1.6 Opioids: analgesia and pain-perception

Pain perception is sometimes (and erroneously) seen as a simple and reflexive process from the painful stimulus to the perception of it. But pain perception is profoundly shaped by emotional and affective processes happening in the frontal cortex (for review of the PFC and pain, see (Devinsky et al., 1995; Hardy & Leichnetz, 1981; Lasagna,

1964; Ong et al., 2019; Petrovic, Kalso, Petersson, & Ingvar, 2002; Royce, 1983;

Westlund & Willis, 2015). Expectations, moods, and dispositions all affect the way we perceive and express pain. For instance, the analgesic effects of placebos are mediated by the ACC, PFC, insula (Choi et al., 2016; Qiu, Wu, Xu, & Sackett, 2009; Zubieta et al.,

2001), which shows that pain perception is shaped by complex processes in the neocortex.

Much of pain modulation is carried out by the PAG, which has control over descending fibers that can interfere with the transmission of pain in the spinal cord.

Exogenous/endogenous opioids, drugs, and other neurotransmitters can enhance PAG activity, which then modulates the RVM. The RVM projects serotonergic efferent fibers down the spinal cord, along with noradrenergic fibers from associated regions. Thus, opioids can have a tremendous effect on descending pain modulation and reducing the pain signal that enters the brain. But opioids can reduce the affective expression of pain but leave the actual pain perception component intact (Amir & Amit, 1978; Oertel et al.,

2008). Therefore, animals may feel the pain, but simply just care less about it. Several

16 regions of the brain can modify the activity of the RVM and PAG to influence pain perception, including the PFC (Bingel et al., 2011).

The PAG receives most of its afferent input from the mPFC, with smaller input from the somatosensory, insular, and cingulate cortices (Hardy & Leichnetz, 1981; Ong et al., 2019; Y.-f. Xie et al., 2009). This pathway appears to be specifically activated when treatments are expected to lead to pain relief (including placebos) (Choi et al.,

2016). Placebo and nocebo effects appear to have opposite polarity in opioid and dopamine release in cortical areas and the PAG which demonstrates a relationship between these two regions, evidence for cortical processing of pain perception, and the relationship between opioids and dopamine (Scott et al., 2008).

To investigate the correlational studies from brain imaging, scientists have done experiments with microinfusions in animal models. For instance, infusion of opioid agonists into the ACC of rats blocks pain-induced aversion during conditioned place preference tasks (Navratilova et al., 2015). Infusions of opioid agonists into the ventrolateral cortex of rats induces a naloxone-reversible inhibition of their tail-flick reflex in response to pain (X. Huang, Tang, Yuan, & Jia, 2001). Therefore, it appears that opioid receptors in neocortical (ACC) regions can modulate our reactions and aversion to pain.

Reactivity and expression of pain is also a very important feature that neocortical regions and the endogenous opioid system contribute to. Fear-conditioned analgesia

(where an animal has a weak reactivity to painful stimuli in a place that has been previously paired with painful stimuli) has been shown to be modulated by endogenous 17 opioids through microinfusions into the amygdala (Butler et al., 2011; Tershner &

Helmstetter, 2000). This is most likely related to projections from the amygdala to the

PAG (Helmstetter, Tershner, Poore, & Bellgowan, 1998). However, in addition to directly projecting to the PAG, the BLA neurons can also project to the mPFC, and PNs in Layer 5 can then project down to the PAG as well (Cheriyan, Kaushik, Ferreira, &

Sheets, 2016). The amygdala has bilateral connections with the cortex that are believed to be involved in fear conditioning an anticipation of pain and strengthened in people with chronic pain (Butler et al., 2011; Cheriyan et al., 2016; Simons et al., 2014). These nociceptive neurons of the mPFC are likely to be related to affective demonstration of pain, or the expectation of pain rather than to detect and discriminate pain (Onozawa,

Yagasaki, Izawa, Abe, & Kawakami, 2011). µORs in the mPFC can modulate reactivity to fear (relayed by input from the basolateral amygdala) (Tejeda et al., 2015).

So then if all of these brain regions collaborate to affect nociception, what then is the unique role of the PFC in this network? It is believed that the PFC uses the mood or expectations of the subject to modify the perception of their pain. Many factors could contribute these moods and expectations - for example, expectations from pharmacotherapy, memories, psychotherapy, and conscious suppression of pain could all modify the perception and expression of pain (Ong et al., 2019).

The mPFC, OFC and ACC are therefore involved in pain perception, pain expression, and decision-making. They also have opioid receptors that can be used to affect all these features. Having explained the gross neuroanatomy component of the endogenous opioid system, I will now refocus the attention on the neocortex by

18 introducing the neurons in that region. By doing this, we can then begin to understand how endogenous opioids exert their effects on neocortical networks and individual neocortical neurons.

1.7 Introduction to neocortical neurons

Neurons of the neocortex are composed of 2 main categories of neurons that interact and form a network to carry out the functions of the neocortex: the excitatory neurons and inhibitory neurons. These excitatory and inhibitory neocortical neurons form synapses on other neurons, which allows them to influence the firing on the neuron that they synapse onto. Any given neocortical neuron may receive input from both inhibitory and excitatory neurons, as well as afferent fibers from distant brain regions that I have previously discussed here. Since these neurons receive both excitatory and inhibitory input, their own firing tendency (excitability) depends on which of the two types of input are stronger in any given moment.

The most common type of neocortical neuron (about 80%) are the excitatory

(glutamatergic) neurons that can synapse on nearby neurons, as well as neurons that are far away (DeFelipe & Fariñas, 1992; Noback, Strominger, Demarest, & Ruggiero, 2005).

The other 20% are inhibitory (GABAergic) which synapse only on nearby neurons.

The inhibitory neurons will suppress the nearby excitatory neurons by secreting

GABA (γ-aminobutyric acid). Meanwhile the excitatory neurons synapse onto both near and far neurons and secrete the excitatory neurotransmitter glutamate. Both neuronal types may also produce other neurotransmitters, including endogenous opioids, which

19 may also produce other effects on their targets. These two neuron types interact heavily to regulate each other’s excitability (tendency to fire).

1.8 Excitatory cortical neurons

Most cortical excitatory neocortical neurons have a pyramidal shape and therefore are called Pyramidal Neurons (PNs). A smaller group of excitatory neurons, called granule or stellate neurons, can be also found, but they are mostly found in sensory cortex (Johns,

2014).

Pyramidal Neurons (PNs) are glutamatergic neurons found in all cortical layers except Layer I (DeFelipe & Fariñas, 1992); Noback et al., 2005). However, they tend to stratify in Layers 2/3 and Layer 5. Their projections can be directed locally or directed to other parts of the cortex and subcortical areas. They form asymmetric synapses at their targets, which include interneurons and dendritic spines of other PNs. Although these cells are relatively homogenous when compared to interneurons, PNs can assume a variety of shapes (DeFelipe & Fariñas, 1992). Generally, though, they have a pyramidal- shaped soma between 20-120 µM wide and have elaborate branches directly off their soma. Their apical dendrites project into Layer I of the neocortex and disperses into branches. All PNs bear ~10,000 spines along their apical dendrites which represent an excitatory postsynaptic density – which is where another excitatory neuron has synapses onto them. They may also have smaller, basal dendrites at the other 2 angles of their pyramidal soma. Their axon projects from the basal part of the neuron, opposite the apical dendrite (Johns, 2014).

Excitatory neurons develop from the dorsal telencephalon (the pallium) and migrate radially to the cortical plate from the germinal zone during embryonic development(Campbell, 2005). PNs in the deep layers develop first and then are bypassed by newer progenitors that create the superficial cortical layers (Campbell, 2005).

Therefore, the Layer 5 PNs usually develop earlier than Layer 2/3 PNs.

Some research in the neocortex using immunocytochemistry (ICC; an antibody- based technique to identify the localization of specific proteins) have specifically looked for expression of the µOR in PNs. For the most part, PNs in neocortical regions do not express the µOR. For example, 96.9% of the µOR-expressing (µOR+) neurons in the neocortex have been found to produce GABA (Taki, Kaneko, & Mizuno, 2000). On the other hand, some research teams have reported µOR-immunoreactivities on neocortical

PNs during their immunocytochemistry (ICC) experiments – so there is a possibility that some do express the µOR (Schmidt et al., 2003).

1.9 Inhibitory cortical neurons

Cortical interneurons make up the remaining 20% of neocortical neurons and they can belong to smaller subcategories of interneurons. All the inhibitory neurons secrete the neurotransmitter GABA, which binds to postsynaptic/extrasynaptic GABAA and

(sometimes) GABAB receptors, causing inhibition in their targets. Although these inhibitory neurons can be targeted by other neurons in distant parts of the brain, but they only synapse on local neurons. Therefore, these inhibitory cortical neurons are also called interneurons or, more precisely, GABAergic interneurons which excludes the glutamatergic interneurons.

Cortical interneurons have a short axon that projects locally, and they often receive afferent fibers from cortical and noncortical areas. For example, the mediodorsal nucleus of the thalamus projects to the rat mPFC and synapses on mostly interneurons over PNs (Rotaru, Barrionuevo, & Sesack, 2005), These axons can arise from the soma or principal dendrites of the interneuron (The Petilla Interneuron Nomenclature Group,

2008). Beyond that, their morphology can differ dramatically from fellow interneurons and is not an unambiguous indicator of cell type. Given their wide range of properties, scientists were driven to categorize different interneurons in order to understand their collective role in the brain.

However GABAergic interneurons are notoriously difficult to subclassify, because of the combinations of morphology, expression markers, and physiological properties that they can possess. But these distinctions are important because different subsets of interneurons have different roles, and there are some general patterns and distinctions. The lack of a formal criteria for classification led to formation of a consortium of scientists, the Petilla Interneuron Nomenclature Group (PING), that was established to consolidate research and to provide a standard classification of interneurons for neuroscientists (Batista‐Brito & Fishell, 2009; The Petilla Interneuron

Nomenclature Group, 2008). Although individual labs have their own internal preferred categorization systems, most classification schemes are similar to PING’s scheme.

Pioneering studies in rodent cortical interneurons have initially grouped all cortical interneurons into 3 distinct and non-overlapping classes based on expression of protein markers (Kawaguchi & Kubota, 1997; McCormick, Connors, Lighthall, & Prince,

1985). These classes are parvalbumin-positive (PV+) interneurons, vasoactive intestinal peptide-positive (VIP+) interneurons, and somatostatin-positive (SST+ or SOM+) interneurons. However, in the years since then, a much more complicated portrait emerged; interneurons are difficult to group into discrete classes because they sometimes express more than one of those definitive markers, or apparently don’t express any of those markers (Markram et al., 2004; Parra, Gulyas, & Miles, 1998). The variability in interneuronal properties could obfuscate results of experimental error (Markram et al.,

2004). Nevertheless, the conventional categorization scheme provides useful (though flawed) heuristics for inferring properties based on distinct cell markers.

The proportions of each group to the total interneuron population in the somatosensory cortex are roughly as follows: 40% for PV+ interneurons; 30% for SST+ interneurons; and 30% for the serotonin receptor 5HTa3R+. The latter group encompasses VIP+ interneurons and, some have argued, is a more parsimonious interneuronal marker than VIP (Bernardo Rudy, Fishell, Lee, & Hjerling-Leffler, 2011).

The subdivisions of interneurons could be further divided into groups based on the presence or absence of other markers, and they can have alternative cell markers (Kubota

& Kawaguchi, 1997; Bernardo Rudy et al., 2011; Taki et al., 2000; The Petilla

Interneuron Nomenclature Group, 2008). Cortical interneurons are often described in very literal terms; a given neuron might be described as “fast-spiking” during electrophysiological studies or “parvalbumin-positive” (PV+) during ICC studies - even though the properties tend to be mutually inclusive (i.e., fast-spiking neurons are almost always PV+, and vice versa).

Cortical interneurons have numerous properties that distinguish them from PNs.

Experiments in the hippocampus show that many of these differences maximize their speed and reliability. Some cortical interneurons, such as PV+ basket cells, also innervate other interneurons of the same type through both electrical and chemical synapses

(Galarreta & Hestrin, 1999, 2002; Gibson, Beierlein, & Connors, 1999; Juszczak &

Swiergiel, 2009). These adaptations are predicted to enable interneurons to synchronize reliably and orchestrate network activity that are important for executive function (for review, see Miles (2000)) (Miles, 2000). The importance of synchronization will be discussed later.

The presence of the µOR and endogenous opioids within interneurons have been reported extensively with ICC in studies of the hippocampus and neocortex (Arvidsson et al., 1995; Suzanne B Bausch & Chavkin, 1995; S. B. Bausch et al., 1995; Drake &

Milner, 1999, 2002, 2006; Taki et al., 2000). These experiments also allow us to predict which categories of neurons are most likely to express the µOR, and further allow us to predict how this relates to the effects of the µOR given the groups that the µOR is found in.

1.10 µOR-expressing cortical interneurons

The rate of expression of the µOR has usually been determined through ICC stains of neocortical or hippocampal slices. However, antibodies generated for the µOR have shortcomings for reasons that I will address later. Another confounding feature is that different methods of localizing the µOR produces slightly different rates of expression.

Finally, the hippocampal formation and the neocortex appear to show some differences

24 distributions of the receptor. While obviously neocortical data is more applicable to experiments of the µOR in the neocortex, a great deal of data about cortical µORs is derived from studies done in the hippocampal formation. These points of disagreement will be commented on throughout this literature review.

Yet despite the limitations of primary antibodies for the µOR, they can still provide a useful starting point for understanding how often (and where) the receptor may be found; in the neocortex, ICC data and an experiment with a fluorescent ligand suggests that it is expressed by only 5-20% of neocortical interneurons (or 3-5% of all neurons) (M. C. Lee, Cahill, Vincent, & Beaudet, 2002; Taki et al., 2000).

Researchers have also delved further to identity of these µOR+ neurons (µOR+ are neurons “positive” for the µOR) based on their coexpression of different markers for interneurons. ICC data on rat neocortical slices show that about 97% of µOR+ produce

GABA, showing that µOR are mostly GABAergic. The majority (92%) of µOR+ had also expressed VIP indicating that the VIP+ pool of interneurons is the biggest contributor to the µOR+ interneurons. Inversely, 70.2% of VIP+ neurons showed presence of the µOR. Meanwhile, 8.2% showed presence of PV; and 2.9% showed presence of SST. Inversely, only 1.1% of PV+ neurons and 0.6% of SST+ neurons showed the presence of µOR puncta suggesting that these neurons rarely express the

µOR – or at least not in their somata; this technique may not be relied upon to identify axonal expression of the µOR due to equipment and practical limitations (Taki et al.,

2000).

But electrophysiology-based experiments (which probably also include sampling bias towards certain interneurons) show much higher estimates than ICC experiments in neocortical slices; one group found a DAMGO response in 2/3rds of the neocortical interneurons that they surveyed with electrophysiology (while none of 18 PNs showed a response) (Witkowski & Szulczyk, 2006).

Another electrophysiology-based experiment that also utilized sc-PCR tested various neocortical neurons for µOR mRNA and probed for a response to DAMGO; they found a DAMGO response (indicated by hyperpolarization and a decrease in input resistance) in 86 interneurons out of 169 recorded neurons. However, they only detected presence of µOR mRNA in 62.5% of the neurons that had responded to DAMGO, indicating that the DAMGO exposure was the more sensitive technique for detecting the receptor. Most of their responders were regular or irregular-spiking and had a high incidence of expression of VIP. They did not appear to sample neurons randomly, for example, only 98 of 360 recordings were PN, despite constituting 80% of the neurons of the cortex. All DAMGO responders tested positive for GAD65 or GAD67 mRNA, and none of the PNs hyperpolarized to the DAMGO application or showed presence of µOR mRNA (Férézou et al., 2007).

Interestingly the Ferezou et al. (2007) noted an exceptionally high rate of

DAMGO responders were found in the Layer 1 of the cortex and conformed to the

NGFCs morphology. This indicates that these neurons likely express high rates of the

µOR (Férézou et al., 2007). This receptor has been observed in the NGFCs of the hippocampus as well (Krook-Magnuson, Luu, Lee, Varga, & Soltesz, 2011; C. J. Price,

Scott, Rusakov, & Capogna, 2008). Taken together, these studies indicate that cortical neurogliaform neurons are likely to express this receptor.

Although hippocampal studies can often be applied to neocortical studies (in that the neuron types are strongly similar), the electrophysiology and ICC experiments in both locations has raised some debate over the localization of the µOR in the PV+ neurons of the neocortex. As I’ll explain next, several groups have observed the µOR in the PV+ interneurons hippocampus, but other research groups have failed to find it in the PV+ interneurons of the neocortex. While this may be pedantic, PV+ interneurons are the largest group of neocortical interneurons and are important to the regulation of neocortical PNs. Thus, their susceptibility to opioidergic activity is not a trivial matter and could be very consequential to neocortical activity.

Expression of the µOR has been reported extensively in the somata of PV+ interneurons of hippocampal cortex, as indicated by immunostaining (Bartos & Elgueta,

2012; Suzanne B Bausch & Chavkin, 1995; Drake & Milner, 1999, 2002, 2006; Stumm et al., 2004; Torres-Reveron et al., 2009). Studies have also shown that some PV+ interneurons of the hippocampus have also found a stereotypical hyperpolarizing response to DAMGO (Glickfeld, Atallah, & Scanziani, 2008; K. R. Svoboda, Adams, &

Lupica, 1999). These experiments have led some to infer that they are also expressed by

PV+ interneurons in the neocortex as well (Volk et al., 2012). However, the best current information suggests that this is rare (Taki et al., 2000). Additionally, in the report I described earlier, there was a lack of reported DAMGO responses in fast-spiking interneurons, even though 38 were tested for it (Férézou et al., 2007). Yet recent work

27 using electrophysiology and optogenetics in the mPFC cortices of mice found that the

µOR regulates the amplitude and duration of inhibitory postsynaptic potentials in PNs with PV+ presynaptic terminals (C. Jiang et al., 2019). An even more-recent study likewise reported similar results in some parts of the OFC (Lau et al., 2020). Taken together, it appears likely that the µOR+ is sometimes found in PV+ neurons in the neocortex - as it is in the hippocampus. However, there does not appear to be perisomatic expression of the µOR in PV+ neurons of the neocortex (as there is in the hippocampus, to a high degree), which makes the frequency of colocalization in PV+ neurons difficult to determine. With the lack of perisomatic expression of the µOR+, there may not be a characteristic hyperpolarizing DAMGO response from recordings made in the somata of fast-spiking (mostly PV+) interneurons of the neocortex.

So, it appears that neocortical expression of the µOR may be found in many VIP+ neurons, neurogliaform neurons, and likely PV+ terminals as well, but ICC perhaps is not the best way to definitively determine its localization, due to nonsomatic expression and the general difficulties of producing antibodies for GPCRs (Jo & Jung, 2016).

Furthermore, efforts using OPRM1-driven (the gene for the µOR) expression of fluorescent proteins may not be a reliable indicator for the same reason (Erbs et al., 2015;

Gardon et al., 2014). To address this problem, efforts are underway to leverage the

Cre/Lox recombinase system to reliably localize this receptor (Bailly et al., 2020). Until then, it appears that electrophysiology is the most sensitive technique for finding this receptor.

While the pattern of expression can be roughly deduced from the combination of

ICC and electrophysiology data shown here, it does not explain what these neurons do in circuits or an intact brain. To understand the effects of the µOR in the neocortex, I must next explain the roles of the PV+, VIP+, and neurogliaform neurons. From that information, we can make more informed predictions on the net result of µOR application on neuronal networks.

1.11 Roles and features of µOR+ interneurons: PV+ and VIP+

The PV+ interneurons of the neocortex are powerfully involved in the regulation of the excitation/inhibition (E/I) balance in this region. Their inhibitory control over PNs is illustrated by data that show nearly every PV+ interneuron synapses with every nearby

PN (Packer & Yuste, 2011). PV+ have strong control over the firing of PNs, and their synchronized network carefully controls the timing of PN neuron spiking (Agmon &

Connors, 1991). These neurons are implicated in executive function of the neocortex

(David A Lewis, 2014), and are hypoactive in psychiatric conditions, such as schizophrenia (Glausier, Fish, & Lewis, 2014; Volk et al., 2002) and autism (McNally &

McCarley, 2016; Rojas & Wilson, 2014). Therefore, there is a strong link between the role of PV+ interneurons and overall cortical function due to their tight regulation of excitatory activity (B. R. Ferguson & Gao, 2018). For review on the role of PV+ interneurons on neuropsychiatric illnesses, see (B. R. Ferguson & Gao, 2018).

VIP+ interneurons that express the µOR have been shown to express nicotinic receptors (Férézou et al., 2007) and often also express ChAT (Taki et al., 2000). This finding is consistent with other research showing that VIP+ interneurons respond to

29 acetylcholine (Fu et al., 2014). On a larger scale, there is a relationship between the nicotine and the endogenous opioid system, namely, that nicotine seems to stimulate release of these endogenous opioids (Mandillo & Kanarek, 2001; Mathieu-Kia, Kellogg,

Butelman, & Kreek, 2002; Pomerleau, 1998).

VIP+ interneurons have been shown to inhibit nearby interneurons(Fu et al.,

2014; Jackson, Ayzenshtat, Karnani, & Yuste, 2016; Pi et al., 2013). Yet this raises an interesting issue; how can opioids, which suppresses VIP+ neurons, lead to disinhibition of PNs while high activity of VIP+ interneurons also lead to disinhibition of the PNs?

This question does not seem to be thoroughly addressed in the literature. One of the explanations may be that, despite the high degree overlap between the µOR and VIP expression, the influence of other cortical neurons, such as PV+ interneurons and neurogliaform cells, are more powerful.

1.12 Roles and features of µOR+ interneurons: NGFCs

The neurogliaform cells (NGFCs) are perhaps one of the more idiosyncratic groups of interneurons. Santiago Ramon y Cajal provided the earliest published description of these neurons by noting their small soma, short but radiating dendrites, and short axons

(Ramon y Cajal, 1995). They are multipolar interneurons that do not express the classic interneuronal markers, but usually do express Reelin and Neuropeptide Y (Jasper, 2012;

Niquille et al., 2018). They do also express the serotonin 5-hydroxytryptamine 3A

(5HT3a) receptor. This receptor is not expressed by SST+ or PV+ neurons but is generally expressed by most all VIP+ neurons and therefore not a specific cell marker for neurogliaform neurons (S. Lee, Hjerling-Leffler, Zagha, Fishell, & Rudy, 2010).

NGFCs have unique electrophysiological characteristics. Classically, they are known for their “late-spiking” firing pattern, i.e., when a sustained depolarizing current is injected, there is a short delay before firing their first spike (Chu, Galarreta, & Hestrin,

2003; Kawaguchi & Kubota, 1997). Their RMP is typically around -60mV and they have low input resistances (Tricoire et al., 2010). Although they can most easily be found in

Layer 1 of the neocortex, they are not exclusively found in Layer 1, and nor are they the only interneurons found in Layer 1 (Férézou et al., 2007; S. Lee et al., 2010; Miyoshi et al., 2010; Olah et al., 2007).

NGFCs are well-known for producing dense axon terminals in extrasynaptic spaces and inhibiting targets through tonic GABA secretion (Overstreet-Wadiche &

McBain, 2015). The effects of NGFCs on their targets has unique consequences.

Consistent with their axon terminals in extrasynaptic space, NGFCs secrete a considerable amount of GABA into the extracellular environment which results in tonic inhibition of nearby neurons through volume transmission, in addition to the faster inhibition mediated by synaptic GABAA receptors (Karayannis et al., 2010; Olah et al.,

2009; Tricoire et al., 2010). Conversely, the tonic secretion of GABA may also desensitize their targets; NGFCs have been observed to produce rapid desensitization

(Karayannis et al., 2010). The lack of spatial and temporal specificity makes the role of

NGFCs difficult to predict (for review, see (Overstreet-Wadiche & McBain, 2015)).

The GABAA receptor is not the only target of NGFCs; NGFCs uniquely secrete

GABA onto GABAB receptors on PNs and interneurons, which can be located both on somata and glutamatergic terminals (Karayannis et al., 2010; Olah et al., 2009;

Christopher J. Price et al., 2005; Tamás, Simon, & Szabadics, 2003). NGFCs are the only neocortical neurons that have been observed to stimulate GABAB receptors with a single

AP - whereas other interneurons usually need trains of APs to activate GABAB receptors on their targets (Christopher J. Price et al., 2005; C. J. Price et al., 2008; Tamás et al.,

2003). Therefore, the recruitment of these neurons can create long-lasting inhibitory postsynaptic potentials in interneurons through the stimulation of GABAB receptors (Olah et al., 2007; Tamás et al., 2003). Consistent with the inhibitory role for µOR and its localization on neurogliaform neurons, DAMGO has been found to decrease activation of the GABAB receptor in neocortical neurons (Lafourcade & Alger, 2008; McQuiston,

2008). This effect appears to be partially mediated by voltage-gated calcium channels – a known target of the µOR, particularly at axon terminals (M. Connor, Schuller, Pintar, &

Christie, 1999; Law, Wong, & Loh, 2000; Rubovitch, Gafni, & Sarne, 2003).

NGFCs frequently interlink with interneurons and PNs through chemical and electrical synapses (X. Jiang et al., 2015; Olah et al., 2009). While NGFCs are known to electrically couple with other NGFCs, there is conflicting evidence that they may couple with non-neurogliaform interneurons as well (Chu et al., 2003; Fukuda, Kosaka, Singer,

& Galuske, 2006; Hestrin & Galarreta, 2005; Anna Simon, Oláh, Molnár, Szabadics, &

Tamás, 2005; Zsiros & Maccaferri, 2005).

NGFCs also show a type of slow integration (perhaps presumptuously) called retroaxonal barrage firing; in the 10s of seconds following a sustained depolarizing current, the neurons will fire many action potentials after the stimulus has already ended

(Overstreet-Wadiche & McBain, 2015; Norimitsu Suzuki, Tang, & Bekkers, 2014).

Documented reports of this type of firing in the neocortex is presently restricted to

NGFCs and PV+ interneurons in slices and in vivo (N. Suzuki & Bekkers, 2010). While a similar phenomenon has been reported in other neurons before, retroaxonal barrage firing in the cortex can be blocked with chemical gap junction inhibitors, but not GABAergic or glutamatergic inhibitors, and therefore implicates gap junctions as transducing this firing

(Overstreet-Wadiche & McBain, 2015; M. E. J. Sheffield, Best, Mensh, Kath, &

Spruston, 2011). To explain the mechanism inherent in its name, some have suggested that axon terminals of NGFCs can become tonically active during the depolarizing current injection and conduct APs through the electrically linked network (M. E. J.

Sheffield et al., 2011; Norimitsu Suzuki et al., 2014). Regardless of the mechanism, some data has shown that DAMGO application reduces retroaxonal barrage firing and provides additional corroborating data that NGFCs express this receptor and reduces this barrage firing phenomenon (Krook-Magnuson et al., 2011; Norimitsu Suzuki et al., 2014).

1.13 µOR disinhibition

The ICC and autoradiography data I have discussed mostly indicate that interneurons express the µOR with some ambiguity about whether PNs do express it. To take a closer look at function of this receptor, researchers have taken electrophysiological recordings from neurons and applied µOR-agonists.

In contrast to autoradiography and ICC which simply inform us of the location of the µOR, electrophysiology gives us a better idea of the effect of the µOR on individual neurons. The electrophysiology data mostly continue this theme and show that there is little doubt that some cortical interneurons react to µOR-agonism with inhibition

(Férézou et al., 2007; Glickfeld et al., 2008; K.-W. Kim et al., 2000; Madison & Nicoll,

1988; Pang & Rose, 1989; K. R. Svoboda et al., 1999; Kurt R Svoboda & Lupica, 1998;

Wimpey & Chavkin, 1991; Witkowski & Szulczyk, 2006; Zieglgansberger, French,

Siggins, & Bloom, 1979). However, there is also some data that suggest that some PNs are directly excited by µOR-agonism (Przewlocki et al., 1999), while others have reported inhibition of PNs with µOR agonism (E. Tanaka & North, 1994). Naturally, one must be cautious about the bias towards reporting positive results, for instance, some groups have used electrophysiology and sc-PCR to examine the µOR’s effect on PNs and observed none (Férézou et al., 2007; Witkowski & Szulczyk, 2006).

Researchers have also investigated whether µOR agonists enhance cortical excitation through disinhibition (that is, creating greater excitatory activity by inhibiting the suppressive effects of GABAergic interneurons). Several studies indicate that excitation in PNs is secondary only through inhibiting the nearby GABAergic neurons

(Homayoun & Moghaddam, 2007; C. Jiang et al., 2019; H. K. Lee et al., 1980; Madison

& Nicoll, 1988; Pang & Rose, 1989; Zieglgansberger et al., 1979). However, several other studies suggest that the µOR upregulates activity of PNs, thereby directly activating them (Przewlocki et al., 1999; Rola, Jarkiewicz, & Szulczyk, 2008; Schmidt et al., 2003;

Witkowski & Szulczyk, 2006). This matter remains unresolved.

1.14 Excitation/inhibition balance

If the µOR inhibits interneurons and disinhibits PNs – why would that matter? A continuous loop of excitatory-only neurons could never organize and would likely be nearly impossible to properly regulate, especially given the afferent pathways that tend to

34 converge on any neuroanatomical structure and may compete for control. GABAergic interneurons fulfill an important purpose: properly regulating the output of brain structures and enabling flexibility (Isaacson & Scanziani, 2011; Klausberger et al., 2003).

Therefore, persistent conditions that reduce the excitability of GABAergic interneurons can in turn result in neurological defects (B. R. Ferguson & Gao, 2018).

The excitation of cortical networks is kept in careful balance by inhibition, and vice versa. This balance must be maintained for the coordinated activity to be sustained during cognitive tasks (Yizhar et al., 2011). Several psychiatric disorders, including; autism (Bozzi, Provenzano, & Casarosa, 2018; Nelson & Valakh, 2015; Orekhova et al.,

2007; Rubenstein & Merzenich, 2003), Rett syndrome (V. S. Dani et al., 2005), schizophrenia (Fazzari et al., 2010; D. A. Lewis, Hashimoto, & Volk, 2005; Yizhar et al.,

2011) and epilepsy (Bozzi et al., 2018; Engel, 1996) are believed to result from an improper excitation/inhibition E/I balance. These syndromes may appear from abnormalities during interneuronal migration, dysregulated postnatal apoptosis (Wong et al., 2018), synaptogenesis (Fazzari et al., 2010; J.-M. Yang et al., 2013), neurotransmitter release (Bozzi et al., 2018; McNally & McCarley, 2016), or temporary states from drug use (Berke, 2009; Premoli et al., 2017).

1.15 Cognitive function and band oscillations

The invention of the electroencephalogram allowed researchers to record the mass activity of neurons firing synchronously due to their effect on the electrical field around the subject’s head. These band oscillations (BOs) (Berger, 1929; Steriade, 2001).

Eventually, researchers were able to induce large-scale neuronal activity with transcranial

35 magnetic stimulation (Barker, Jalinous, & Freeston, 1985). These inventions allow us to measure brain activity and a limited means to affect that activity. They were also some of the earliest methods to understand how neurons in specific regions collaborate to fulfill a collective function.

The activity of neurons produces a measurable electric field that, when graphed against time, creates a pattern of high-frequency oscillations. Known mammalian oscillatory frequency can occur between 0.5 MHz and 500 MHz. While a given structure can have a prevailing frequency depending on their brain state, task and neuroanatomical structure, a given structure usually has a composite of multiple BOs contributing to the net oscillation (Steriade, 2001). Slower frequencies of oscillations (which are usually more widespread) have a strong modulatory role, whereas faster band oscillations are more confined in space and relatively easily disrupted/modulated (Buzsáki & Draguhn,

2004; Buzsáki, Geisler, Henze, & Wang, 2004; Csicsvari, Jamieson, Wise, & Buzsáki,

2003; Sirota, Csicsvari, Buhl, & Buzsáki, 2003).

The Gamma Band Oscillations (GBOs) are a particularly relevant band oscillation to the study of cortex. These GBOs are believed to contribute to proper memory, attention, and sensory binding (which links senses together) (Bragin et al., 1995;

Csicsvari et al., 2003; Fell et al., 2001; Fries, Reynolds, Rorie, & Desimone, 2001; Gray,

1994; Llinás, Ribary, Contreras, & Pedroarena, 1998; Singer, 1993; Varela, Lachaux,

Rodriguez, & Martinerie, 2001). Although the features of these bands may vary by region and frequency, the GBOs of the hippocampus and neocortex appear to involve both the

PNs and the cortical interneurons interacting to produce BOs as detected through an EEG

(Blatow et al., 2003; Csicsvari, Hirase, Czurkó, Mamiya, & Buzsáki, 1999; Csicsvari et al., 2003; Suffczynski, Crone, & Franaszczuk, 2014; Traub, Jefferys, & Whittington,

1997; M. A. Whittington, Traub, & Jefferys, 1995).

A primary role for cortical interneurons is to provide synchronization inhibition for the PNs, and permit PNs to then synchronize their activity in the transient gaps of inhibition (Blatow et al., 2003; Cobb, Buhl, Halasy, Paulsen, & Somogyi, 1995;

Hartwich, Pollak, & Klausberger, 2009; Lytton & Sejnowski, 1991; M. A. Whittington et al., 1995). Perturbations that disrupt the balance of excitation and inhibition can lead to deleterious consequences to the cognitive ability of mammals. Disorganized BOs is one of the features of numerous mental illnesses, such as schizophrenia (Cho, Konecky, &

Carter, 2006; Hirano et al., 2015; David A. Lewis et al., 2012; Minzenberg et al., 2010), which also coincides with abnormalities that downregulate activity of their inhibitory neurons (Lodge, Behrens, & Grace, 2009). The relationship between BOs and their symptoms is not presently clear but may be related to their disrupted sensory perception and reduced cognitive ability (Green & Nuechterlein, 1999). However strong and persistent GBOs are not always beneficial, for instance these GBOs are particularly strong in people with chronic pain and tend to be weaker after therapy (Ebrahimian et al.,

2018). The use of opioid drugs disrupts cortical BOs (Gulyás et al., 2010; M. Whittington et al., 1998; Y. Zhang et al., 2019).

These band oscillations are decent correlates of the task the subject is handling, but they do not lend a mechanistic understanding of how the µOR is affecting cortical processes. For instance, while the µOR is shown to increase net cortical excitation, EEGs

37 cannot immediately relate that to glutamatergic activity. To accomplish that, I will refer to a different kind of oscillation: calcium oscillations. Whereas band BOs are phenomena produced by synchronous depolarizations of massive amounts of neurons in an intact brain, calcium oscillations are smaller-scale events to visualize glutamatergic activity that can be seen in neuronal cultures.

1.16 Calcium transients and excitability

Data from microelectrode array experiments in dissociated neurons demonstrate that interventions that increase excitability of the neurons leads to global coupling and uniform activity (Nakanishi & Kukita, 1998; Penn, Segal, & Moses, 2016). But to more easily monitor network activity in many culture neurons at once, calcium-imaging has been used to visualize excitation as synchronous calcium oscillations (calcium transients)

(Bacci, Verderio, Pravettoni, & Matteoli, 1999; Inglefield & Shafer, 2000; Robinson et al., 1993). This produces a sporadic “flickering” effect in cultured neurons from the transient binding of calcium the calcium-indicator dye loaded into neurons. These flickering events are often synchronized in multiple neurons at once, but this is not always the case. Several groups have done experiments to determine the cause of these calcium oscillations, and these oscillations relate to neuronal activity.

To investigate the cause of these calcium oscillations, various groups have simultaneously recorded electrophysiology and calcium-dye fluorescence to correlate the two. They found that these oscillations correspond to bursts APs on top of a feature that resembled an EPSPs (Bacci et al., 1999; Muramoto, Ichikawa, Kawahara, Kobayashi, &

Kuroda, 1993; T. H. Murphy, Blatter, Wier, & Baraban, 1992; Przewlocki et al., 1999;

Shen, Piser, Seybold, & Thayer, 1996; Sombati & Delorenzo, 1995).

The oscillations can be prevented by blocking depolarizations with TTX, or by blocking all glutamate receptors (Dravid & Murray, 2004; Lo, Fallert, Piser, & Thayer,

1992; T. H. Murphy et al., 1992; Nuñez, Sanchez, Fonteriz, & Garcia‐Sancho, 1996;

Przewlocki et al., 1999). Blocking voltage-gated calcium channels generally does little to weaken the oscillations, which indicates that these channels are not a major route for calcium entry in this phenomenon (Cao et al., 2014; Dravid & Murray, 2004). In some instances, inhibition of Phospholipase C abolished the waves indicating that calcium- induced calcium release was a large component of this effect (Dravid & Murray, 2004).

However, groups are split on whether the major source of calcium to activate PLC is through NMDA receptors that conduct extracellular Ca2+ into the cells, or from the internal stores by an mGluR coupled to a Gq protein (Dravid & Murray, 2004; Przewlocki et al., 1999).

The identity of the glutamate receptor responsible for the calcium oscillations, and the contribution of calcium-induced calcium release has become a point of disagreement for researchers. This disagreement at least partially stems from the [Mg] used in the

ACSF, with some groups preferring physiological concentrations while others use no-Mg

ACSF to amplify and synchronize the oscillations; some researchers have tested these oscillations in presence of low [Mg2+] and have proposed that these calcium oscillations are due to the calcium-conducting properties of the NMDA receptors (Canepari, Bove,

Maeda, Cappello, & Kawana, 1997; Hemstapat, Smith, & Monteith, 2004; Inglefield &

Shafer, 2000; Lawrie, Graham, Thorn, Gallacher, & Burgoyne, 1993; Nuñez et al., 1996;

Penn et al., 2016; Shen et al., 1996; T. Tanaka, Saito, & Matsuki, 1996; X.-s. Wang &

Gruenstein, 1997), but others have tested these oscillations at physiological [Mg2+] and argued that calcium oscillations are mostly caused by calcium-conducting AMPA receptors and Group I metabotropic Glutamate receptors (mGluRs) coupled to calcium- induced calcium release (Bacci et al., 1999; Cao et al., 2014; Dravid & Murray, 2004;

Flint, Dammerman, & Kriegstein, 1999).

Somewhat confusingly, “calcium oscillation” and “calcium wave” are frequently used interchangeably through the literature I have cited here; a distinction should first be made regarding the durations of these events and the cell types that they occur in. One well-known calcium-imaging phenomena are the slow calcium waves which are usually observed in non-excitable cells (e.g., glia, HEK293 cells), though they can be relayed to neurons as well (Berridge, 1993; Kawabata et al., 1996). These slow calcium waves typically have long rise and decline phases; they can take 10s of seconds to rise and can last for several minutes (J. W. Dani, Chernjavsky, & Smith, 1992). Calcium waves are believed to be a phenomenon that results from gap junctions and the slow spread of IP3

(which can stimulate the release of Ca2+ from internal stores) through them (Balaji et al.,

2017; Leybaert & Sanderson, 2012). However, calcium waves can be instigated by glutamatergic activity, thus there may be a causal relationship between the 2 phenomena

(Zur Nieden & Deitmer, 2005). Calcium oscillations are on the shorter scale; they consist of calcium spikes that can rise in less than a second, and then taper off back to baseline over the course of a few seconds. These shorter oscillations have been found previously

40 in cultured neocortical neurons and correspond to activation of postsynaptic glutamate receptors through glutamate release. Though IP3 may very well be involved in these oscillations, there was no evidence that it was spreading through gap junctions in neurons

(Dravid & Murray, 2004; T. H. Murphy et al., 1992).

These calcium oscillations are linked to development and gene expression in neurons (Dolmetsch, Xu, & Lewis, 1998; Spitzer, Olson, & Gu, 1995). Oscillations may

2+ also have a neuroprotective effect against rising [Ca ]in that follow neurological trauma

(possibly by dismounting NMDA receptors from the actin cytoskeleton) (Geddes-Klein,

Serbest, Mesfin, Cohen, & Meaney, 2006).

This model has been used previously to study activation of luteinizing hormone- releasing neurons (Terasawa, Schanhofer, Keen, & Luchansky, 1999), and their response to estrogen (Abe, Keen, & Terasawa, 2008); the effects of the kappa opioid receptor on glutamate release in spinal cord cultures (Kelamangalath, Dravid, George, Aldrich, &

Murray, 2011); the excitatory effects of nicotine on neocortical neuron cultures (J. Wang et al., 2016); cannabinoid regulation of glutamate release in the hippocampus (Shen et al.,

1996); HIV-1 envelope protein and cytotoxicity (Lo et al., 1992); environmental toxins on neurons (Inglefield & Shafer, 2000); and dopamine receptor 4 on hippocampal cultures (Y.-L. Wang et al., 2017). Efforts are also underway to develop this model to study seizure-related drugs; the hypersynchrony observed under certain conditions

(including low Mg2+) resemble that of epileptiform patterns (Cao et al., 2015; Pacico &

Mingorance-Le Meur, 2014).

While the effects of the µOR on network activity is one thing, the receptor’s mechanism inside the neuron is another. Due to the boom in interest in the µOR, there is much known about this receptor’s intracellular cascade and target ion channels. However, there is a great deal that is not known because of the µOR’s vast complexity. In order to explain the unknown features of this receptor, I will provide more details about the receptor, its ligands, and its intracellular signaling pathway.

1.17 Opioid receptors

There are 3 known types of opioid receptors the Δ Opioid Receptor (DOR), the µ

Opioid Receptor (µOR), the Κ Opioid Receptor (KOR) (Al-Hasani & Bruchas, 2011). All these ORs are Gi-coupled Protein Receptors. Several other receptors were initially proposed as ORs (e.g., sigma and opioid-like receptor 1) but were subsequently ruled out as classical opioid receptors due to their wide binding profile to non-opioid neurotransmitters (Sauriyal, Jaggi, & Singh, 2011; Stein, 2016). But these receptors may still play a modulatory role in the effects of the 3 classical opioid receptors (F. J. Kim et al., 2010; Kobayashi, Ikeda, Ichikawa, Togashi, & Kumanishi, 1996).

Opioid receptors can be found in both peripheral and central locations; besides the brain, they can also be found on immune cells, neuroendocrine organs and some ectodermal cells. Opioid receptors are subjected to alternative splicing posttranslational modifications to introduce variety in receptor properties. Generally, both the DOR and

µOR can be found throughout most of the neocortex. The KOR can also be found in the deep layers of the neocortex but is less concentrated than the other ORs (Hiller & Fan,

1996; Merrer, Becker, Befort, & Kieffer, 2009).

Opioid receptors are also believed to form homodimers or heterodimers with other ORs in vitro and in vivo, though this view is not universally accepted (Cvejic &

Devi, 1997; Y.-X. Pan, Bolan, & Pasternak, 2002; Prinster, Hague, & Hall, 2005; Rios,

Jordan, Gomes, & Devi, 2001; D. Wang, Sun, Bohn, & Sadée, 2005). While all the implications of heterodimerization of ORs is not immediately clear, this possibility introduces more variability into the response of these receptors to agonist binding and the receptors may therefore mix their signaling pathways.

The KOR earned a lot of attention as a potential target for therapy since it was an opioid receptor that is not associated with addiction. Unfortunately, the KOR agonists are not clinically useful for pain relief, though their antagonists might be. The KOR tends to have the opposite effects of the µOR and DOR. While µOR-acting drugs are euphoric, hedonic, and addictive; KOR agonists are usually just unpleasant (dysphoric), hallucinogenic, and dissociative - and therefore not drugs of abuse (Chavkin, 2011;

Shippenberg, 2009). But this aversion and dysphoric is not always seen in KOR-acting drugs (Butelman & Kreek, 2015). These receptors can be found in the PFC, where they inhibit dopamine release from VTA afferents there (Tejeda et al., 2013). However, there is still interest in developing KOR agonists to treat itchiness and opioid abuse (Chavkin,

2011), or KOR antagonists to treat chronic stress and anxiety (Chavkin, 2011; Knoll &

Carlezon, 2010).

The DOR has attracted interest as a potential target for therapy and specific peptidergic drugs have developed to study this receptor (Gavériaux-Ruff & Kieffer,

2011). However, this receptor has not benefitted from the boom in interest that µOR-

43 acting drugs have brought to that receptor. DOR knockout mice show relatively subtle changes in pain response (Martin, Matifas, Maldonado, & Kieffer, 2003). Interestingly, whereas the µOR is useful for acute pain and acute mood problems, the DOR may be more useful for chronic pain and depression (Filliol et al., 2000; Gavériaux-Ruff &

Kieffer, 2011; Nadal, Baños, Kieffer, & Maldonado, 2006; Scherrer et al., 2006;

Valentino & Volkow, 2018). However, they tend to also have convulsive effects and research has progressed only slowly and cautiously (Pradhan, Befort, Nozaki, Gavériaux-

Ruff, & Kieffer, 2011; Saitoh & Yamada, 2012).

1.18 µ opioid receptor

The µOR is the principal site of exogenous opioidergic drugs (Valentino & Volkow,

2018). The µOR has benefitted greatly from the surge in interest of opioid receptors.

To discover the role of the µOR in adults, scientists performed gene knockouts in mice of the µOR and tested their analgesia, addictive tendencies, motivation, and maternal attachment. Mice that lack the µOR are resistant to morphine addiction and analgesia (Matthes et al., 1996; Merrer et al., 2009). µOR knockout mice also suffer from reduced motivation to eat and reduced maternal attachment (Kas et al., 2004; Papaleo,

Kieffer, Tabarin, & Contarino, 2007). Therefore, this specific receptor appears to be the primary OR in mediating the effects of exogenous opioidergic drugs, as well as the endogenous means for motivation and pain suppression.

The µOR can be found throughout much of the neocortex, except in visual regions at the posterior side of the brain. It can be found in a roughly caudal-to-rostral gradient and, consequently, is more concentrated in frontal regions of the brain. Its 44 staining is relatively dense in Layer 3, followed by Layers 1 and 2 (Hiller & Fan, 1996).

The µOR is expressed at birth, but expression gradually increases for the first several postnatal weeks (Clendeninn, Petraitis, & Simon, 1976; Coyle & Pert, 1976; Garcin &

Coyle, 1977). I have previously alluded to the difficulty of developing antibodies to

GPCRs and the µOR, which I will now explain in more detail.

One of the most remarkably-paradoxical features of the µOR, is that this expression of this receptor is driven by only one gene – the OPRM1 (oprm1 in rodents, according to IUPHAR nomenclature) gene – and yet its effects and regulation introduce a spectrum of features that make it difficult to study (Classification, 2020). This receptor is extremely variable due to splice variants, RNA editing, post-translational modifications, presumed multiple agonist binding sites, functional selectivity, and other characteristics that introduce variability into their structure and binding profile (Groer et al., 2007;

Gavril W. Pasternak & Pan, 2013). Its property of functional selectivity (also called biased agonism) is a particularly problematic aspect to understand, because this receptor can respond differently to different agonists.

This variability is on full display with opioid-acting drugs causing vastly different opioid effects on different patients (Cherny et al., 2001) and rodent strains (Mogil, 1999;

Reith, Sershen, Vadasz, & Lajtha, 1981). In other instances, µOR antagonists are unable to reverse all µOR agonist effects in patients when they are taken concurrently (Andoh et al., 2008). The variable states of the receptor may be the factor that allows patients to cycle drugs to mitigate the effects of opioid tolerance to specific µOR-acting drugs

(Cherny et al., 2001; Moulin, Ling, & Pasternak, 1988).

Localization of this receptor has primarily been driven by autoradiography and

ICC, but both techniques have shortcomings and points of disagreement due to the nature of this receptor. The µOR is one of the few receptors known to have distinguishable agonist and antagonist conformations; radiolabeled antagonists tend to label the receptor more widely than agonists, and therefore antagonist conformation usually assumed to be the more stable conformation (Bruno Cauli et al., 1997). These conformations may be strongly influenced by presence by ions, including Na+ and Mg2+ (G. W. Pasternak,

Snowman, & Snyder, 1975).

Antibodies for the µOR have primarily been limited to C-terminus epitopes for the µOR because of the receptor’s susceptibility to glycosylation on the external side

(Gavril W. Pasternak & Pan, 2013). However, the C-terminus of the µOR is also the primary source of variation between splice forms, which prevents primary antibodies from recognizing all forms of the µOR (Abbadie, Pan, Drake, & Pasternak, 2000; Gavril

W. Pasternak & Pan, 2011). For a review of these characteristics, see (Gavril W.

Pasternak & Pan, 2013).

There are several known splice variants of the OPRM1/oprm1 gene, which have unique regional distributions through the brains of rodents (Abbadie et al., 2000; Y. X.

Pan et al., 1999), Studies have found OPRM1/oprm1 to be well-conserved in rodents and humans, which at least lends itself to study in rodent models (J. Xu, Xu, Rossi, Pasternak,

& Pan, 2011).

The µOR shows the properties of ‘biased agonism’ or ‘functional selectivity’ and makes the µOR’s potential effects more variable. This property allows the receptor to 46 respond differently to different agonists. This has implications for trafficking, desensitization, and intracellular mechanisms of the receptor which will be discussed later (von Zastrow, 2010). While precise terms (e.g. “full agonist”) has traditionally been applied to particular ligands, the influence of functional selectivity enables receptors to have much more varied responses than this general terminology permits (Raehal, Schmid,

Groer, & Bohn, 2011).

1.19 Endogenous opioid ligands

Endogenous opioids are all peptide neurotransmitters that are derived from pre- propeptides. These pre-propeptides are processed through posttranslational modifications, splicing, and proteolytic cleavage to yield several different peptide neurotransmitters from the same family (McLaughlin, 2013). Although there may be dozens of individual endogenous opioids, they can be grouped into just a few categories based on a common epitope. Enkephalins (µOR and DOR) are mostly derived from preproenkephalin-A gene, dynorphins (KOR) are derived from the Dynorphin-A gene, and β-endorphins (µOR and

DOR) are derived from the prohormone proopiomelanocortin (POMC) (McLaughlin,

2013; McLaughlin & Zagon, 2013; Spampinato, Bedini, & Baiula, 2013; Taki et al.,

2000). However, POMC is not transcribed in the neocortex and therefore will not be discussed here further (Gee, Chen, Roberts, Thompson, & Watson, 1983). There is also a fourth group of opioid peptides – the endomorphins which are products of a presently- unidentified pre-propeptide. Endomorphins 1 and 2 (µOR) are especially powerful and specific agonists for the µOR. These peptides are a product of an unknown precursor

(Terskiy et al., 2007).

The Pre-proenkephalin A gene produces 6 peptides from the 2 major varieties of enkephalin (Enk); Met-enkephalin (Met-Enk) and Leu-Enkephalin (Leu-Enk). This gene produces multiple forms of Met-Enk, including an octapeptide and heptapeptide, but only one form of Leu-Enk (McLaughlin, 2013). Meanwhile Pre-prodynorphin A transcripts, which mostly produces KOR-acting peptides, also contains a sequence for Leu-Enk on a decapeptide. Although this Leu-Enk can stimulate the µOR, other dynorphins products do not (Spampinato et al., 2013). Pre-prodynorphin and Pre-proenkephalin proteins have been observed extensively in both interneurons and some PNs, but mostly interneurons

(Alvarez‐Bolado, Fairén, Douglass, & Naranjo, 1990; Fallon & Leslie, 1986; Olenik &

Meyer, 1997). Interestingly, these pre-propeptides are often produced by the same neurons that express the µOR, suggesting that these are autocrine and paracrine regulators as well (Férézou et al., 2007; Taki et al., 2000).

Enkephalins are highly expressed in the neocortex by several types of neurons, though localizing it to specific neurons has been a problem despite its high expression there. As pointed out by Férézou et al. (2007), the concentration of enkephalins in the neocortex is 4 times more than the concentration of VIP – which is considered to be a major marker for many interneurons (Crawley, 1985; Lindberg, Smythe, & Dahl, 1979;

Rossier et al., 1977). Yet experiments generally show relatively restricted expression of the precursor to µOR+ neurons; ICC experiments in the neocortex show a very high degree (80-90%) of overlap between expression of the µOR and pre-proenkephalin within the same interneurons (Taki et al., 2000). On the other hand, experiments that used colchicine treatment – as well as Ferezou et al.’s own sc-PCR data – show that PNs do

48 often have low levels of the neurotransmitter which may not be detectable with the ICC experiments used extensively (F, van der Kooy, & Bloom, 1984; Férézou et al., 2007).

Genetic deletion of pre-proenkephalin A in mice produces offspring that are surprisingly healthy and seemingly normal – they even have normal stress-induced analgesia. However, they are also hyper-aggressive and anxious (König et al., 1996).

These knockout mice, interestingly, are also resistant to the effects of chronic stress and depression (Bilkei-Gorzo, Michel, Noble, Roques, & Zimmer, 2007; Melo, Drews,

Zimmer, & Bilkei-Gorzo, 2014).

Endomorphins 1 and 2 (EM-1 and EM-2) are very highly selective and have high affinity for the µOR, but less is known about those neurotransmitters (Hackler, Zadina,

Ge, & Kastin, 1997; Martin-Schild, Gerall, Kastin, & Zadina, 1999; Mizoguchi,

Sakurada, & Sakurada, 2013; Terskiy et al., 2007; J. Zadina, Kastin, Ge, & Hackler,

1994; J. E. Zadina, Hackler, Ge, & Kastin, 1997). They are products of an unidentified precursor protein which unfortunately has made it difficult to localize to particular neurons (Terskiy et al., 2007). EM-1 can be found abundantly in the cerebral cortex, while EM-2 is reportedly rarer in the cortex according to ICC data (Gu et al., 2017;

Martin-Schild et al., 1999; Schreff, Schulz, Wiborny, & Höllt, 1998). However, both endomorphins have been reported in human cortical brain lysates in reportedly high concentrations, though a direct comparison with other neurotransmitters is difficult to make with these limited data (Hackler et al., 1997).

Therefore endomorphins 1&2 and enkephalins are the endogenous neurotransmitters for the (neocortical) µOR. While they have a spectrum of selectivity 49 and affinity, and given this wide variety of µOR agonists, efforts were devoted to selecting a highly selective and high-affinity agonist for the µOR.

1.20 DAMGO

To study the physiological effects of the µOR, obviously, an agonist is needed. While there are many opioid peptides, they have varying levels of specificity and affinity for the

µOR. The endogenous forms of enkephalins have off-target binding to other ORs (mostly to the DOR) (Merrer et al., 2009; Takahashi, 2016). Considering that endogenous opioids already consist as many different peptides, efforts were made to produce a specific, stable, high-affinity µOR agonist that was structurally derived from the endogenous peptide and therefore had similar receptor effects as endogenous forms of enkephalins

(Handa et al., 1981).

The synthetic peptide [D-Ala2,N-MePhe4,Gly-ol5]enkephalin (DAMGO) was discovered to be a selective agonist for the µOR and is used extensively for µOR-related research ranging from brain slices (Bushell et al., 2002; Chiou & Huang, 1999; Faber &

Sah, 2004; Férézou et al., 2007; Finnegan, Chen, & Pan, 2006; McQuiston, 2008; Qu et al., 2015; K. R. Svoboda et al., 1999; Witkowski & Szulczyk, 2006), to in vivo testing

(Castro & Berridge, 2014; Helmstetter et al., 1998; Kekesi et al., 2011; Ma et al., 2020;

Newman, Pascal, Sadeghian, & Baldo, 2013; Selleck et al., 2015), to cell culture (Patel et al., 2006; Piros, Prather, Law, Evans, & Hales, 1996; Piros et al., 1995; Przewlocki et al.,

1999; Rubovitch et al., 2003; E. Tanaka & North, 1994). The currently favored pharmacological antagonist for the µOR is CTAP D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-

Thr-NH2 (CTAP) for its specificity and low promiscuity (Carrero, Kaigler, Hartshorn,

Fadel, & Wilson, 2019; B. Chieng, Connor, & Christie, 1996; Falk et al., 2020; Ma et al.,

2020).

1.21 Opioid receptor intracellular cascades

The µOR has a complex intracellular cascade that can terminate with regulation ion channel proteins, as well as transcription factors. Unfortunately, due to the complexity of this receptor, the details for its mechanism in regulating potassium channels are not completely well-understood. However, experiments done in heterologous systems and neurons show some key components that will be discussed here. I will outline its basic intracellular cascade in this section. But later, I explain its interaction with specific ion channels.

The superfamily of opioid receptors is composed of seven transmembrane spanning (7TM) segments and signal through Gi/Gs signaling pathways. They are categorized as class A (Rhodopsin) Gi/Go G-protein coupled receptors with an extracellular N-terminal and an intracellular C-terminal end (Katritch, Cherezov, &

Stevens, 2013). They are usually coupled to pertussis toxin-sensitive heterotrimeric Gi/Go proteins (Law et al., 2000). Binding of agonists results in activation of both α and βγ subunits of the G-protein complex, which can interact with separate signaling pathways and can itself deactivate some ion channels through a direct interaction. The Gβγ subunits interacts with calcium channels directly (more about ion channels later), while the Gαi subunit typically inhibits adenylyl cyclase and suppresses cAMP formation (Zamponi &

Snutch, 2002). Diminished levels of cAMP can downregulate PKA and lead to decreased levels of phospho-CREB, which then causes upregulation or downregulation of different

51 genes (Nonnemacher et al., 2017; Purves, 2008). While this may cause acute effects, the most consequential gene regulation changes may occur when cellular tolerance creates chronic hyperactivation of PKA and cAMP, which then causes gene expression changes

(Bilecki & Przewlocki, 2000; Noble & Cox, 1996; Terwilliger, Beitner-Johnson,

Sevarino, Crain, & Nestler, 1991). However, the route to ion channel regulation is more convoluted.

The downregulated PKA releases Phospholipase AA (PLA2) from inhibition (Zor

& Reiss, 1991). Activated PLA2 metabolizes the esterified form of arachidonic acid, found in cell membranes, and the metabolized product can participate in at least 3 pathways; the cyclooxygenase, lipoxygenase and expoxygenase (cytochrome p450 pathway) (Piomelli & Greengard, 1990; Vaughan, Ingram, Connor, & Christie, 1997).

The cyclooxygenase pathway leads to formation of prostaglandins, prostacyclin, and thromboxane A2. The lipoxygenase pathway results in formation of 12-lipoxygenase products. Inhibition of the cyclooxygenase and 5-lipoxygenase pathways potentiated the

µOR’s neurophysiological inhibition of hippocampal neurons, while inhibition of 12- lipoxygenase blocked much of it (Vaughan et al., 1997). These data suggest that the 12- lipoxygenase pathway likely mediates some of the µOR’s inhibitory effect in hippocampal interneurons. The authors also found that inhibition of PLA2 likewise blocked the effect of DAMGO and it appears that PLA2 was the enzyme that was cleaving arachidonic acid (Vaughan et al., 1997). This pathway likely also contributes to the synergistic effects of NSAIDs (which block the cyclooxygenase pathway) and opioid drugs by directing AA metabolites towards the 12-oxygenase pathway (MacDonald J

Christie, Connor, Vaughan, Ingram, & Bagley, 2000; M. J. Christie, Vaughan, & Ingram,

1999).

There is also strong evidence that Gβγ subunits from Gi receptors can activate phospholipase C (PLCβ). This mobilizes intracellular calcium release and increase

Protein Kinase C activity and IP3 levels, even though this mechanism is typically associated with Gq receptors rather than Gi receptors (C. L. Huang, Feng, & Hilgemann,

1998; Law et al., 2000; Rubovitch et al., 2003; Smart & Lambert, 1996; Smart, Smith, &

Lambert, 1994; Tsu, Chan, & Wong, 1995; Zimprich, Simon, & Höllt, 1995). While this finding may be somewhat controversial, a group did find that PKC (and PKA) may mediate some of the µOR’s inhibitory effects in neocortical interneurons (Witkowski &

Szulczyk, 2006). Other research has shown that DAMGO can upregulate the activity of

PLCβ, and that PLCβ is necessary to mediate the antinociceptive effects of morphine in mice (W. Xie et al., 1999). Consistent with this, PLCβ I implicated in increasing the activity of PLA2, which I previously mentioned cleaved arachidonic acid to mediate the

µOR’s regulation of voltage-sensitive K channels in hippocampal neurons (Zor & Reiss,

1991). Additionally, µOR activation of Gαq proteins has also been reported as mediating some of the supraspinal analgesic effects of opioids as well (Sánchez-Blázquez, Gómez-

Serranillos, & Garzón, 2001).

The µOR may have more direct and indirect interactions with many receptors and ion channels; it co-immunoprecipitates with the NR1 subunit of NMDA receptors

(Rodríguez-Muñoz, Sánchez-Blázquez, Vicente-Sánchez, Berrocoso, & Garzón, 2012); the µOR, intracellular calcium, and NMDA receptors may work collaboratively to alter

53 gene transcription (Chartoff & Connery, 2014; Hyman, Malenka, & Nestler, 2006;

Konradi, Cole, Heckers, & Hyman, 1994). NMDA receptor antagonism has also been shown to reverse the effects of morphine tolerance, suggesting some upregulation of

NMDARs occurs, though it may not be clearly demonstrated to occur as a direct regulation – as opposed to a disinhibitory effect and increase in glutamatergic outflow

(Trujillo & Akil, 1991).

On the topic of biased agonism and functional selectivity, researchers have observed that different agonists can differentially regulate intracellular signaling pathways. Application of the label-free integrative pharmacology on-target (iPOT) approach demonstrated that dozens of µOR agonists could bias the receptor towards different pathways (Morse, Sun, Tran, Levenson, & Fang, 2013). This has also been demonstrated on a smaller scale and appears to result from biased activation of different

Gαi/o proteins associated with the receptor (S. Allouche, Polastron, Hasbi, Homburger, &

Jauzac, 1999; Massotte, Brillet, Kieffer, & Milligan, 2002; Raehal et al., 2011). There are also many studies that have found that different agonists can induce phosphorylation of the µOR at several different sites (Grecksch et al., 2011; McPherson et al., 2010; Rivero et al., 2012; Y. Yu et al., 1997).

Naturally, this raises the issue of how these different effects relate to the intracellular signaling cascade, and eventually, to different neurophysiological effects of different ligands. Unfortunately, limitations to understanding this phenomenon are imposed by current technology, different cell types that express the µOR, and the dozens

54 of exogenous and endogenous ligands available to screen. Some of these repercussions are discussed later.

1.22 Opioid desensitization

In addition to the addictive potential of opioids, the µOR is known to desensitize and accumulate tolerance to follow-up exposure to agonists. This reduces the effect of opioid agonists over the course of sequential doses and can ultimately require higher doses to achieve the same effect. Adaptations that occur over longer-term exposure are usually classified as tolerance, while shorter term adaptive responses to acute effects are classified as desensitization. Desensitization can happen within minutes (Stéphane

Allouche, Noble, & Marie, 2014). Tolerance can develop against all actions of opioids, but it accumulates at different rates for different aspects (Blanchet & Lüscher, 2002;

Gavril W. Pasternak & Pan, 2013). Similarly, different biochemical effects of ORs can desensitize at different times (Stéphane Allouche et al., 2014).

Studies using in vitro models to research the µOR report a wide variety of desensitization properties. With DAMGO, desensitization is sometimes visible by 5 minutes of exposure with observable reductions in DAMGO effects, but usually takes 10-

20 minutes – or even longer than that (Koch et al., 2005; Law et al., 2000; Raehal et al.,

2011). This varies system-to-system, and agonist-to-agonist. Morphine is highly desensitizing, while DAMGO is less desensitizing; co-applying DAMGO with morphine can reduce morphine-alone desensitization, suggesting that this principle may have clinical applications (Finn & Whistler, 2001; Ma et al., 2020). Different pathways can also desensitize along different timelines, for instance; the presynaptic effects

(interactions with voltage-sensitize calcium channels) is stable for at least 15 minutes after exposure to DAMGO, but the postsynaptic effects are significantly reduced at that timepoint (Blanchet & Lüscher, 2002).

Many explanations and mechanisms for the development of desensitization (and ultimately tolerance) revolve around the trafficking and regulation of the µOR after it has bound a ligand, therefore the endocytic cycling process will be discussed next.

1.23 µOR endocytic cycling

Opioid receptors, like many GPCRs, can be rapidly endocytosed by clathrin-coated pits.

β-arrestin, which binds to activated ORs, facilitates this process (S. S. Ferguson et al.,

1996; Goodman et al., 1996). Endocytosis can also be facilitated by phosphorylation by

GRKs (Kovoor, Celver, Abdryashitov, Chavkin, & Gurevich, 1999). Consistent with the functional selectivity of the µOR, this process is remarkably influenced by the ligand; morphine usually does not lead to internalization of the receptor, but DAMGO (including in cultured rat neocortical neurons), etorphine, methadone, and fentanyl all do rapidly

(Finn & Whistler, 2001; L. He, Fong, Von Zastrow, & Whistler, 2002; Keith et al., 1998;

Keith et al., 1996; Schulz et al., 2004; Trafton & Basbaum, 2004). DAMGO can also recruit β-arrestins more consistently than morphine is able to (Bohn, Dykstra, Lefkowitz,

Caron, & Barak, 2004; Whistler & Von Zastrow, 1998, 1999; Jie Zhang et al., 1998).

Recruitment of β-arrestins does not halt all of the intracellular signaling pathway, though some branches of the pathway may stop (Williams et al., 2013). Failing to induce rapid internalization of the µOR may lead to quick desensitization of the receptor (Williams et

56 al., 2013). For a review of the µOR, functional selectivity and desensitization, see

(Raehal et al., 2011) or (Williams et al., 2013).

Once endocytosed, the ligand dissociation can occur and the receptor can be sorted to its new destination; the µOR can potentially be shuttled back to the cell membrane, or to a lysosome for destruction (von Zastrow, 2010). Trafficking it back into the membrane can promote resensitization by enabling it to activate again, while the lysosome will lead to its proteolytic destruction (Ma et al., 2020; von Zastrow, 2010).

Numerous factors can influence this fate, for instance ubiquitylation has been shown to exert an influence on the fate of endocytosed ORs (J.-G. Li, Haines, & Liu-Chen, 2008), or the splice variant of the µOR (Tanowitz, Hislop, & von Zastrow, 2008; Tanowitz & von Zastrow, 2003). Coactivation of other receptors can influence recycling of the receptor, for instance, the NK1 neurokinin receptor (activated by substance P) can reduce arrestin activity and receptor endocytosis (Y. J. Yu, Arttamangkul, Evans, Williams, &

Von Zastrow, 2009).

Some have proposed that agonist-induced endocytosis enhances sensitivity and allows drugs like DAMGO (which promotes endocytosis) to be less desensitizing than drugs like morphine, which typically does not promote endocytosis (Finn & Whistler,

2001; Ma et al., 2020). Others have argued against this view by claiming endocytosis is not the causal factor, though this may be a heuristic to relate endocytosis to desensitization (Finn & Whistler, 2001; Gavril W. Pasternak & Pan, 2013).

Although the existence of µOR/DOR dimers has not been proven, there is evidence that these dimers may promote the development of tolerance. Heterodimers of 57 the mu/delta opioid receptors are suspected to play a role in development of tolerance (S.

Q. He et al., 2011). DOR agonism promotes µOR desensitization (S. Q. He et al., 2011), while disrupting µOR domains suspected of linking µOR/DOR dimers also disrupts tolerance (S. Q. He et al., 2011). Some have pointed out that the DOR appears to have the opposite tolerance/trafficking pattern of the µOR; endocytosing the DOR correlates with greater tolerance (Scherrer et al., 2006; von Zastrow, 2010). For a review of opioid heterodimers and tolerance, see (Gavril W. Pasternak & Pan, 2011).

1.24 Basics of neurophysiology

Ohm’s Law states that V=IR and therefore generation of a voltage requires a current to move through a resistor. Neurons use their plasma membrane (PM) is a natural resistor and conduct ionic currents through the plasma membrane, which then produces its membrane voltage (Vm). Different ion currents can create different voltage magnitudes and polarities; therefore, a neuron can modulate its Vm by controlling which (and when) ions can cross the PM. When a neuron sends its signal, it does so by rapidly changing its polarity, and then rapidly repolarizing to its initial starting point. It does this by rapidly changing the identity of the dominant ion that can pass through the PM. This depolarization is called an Action Potential (AP). Neurons normally start off with a negative polarity compared to their extracellular environment. During an AP (also called a spike), their polarity rapidly shifts from negative to positive. They can do this by rapidly altering the permeability of their PM to different ions and therefore change the rapidly ionic current through their PM.

Differential distributions of ions on the inside and outside of cells creates an ion gradient. Given the opportunity (permeability through the PM), the ions will diffuse down their concentration gradient. Sodium and calcium ions are highly concentrated on the outside of the cells and therefore will diffuse into the inside of cells through sodium ion channels and calcium ion channels. This creates a depolarizing effect on the neuron and a transiently-positive Vm during an AP. Importantly, these depolarizations also allow calcium ions to enter and interact with various biochemical pathways – for instance, by enhancing neurotransmitter release at synaptic terminals or activating gene transcription and synaptogenesis. Potassium is highly concentrated on the inside of the cells. The outward potassium current has a hyperpolarizing effect on cells. Neurons at rest (not during an action potential) have a relatively high permeability to potassium and therefore restricts their excitability. Neurons have very high resistance and therefore can maintain and alter their polarity with very little current (very little ion flow).

Each cell has a resting membrane potential (RMP) produced to the slow outward

K current and a little inward sodium current. GPCRs, such as the µOR, can add or subtract leak channels from the membrane. In the µOR’s case, it can make the RMP more negative (more hyperpolarized) by opening more leak channels. The details and identities of these channels will be given later, but the µOR is also implicated in trafficking voltage-sensitive channels, which are more complex.

GPCRs can move voltage-sensitive ion channels (VSICs) that are mostly not open at rest. These VSICs but can open/close in response to a change in Vm. After opening,

59 they may rapidly inactivate. But others do not inactivate and simply stay open until the

Vm moves out of their preferred range.

These voltage-sensitive ion channels, which affect Vm, can themselves also be triggered by changes in Vm, and this sets up a potential positive feedback loop. Voltage- sensitive sodium channels, for instance, are activated by depolarized Vm, but themselves also create a more-depolarizing Vm before they inactivate and close. These Voltage- sensitive sodium channels create sharp APs (spikes) as they open and then inactivate.

Voltage-sensitive calcium channels act similarly but are even more depolarizing, and the calcium ions they conduct can interact with biochemical pathways. Voltage-sensitive potassium channels, meanwhile, can dampen this chain reaction (dampen APs or smaller depolarizations that don’t rise to the level of an AP) by activating at depolarized Vm but conduct an outward potassium current to repolarize the membrane and, thus, they counteract the sodium and calcium channels by maintaining a strongly negative Vm and preventing the sodium and calcium channels from opening.

There are also receptors, such as the NMDA and AMPA receptors that bind the excitatory neurotransmitter glutamate and allow sodium (and sometimes calcium) to enter and produce a transient depolarization. The NMDA receptor, though, is occluded by a

Mg2+ ion and frequently requires the AMPA receptors to depolarize the cell and dislodge that ion. Therefore, they are channels that can initiate the process of an AP by binding glutamate and causing a transient depolarization. The NMDA receptor (and some AMPA receptors) can conduct calcium, and therefore also can cause gene transcriptional and biochemical changes in the neuron.

On the other side of the spectrum are the GABAA and GABAB receptors, which are usually inhibitory and bind the inhibitory neurotransmitter GABA. GABAergic interneurons activate these. These receptors, respectively, conduct an inward chloride current, or an outward potassium current which both hyperpolarize the neuron and counteract glutamatergic excitation. GABAA receptors are faster to open than GABAB receptors, because GABAB receptors are GPCRs, that act indirectly – like the µOR does.

To put this together as an example, a typical resting neuron may receive numerous, transient excitatory postsynaptic potentials (EPSPs) a result of their NMDA and AMPA receptors and inhibitory postsynaptic potentials (IPSPs) from activation of their GABAA and GABAB receptors. The outcome of those 2 forces depends on the activity of the neuronal network as the. But eventually that Vm may become so positive that it triggers a critical amount of voltage-dependent sodium channels to open causing a spike (an AP). Voltage-sensitive potassium channels, which are slower to open, then bring the Vm back down to a hyperpolarized Vm value, ending the AP. Neurons have axons that these APs can be conducted along, and at the terminal ends of axons are neurotransmitter vesicles which could contain glutamate, GABA, or other neurotransmitters depending on the identity of the neuron. At the axon terminals, voltage- dependent calcium channels can open during APs and allow calcium stimulate release of neurotransmitter onto the next neuron. They will then release their glutamate, GABA, enkephalins, and the cycle continues into the next neuron.

1.25 Hyperpolarization by the µOR

The µOR has been observed to be strongly inhibiting on neurons with the receptor by hyperpolarizing the neuron and diminishing its capability to fire APs, including with the synthetic agonist DAMGO (Brunton & Charpak, 1998; B Chieng & Christie, 1994;

Childers, 1991; Ikeda, Kobayashi, Kumanishi, Niki, & Yano, 2000; Kovoor, Henry, &

Chavkin, 1995; Stein & Machelska, 2011; Stein & Zöllner, 2009). This is also observed in cortical neurons as well with DAMGO (Férézou et al., 2007; Glickfeld et al., 2008;

Krook-Magnuson et al., 2011; K. R. Svoboda et al., 1999; Wimpey & Chavkin, 1991;

Witkowski & Szulczyk, 2006). This is normally one of the µOR’s faster responses (Law et al., 2000). For a review of OR coupling with GIRK channels, see (Ikeda, Yoshii, Sora,

& Kobayashi, 2003).

Much of this hyperpolarizing effect has been attributed to G-protein coupled inwardly-rectifying channels (GIRK or Kir3 channels), which are open at rest and provide the cell membranes with a constant hyperpolarization and decreased input resistance (A. M. Brown & Birnbaumer, 1990; Harris & Williams, 1991; Ikeda et al.,

2000; Kobayashi et al., 1996; Lüscher & Slesinger, 2010; Nockemann et al., 2013; R

Alan North, 1989; Signorini, Liao, Duncan, Jan, & Stoffel, 1997). Activation of these channels is believed to be mediated by the Gβγ subunits of the G-protein complex (Raveh,

Cooper, Guy-David, & Reuveny, 2010). The µOR-mediated upregulation of GIRK channels is the first interaction to desensitize, with significantly reduced upregulation of

GIRK channels occurring around 15 minutes (Blanchet & Lüscher, 2002). Tis subunit’s regulation of GIRK channels is believed to be a direct interaction and can be terminated

62 when the Gβγ subunit re-associates with the Gαi/o-GDP protein (Schreibmayer et al.,

1996).

There are 4 GIRK channels (GIRK1-GIRK4) but only GIRK1-GIRK3 are highly expressed in the neocortex; GIRK4 is restricted to deeper layers and only found in small amounts (Karschin, Dißmann, Stühmer, & Karschin, 1996; Murer et al., 1997). GIRK channels form tetramers, and typically heterotetramers (Lüscher & Slesinger, 2010).

To understand the colocalization of the µOR and GIRK channels (GIRK1 specifically) researchers have used ICC with antibodies to both proteins in various parts of the brain, including the neocortex. They found a high degree of overlap between these

2 proteins there in the soma and proximal dendrites of cortical neurons. However, the overlap was not complete; distal puncta (presumed to be axon terminals) occasionally stained for the µOR, but not GIRK1, indicating that the µOR’s inhibitory effect in distal processes are not associated with hyperpolarization there (Suzanne B Bausch & Chavkin,

1995). Unfortunately, it appears that similar studies have not been attempted with GIRK1 or GIRK3.

Another important issue to address is whether µOR+ interneurons always hyperpolarize when they are exposed to the µOR agonist, DAMGO. This is difficult to answer since electrophysiology seems to be the most sensitive technique, for instance

Férézou et al. (2007) (Férézou et al., 2007) found µOR mRNA in only 20 out of 32 neurons (with sc-PCR) that hyperpolarized with DAMGO (Férézou et al., 2007).

According to Wimpey & Chavkin (1991) (Wimpey & Chavkin, 1991), most of their

DAMGO responders in the CA1/subiculum region of the hippocampal formation did not 63 hyperpolarize, but instead activated unidentified voltage-sensitive K channels (VSKCs)

(Wimpey & Chavkin, 1991). It appears that while electrophysiology is still the most sensitive technique, but hyperpolarization may not be the only indicator of a DAMGO response (Férézou et al., 2007). It should also be noted that the GIRK1/µOR colocalization study discussed earlier analyzed at least 20 different brain regions and found the greatest degree of discordance of coexpression in the hippocampal formation and therefore the neocortical µORs may be more likely to induce somatic hyperpolarization than hippocampal µORs (Suzanne B Bausch & Chavkin, 1995).

In summary, it appears that a strong hyperpolarization may not be the only indicator of a response to DAMGO and neurons. In addition to Wimpey & Chavkin

(Wimpey & Chavkin, 1991), there are numerous reports that cortical and noncortical

µORs can upregulate VSKCs in response to DAMGO (Faber & Sah, 2004; Finnegan et al., 2006; Vaughan et al., 1997; Wimpey & Chavkin, 1991) and Met-Enk (Vaughan et al.,

1997) which is the next topic.

1.26 Voltage-sensitive K channels

Voltage-sensitive K channels (VSKCs) can be divided into several different groups based on their shared properties. VSKCs can mediate either A-type current, which activates and deactivates rapidly, or delayed rectifier current that activate slowly and deactivate slowly or do not inactivate at all. They are also grouped into families based on their genetic similarity; channels within the family typically generate the same type of current, but sometimes they do not. These channels are expressed widely throughout the brain, including the neocortex.

The Shaker (Kv1) family of Kv channels comprise the largest group of voltage- sensitive K channels. They are constituted of 4 α-subunits and sometimes auxiliary β subunits. There are 6 known members expressed in the brain and are numbered Kv1.1-

Kv1.6. These peptides may homotetramerize with identical subunits, or they more form heterotetramers with each other which can lead to various combinations of properties.

They are typically slower to inactivate, and they deactivate slowly, or not at all. They tend to be triggered at more negative potentials when compared to other VSKCs (B Rudy et al., 2009). These channels can be blocked with 4-aminopyridine (4-AP). The protein

αDendrotoxin (αDTX) selectively blocks Kv1.1, Kv1.2 and Kv1.6 subunits; possession of even a single Kv1.1, Kv1.2 or Kv1.6 subunit in heterotetramers will result in channel blockade. Their current is called D-type (ID) current due to their susceptibility to dendrotoxin (B Rudy et al., 2009). The tendency of these channels to heterotetramerize, complex with auxiliary subunits, and capability to undergo post-translation modifications in their α and auxiliary subunits makes the function and properties of ID current extremely variable. While the Kv1 channels are usually delayed rectifiers, Kv1.4 channels and other

Shaker modified Kv1 channels may mediate A-type current (Carrasquillo, Burkhalter, &

Nerbonne, 2012). The αDTX-sensitive current in has been reported to modulate threshold voltage for action potentials, afterhyperpolarization, spike frequency, and small effects on the AP repolarization kinetics rat visceral sensory neurons (Glazebrook et al., 2002).

This Shaker class of channels is believed to exert a strong influence over the excitability of neocortical PVBCs (X. Li, Surguchev, Bian, Navaratnam, & Santos-

Sacchi, 2012). The high sensitivity to voltage of these channels allows them to control

65 repeated firing (J. Connor & Stevens, 1971). These channels can be found in neocortical

PNs to restrict firing rates, however expression of Kv1 channels tends to be strongly influenced by layer and region (Bekkers & Delaney, 2001; Martina, Schultz, Ehmke,

Monyer, & Jonas, 1998). The αDTX-sensitive subset of Shaker channels is also implicated in duration of afterhyperpolarizations in hippocampal PNs (Golding, Jung,

Mickus, & Spruston, 1999).

The Shab (Kv2) family of Kv channels has two known members: Kv2.1 and

Kv2.2, and they typically function as delayed rectifiers, but in some instances, they can influence the repolarization phases of APs (A-type) (Liu & Bean, 2014). These channels can likewise heterotetramerize and undergo posttranslational modifications to alter their kinetics (B Rudy et al., 2009). These channels can be found at axon initial segments of hippocampal PNs (Sarmiere, Weigle, & Tamkun, 2008) and is generally a common feature of both interneurons and PNs (Murakoshi & Trimmer, 1999).

The Shaw family (Kv3) are a group of A-type channels that typically activate and deactivate quickly. They also tend to be activated ay more depolarized voltages than the previously mentioned channels. It’s believed that this channel manages the duration and repolarization phases of action potentials due to these fast kinetics, but unfortunately this family is difficult to investigate due to the lack of specific blockers for Kv3 channels

(Labro, Priest, Lacroix, Snyders, & Bezanilla, 2015; B Rudy et al., 2009). RT-PCR results from hippocampal experiments show that these channels are expressed in at least some types of interneurons and a minority of PNs (Martina et al., 1998). Blocking these channels with the non-specific channel blocker 4-aminopyridine reduces their firing rates

(Martina et al., 1998). These channels can be found in Layer 1 interneurons (Weiser et al., 1994) They can enable cortical neurons to repolarize more quickly, and fire more APs

(Cao et al., 2014).

The Shal (Kv4) family of K channels are another group of A-type channels that are similar to the Shaw family, but are generally capable of activating at more hyperpolarized voltages. They are believed to modulate repolarization phases of action potentials and possibly interspike intervals. Although Shal Kv channels do tend to inactivate quickly, they are believed to recover from inactivation relatively rapidly.

Immunofluorescence of rat visual cortex shows that Kv4 channels can be found in both cortical interneurons and PNs. This includes GABAergic post-synaptic sites, somata, dendrites and dendritic spines (Burkhalter, Gonchar, Mellor, & Nerbonne, 2006). These channels are believed to modulate the duration and frequency of APs in neocortical PNs

(Carrasquillo et al., 2012).

Finally, the Kv7 (KCNQ) channels are a group of channels that mediate the IM current, which is targeted by many neuropeptides. This current is active between RMP and threshold for APs. While it is somewhat slow to activate, it either does not deactivate or deactivates very slowly. Therefore, these channels are unlikely to influence individual

APs, but may suppress spike frequency adaptation and interspike interval (D. Brown,

1988; Storm, 1987). This IM current is believed to be modulated specifically by Gq

Coupled Protein Receptors, like the eponymous muscarinic receptor that the current is named after. The likely mechanism is through mobilization of PLCβ and regulation of

PIP2 (B Rudy et al., 2009). These channels are found in hippocampal PNs and

67 interneurons and regulate interspike interval (Hu, Vervaeke, & Storm, 2007; Lawrence et al., 2006). These channels have been demonstrated to show powerful effects on cortical excitability and performance (Leão, Tan, & Fisahn, 2009; Peters, Hu, Pongs, Storm, &

Isbrandt, 2005).

To summarize, there are many families of potassium channels that can mediate K currents with slightly different properties. Delayed rectifiers tend to be associated with ability to fire APs, while the faster A-type current is more likely to modulate the duration of the spikes. To fully assay the potassium channels in current clamp mode (which enables APs to happen as they would in neurons) there are several important measures.

Specifically, the firing frequency or interspike interval, the resting membrane potential, input resistance, repolarization rate or duration of an AP, and the magnitude of afterhyperpolarization.

1.27 µOR regulation of voltage-sensitive channels

The µOR is a pertussis toxin sensitive Gi coupled receptor that generally acts through two divergent mechanisms (Al-Hasani & Bruchas, 2011; Ingram & Williams, 1994). Upon binding of an agonist to an opioid receptor the Gαi and Gβγ subunits dissociate from each

2+ other. The Gαi activates local inhibits adenylyl cyclase, meanwhile the Gβγ inhibits Ca conductance by binding to, and inactivating, local voltage-gated calcium channels

(Capogna, Gähwiler, & Thompson, 1993; M. Connor et al., 1999; Glickfeld et al., 2008;

Rusin, Giovannucci, Stuenkel, & Moises, 1997; Zamponi & Snutch, 2002). Upregulation of potassium channels appears to be cAMP-dependent in at least some cases (Chen & Yu,

1994), but not in others (Sánchez-Blázquez et al., 2001; Witkowski & Szulczyk, 2006).

The µOR is also believed to negatively regulate calcium (and some sodium) currents as well. This includes: suppression of N-type Ca2+ channels (Law et al., 2000;

Rubovitch et al., 2003); suppression of L-type Ca2+ channels (Piros et al., 1996; Piros et al., 1995); P/Q-type Ca2+ channels and G protein-activated inwardly-rectifying K channels (M. Connor et al., 1999; Henry, Grandy, Lester, Davidson, & Chavkin, 1995;

Jeong & Ikeda, 1998). µORs are also believed to downregulate activity of NMDARs (C.

W. Xie & Lewis, 1997). µORs have also been reported to regulate a variety of other channels in DRG neurons, including inhibiting TRPV1 channels (Endres-Becker et al.,

2007; Spahn et al., 2013), inhibiting acid-sensing ion channels (Cai et al., 2014).

Yet not all channels that the µOR modulates in cortical neurons have been identified. In dorsal striatal neurons, researchers found that multiple K currents were being modulated by DAMGO (Ponterio et al., 2013). Researchers found that an unknown, voltage-dependent K current was activated with DAMGO as well in hippocampal neurons (Wimpey & Chavkin, 1991). Follow-up research with both the endogenous ligand Met-Enk and DAMGO indicated these agonists were upregulating a K current that was 4-AP sensitive and likely belong to the Shaker family of K channels

(Vaughan et al., 1997). Limited knowledge of voltage-sensitive K currents at the time may have prevented further inquiry about the identity of these channels in cortical neurons, but more recent research from other parts of the brain provides more insight about the identity these channels.

Findings from the lateral amygdala by Faber et al. (2004)(Faber & Sah, 2004) indicated that the µOR was upregulating a voltage-sensitive K current to increase spike

69 frequency adaptation – a feature that suppresses the number of evoked action potentials and increases the time between spikes (interspike interval). They likewise found that this effect was blocked by inhibiting phospholipase A2 with AACOCF3. Most interestingly they identified this current as sensitive to both αDTX and Tityustoxin-Kα (TsTx-Kα), which is a feature of the Shaker-family channel Kv1.2. In a similar study on basolateral amygdala neurons that project to the central nucleus of the amygdala, Finnegan et al

(2006) (Finnegan et al., 2006) found that DAMGO reduces the amplitude of evoked

IPSCs and reduced the frequency of mIPSCs. They found that 4-AP, αDTX, TsTx-Kα, and Dendrotoxin-K (blocker of Kv1.1, another Shaker channel). These findings in amygdalar structures suggest that the µOR upregulates the Shaker channels Kv1.1 and

Kv1.2 to inhibit neurons in a PLA2-dependent manner. These results closely mirrored findings in cortical neurons that suggested that unidentified VSKCs were being upregulated by the µOR (Wimpey & Chavkin, 1991).

Therefore, these mechanisms and data collected from the amygdala implicate

αDTX-sensitive current as a target of modulation, which can then modulate the number of APs that are evoked from µOR+ neurons. The relationship between these αDTX- sensitive channels and the µOR has not presently been explored in neocortical neurons but could be one of the mechanisms that the µOR uses to suppress neocortical them.

1.28 Measures of excitability in individual neurons

Cortical neurons have a variety of mechanisms that they can leverage to modulate their output. These mechanisms are used to adjust their firing rates, probability of an action potential (AP), and their neurotransmitter release at synaptic terminals.

Perhaps the simplest way is for neurons to increase their permeability to potassium, which moves their Vm closer to EK and hyperpolarizing the cell – moving it away from its threshold for an action potential. This can be accomplished by opening K leak channels which provide their PM with a constant hyperpolarization. Therefore, their

RMP will become more negative. Opening more channels also decreases their input resistance, as measured by a recording electrode. This is potentially more challenging than measuring RMP, since measuring input resistance necessarily requires inputting a current. A depolarizing current may cause the neuron to fire, and a hyperpolarizing current may not fully capture the breadth of voltage-gated channels that would be activated by a depolarizing current. The µOR is consistently shown the hyperpolarize neurons.

Neurons may also modulate their threshold for an AP – a voltage range at which voltage-sensitive Na channels are likely to begin to open and create the positive-feedback loop that creates an AP. This could be done by utilizing posttranslational modifications or trafficking of voltage-sensitive K channels to the cell membrane and make them more likely to open upon a depolarizing stimulus. On the other hand, cells may also downregulate their voltage-sensitive Na channels to make them less able (by trafficking or posttranslational modifications) to open upon a depolarizing stimulus. The µOR is not necessarily known to modulate AP threshold (Faber & Sah, 2004), though it is relatively easy to measure since the spikes evoked at threshold is normally minimal and can produced a stable change in Vm.

Cortical neurons rarely fire only one AP in response to a depolarizing stimulus – they often fire several. The pattern, frequency, and number of APs is subject to modulation to alter their excitability. Interspike interval (ISI) – the length in time between

APs – is a property that is inversely related to the frequency of APs. The µOR has been shown to increase the ISI in amygdalar neurons, which is reversible with αDTX (Faber &

Sah, 2004). Therefore, this is an important aspect to measure. Similarly, the number of evoked APs is closely related to the ISI.

The duration of an action potential that neurons can modulate too and may lead to variable physiological effects. For instance, neurons may reduce the activity of their A- type K currents to make APs longer in duration. This may slow down the rate of APs – because the Vm will take longer to reset down to RMP and allow the VSNCs channels to shift into the closed conformation. However, at a presynaptic terminal, the longer- duration AP may make it more likely that Ca channels will open and cause release of neurotransmitter.

1.29 Cortical neurodevelopment and Dlx genes

Discriminating PNs from GABAergic interneurons is a common issue with any testing preparation of neocortical neurons. This distinction may be important to understanding how PNs are affected by DAMGO exposure or targeting a specific population for electrophysiology. Luckily, this obstacle can be overcome with the help of transgenics.

Excitatory and inhibitory neocortical neurons have different embryonic origins.

They therefore are hallmarked by expression of different genes, which can be used to identify them. Utilizing unique homeobox genes expressed by either excitatory or 72 inhibitory neurons allows researchers to identify them in virtually any experimental system.

All cortical neurons, whether inhibitory or excitatory, are produced from temporary proliferative zones (the ventricular zone and subventricular zone) and migrate into the prenatal cortex. The deeper layers of the cortex migrate the earliest and Layer 2 developing the latest – Layer 1 is an exception; the most superficial cortical layer appears to develop first and develop persistently through the postnatal period (Rakic, 2009)

While PNs migrate radially into the cortex from the pallium, the cortical interneurons are generated in the prenatal, basal telencephalon (the sub pallium) and migrate tangentially into the cortex. The subpallium consists of several domains, including the medial ganglionic eminence (MGE), the lateral ganglionic eminence

(LGE), the preoptic area (POA). The caudal ganglionic eminence (CGE), meanwhile, is located at a caudal fused area of the MGE and LGE. Fate-mapping experiments show that interneurons are born in the ventricular zone of these ganglionic eminences, while PNs are born in the VZ of the cortex (Stewart A Anderson, Marín, Horn, Jennings, &

Rubenstein, 2001). Mouse interneurons begin this migration on about day E12 and the neurons populate all cortical areas (including hippocampal cortex and neocortex) (Flames

& Marin, 2005). Migration of progenitors from the subcortical telencephalon appears to be complete by around E17 in rodents – including in rat models (S. A. Anderson,

Eisenstat, Shi, & Rubenstein, 1997; Metin, Baudoin, Rakic, & Parnavelas, 2006; Q. Xu,

Cobos, De La Cruz, Rubenstein, & Anderson, 2004).

While some researchers have proposed that interneurons are granted their unique identity based on where they happen to stratify, fate-mapping data suggest that their identities are determined prior to migration (Stewart A Anderson et al., 2001; Batista-

Brito, Machold, Klein, & Fishell, 2008; Butt et al., 2005). The MGE interneuronal pool gives rise to most PV+ and SST+ interneurons, while the CGE gives rise to other classes

(Butt et al., 2005).

Interneuron migration and development is supported by unique expression of the homeodomain transcription factors Dlx of the family which can be used to identify interneurons from PNs (Flames & Marin, 2005; Pla et al., 2017). Six known Dlx orthologs are known, numbered 1 through 6 are expressed by prenatal cortical interneurons but data derived from a Dlx 5/6 BAC suggests that expression of Dlx5/6 persists in mature cortical interneurons (Y. Wang et al., 2010). Using transgenic expression of β-galactosidase, researchers have shown that Dlx5 or Dlx6 are expressed by only, and all, cortical interneurons. Sc-PCR analyses revealed that these neurons represent all the major expression categories of neurons (SST+, 5HTr3a+, and PV+) (de

Lombares et al., 2019). Furthermore, experiments using Dlx5/6-Cre mice show that these neurons differentiate into GABAergic interneurons, while the excitatory neurons are instead predicted by expression of Emx1 (Batista-Brito et al., 2008; Gorski et al., 2002).

A Dlx5/6-Cre construct has also been used to disrupt the excitatory drive into mouse interneurons in vivo and thereby create behavioral defects that resemble a schizophrenic endophenotype (Fazzari et al., 2010).

Unfortunately, the embryonic lethality of Dlx gene knockout makes it difficult to determine the exact roles of these genes, but in mature interneurons Dlx5/6 appears to promote expression of the GABA-synthesizing enzymes GAD65 and GAD67; ectopic expression of these genes can induce production of GABA (de Lombares et al., 2019).

1.30 Primary neuron culture

The need to conduct experiments on neurons in an accessible environment led to development of several models to investigate neurons that are grown in vitro. Some models of neurons are derived from tumors (Biedler, Helson, & Spengler, 1973), but the primary neuron culture model enables researchers to extract otherwise-normal neurons from the brains of rodents and culture them in a controlled environment for experiments

(Gordon, Amini, & White, 2013). The present-day neuron culture protocol has evolved for 100 years (Harrison, 1910) to enhance their validity and study the phenomena referenced in this literature review. Modern iterations of this technique have developed to maximize their similarity with neurons in situ.

Primary Neuron Cultures are popular models for use with electrophysiology due to the ease-of-access and separation from tonic factors secreted by neocortical neurons

(Brewer, Torricelli, Evege, & Price, 1993; Dichter, 1978; Mains & Patterson, 1973). This simplifies many of interactions between neurons.

Primary neuron cultures, as described above, have been used extensively to study calcium oscillations for the ease of imaging a network of neurons in a single plane

(Dravid & Murray, 2004; T. H. Murphy et al., 1992; Przewlocki et al., 1999; J. Wang et al., 2016). These cultures are highly amenable for these studies for their ease of use and 75 the ability to resolve single neurons without obfuscation from additional tissue layers.

They can enable the collection of a large volume of data that may be required to sample a heterogenous population, such as neocortical interneurons.

Primary cultures of rat neocortical neurons have been used extensively to investigating µOR’s intracellular signaling pathway, mechanisms (Arttamangkul,

Torrecilla, Kobayashi, Okano, & Williams, 2006; M. C. Lee et al., 2002; Junhui Zhang,

Qian, Zhao, Hong, & Xia, 2006), and electrophysiology (E. Tanaka & North, 1994) - including the µOR’s disinhibitory effect in non-cortical primary cultures (Crain, Shen, &

Chalazonitis, 1988). Therefore, primary neuron cultures are a powerful tool to investigate the phenomena outlined in this literature review.

CHAPTER 2: NEOCORTICAL µ OPIOD RECEPTORS ENHANCE SPONTANEOUS CALCIUM OSCILLATIONS THROUGH DOWNREGULATION OF GABAA AND GABAB RECEPTORS

2.1 Introduction

The endogenous opioid system of the cerebral cortex mediates antinociception, reward valuation and reward-seeking behaviors. Dysregulation of this system is believed to contribute to pathological and compulsive behaviors, such as behaviors eating disorders, pathological gambling, and drug-seeking (B. Bencherif et al., 2005; Badreddine

Bencherif et al., 2005; Joutsa et al., 2018; Mick et al., 2016; Mitchell et al., 2012; Qu et al., 2015; Zubieta et al., 1996). Experimentally, impulsive behaviors and binge-eating can be induced through infusion of the µ opioid receptor specific (µOR) agonist [D-Ala2, N-

Me-Phe4, Gly5-ol]-Enkephalin (DAMGO) into the frontal cortex of animal models

(Mena et al., 2011; Selleck et al., 2015). Blockade of the µOR with naltrexone inhibits these compulsive behaviors (Bartus et al., 2003; Blasio et al., 2014). These aberrancies are believed to result from disruption of the activity of cortical networks (Haider, Duque,

Hasenstaub, & McCormick, 2006; M. Whittington et al., 1998; Y. Zhang et al., 2019).

The µOR is believed to suppress GABAergic release through its expression primarily on cortical GABAergic interneurons, and therefore lead to overactivity of the targets of their inhibition, the glutamatergic Pyramidal Neurons (PNs) (Drake & Milner,

1999, 2002, 2006; Férézou et al., 2007; Huo et al., 2005; Madison & Nicoll, 1988;

Stumm et al., 2004; Taki et al., 2000; Witkowski & Szulczyk, 2006; Zieglgansberger et al., 1979). But some research suggests that the µOR µOR may localize to PNs and directly activate them as well (Przewlocki et al., 1999; Rola et al., 2008; Schmidt et al.,

2003; Witkowski & Szulczyk, 2006). Therefore, it is unclear whether excitatory effects of µOR agonism is due to disinhibition or excitation of Pyramidal Neurons, or whether the µOR can perform both functions. These conflicting results have presented a challenge to understanding how the µOR exerts its effects on cortical circuits. Furthermore, several lines of evidence suggest that the µOR suppresses the neurogliaform class of neurons, which uniquely activate GABAB receptors on excitatory Pyramidal Neurons (Krook-

Magnuson et al., 2011; Olah et al., 2007; Tamás et al., 2003) – whereas other cortical interneurons generally activate only GABAA receptors. This positions µOR+ interneurons to play an important role on slower-acting and longer-lasting inhibition on cortical networks by regulating postsynaptic activity of GABAB receptors.

The dynamics of excitatory and inhibitory network activity have been explored in cultured neurons, where neurons engage in frequent spontaneous activity and synchronized depolarizations due to synaptic interlinkage (Canepari et al., 1997;

Chiappalone, Bove, Vato, Tedesco, & Martinoia, 2006; Nakanishi & Kukita, 1998). To easily monitor network activity in many culture neurons at once, calcium-imaging has been used to visualize large-scale network events in cultured neurons; excitation in cultured neurons can be visualized as spontaneous synchronized calcium oscillations

(SCOs) resulting from synchronized activation of glutamatergic receptors (Bacci et al.,

1999; Muramoto et al., 1993; Nakanishi & Kukita, 1998; Robinson et al., 1993). These oscillations can be prevented by blocking depolarizations with TTX, or by blocking glutamate receptors (Dravid & Murray, 2004; Kelamangalath et al., 2011; T. H. Murphy et al., 1992; Nuñez et al., 1996; Przewlocki et al., 1999). Within individual neurons, these calcium transients correspond with excitatory postsynaptic potentials as well as bursts of action potentials (Muramoto et al., 1993; T. H. Murphy et al., 1992). Some researchers have used this model in low concentrations of Mg2+ and found that these calcium oscillations and bursts of action potentials are due to the calcium-conducting properties of the NMDA receptors (Canepari et al., 1997; Penn et al., 2016; Przewlocki et al., 1999; T.

Tanaka et al., 1996), but others argue that calcium oscillations in physiological [Mg2+] are actually caused by calcium-conducting AMPA receptors and mGluRs coupled to calcium-induced calcium release (CICR) (Dravid & Murray, 2004; Flint et al., 1999). Yet other authors have suggested that voltage-sensitive calcium channels play a large or predominant role in generating SCOs (Bacci et al., 1999; Dravid & Murray, 2004;

Inglefield & Shafer, 2000). Although there is disagreement over the identity of the 78 glutamatergic receptor or ion channel based on the type of neuronal culture and experimental condition, these calcium oscillations appear to be related to activity in excitable cells.

The effect of DAMGO on SCOs was explored in previous experiments by another lab, and it was suggested that the µOR directly stimulates NMDA receptors on hippocampal Pyramidal Neurons (Przewlocki et al., 1999), which may throw doubt on the view that the µOR excites cortical networks by suppressing interneurons. On the other hand, electrophysiological experiments done by our lab (Dutkiewicz & Morielli,

2020), and another lab (Férézou et al., 2007), suggest that neocortical interneurons are suppressed by DAMGO stimulation. Although these mechanisms are not mutually exclusive, these lines of evidence have presented a challenging research issue, whether the µOR’s role in exciting cortical networks is mainly driven by suppressing cortical interneurons, or by directly activating the Pyramidal Neurons. Furthermore, the evidence that the µOR localizes to neurogliaform neurons also raises the possibility that DAMGO stimulation may reduce GABA release onto both GABAA and GABAB receptors, and thus DAMGO-induced disinhibition may be more consequential to network activity than previously considered.

We hypothesized that the µOR’s excitatory effect on cultured neurons was primarily by suppressing subsets of GABAergic interneurons (i.e., disinhibition) rather than by directly exciting the PNs themselves. We further predicted that DAMGO would suppress GABA release onto both GABAA and GABAB receptors due to the research

79 suggesting that neurogliaform neurons express the µOR. We therefore tested DAMGO in the presence of GABAR-blockers to prevent the DAMGO-induced SCO changes.

2.2 Materials and methods

2.2.1 Cell cultures

All procedures were authorized by the Institutional Animal Care and Use Committee

(IACUC) at University of Vermont. We dissected CD® IGS Sprague-Dawley (Charles

River) pregnant rat dams to harvest neocortical neurons from the frontal cortices of E21 rat embryos. Brain cortices were rinsed in Hibernate A (Gibco, Grand Island, NY), dissociated with papain (Worthington, Columbus, OH) and mechanically separated through gentle trituration with a pipette. Dissociated cortical neurons were and cultured on round 12mm glass, PEI-coated (Sigma-Aldrich, St. Louis, MO) coverslips at an approximate density of 3x104 cells/cm2. We maintained the neurons in Neurobasal A,

B27, Penicillin-Streptomycin and Glutamax (All from Gibco), in a humidified 37° C incubator with 5% CO2. For experiments where we utilized the interneuron marker, we transformed those cultures on DIV0 with AAV-mDlx-NLS-mRuby2, which was a gift from Viviana Gradinaru (Addgene viral prep # 99130-AAV1); http://n2t.net/addgene:99130; RRID:Addgene_99130).

2.2.2 Calcium imaging

We removed coverslips for their growth media and incubated them in the dark, at ambient temperature in a combination of a 1:50 dilution of B27 and Hibernate A. We prepped the fluo-4 AM dye (Molecular Probes, Eugene OR) by combining 2uLs of a 1

µg/µL (in DMSO) stock solution with 2µLs of 20% Pluronic F-127 (Molecular Probes).

After ensuring the dye was dissolved in the Pluronic F-127, we combined this with the neuron’s Hibernate A (Gibco) buffer at a final concentration of 1 µg fluo-4, AM/1 mL of

Hibernate A+B27. The neurons were incubated in this solution for 1 hour before imaging.

The drugs were purchased from two main sources: [d-Ala2,N-Me-Phe4,Gly5-ol]-

Enkephalin] (DAMGO), D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP), picrotoxin, and CGP55845 were purchased from Sigma (St. Louis, MO); 6-Cyano-7- nitroquinoxaline-2,3-dione disodium (CNQX) and D-(-)-2-Amino-5-phosphonopentanoic acid (dAP5) were from Tocris (Bristol, United Kingdom).

After dye loading, we transferred them into a recording chamber and perfused them with ACSF (in mM): NaCl, 126; KCl, 5; NaH2PO4, 1.25; CaCl2, 2; MgCl2, 1;

NaHCO3, 26; Glucose, 20; and pyruvate, 5(Férézou et al., 2007). The ACSF was warmed

o to 30 C and constantly bubbled with a mixture of 95% O2 and 5% CO2. Neurons were visualized and recorded on a Nikon Eclipse E600 fluorescence microscope. We used a

40x objective dipping lens and acquired an imaging area with at least 15 neurons. During image capture, we illuminated the neurons with an X-Cite 120 fluorescence illuminator to activate green fluorescence (Excelitas, Waltham, MA). In instances where we used red- fluorescent interneurons, we captured red fluorescence (mRuby2) with one image prior to recording the green (fluo-4) channel. Fluorescence was captured with a Photometrics

CoolSnapES Monochrome Camera (Roper Scientific, Sarasota Florida) and imaged with an exposure time of 250ms with either Metavue 6.1 (Molecular Devices, San Jose, CA) software or µManager (Edelstein et al., 2014).

Our preliminary experiments indicated that we could capture at least 125s of images at 4 frames/second before significant bleaching occurred. Therefore, we recorded two 62.5s stacks; one before the drug (slot 1) and one after the drug (slot 2).

Although previously published studies by other labs have sometimes opted to use low or no [Mg2+] ACSF, which enhances spontaneous synchronized calcium oscillations

(SCOs), we observed spontaneous calcium oscillations in the presence of physiological levels (1 mM Mg2+), which has also has been reported previously in rodent cortical cultures (Dravid & Murray, 2004; T. H. Murphy et al., 1992). We therefore continued to use 1mM Mg2+ ACSF.

For the first dataset (Figure 7), we recorded .tiff image timestacks under a pre- drug condition, and then applied the drug buffer for 62.5s and subsequently recorded an image time stack for the post-drug condition while continuously applying the drug

(Figure 5). However, the drug sequence varied depending on which data set it was collected from; in the first dataset, slot1 always corresponded with vehicle ACSF (saline), and timeslot 2 always corresponded to a combined drug bolus. This drug combination is indicated on the graphs.

In the second dataset (Figure 11), picrotoxin and CGP55845 were preincubated for 15-25 minutes before exposure to DAMGO, such that DAMGO was applied in a background of those drugs. In the case of Figure 12, CGP55845 was similarly applied in a background of picrotoxin.

2.2.3 Calcium-imaging analyses

Data initially were contained in the form a brightfield image and a stack of 250 green- channel images. Cell counting and image analyses were done on ImageJ 1.52p

(Schindelin et al., 2012) prior to running the Python script. We used ImageJ to draw ROIs separately around nonred or red (where applicable in the second data set) with the Timer

Series Analyzer V3 plugin, which produced a table of calcium-indicator fluorescence over the 250 images. That table was copied over onto separate Microsoft Excel

Spreadsheets into labelled data folders. Those labelled folders and the spreadsheets therein served as the data input for the calcium-imaging Python scripts. These Excel spreadsheets were analyzed directly by the ca_transient_peak_finder_14.5.py script. in the first dataset (for Figure 6 and Figure 7) which was programmed to generate xy plots for the peaks (Figure 1 and Figure 2).

Figure 1. Representative example of calcium peaks in one neuron.Spontaneous calcium oscillations were captured in each neuron over 62.5s (4 frames / second) and graphed on xy plots. The y axis represents relative proportional units of calcium indicator fluorescence ((F-F0)/F0), where the highest peak in the

83 timestack (including all of the neurons/ROIs not shown here) is set to an amplitude of 1.00, and these peaks are scaled against the tallest peak. X axis represents frames at (4 frames/sec). On the calcium peaks: yellow lines denote the base of each recognized peak, green horizontal lines are the measures of halfwidth for the peaks, and the red dot at the tips of the peaks are noting that the peak has been successfully culled for analysis by the Python script.

The data for the second dataset (for Figure 11) utilized the auto_mask_red_cell_v2.5.py script to automatically generate ROIs in neurons. It created two spreadsheets, one for red neurons and one for nonred neurons. These spreadsheets were then analyzed separately in the same manner as described above.

We observed that groups of neighboring neurons captured within imaging frames tended to experience apparently synchronous SCOs, and therefore opted not to use individual neurons as members of the sample (Figure 2). Instead, we measured all SCOs occurring in all neurons within the imaging frame and derived median values of the SCOs in each neuron within the frame. We then further determined the mean value of these median values, to calculate the average SCO from each imaging stack. Through these calculations, we compared the SCOs before and after various drug treatments to determine how the drugs impacted the average SCO occurring within the imaging frame.

Thus, each member of our sample represented the average (mean of median) SCO occurring in the imaging frame during the 62.5s recording.

Saline DAMGO

Calcium (F-F0)/F0

Time

Figure 2. Neurons in image field have synchronous spontaneous calcium oscillations. A temporal alignment of all neurons in an imaging field (Left) before DAMGO, and the same neurons (Right) after DAMGO. Y axis shows calcium indicator fluorescence with each separate neuron arranged in a separate row. Each neuron’s highest peak is set to a value of 1.00 to standardize the scale across neurons. X axis represents a period of 80s in each of the two conditions. We observed that neurons in the imaging field had synchronized spontaneous calcium oscillations (SCOs). Therefore, we derived median peak values for each neuron, and then derived the mean of the median values across all the neurons in the field for each of the 2 conditions, in order to derive the average peak value before and after the drug. DAMGO’s enhancement of calcium peak halfwidth is also evident from these graphs.

2.2.4 Validation for auto_mask

To validate the auto_mask script that automatically drew ROIs in the second dataset, we hand-drew ROIs in 214 spreadsheets that were also analyzed using the auto_mask script.

We performed a bivariate correlation in IBM SPSS 27.0.0.0. to correlate the mean SCO amplitude, halfwidth, and number between the script-drawn and manually drawn ROIs.

Measure Bivariate Correlation Amplitude R = 0.790, p < 0.001 Halfwidth R = 0.909, p < 0.001 Number R = 0.856, p < 0.001 Table 1. Validation for auto_mask ROI generator. To determine whether the automatic generator could derive values positively correlated with hand-drawn ROIs (in ImageJ), we utilized the script and hand- generated ROIs on 214 image stacks across all drug conditions. We determined that all measures featured here were positively and significantly correlated with each other, indicating that the auto_mask script ROIs were consistent with the manually drawn ROIs. 2.2.5 Whole-cell recordings

Patch pipettes with resistances of 5-10MΩ were fabricated from borosilicate glass capillaries and filled with intracellular saline containing (in mM) K-gluconate, 144;

MgCl2, 3; ethyleneglycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid, 0.5; 4-(2- hydroxyethyl)-1-piperazineethanesulfonic acid, 10 (Férézou et al., 2007). The pH was adjusted to 7.2 with potassium hydroxide and confirmed for an osmolarity of 285/295 mosm. Whole-cell recordings were made with an Axopatch 200B (Axon Instruments) amplifier and Clampex 9.2.1.9 (Molecular Devices) software. Signals were sampled at

10kHz with a Digidata 1322a (Axon Instruments) DA converter. Calcium-imaging was aligned with electrophysiology through a green light emitting diode that was linked to the

Digidata 1322a converter and controlled by the Clampex software; the LED emitted a start sequence of light flashes which were visible in the first 16 seconds of recording and removed prior to analysis.

2.2.6 Sample sizes and statistics

Sample sizes were initially calculated based on preliminary data on DAMGO’s effect in a large test group of 21 saline and 21 DAMGO recordings. This test group was used to 86 determine future sample sizes and the polarity of DAMGO effects for statistical testing

(one-tailed versus two-tailed). Although we measured several properties of SCOs, we focused on DAMGO’s effect on SCO halfwidth, since preliminary data indicated that it was the most frequent DAMGO effect, and also the only persisting effect in the presence of picrotoxin. In this large pool of preliminary recordings, we found that fold changes of halfwidth in the DAMGO recordings (N = 21, M = 1.80, SD = 0.78) were significantly greater (t(40) = 4.68, p < 0.001) than the fold changes in saline controls (N = 21, M =

0.97, SD = 0.78). A posthoc analysis of that preliminary data pool revealed that the effect size d was a value of 1.44. We used G*Power software (Faul, Erdfelder, Buchner, &

Lang, 2009; Faul, Erdfelder, Lang, & Buchner, 2007) to determine that the required sample size was an n = 12 for a one-tailed t test. For the first dataset, we therefore used groups of n = 12 for each group. In the second dataset, we used an n = 13 to provide a +1 sample number in case of technical difficulties, which did not occur, and we thus used all

13 recordings.

The mean scores for calcium peak halfwidth, number of peaks, and amplitude in all drug conditions were graphed and analyzed in GraphPad Prism version 8.4.3 for

Windows, GraphPad Software, San Diego CA, USA www.graphpad.com. Statistical testing on the first dataset commenced by testing for a normal distribution with a

D’Agostino-Pearson’s omnibus test. When data within a group were significantly different from a normal distribution (p < 0.05), then that group was compared with a

Mann-Whitney U test. Data that were not significantly different from normal distributions were instead tested with a one-tailed unpaired t-test. Statistical testing on

(Figure 11), where 4 groups were included, we instead tested with ordinary one-way

ANOVAs and Sidak’s test for multiple comparisons in GraphPad Prism.

2.3 Results

2.3.1 Measuring synchronized spontaneous calcium oscillations

To measure spontaneous calcium oscillations, we directed our analyses on 3 basic properties of the calcium spikes to comprehensively analyze SCOs: peak height, duration, and the number of calcium peaks. Respectively, these represent the y dimension of the calcium peak, the x dimension, and the quantity of SCO peaks (Figure 1).

2.3.2 SCOs require neuronal activity and glutamatergic signaling

Other labs have reported SCOs similar to our observations (Figure 1) and found that these oscillations correspond to excitatory postsynaptic potentials and bursts of action potentials in the neurons (Bacci et al., 1999; T. H. Murphy et al., 1992; Przewlocki et al.,

1999; Robinson et al., 1993). To investigate whether SCOs in our preparation correlated with neuronal activity, we counted SCOs in a variety of conditions designed to prevent glutamatergic drive and action potentials,

Other labs have found that SCOs are produced by activation of postsynaptic glutamate receptors (Bacci et al., 1999; Dravid & Murray, 2004; T. H. Murphy et al.,

1992; Przewlocki et al., 1999). To test this in our preparation, we counted calcium oscillations occurring during pharmacological blockade of glutamatergic receptors. We therefore applied the combination of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic

88 acid receptor (AMPA) and kainate blocker 6-Cyano-7-nitroquinoxaline-2,3-dione

(CNQX; 20 µM) and the N-methyl-D-aspartate (NMDAR) inhibitor D-(-)-2-Amino-5- phosphonopentanoic acid (AP5; 50 µM), which eliminated all SCO activity (Figure 3).

The finding that, SCOs depend on intact glutamatergic signaling is in agreement with previous reports of SCOs in primary cortical cultures (Cao et al., 2015; Dravid & Murray,

2004; Inglefield & Shafer, 2000; T. H. Murphy et al., 1992; Nuñez et al., 1996;

Przewlocki et al., 1999; Shen et al., 1996).

To test whether SCOs were dependent on spontaneous action potentials, we applied 1 mM of the sodium channel inhibitor tetrodotoxin (TTX) to block them. We found that this drug also prevented SCOs from occurring (Figure 3). These findings were again consistent with previous research from other labs which have found TTX suppresses SCOs (Dravid & Murray, 2004; T. H. Murphy et al., 1992; Przewlocki et al.,

1999; Shen et al., 1996) (Inglefield & Shafer, 2000).

Average Peaks

CNQX+AP5 ns

TTX ns

Saline ****

0 5 0 5 1 1 Peaks (neuron x min)-1

Figure 3. Neuronal activity and glutamatergic signaling required for SCOs. We counted calcium peaks occurring in a 160s period in 3 groups: saline controls; voltage-sensitive sodium channel inhibitor tetrodotoxin (TTX; 1 µM); and the combination of (20 µM) AMPA/kainate blocker 6-Cyano-7- nitroquinoxaline-2,3-dione (CNQX) and the NMDAR inhibitor (50 µM) D-(-)-2-Amino-5- phosphonopentanoic acid (AP5). In the saline condition, neurons exhibited a median of 6.93 oscillations (range 0.25 to 12.66) per neuron, per minute. In the TTX condition, neurons exhibited a median of 0.00 (range 0.00 to 0.10). In the CNQX+AP5 condition, neurons exhibited a median of 0.03 oscillations (range 0.00 to 0.24) per neuron, per minute. We found that the number of calcium oscillations in the TTX (t(11) = 1.45, p = 0.17) and CNQX+AP5 (t(11) = 2.19, p = 0.05) condition were not significantly different from a test value of 0 (one-sample t tests), while the saline group was significantly different from a test value of 0 (t(11) = 10.01, p < 0.0001). Therefore, spontaneous calcium oscillations are virtually undetectable in TTX and the combination of CNQX+AP5.

We next performed simultaneous electrophysiology and calcium imaging to determine whether the SCOs corresponded to neuronal activity, and thus determine whether SCOs could be used to indicate bursts of activity (Figure 4). Previous studies on

SCOs have found that they correlate with large depolarizations, which appear to be excitatory postsynaptic potentials, and bursts of action potentials - but not single, isolated action potentials (Bacci et al., 1999; T. H. Murphy et al., 1992; Przewlocki et al., 1999;

Robinson et al., 1993). We similarly found that the SCOs correlated with large depolarizations that resembled excitatory postsynaptic potentials. Importantly, we also found that these SCOs do not correspond to isolated action potentials, but instead bursts.

We found that depolarizing current injections (Figure 4) were capable of causing calcium peaks, spontaneous bursts of activity were sometimes greater. Thus, bursts of action potentials were correlated with SCOs.

Figure 4. SCOs correspond to bursts of activity. We simultaneously performed electrophysiology and recorded calcium indicator fluorescence, in order to correlate the SCOs with bursts of activity. Every 3 seconds, we applied a 1s suprathreshold current to induce action potentials, to determine whether the bursts of activity were causal or correlation. (Red Line) Vm over time, showing action potentials and RMP. (Gray line) Applied current over time. (Blue line) Calcium indicator fluorescence. This image indicates that calcium peaks are generally produced through spontaneous activity in this neuron.

Taken together, SCOs were dependent on intact spontaneous glutamatergic signaling and action potentials, and thus necessarily dependent on excitatory drive and

91 spontaneous neuronal activity. Our observations of the simultaneous electrophysiology and pharmacological experiments were consistent with previous experiments of other labs; we similarly saw large depolarizations and bursts of action potentials that corresponded to synchronized SCOs. These large depolarizations appeared consistent to previous descriptions that suggested that they were excitatory postsynaptic potentials.

Having found that the SCOs in our preparation were correlated with bursts of activity, and that they are dependent on spontaneous APs and availability of glutamatergic receptors, we next conducted a series of pharmacological experiments to elucidate the nature and mechanism of DAMGO’s effects on SCOs.

2.3.3 First dataset experimental design

To assay the effects of drugs on the SCOs, we designed an experiment to measure SCOs before and after drug combinations to determine how drugs altered SCOs (Figure 5).

Continuous Application

of Drug Combo

) )

d d

r r

o o

c c

e e

r r

t t

r r

d d

a a

n n

t t

s e s e

( ( ( (

1 2

t t t t

o o o o

l l l l

S S S S

0 60 120 180 240 300 360 420 480 540 600 660 720 780 slot1 slot2

Fold Change = slot 2/slot 1 (postdrug/predrug) Time (s)

Figure 5. Timecourse of experiment. We recorded calcium transients during an 80s period before exposure to the drug (slot 1) and after exposure to the drug or drug combination (slot 2) to study the effects of these drugs on SCOs. To measure the change in the properties of these calcium oscillations, we divided the postdrug value by the predrug value to calculate the fold-change in that property due to the drug.

2.3.4 µOR agonism enhances spontaneous Ca2+ oscillations Given our result that SCOs correlate with neuronal activity, we next applied DAMGO and measured SCOs before and after this drug to find whether µOR-agonism enhanced their SCO amplitude, duration

(halfwidth), and/or frequency. This finding would support the µOR’s known role in exciting cortical circuits and could provide a basis for determining whether the effect of

DAMGO was disinhibitory or providing direct excitation to cortical cultures.

We found that DAMGO (3 µM, 80s) induced a significant increase (t(22) = 2.74, p = 0.006) in SCO amplitude (M = 1.80, SD = 0.94) when they were compared to negative controls (M = 0.93, SD = 0.58). We also found a significant (t(22) = 2.69, p =

0.007) increase the halfwidths of DAMGO-exposed neurons (M = 1.71, SD = 0.69) when compared to saline controls (M = 1.09, SD = 0.42). However, we did not find significant changes (U = 43, p = 0.050) in the number of calcium transients in the DAMGO (Mdn =

1.17) condition versus the control (Mdn = 0.71). In summary, we found that DAMGO enhances the duration and magnitude of these calcium oscillations, but not their number.

Although these findings substantiated the µOR’s known role in enhancing the activity of neuronal circuits, it was unclear whether DAMGO was directly activating excitatory neurons or inhibiting GABAergic neurons. For instance, our finding that DAMGO enhanced the duration (halfwidth) of SCOs could suggest that µORs directly (on

Pyramidal Neurons) were inducing release of glutamate over a longer period, or it could

93 also suggest that µOR+ inhibitory neurons were slower in being recruited to suppress excitatory neurons, which in turns allows the PNs to release glutamate over longer periods before being sufficiently inhibited to cease secretion of glutamate. However, in the latter (disinhibitory) scenario, GABA antagonists could prevent inhibitory interneurons from intervening during the DAMGO challenge.

a b c d Average Amplitude Average Amplitude Amplitude Amplitude With Saline With DAMGO

0.6 0.4

0.3 DAMGO CTAP+DAMGO 0.4 SCO SCO Amplitude Amplitude0.2 ** ns (Relative) 0.2 (Relative) 0.1 Saline Saline

0.0 0.0 Pre Post Pre Post 0 1 2 3 4 0 1 2 3 Fold change from baseline Fold change from baseline e f g h Average Halfwidth Average Halfwidth Halfwidth Halfwidth With Saline With DAMGO

15 20

15 DAMGO CTAP+DAMGO 10 SCO SCO Halfwidth Halfwidth10 ** ns (Frames) 5 (Frames) 5 Saline Saline

0 0 Pre Post Pre Post

0 1 2 3 4 0 5 0 5 0 5 0. 0. 1. 1. 2. 2. Fold change from baseline Fold change from baseline i j k l Average Number of Peaks Average Number of Peaks Number of Peaks Number of Peaks With Saline With DAMGO

15 15

DAMGO CTAP+DAMGO 10 10 Number Number of of ns ns Peaks 5 Peaks 5 Saline Saline

0 0 Pre Post Pre Post 0 1 2 3 0 1 2 3 Fold change from baseline Fold change from baseline

Figure 6. DAMGO enhances calcium oscillation duration and amplitude. We measured the average amplitude, halfwidth, and number of calcium transients after incubation with (n = 12) vehicle saline or (3 µM; n = 12) DAMGO. (Top Row) Amplitude numerical values are taken as a proportion of the tallest peak in the predrug recording of the timestack. (First column) Real values of saline controls for a) amplitude, e) halfwidth, and i) average number of peaks before and after vehicle saline. (Second column) Real values of DAMGO-exposed neurons in (b) amplitude, (f) halfwidth, and (j) average number of peaks before and after DAMGO. (e,f) Frames measured in 4 frames/second. To visualize the change in these properties in individual neurons, we graphed the fold change from baseline in these same properties. We found significant DAMGO-induced increases (p < 0.05) in (c) amplitude and (g) halfwidth of calcium transients relative to the change in the saline group. However, we did not find a significant DAMGO-induced change (p > 0.05) in (k) number of peaks relative to the saline controls. (d,h,l) Validation for the specific effect of

DAMGO by combining it with the µOR-specific antagonist D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP) successfully blocked the effect of DAMGO relative to fold changes in saline controls (p > 0.05)

PARAMETER Mean Difference P value 95% Confidence Interval for Difference (DAMGO-Saline) Lower Upper Amplitude 0.88 0.006 0.21 1.54 Halfwidth 0.63 0.007 0.14 1.11 Number 0.33 0.050 -0.20 0.86 Table 2. DAMGO enhances amplitude and halfwidth of SCOs. We found that DAMGO enhanced the amplitude and halfwidth of SCOs in cultured neurons. However, the number of SCOs occurring within imaging frames were not significantly impacted by DAMGO. Statistical testing was performed with unpaired t-tests that compared fold changes in the DAMGO group with fold changes in the saline control group, after first being tested for normality with D’Agostino-Pearson’s test. Peak Number was significantly different from a normal distribution, and therefore tested with a Mann-Whitney U test.

2.3.5 The effect of DAMGO is mediated through both GABA receptors

Prior investigations by other groups (Drake & Milner, 1999, 2002, 2006; Férézou et al.,

2007; Huo et al., 2005; Madison & Nicoll, 1988; Stumm et al., 2004; Taki et al., 2000;

Witkowski & Szulczyk, 2006; Zieglgansberger et al., 1979) and our own lab (Dutkiewicz

& Morielli, 2020) into the mechanism of the µOR have suggested that the µOR acts primarily by inhibiting activity of inhibitory interneurons. However, other groups have indicated that the µOR is found on excitatory neurons as well (Rola et al., 2008; Schmidt et al., 2003; Witkowski & Szulczyk, 2006), including a series of experiments that used the SCO model in cultured hippocampal neurons (Przewlocki et al., 1999). Therefore, it was unclear whether DAMGO was suppressing inhibitory neurons and enhancing SCOs through disinhibition, or whether DAMGO was directly promoting the activity of excitatory neurons and enhancing SCOs through a more direct mechanism.

Our previous study showed that µOR inhibits primary culture neocortical interneurons (Dutkiewicz & Morielli, 2020). We therefore predicted that DAMGO augmented spontaneous calcium oscillations by through its effects on GABAergic 95 interneurons. In this disinhibitory model, we anticipated that blocking GABA receptors would prevent DAMGO from exerting its effects on SCOs. In contrast, if DAMGO acts by directly enhancing the activity of PNs, GABA receptor inhibitors should fail to prevent DAMGO from enhancing the calcium oscillations.

We tested this first by applying the GABAA antagonist, picrotoxin; if DAMGO was inhibiting release of GABA onto GABAA receptors, this drug should mimic

DAMGO in polarity. Picrotoxin alone (M = 2.31, SD = 1.64) significantly enhanced SCO amplitude (t(22) = 2.76, p = 0.006) when compared to the saline controls (M = 0.93, SD =

0.58). Picrotoxin also significantly enhanced calcium oscillation halfwidth compared to the change in control. However, picrotoxin (Mdn = 0.69) did not significantly (U = 62, p

= 0.295) enhance SCO number compared the change in control (Mdn = 0.71). Therefore, the effects of picrotoxin were similar in polarity to the effects of DAMGO and could suggest that the µOR downregulated GABAA receptors. To test this, we coapplied

DAMGO and picrotoxin together to determine whether DAMGO had an additive effect

(Figure 6). Effects of DAMGO with picrotoxin would suggest that DAMGO’s effects were not solely mediated by downregulated GABAA receptors.

We found that the combination of picrotoxin and DAMGO (M = 1.81, SD = 1.08) did not result in a significantly different (t(22) = 0.89, p = 0.192) change in SCO amplitude versus picrotoxin alone (M = 2.31, SD = 1.64). However, the combination of picrotoxin and DAMGO (M = 4.51, SD = 2.10) did significantly (t(22) = 1.86, p = 0.039) enhance SCO halfwidth compared to the fold change in picrotoxin alone (M = 3.10, SD =

1.61). This result suggested that other factors may be involved than reduced GABA

96 release onto GABAA receptors, which picrotoxin should have blocked. The combination of picrotoxin and DAMGO (M = 0.55, SD = 0.26) did not significantly (t(22) = 0.89, p =

0.191) change SCO number when compared to picrotoxin alone (M = 0.68, SD = 0.42).

Therefore, neither the µOR nor GABAA receptors do not to appear to alter the number of

SCOs (Figure 6k and Figure 7e).

Picrotoxin with DAMGO failed to completely block the DAMGO effect; this could mean that the µOR directly excites PNs, or that the effect of DAMGO is mediated through both GABAA and GABAB receptors. We hypothesized that blocking both

GABAA and GABAB receptors would block the DAMGO effect based on studies by other labs that suggest that the µOR is expressed by GABAB-stimulating neurogliaform neurons (Lafourcade & Alger, 2008; Olah et al., 2007).

To determine whether DAMGO was mediating its excitatory effect through downregulation of GABAA + GABAB receptors or direct excitation of PNs, we tested

DAMGO in combination of picrotoxin and the GABAB antagonist CGP55845. We found that this combination of GABAA receptor blockade and GABAB receptor blockade

DAMGO (M = 2.65, SD = 1.04) did not result in significantly different (t(22) = 1.71, p =

0.51) fold change in halfwidth versus CGP55845 and picrotoxin themselves (M = 2.04,

SD = 0.65), which indicated that this drug combination could block the DAMGO effect where picrotoxin alone did not (Figure 7c,d). This finding was consistent with our hypothesis, that DAMGO was mediating its SCO-enhancing effect by suppressing interneuronal release of GABA onto both GABAA and GABAB receptors, and not

GABAA receptors alone.

The combination of CGP55845 + picrotoxin + DAMGO (M = 1.76, SD = 0.65) also did not result in a significantly different change (t(22) = 1.28, p = 0.106) in SCO amplitude versus CGP55845 + picrotoxin (M = 2.25, SD = 1.14; Figure 7b). This was consistent with Figure 7a, because picrotoxin (without CGP55845) itself could prevent this DAMGO effect that we observed. Unexpectedly, we also found that CGP55845 + picrotoxin + DAMGO (M = 0.40, SD = 0.24) resulted in a significantly fewer (t(22) =

2.03, p = 0.027) SCOs versus CGP55845 and picrotoxin alone (M = 0.82, SD = 0.68).

GABAA GABAA+GABAB Blockade Blockade b) a) Amplitude Amplitude

Picrotoxin+DAMGO Picrotoxin+CGP+DAMGO ns ns Picrotoxin Picrotoxin+CGP ** *** Saline Saline

0 2 4 6 0 1 2 3 4 5 Fold change from baseline Fold change from baseline c) Halfwidth d) Halfwidth

Picrotoxin+DAMGO * Picrotoxin+CGP+DAMGO Picrotoxin ns *** Picrotoxin+CGP Saline *** Saline

0 5 0 5 1 1 Fold change from baseline 0 1 2 3 4 5 Fold change from baseline e) Number of Peaks f) Number of Peaks

Picrotoxin+DAMGO Picrotoxin+CGP+DAMGO ns * Picrotoxin Picrotoxin+CGP ns ns Saline Saline

0 1 2 3 0 1 2 3 Fold change from baseline Fold change from baseline

Figure 7. Blockading both GABARs prevents DAMGO-enhancement of SCOs. We tested for the effect of DAMGO when GABA receptors were blockaded. (Left column) We co-administered (100 µM) picrotoxin with DAMGO. We found that picrotoxin alone produced significant increases in (a) peak amplitude and (c) halfwidths of calcium peaks. Co-administering DAMGO with picrotoxin blocked the effect of DAMGO in (a) amplitude not in (c) halfwidth. (Right column) We therefore coapplied DAMGO with both picrotoxin and the GABAB antagonist CGP55845 and found that this drug combination blocked the DAMGO effect (p in both (b) amplitude as well as (d) halfwidth. (e,f) Meanwhile, we found no significant effects in number of calcium peaks in either condition. (First and Second column) Lines connect the same neuron’s pre and postdrug value for each individual replicate. (Third and Fourth column) Edges of the boxes are drawn at the 1st and 3rd quartiles while the “whiskers” connect the largest and smallest values for that group. We tested the data with D’Agostino-Pearson omnibus K2 test (p > 0.05) to ensure a normal distribution. Figure (d) was analyzed with Mann-Whitney U test (one-tailed) due to a nonnormal distribution, while the rest of the comparisons were executed with unpaired t tests (one-tailed).

Calcium Calcium (F-F0/F0) (F-F0/F0)

Frames (4 frames/sec) Frames (4 frames/sec) Pre-DAMGO Post-DAMGO

Calcium Calcium (F-F0/F0) (F-F0/F0)

Frames (4 frames/sec) Frames (4 frames/sec) Pre-Picrotoxin Post-Picrotoxin

Figure 8. Representative calcium peaks for various drug conditions. (Top Row) Before and after DAMGO. (Bottom Row) Before and after picrotoxin. Y axes are calcium indicator fluorescence, when the highest- intensity calcium peak in the pre-drug timestack (not pictured) is set to a value of 1.00, and all other peaks are a proportion of the tallest peak. X axes represent imaging frames (4 frames/sec). On the calcium peaks, yellow lines denote the base of each recognized peak, green horizontal lines are the measures of halfwidth

99 for the peaks, and the red dot at the tips of the peaks are indicating that the peak has been successfully culled for analyses by the Python script.

2.3.6 Interneurons and Pyramidal Neurons

Our results suggested that interneurons were being downregulated by DAMGO, which caused the PNs to become hyperactive in response to reduced activity of GABA receptors. If this is true, we might expect activity of interneurons to decrease, as activity of excitatory neurons to increase. However, the data described so far did not discriminate between inhibitory and excitatory types of neurons, and we could not determine if activity of interneurons in this preparation was decreasing despite the net enhancement of

SCOs. We therefore separately labelled interneurons to determine whether their SCOs were differentially regulated from PNs by the drugs in this model.

100

Figure 9. Cultured neurons transformed with AAV-mDlx-NLS-mRuby2. Brightfield image of cultured rat neurons. (Top Right) Fluorescence of mRuby2 in red. (Bottom) Overlay of both images to show the neocortical interneurons in the field. Neurons at DIV (40x magnification).

Neocortical interneurons only constitute 10-25% of all neocortical interneurons, while excitatory neurons constitute the remaining proportion (Beaulieu, 1993; Jones,

1993; Meyer et al., 2011; Ren, Aika, Heizmann, & Kosaka, 1992). We therefore expected that interneurons in these cultures would comprise a similar proportion. To identify interneurons in our model, we exposed the cultured neurons to an AAV that induces expression of a red fluorescent protein in interneurons by using the interneuron-specific enhancer Dlx5/6. We manually counted neurons in 110 images across 3 independent cultures and found that interneurons constituted 432 neurons out of a total of 2428 neurons (17.8%). We also used a separate approach with a Python script that automatically enumerated neurons and found that interneurons constituted 1037 out of

6461 neurons (16.1). Collectively, these data show that interneurons in our culture system comprised the expected 10-25% range (Beaulieu, 1993; Jones, 1993; Meyer et al., 2011;

Ren et al., 1992).

101

2.3.7 Second dataset experimental design

In our first dataset, we observed that picrotoxin alone could significantly enhance SCO halfwidth and amplitude (Figure 7) and its effect outweighed the effects of DAMGO.

While this was expected, the magnitude of picrotoxin’s effect in that model made the effect of DAMGO difficult to separate from the larger effect of picrotoxin. To address this, we instituted a change in the drug exposure protocol by applying picrotoxin for 15-

25 minutes before applying DAMGO (i.e., a “background” of picrotoxin). While this drug application procedure did depart from our earlier method, it also conferred the benefit of testing these effects in a slightly different way to reinforce their validity.

Calcium-Imaging (Second set) Experiment Model

Addition of a single drug to the preincubated background

) )

d d

r r

o o

c c

e e

r r

t t

r r

d d a

Preincubation of Slot 1 Drug a

n n

t t

s e s e

( ( (

(Extends 15-25 minutes prior (

1 2

to slot 1 recording)

t t t t

o o o o

l l l l

S S S S

0 60 120 180 240 300 360 420 480 540 600 660 720 780 slot1 slot2

Fold  Due to Drug (or time) = Slot2/Slot1; Final/Initial Time (s)

Figure 10. Second dataset experimental design. For our next dataset, we administered a preincubated background of (100 µM) GABAA-specific inhibitor picrotoxin or picrotoxin + the GABAB specific inhibitor CGP55845 (10 µM), and then applied (3 µM) DAMGO. This mitigated the more drastic effects of

102 combining DAMGO with the GABAR blockers, in order to further isolate the effect of DAMGO from the effects of picrotoxin and CGP55845.

2.3.8 Changes similar in interneurons and excitatory neurons

Based on our data, we believed that interneurons were specifically being suppressed, which allowed excitatory PNs to enhance their activity. We therefore hypothesized that interneurons would have diminished SCOs when compared to enhanced SCOs in PNs. To determine whether DAMGO was differentially affecting inhibitory versus excitatory neurons, we tested for DAMGO effects we previously found in here separately in the populations of neurons (Figure 11).

Post-hoc comparisons using Tukey’s multiple comparisons test revealed that significant effects of DAMGO on SCO halfwidth could be found in both interneurons (M

= 1.98, SD = 0.58) and noninterneurons (M = 1.72, SD = 0.64) relative to inhibitory and excitatory controls exposed to vehicle saline (M = 1.09. SD = 0.25; M = 1.20, SD = 0.35).

Therefore, the effect of DAMGO was not significantly different between excitatory and inhibitory neurons. Unexpectedly, the SCOs in interneurons and PNs were similarly upregulated, and augmentation of SCO halfwidth could be found in both neuronal types.

We next investigated whether this DAMGO enhancement of SCO halfwidth was again distinguishable in picrotoxin, and absent in the presence of picrotoxin and

CGP55845. These findings would replicate our conclusion that DAMGO suppressed

GABA release onto GABAA and GABAB receptor. Consistent with this, administering

DAMGO in a background of picrotoxin resulted in significantly larger halfwidths in both interneurons (M = 1.33, SD = 0.39) and noninterneurons (M = 1.49, SD = 0.41) relative to

103 their no-DAMGO vehicle controls (M = 0.99, SD = 0.16; M = 1.13, SD = 0.14). This showed again that the effect of DAMGO was not solely mediated by downregulation of

GABAA receptors. We therefore tested whether DAMGO-enhancement of SCO halfwidth was blocked in a background of picrotoxin and CGP55845.

The effect of DAMGO was blocked only in the presence of both picrotoxin and

CGP55845 in both interneurons (M = 1.01, SD = 0.29) and excitatory neurons (M = 1.15,

SD = 0.27) in comparison to their respective controls (M = 1.25, SD = 0.56; M = 1.28, SD

= 0.56).

In summary, found that interneurons and excitatory neuronal SCO durations were similarly enhanced in response to DAMGO. However, we also found here that DAMGO enhancement of SCO halfwidth was still present in a background of picrotoxin, but not a background of picrotoxin + CGP55845. Therefore, we were able to replicate our main finding from the first dataset – that DAMGO’s effects were mediated through cortical interneurons and both GABAA and GABAB receptors.

104

DAMGO DAMGO DAMGO in GABA +GABA blockade in GABAA blockade A B b) a) Change in Amplitude Change in Amplitude c) Change in Amplitude

n Blank (Excitatory) n Blank (Excitatory)

Blank (Excitatory) n

s Blank (Interneurons) s Blank (Interneuron) Blank (Interneuron)

✱✱

DAMGO (Excitatory) DAMGO (Excitatory) n

DAMGO (Excitatory) n

DAMGO (Interneurons) DAMGO (Interneuron) DAMGO (Interneuron)

0 1 2 3 4 5 0 5 0 5 0 1 2 3 0. 0. 1. 1. Fold change from baseline Fold change from baseline Fold change from baseline

d) Change in Halfwidth e) Change in Halfwidth f) Change in Halfwidth

Blank (Excitatory) n Blank (Excitatory) n

Blank (Excitatory) n

✱

✱ Blank (Interneurons) Blank (Interneuron) Blank (Interneuron) s

✱✱✱

✱

DAMGO (Excitatory) n DAMGO (Excitatory) n

DAMGO (Excitatory) n

DAMGO (Interneurons) DAMGO (Interneuron) DAMGO (Interneuron)

0 1 2 3 4 0 1 2 3 0 1 2 3 4 Fold change from baseline Fold change from baseline Fold change from baseline

g) Change in Number of Peaks h) Change in Number of Peaks i) Change in Number of Peaks

Blank (Excitatory) n Blank (Excitatory) n

Blank (Excitatory) n

n n

s Blank (Interneurons) s Blank (Interneuron) Blank (Interneuron)

n n

DAMGO (Excitatory) n CGP55845 (Excitatory) n

DAMGO (Excitatory) n

DAMGO (Interneurons) DAMGO (Interneuron) CGP55845 (Interneuron)

0 1 2 3 4 0 5 0 5 0 0 5 0 5 0 0. 0. 1. 1. 2. 0. 0. 1. 1. 2. Fold change from baseline Fold change from baseline Fold change from baseline

Figure 11. DAMGO enhancement of SCOs observed in both neuron types. To test whether DAMGO- induced changes in SCOs were similar in both neuronal types, we tested DAMGO in the presence of GABAergic blockade. (Left column) Changes in a background of saline. (Middle column) Changes due to DAMGO in a 10-minute preincubation with the GABAA-antagonist picrotoxin. (Right column) Changes due to DAMGO after a 10-minute preincubation with picrotoxin and the GABAB-antagonist CGP5845. We found DAMGO enhanced (d) SCO halfwidth, and not (g) number of SCOs. Unlike the previous dataset, however, there were not significant differences in SCO amplitude in this dataset. (e) In the presence of picrotoxin, DAMGO again enhanced SCO halfwidth, and failed to enhance their (b) amplitude and (h) number which recapitulated our earlier dataset. In the presence of picrotoxin and CGP55845, DAMGO once again failed to change (c) amplitude (g) halfwidth, and (j) number, which also recapitulated our earlier findings. All groups include the same number (n = 13) trials. Mean fold changes were tested by comparing mean fold changes in properties relative to other cell types and the same cell type in the vehicle condition. Statistics were analyzed using Sidak’s test for multiple comparisons.

105

2.3.9 Effect of DAMGO in picrotoxin similar to GABAB blockade

Our data had indicated that DAMGO still enhances SCO halfwidth in the presence of picrotoxin in both datasets, and that the DAMGO enhancement was found in both excitatory and inhibitory neurons. We next investigated whether CGP55845 in a background of picrotoxin resemble the effect of DAMGO in the same background, which would be consistent with our hypothesis that DAMGO reduces activity at both GABAA and GABAB receptors.

a) Change in Amplitude

Blank (Excitatory)

✱

Blank (Interneuron) s

CGP55845 (Excitatory) n

s CGP55845 (Interneuron)

.0 .5 .0 .5 .0 0 0 1 1 2 Fold change from baseline b) Change in Halfwidth

Blank (Excitatory) n

✱✱✱

s Blank (Interneuron)

✱✱✱

CGP55845 (Excitatory)

s CGP55845 (Interneuron)

0 1 2 3 4 Fold change from baseline

c) Change in Number of Peaks

Blank (Excitatory) n

✱✱✱

s Blank (Interneuron)

✱✱✱

CGP55845 (Excitatory) n

s CGP55845 (Interneuron)

.0 .5 .0 .5 0 0 1 1 Fold change from baseline

Figure 12. Effect of GABAB inhibition. We found that application of the GABAB antagonist CGP55845 in a background of picrotoxin resulted in no significant changes in (a) SCO amplitude. However, we found significant increases in (b) SCO halfwidth after CGP55845 in both interneurons and excitatory neurons. We also found significant reductions in (c) SCO number after CGP55845 in both neuronal types. Mean fold 106 changes were tested by comparing mean fold changes in properties relative to other cell types and the same cell type in the vehicle condition. Statistics were analyzed using Sidak’s test for multiple comparisons.

In a background of picrotoxin, CGP55845 produced no significant change in SCO amplitude in interneurons (M = 0.57, SD = 0.24) or excitatory neurons (M = 0.82, SD =

0.23) relative to their vehicle controls (M = 0.50, SD = 0.25; M = 0.81, SD = 0.39). We found that addition of CGP55845 produced a significant increase in SCO halfwidth relative in both interneurons (M = 1.80, SD = 0.65) and excitatory neurons (M = 1.91, SD

= 0.81) relative to controls (M = 0.98, SD = 0.17; M = 1.04, SD = 0.14). These effects were consistent with our DAMGO findings in a background of picrotoxin (Figure 7c), which showed that DAMGO produces an increase in SCO halfwidth.

We observed a consistent and significant trend towards fewer SCOs upon application of CGP55845 (Figure 11c). While SCO number did not generally change throughout our previous experiments, the number of SCOs was significantly reduced in both interneurons (M = 0.60, SD = 0.28) and excitatory neurons (M = 0.58, SD = 0.21) versus their respective controls (M = 0.96, SD = 0.21; M = 0.96, SD = 0.21) under this condition.

Taken together, these results (Figure 11) do show similar effects of DAMGO in a background of picrotoxin; enhancement of SCO halfwidth but not amplitude, which is further evidence that DAMGO reduces GABA release onto both GABAA and GABAB receptors in neocortical cultures.

107

2.4 Discussion

We report here that µOR agonism upregulates spontaneous activity in primary neocortical cultures; on average, we observed that DAMGO enhanced SCO duration by

63% (Table 2). These results therefore strongly suggest that the µOR excites neocortical circuits, in part by enhancing durations of bursts of activity being indicated by SCOs. In addition, our results also implicate the µOR in a disinhibitory role by suppressing the activity of interneurons, as shown by our findings that the µOR’s enhancement of SCO can be blocked by coapplication with GABA inhibiters (Figure 7), or in a background of

GABA inhibitors (Figure 11). Given the ambiguity in the field about whether DAMGO provides disinhibition, or direct excitation to PNs, these experiments provide additional evidence that the main network effect of µOR stimulation is due to inhibition of interneurons.

We speculate that this enhancement may indicate that cortical interneurons were slower in counteracting bursts of action potentials (being indicated by the SCO), which allowed pulses of excitatory activity to continue for longer before being suppressed by interneurons. This finding could suggest that the µOR may make the cultured networks of interneurons less responsive and able to provide feedback inhibition to local excitatory neurons. Therefore, these results may suggest that this receptor dysregulates networks of interneurons by reducing their ability to respond quickly to bursts of glutamatergic activity.

A previous study by another lab has found that DAMGO enhances SCOs in hippocampal cultures, but our findings contract it in 2 major ways (Przewlocki et al.,

108

1999). We performed simultaneous GABAA and GABAB receptor blockade to show that effects of DAMGO are ameliorated only when both receptors are blocked, and GABAA blockade fails to do that. Therefore, our main conclusions that DAMGO disinhibits neocortical networks appears to be well-supported in both datasets (Figure 7 and Figure

11). Our results also departed from those earlier findings in another way – we obtained fairly inconsistent results on DAMGO enhancement of SCO amplitude, but instead we observed that DAMGO consistently enhanced the duration (halfwidths) of the SCOs.

Our electrophysiological recordings indicate that SCOs generally are not caused by isolated APs, which appear to correspond to no detectable SCO activity (Figure 4).

Instead SCOs generally correspond to longer-lasting depolarizations that appear to be excitatory postsynaptic potentials. Our results therefore are similar to previous reports that used simultaneous electrophysiological recordings to show that SCOs correspond to excitatory postsynaptic potentials (Bacci et al., 1999; T. H. Murphy et al., 1992;

Przewlocki et al., 1999). In addition to that, we also found that depolarizing current applications are capable of causing calcium peaks, though spontaneous activity could trump those induced calcium peaks (Figure 4). This suggests that the causal arrow could point both ways; bursts of action potentials can cause SCOs, while our data discussed to this point also suggest that EPSPs themselves can induce both bursts of action potentials and SCOs that correlate with them. We also observed a possible refractory period (Figure

4) towards the end of the calcium indicator fluorescence traces, which could suggest that

CICR phenomena may be involved with the calcium peaks being induced by the current

109 application. In summary, SCOs correspond to bursts of activity, but SCOs can be proximally caused by different phenomena.

We investigated whether this model can be used to show inhibition of interneurons and overactivity of PNs upon application of DAMGO. Yet, our results showed uniformly that the SCOs were not significantly different between the two types in various drug conditions. How do we reconcile this with disinhibition? On one hand, some authors show that SCOs are constituted of a voltage-sensitive calcium channel component (Bacci et al., 1999; Dravid & Murray, 2004; Inglefield & Shafer, 2000).

However other authors show that SCOs are mostly a result of glutamatergic input (Cao et al., 2015; Dravid & Murray, 2004; Inglefield & Shafer, 2000; T. H. Murphy et al., 1992;

Nuñez et al., 1996; Przewlocki et al., 1999; Shen et al., 1996). Our own results support the latter model – that the SCO is a result of excitatory drive being input into each neuron, which would explain why SCOs are similar in interneurons and excitatory neurons, because nearby neurons receive the same input. This explanation would further suggest that interneurons are unable to provide adequate inhibition despite the increase in excitation going into both neuronal types; electrophysiological experiments from neocortical interneurons show that DAMGO reduces frequency and duration of action potentials elicited by same-magnitude depolarizations from an electrode(Dutkiewicz &

Morielli, 2020; Férézou et al., 2007). Therefore, these electrophysiological data are consistent with the notion that interneurons are downregulated, despite experiencing greater glutamatergic input being indicated by the SCOs. Although GABAA receptors are the main cortical GABA receptor, our results indicated that the DAMGO-enhancement of

110 duration calcium oscillations was successfully blocked only by combined GABAA and

GABAB receptor inhibitors and failed to be blocked by only GABAA blockade in both of our datasets (Figure 7 and Figure 11). While cortical GABAA receptors are plentiful,

GABAB receptors are confined to synaptic and extrasynaptic spaces in neurons that are most strongly activated by neurogliaform neurons (Olah et al., 2007; Tamás et al., 2003).

These neurogliaform neurons are shown to express high rates of the µOR in the cortex

(Férézou et al., 2007) and hippocampus (Krook-Magnuson et al., 2011). Therefore, our data that µOR+ neurons release GABA onto GABAA and GABAB receptors are consistent with previous experiments by other labs (Olah et al., 2007; Tamás et al., 2003) and consistent with studies that suggest that neurogliaform neurons are µOR+ (Férézou et al., 2007; Krook-Magnuson et al., 2011). These experiments suggest that this phenomenon can be reproduced in the SCO model.

We found that picrotoxin blocked the DAMGO-enhancement of calcium oscillation amplitude, even though the DAMGO effect of picrotoxin on duration

(halfwidth) persisted. One possible reason is that picrotoxin had so profoundly amplified the intensity of calcium oscillations, that DAMGO did little to add to the height of the calcium peaks. This may happen, for instance, if the picrotoxin condition fully recruited the NMDA receptors by sufficiently releasing glutamatergic neurons from enough inhibition to promote glutamate release, which itself may in turn further activate other glutamatergic neurons. Previous experiments by other labs have related SCO amplitude to the activity of NMDA receptors, which is expected to promote greater Ca2+ entry during postsynaptic potentials (Bacci et al., 1999; Inglefield & Shafer, 2000; T. H.

111

Murphy et al., 1992; Nuñez et al., 1996; Przewlocki et al., 1999; Shen et al., 1996). This can potentially pose a limitation to this study, because any potential positive regulation of

NMDA receptors may be overlooked by the overwhelming enhancing effect of picrotoxin on SCO amplitude (Figure 7a), though this does not appear to mask the effect of

DAMGO on SCO halfwidth (Figure 7c and 11e).

There were several effects that we found in our first data set (Figure 6 and Figure

7) but were not found when a similar experiment was repeated in the second dataset

(Figure 11). Firstly, we initially found an effect on SCO amplitude due to DAMGO

(Figure 6c), however this effect was not present in the second data set (Figure 11a).

DAMGO’s effect on SCO amplitude appeared to be weaker and less frequent than its effect on SCO halfwidth, which may have led to the inconsistent results. Secondly, we initially found a statistically significant reduction in SCO number in the first dataset

(Figure 7f). We initially believed that DAMGO’s effect under the combination may be limited to cortical interneurons and suppressed SCO number in interneurons, however the effect was not found in interneurons (nor in noninterneurons) in the second dataset

(Figure 11i). Therefore, it seems more likely that this was a statistical anomaly. While the failure to replicate these findings present a more complicated picture of DAMGO’s effects, the repetition of these experiments at least may provide more validity and insight for follow-up studies. For example, the effect of DAMGO on SCO amplitude may be less consistent. However, both datasets show clearly that DAMGO shows a picrotoxin- resistant enhancement of SCO duration (Figure 7c and Figure 11e), and that all effects of

DAMGO are blocked by coapplication (Figure 7d) or preincubation (Figure 11f) with a

112 combination of picrotoxin and CGP55845. Therefore, our conclusions that DAMGO promotes excitation secondary to inhibition of interneurons appears to be well-supported by our data, and by data from other labs in different models that suggests µORs localize to GABAergic interneurons (Drake & Milner, 1999, 2002, 2006; Férézou et al., 2007;

Huo et al., 2005; Madison & Nicoll, 1988; Stumm et al., 2004; Taki et al., 2000;

Witkowski & Szulczyk, 2006; Zieglgansberger et al., 1979).

One lingering issue is the neurophysiological mediators of SCO amplitude and duration (halfwidth). Previous experiments on SCOs, which generally use low ACSF

[Mg2+], have consistently linked SCO amplitude with activity of various postsynaptic glutamate receptors (Bacci et al., 1999; Nuñez et al., 1996; Przewlocki et al., 1999; Shen et al., 1996). SCO duration (halfwidth, in our experiment) has only occasionally been quantitatively analyzed but, it too, may be modulated by NMDAR activation because

[Mg2+] manipulations can modify it (T. H. Murphy et al., 1992). Our data show that

GABAB receptors modulate SCO duration, though the mechanism requires some speculation. One conceivable explanation is that the prolonged glutamatergic stimulation indicated by the SCO activates cortical interneurons which provide an immediate GABA release. However, the background of picrotoxin prevented GABAA receptors prevented a quick counterbalance, and the eventual termination of the SCO was brought about through the slower action of GABAB receptors. This explanation is corroborated by our data that show that DAMGO or CGP55845 in a background of picrotoxin both significantly enhance the SCO halfwidths (Figure 11 and Figure 12), and support our main conclusion that DAMGO is downregulating GABA release onto both GABAA and

113

GABAB receptors, which is consistent with their known role in suppressing the GABAB- stimulating neurogliaform neurons of the neocortex (Craig & McBain, 2014; Olah et al.,

2007; C. J. Price et al., 2008).

The downstream physiological origin of these Ca2+ oscillations is presently unknown, and the various mechanisms proposed to contribute to the calcium transients may be responsible for this uncertainty. Calcium transients have previously been linked the development and gene expression in neurons (Dolmetsch et al., 1998; Spitzer et al.,

1995). They are also believed to confer a neuroprotective role against trauma (Geddes-

Klein et al., 2006).

CHAPTER 3: µ OPIOID RECEPTORS MODULATE ACTION POTENTIAL KINETICS AND FIRING FREQUENCY IN NEOCORTICAL INTERNEURONS

3.1 Introduction

The endogenous opioid system of the cerebral cortex mediates antinociception and reward valuation (Choi et al., 2016; Ong et al., 2019; Qiu et al., 2009; Zubieta et al.,

2001). Dysregulation of this system is believed to contribute to pathological and compulsive behaviors such as eating disorders, pathological gambling, and drug-seeking

(Ashok et al., 2019; B. Bencherif et al., 2005; Giuliano & Cottone, 2015; Joutsa et al.,

2018; Mick et al., 2016; Mitchell et al., 2012; Qu et al., 2015; Zubieta et al., 1996).

Experimentally, impulsive behaviors and binge-eating can be induced through infusion of the µ opioid receptor (µOR) specific agonist [D-Ala2, N-Me-Phe4, Gly5-ol]-Enkephalin

114

(DAMGO) into the frontal cortex of animal models (Mena et al., 2011; Selleck et al.,

2015). Blockade of µORs with naltrexone inhibits compulsive behaviors (Bartus et al.,

2003; Blasio et al., 2014). These aberrancies are believed to result from the disruption of activity within cortical networks by µORs (Haider et al., 2006; M. Whittington et al.,

1998; Y. Zhang et al., 2019). More specifically, µORs appear to dysregulate cortical networks by suppressing GABAergic activity and thereby altering the balance of excitation and inhibition in the cortex (David A. Lewis et al., 2012; Shi et al., 2020; Volk et al., 2012). The µOR is believed to suppress GABAergic signaling through its expression primarily on cortical interneurons. This leads to overactivity of the targets of their inhibition, the glutamatergic Pyramidal Neurons (PNs) (Drake & Milner, 1999,

2002, 2006; Férézou et al., 2007; Huo et al., 2005; Madison & Nicoll, 1988; Stumm et al., 2004; Taki et al., 2000; Witkowski & Szulczyk, 2006; Zieglgansberger et al., 1979).

However, some research suggests that µOR may activate PNs directly as well (Rola et al.,

2008; Schmidt et al., 2003; Witkowski & Szulczyk, 2006). Immunohistochemical experiments have found high rates of expression of µOR in interneurons that express vasoactive-intestinal peptide (VIP+) (Taki et al., 2000). Electrophysiological and sc-PCR data in neocortical neurons have supported this finding (Férézou et al., 2007) and have also implicated neurogliaform neurons as expressing this receptor (Férézou et al., 2007;

Lafourcade & Alger, 2008; McQuiston, 2008). Evidence for µOR expression in somata of parvalbumin-positive (PV+) interneurons of the hippocampal formation is fairly clear

(Bartos & Elgueta, 2012; S. B. Bausch et al., 1995; Drake & Milner, 1999, 2002, 2006;

Stumm et al., 2004; Torres-Reveron et al., 2009), however perisomatic expression of this receptor in neocortical PV+ neurons has been investigated but, to our knowledge, has not 115 been reported (Férézou et al., 2007; Taki et al., 2000), though recent evidence shows that

µORs are sometimes found in PV+ axon terminals of the frontal cortex (C. Jiang et al.,

2019; Lau et al., 2020) and insular cortex (Yokota et al., 2016). Therefore, neocortical

VIPergic, neurogliaform, and some PV+ interneurons are generally believed to express

µORs – though not necessarily in their somata.

In addition to the uncertainty surrounding the µOR’s localization, there is also some mystery surrounding its mechanism by which µORs affect interneuron activity. The

µOR is a G-protein coupled receptor and has been shown to activate potassium- conducting inwardly rectifying K (GIRK) channels (Henry et al., 1995; Ikeda et al., 2000;

Loose & Kelly, 1990; Marker, Lujan, Loh, & Wickman, 2005). Regulation of GIRK channels and hyperpolarization are characteristic features of various ORs (Ikeda et al.,

2003). Agonist-induced hyperpolarization has been found in cortical µOR+ neurons

(Férézou et al., 2007; Glickfeld et al., 2008; Madison & Nicoll, 1988; E. Tanaka & North,

1994) non-cortical µOR+ neurons (Grudt & Williams, 1993; Harris & Williams, 1991;

Johnson & North, 1992; Kelly, Loose, & Ronnekleiv, 1990; Lagrange, Ronnekleiv, &

Kelly, 1994; Lagrange, Rønnekleiv, & Kelly, 1995; Loose & Kelly, 1990; R. A. North,

Williams, Surprenant, & Christie, 1987; Sugita & North, 1993), and is also a feature of other opioid receptors as well (B Chieng & Christie, 1994; Chiou & Huang, 1999; Grudt

& Williams, 1993; Johnson & North, 1992; Loose & Kelly, 1990; Loose, Ronnekleiv, &

Kelly, 1990; Madison & Nicoll, 1988; R. A. North et al., 1987; Sugita & North, 1993).

However, hyperpolarization does not always occur in response to DAMGO (Faber &

Sah, 2004; Wimpey & Chavkin, 1991), possibly due to an incomplete overlap between

116

GIRK channel expression and µORs, which may be the case in the cortex (S. B. Bausch et al., 1995). Additionally, reductions in spontaneous APs with µOR-agonism is also found in various parts of the brain, though not all µOR+ neurons fire spontaneously

(Loose & Kelly, 1990; Mitrovic & Celeste Napier, 1998; Ponterio et al., 2013).

Hyperpolarization and decreased spontaneous activity are both commonly found in response to the activation of µOR and other opioid receptors, and the two effects often coincide (Chiou & Huang, 1999; Elghaba & Bracci, 2017; Kelly et al., 1990; Loose &

Kelly, 1990; R. A. North et al., 1987). While it is possible that hyperpolarization could induce a decrease in spontaneous APs by increasing the distance to Vm threshold for action potentials, some studies in cortical neurons have found reductions in spontaneous

APs with only small accompanying hyperpolarization (Krook-Magnuson et al., 2011; M.

E. Sheffield et al., 2013; M. E. J. Sheffield et al., 2011; Norimitsu Suzuki et al., 2014).

Therefore, the µOR may induce hyperpolarization of varying magnitude along with reductions in tonic APs, and one effect could occur independently of the other.

In addition to hyperpolarization and reductions in spontaneous APs, research from other parts of the brain suggests that µORs modulate αDendrotoxin-sensitive channels;

αDTX inhibits Kv1.1, Kv1.2, and Kv1.6 channels (Ponterio et al., 2013). Experiments have found that µORs modulates αDTX-sensitive channels in the basolateral amygdala

(Finnegan et al., 2006), periaqueductal gray (Vaughan et al., 1997), as well as thalamocortical terminals within the frontal cortex (Lambe & Aghajanian, 2001).

Although these channels are known to be expressed by several families of neocortical interneurons (Casale, Foust, Bal, & McCormick, 2015; Goldberg et al., 2008; Golding et

117 al., 1999; T. Li et al., 2014; Porter et al., 1998), they have not been directly investigated for mediating the µOR’s inhibitory effect in the neocortical interneurons. We therefore predicted that µORs activate αDTX-sensitive channels to inhibit neocortical interneurons.

Several studies have investigated the role of the αDTX-sensitive channels by analyzing its effects on action potentials. While Kv1.1, Kv1.2, and Kv1.6 channels have been shown to modulate spike frequency and AP threshold, (Bekkers & Delaney, 2001;

Faber & Sah, 2004; Finnegan et al., 2006), and some studies have suggested that they contribute to AP shape as well. Specifically, researchers have found that αDTX-sensitive channels hasten repolarizations and shorten durations of APs (Geiger & Jonas, 2000;

Pathak, Guan, & Foehring, 2016), including in neocortical interneurons (Casale et al.,

2015). We therefore predicted that part of the mechanisms by which the µOR act in neocortical interneurons is through one or a combination of these αDTX channel- mediated electrophysiological effects. To investigate this, we cultured neocortical neurons and performed patch-clamp electrophysiology on interneurons in current-clamp mode to stimulate and measure their APs. To measure the kinetics of APs, we created

Python scripts to measure 54 membrane properties, AP kinetic properties, and ratios. We were primarily interested in 7 properties that have previously been implicated in mediating DAMGO or αDTX effects in neurons. We predicted that DAMGO and αDTX would modulate (in opposite polarity) resting membrane potential, AP threshold, number of evoked APs, interspike interval, AP halfwidth, maximum repolarization rates, amplitude of afterhyperpolarizations.

118

3.2 Materials and methods

All procedures were authorized by the Institutional Animal Care and Use Committee

(IACUC) at University of Vermont. We dissected CD® IGS Sprague-Dawley (Charles

B27, Penicillin-Streptomycin and Glutamax (All from Gibco), in a humidified 37° C incubator with 5% CO2. Cultures were transformed on day in vitro 1 with an AAV-mDlx-

NLS-mRuby2 to induce expression of a red fluorescent protein in cortical interneurons.

Culture media was half-replaced every 3 days. Experiments commenced between day in vitro 16-37. AAV-mDlx-NLS-mRuby2 was a gift from Viviana Gradinaru (Addgene viral prep # 99130-AAV1); http://n2t.net/addgene:99130; RRID:Addgene_99130).

3.2.2 Whole-Cell Recordings

Cultured neurons were transferred from their growth media into a chamber perfused with

ACSF (in mM: NaCl, 126; KCl, 5; NaH2PO4, 1.25; CaCl2, 2; MgCl2, 1; NaHCO3, 26;

Glucose, 20; and pyruvate,5 (Férézou et al., 2007). The ACSF was warmed to 30oC and constantly bubbled with a mixture of 95% Oxygen and 5% CO2. When indicated, both

CNQX (20 µM; Sigma-Aldrich) and dAP5 (50 µM; Sigma-Aldrich) were included in the

ACSF, which was then constantly perfused through the recording chamber. Saline

119 controls were done alternately with drug condition recordings throughout the course of the experiments. α-Dendrotoxin (αDTX) was purchased from Alomone Labs (Jerusalem,

Israel).

Patch pipettes with resistances of 5-10MΩ were fabricated from borosilicate glass capillaries and filled with intracellular saline containing (in mM) K-gluconate, 144;

10kHz with a Digidata 1322a (Axon Instruments) DA converter. Passive membrane and

AP kinetics measurements were automated using a custom Python program

(cc_analysis.py) the source code for which is available at https://github.com/moriellilab/cc_analysis_46.

All 7 measures that we investigated in the hypotheses were manually tested- against script generated values. In all 7 cases, they significantly and positively correlated to values determined manually with the threshold peak-detector of Clampfit 10.7

(Molecular Devices). All analyses used measurements generated by the cc_analysis program, with the exception of spiking pattern, which was always determined manually.

Measure Correlation RMP r(7) =0.998, p < 0.001 AP Threshold r(7) = 0.982, p < 0.001 Number of APs r(7) = 1.000, p < 0.001 ISI r(7) = 0.999, p < 0.001 120

Halfwidth r(7) = 0.963, p < 0.001 Maximum AP Repolarization Rate r(7) = 0.998, p < 0.001 AHP Amplitude r(7) = 0.921, p < 0.001 Table 3. Manual measurements positively correlate with script-derived values. We validated the script- derived values by randomly selecting 3 recordings for validation by using all 3 of their timeslots. Therefore, an N = 9 files were analyzed in the script and with manual measurements using Clampfit 10.7. In all 7 measurements, values derived with the script were significantly and positively correlated with manual measurements. Values compared with a bivariate correlation and tested for significance (two-tailed; α = 0.05).

3.2.3 Current-clamp protocol

We selected neurons for recording that were fluorescent (red), reflecting their exposure to

AAV-mDlx-NLS-mRuby2 during culturing. We adjusted membrane potentials for a junction potential of -11mV. Only neurons with a healthy appearance were selected for recording. We generally used neurons that were more hyperpolarized than -45mV for recording, though primarily the inclusion criterion was primarily the ability to fire APs from their natural RMP after the application of a depolarizing current.

A diagram of the experiment is illustrated on Figure 15. Upon acquiring a whole- cell configuration, we recorded for 3 minutes to ensure that the RMP was stable and more hyperpolarized than -45mV, and that the neuron was capable of firing APs with a depolarizing current from its RMP. The 20 current steps (1s duration) we used to elicit

APs were adjusted for each recording during this waiting period, because neurons had slightly different RMPs, thresholds, and resistances that required modulating the current magnitude. Once calibrated for the neuron, we kept that setting constant for all 3 drug conditions for all timeslots for that neuron. We subsequently made the pre-drug recording, and then we applied these drugs through bath perfusion for 80s to ensure the concentration and response was stable. We then made the “post-DAMGO” (slot 2)

121 recording after these 80s of incubation, then applied the next drug buffer for 80s (a combination of DAMGO and αDTX), and then made the post-αDTX+DAMGO recording (slot 3). Neurons were passively recorded from between timeslots for Vm changes, but these data were not considered in the analyses. Coverslips were discarded and replaced after a single use.

We quantified spontaneous APs by counting APs that occurred outside of the 1s current pulse. This comprised a noncontiguous period totaling 60 seconds in each timeslot. Spontaneous APs were not analyzed beyond counting them because most

DAMGO-responders stopped firing spontaneous APs after DAMGO.

3.2.4 Measurements of AP and membrane properties

APs were evoked with a 1s current pulse of varying amplitude. These pulses occurred 3s apart. Resting membrane potential (RMP) was determined from the median Vm measured during a 330 ms period immediately before the 1s current pulse. when the 1s current pulse was not being applied. This was composed of a noncontiguous period surrounding the current pulse, totaling 6.6s.

AP Threshold was determined from steady-state Vm at the first current application that induces an action potential. Mean interspike intervals were likewise derived from measuring the mean value for the space between spikes. Number of evoked

APs was measured by averaging the number of APs also during episodes 15-18. The remaining parameters were likewise determined from the average value of episodes 15-

18.

122

AP kinetic values (halfwidth and maximum repolarization rates) were also derived from episodes 15-18 of the 20 current steps. We averaged the values for APs 3-5 on each of the 4 current steps, and then averaged across the current steps to derive the final value. The halfwidth was determined as the width (in ms) of each AP at half the height of the AP.

Figure 13. Sample AP diagram. AP kinetics were measured by the script and an example is shown here to illustrate the derivation of the measurements (Upward-facing blue arrow) Start of the AP demarcated at 10% of the maximum depolarization rate. Vm at this point represents AP threshold. (First Green Circle) Maximum rate of depolarization of the AP. (Second Green Circle) Maximum rate of repolarization of the AP. (Downward-facing blue arrow) 10% of maximum rate of repolarization, marking the end of the AP. (Yellow Diamond) Lowest point of the afterhyperpolarization. Afterhyperpolarizations were measured as the difference between the y value of the start of the AP with the lowest point of the afterhyperpolarization, out to a maximum distance of 2ms from the point where the repolarization phase returns to the same Vm as the start of the AP at the upward blue arrow. (Pink Horizontal Line) The halfwidth of the AP was measured at half the height of the AP from the threshold Vm of the AP at the upward blue arrow.

3.2.5 Data Display

123

The cc_analysis.py analysis program produced a Microsoft Excel spreadsheet with all measurements. To render the spreadsheet compatible with SPSS, we sorted the rows of recordings by filename and added in a row of column labels. SPSS syntax is compatible with the master spreadsheet included here. Additional columns were added to the excel spreadsheet that divided the post-DAMGO (or post-saline) scores by their initial (pre- drug) condition for each particular measure, which were ultimately graphed on box-and- whisker plots throughout this text. Every change in the post-DAMGO condition was some factor of 1 (initial) to show the fold change. Thus, the changes from baseline were determined for the DAMGO recordings and saline controls. While we had either 11 (H- responders) and 10 (S-responders), we randomly selected a group of comparable size to compare changes in the drug condition with entropic changes in the saline controls to identify the drug effect. In the box-and-whisker plots where a subset of 10 or 11 saline controls were required, we used Excel’s built-in random number generator function was used to sort the saline recordings for comparison groups.

3.2.6 Statistics

Sample sizes were initially chosen under the expectation that only 5-20% of neurons would show a hyperpolarizing effect with DAMGO. We estimated that around 55

DAMGO recordings should enable us to record from ~10 DAMGO responders. We also recorded from 21 saline controls to enable us to compare the changes in DAMGO- exposed neurons to negative controls; groups are not of matching sizes because we expected that the whole DAMGO group would not be compared to the saline controls since responders would only compose some of the DAMGO-exposed neurons.

124

Since we were examining for DAMGO or αDTX effects in 7 variables at a time, we first tested the combined variables in mixed-model repeated measures MANOVA tests first to mitigate the high (30.1%) family-wise error rate to an α = 0.05. If the omnibus test failed to produce significant results, we did not test each of the 7 variables individually (main hypotheses). However, we made exceptions for data in the post-hoc analyses of Table 12, since posthoc tests come with some assumptions of false positives.

All mixed-model repeated measures MANOVAs were conducted through IBM

SPSS 27.0.0.0. We defined the independent variable as the timeslot (within-subject variable), and the between-subjects factor as the drug group (either saline or the DAMGO sequence). For these tests, we compared the 11 DAMGO H-responders (or 10 S-

Responders) against all 21 saline-controls. For DAMGO effects we tested timeslot 1 and

2 data. For αDTX effects, we analyzed timeslot 2 and 3 data. We only reported

Slot*Group interaction effects; the effect of Slot (time) most likely reflected entropic decay of the neurons over the course of the recording, while the effect of Group likely reflected stochastic differences of the neurons between the 2 groups of neurons – neither of which were a topic of these analyses. Therefore, in our experimental design, only the

Slot*Group interaction was relevant to whether DAMGO or αDTX effects were present.

In instances where the Slot*Group interaction was significant, we followed up by individually testing each of the 7 measures.

Post-hoc analyses (Table 12) were executed similarly with mixed repeated measures MANOVAs by testing sets of related variables to provide organizational and analytical grouping. APs analyzed in these tests were also measured in episodes 15-18,

125 but first, second, and average (of APs 3-5) were analyzed separately. For instance, the first APs in episodes 15-18 were averaged across the episodes and analyzed separately from the other (second and later) APs in that episode. We excluded the 7 variables previously discussed in the main hypotheses to limit the posthoc analyses to data that were not previously analyzed. Results of their individual omnibus tests are noted in the figure legend for thoroughness, but we provided posthoc analytical data for individual with-subject contrasts regardless of whether their omnibus test was significant.

Statistical testing portrayed on the box-and-whisker plots and summary tables were conducted through in GraphPad Prism version 8.4.3 for Windows, GraphPad

Software, San Diego CA, USA www.graphpad.com. Statistical testing on fold changes proceeded by first testing for skewness and kurtosis using D’Agostino-Pearson omnibus

K2 test in GraphPad Prism. When the statistic was not significantly different from normal

(p > 0.05), we conducted an unpaired t-test (one-tailed) to compare the fold changes in the DAMGO group with the saline group. When the distribution of either group was nonnormal (p < 0.05) we instead did the comparison with a one-tailed Mann-Whitney U- test.

3.3 Results

3.3.1 Identifying responders

Cortical µOR is expressed only in certain subcategories of interneurons; some studies have estimated that only around 3-5% of all neocortical neurons, or 15-25% of

GABAergic interneurons, express µORs (Férézou et al., 2007; M. C. Lee et al., 2002;

Taki et al., 2000), though some have reported as many as 2/3rds of prefrontal GABAergic

126 neurons respond to µOR-agonism (Witkowski & Szulczyk, 2006). To identify and target

GABAergic interneurons in culture, we transformed neurons with an AAV (see Methods) that drove expression of the red fluorescent protein mRuby2 under the interneuron- specific Dlx5/6 enhancer (Batista-Brito et al., 2008; de Lombares et al., 2019; Fazzari et al., 2010).

Figure 14. Cultured neurons with AAV-mDlx-NLS-mRuby2. (Top Left) Brightfield image of cultured rat neurons. (Top Right) Fluorescence of mRuby2 in red. (Bottom) Overlay of both images to show the neocortical interneurons in the field. Neurons at DIV 19 at 40x magnification.

Neocortical interneurons generally only constitute between 10-25% of all neocortical neurons (with the balance being glutamatergic Pyramidal Neurons), depending on the model and methods used to measure the proportions (Beaulieu, 1993;

127

Jones, 1993; Meyer et al., 2011; Ren et al., 1992). To determine whether proportions of interneurons occurring in the culture model used here fell within this range, we manually counted the number of red and nonred neurons across 110 images in 3 independent cultures. We found that red neurons constituted 432 out of 2428 neurons (17.8%). In a second approach, we utilized a Python script that automatically counted the neurons. This automated approach gave similar results, with red neurons constituting 1037 out of 6461 neurons (16.1%). With either counting method, our estimations of the proportions of interneurons to non-interneurons in these dissociated neuronal cultures are typical for previous reports of intact or sliced animal neocortex.

Continuous Application of DAMGO

Continuous Application of DTX+DAMGO

) ) )

d d d

r r r

o o o

c c c

e e e

r r r

t t t

r r r

d d d

a a a

n n n

t t t

e e e

C s s s

( ( (

- ( ( (

1 1 2 2 3 3

t t t t t t

o o o o o o

l l l l l

Waiting and l

S S S S S S W calibration period

0 60 120 180 240 300 360 420 480 540 600 660 720 780 slot1 slot2 slot3

 Due to DAMGO  Due to DTX (or time) (or time) Time (s)

Figure 15. Time course for patch-clamp recordings. Upon achieving whole-cell configuration, we calibrated the amplitude of the current injections to that particular neuron (to factor in its RMP and input 128 resistance) during the waiting and calibration period. (20 µM) CNQX and (50 µM) D-AP5 were constantly perfused to block glutamatergic receptors. We performed the “Initial” (slot 1) recording after this period and applied (3 µM) DAMGO for 80s of incubation and perfusion. We then performed the “post-DAMGO” recording (slot 2) and applied (100 nM) αDTX+DAMGO for another 80s of incubation. After that, we finally collected the “post-αDTX+DAMGO” (slot 3) recording. To evaluate for DAMGO effects, we compared slot 2 with slot 1. To look for αDTX effects in background of DAMGO, we examined the change between slot 3 and slot 2. We also collected recordings from neurons under a saline-control condition to account for the effects of time in the drug group. We recorded from these neurons in an identical fashion (the timeline is the same), but the saline-control neurons only received vehicle buffer, and not DAMGO or αDTX.

Within-Subjects Factor (Slot) Group Slot 1 Slot 2 Slot 3 Between- Saline (N = 21) Saline Saline Saline Subjects Factor Drug Sequence (N = Saline DAMGO αDTX+DAMGO (Group) 55) Table 4. Control and drug groups for electrophysiology. We included 2 groups in the experimental design; a saline-only control and the experimental group that received the sequence of drugs. In both cases, slot 1 recordings are always pre-drug conditions. Slot 2 is synonymously referred to as “post-DAMGO” or as “pre-αDTX” depending on whether we are testing for the effects of (3 µM) DAMGO or (100 nM) αDTX in that particular analysis; the αDTX was always delivered in a background of DAMGO. To detect drug effects, we analyzed the changes between 2 timeslots in the drug group and compared it to the changes in the saline group: slot 1 versus slot 2 for DAMGO effects, and slot 2 versus slot 3 for (100 nM) αDTX effects.

Overall, the DAMGO-exposed neurons (N = 55, M = 1.03, SD = 0.05) were significantly more (t(74) = 3.14, p = 0.001) hyperpolarized than saline controls (N = 21,

M = 0.99, SD = 0.99), which tended to slightly depolarize. However, the µOR is only expressed by a subset of cortical interneurons, and therefore was expected to be represented in only some of the red-fluorescent Dlx5/6-mRuby2 neurons that we recorded from. We anticipated that most neocortical interneurons would be unresponsive to DAMGO. Combining data from all interneurons would thus obscure the effects of

DAMGO (3 µM) and αDTX (100 nM) in neurons that did express had µORs. We therefore screened interneurons for responsiveness to DAMGO prior to subsequent analyses. Hyperpolarization is a common response to DAMGO in neurons expressing opioid receptors (B Chieng & Christie, 1994; Chiou & Huang, 1999; Grudt & Williams,

129

1993; Johnson & North, 1992; Loose & Kelly, 1990; Loose et al., 1990; Madison &

Nicoll, 1988; R. A. North et al., 1987; Sugita & North, 1993), and in particular those expressing µORs (Férézou et al., 2007; Glickfeld et al., 2008; Harris & Williams, 1991;

Kelly et al., 1990; Lagrange et al., 1994; Lagrange et al., 1995; Madison & Nicoll, 1988;

E. Tanaka & North, 1994). We therefore first identified responders by detecting a hyperpolarizing change in resting membrane potential (RMP) after exposure to DAMGO

(Figure 16). By comparing the DAMGO-exposed group’s changes in RMP to the changes in saline controls, we established a cutoff level to separate neurons most affected by

DAMGO. We refer to the DAMGO-exposed neurons above this cutoff as

“hyperpolarizing responders” (H-responders, N = 11). These H-responders hyperpolarized significantly (t(20) = 7.61, p < 0.001) when they (N =11, M = 1.10, SD =

0.03) were compared to a group of random saline controls (n = 11, 푥̅ = 0.98, s = 0.04).

RMP RMP a in Saline Group b in DAMGO Group c Change in RMP d RMP in H-Responders

-20 -20

DAMGO DAMGO

-40 -40 V

V ** m m *** -60 -60 Saline Saline

-80 -80

.8 .9 .0 f .1 .2 .8 .9 .0 .1 .2 Pre Post Pre Post 0 0 1 of 1 1 0 0 1 1 1 ut C Hyperpolarized Fold Change From Initial Fold Change From Initial Hyperpolarized e f g RMP RMP RMP in Saline Controls in H-responders in Nonresponders

-40 -40 0 -45 -45 -20 -50 -50

-55 -55

V V

V -40

m m -60 m -60

-65 -65 -60 -70 -70 -75 -75 -80 Pre Post Pre Post Pre Post

130

Figure 16. H-responders hyperpolarize. The efforts to identify DAMGO-responders and measure the change of RMP in the H-responders began by graphing resting membrane potential (RMP) after the 80s period of incubation with either (3 µM) DAMGO or vehicle saline (a) We plotted the RMP each saline- control neuron (N = 21) individually and (b) each DAMGO-exposed neuron (N = 55) individually. However, these graphs were ineffective in identifying the neurons that hyperpolarized the most under the DAMGO condition. To isolate the change (hyperpolarization), we graphed the fold change in RMP instead of the raw value (c) We plotted the change in RMP (defined as slot 2 RMP/ slot 1 RMP) in saline and DAMGO controls and found a significant hyperpolarization in the DAMGO group compared to the saline group (p < 0.05). To identify true DAMGO responders, we established a cutoff at the most hyperpolarized saline control (vertical dotted line). DAMGO hyperpolarizers (to the right of the cutoff) are hereafter called “H-responders” (N = 11). (d) We extracted these H-responders and a comparison group of random saline controls (n = 11) onto a new plot. This represents the change in RMP among the H-responders, which was significantly (p < 0.05) more hyperpolarized than the random controls. (e) Subset of random saline controls for comparison (n = 11) showing little general trend towards depolarization or hyperpolarization during the slot1 to slot 2 period. (f) A plot of the H-responders (N = 11) RMP in their actual values showing a trend towards hyperpolarization. (g) The RMP of the nonresponders (N = 34) during the same period. (a,b,e,f,g) Dots are values for interneurons and lines connect the same neurons before and after saline or the DAMGO (c,d) Edges of box are 1st and 3rd quartiles for values, dots are the change in RMP for the individual neurons, and whiskers extend to the largest and smallest values. Statistical testing performed with unpaired t tests (one-tailed) after confirming their data had a normal distribution with a D’Agostino-Pearson omnibus K2 test (p > 0.05).

During the recording process, we observed that only 3 of the H-responders were firing spontaneously at rest and, not surprisingly, their rate of spontaneous APs decreased after DAMGO. However, we observed that a sub-population of interneurons in which

DAMGO failed to produce an above cut-off level of hyperpolarization nevertheless underwent reductions in spontaneous APs after DAMGO. We found that 50.0% (22 out of 44) of neurons that fell below the H-responder cut-off level exhibited spontaneous AP activity before DAMGO (Figure 17). Application of DAMGO resulted in a dramatic reduction in the number of spontaneous APs in 45.5% (10 out of 22) of those neurons.

We define this group of spontaneous AP DAMGO-responders (S-responders) by using an arbitrary cutoff of a 50% reduction in spontaneous APs after DAMGO (N = 10).

Spontaneous APs were measured during a noncontiguous period (cumulative of 60s) in between current pulses.

131

Spontaneous AP Rate Spontaneous AP Rate a in Saline Controls b after DAMGO

20 15 d

d 15 n

n 10

e s

s 10

s 5 P

P 5

A A

0 0 Pre Post Pre Post

c d Spontaneous AP Rate e Spontaneous AP Rate Spontaneous APs after DAMGO in S-responders in Nonresponders

15 15

d d n DAMGO n

10 10

o o

c c

e e

s s

/ /

s 5 5 P

Saline P

A A

0 0 0 ff 1 2 3 4 to Pre Post Pre Post u C Fold Change From Initial

Figure 17. Spontaneous APs after DAMGO or saline. Out of the remaining pool of 44 (3 µM) DAMGO- exposed neurons, 22 were spiking spontaneously. Out of the 21 saline controls, 11 were spiking spontaneously. (a) Rate of spontaneous APs after 80s of vehicle saline show no general trend. (b) Rate of spontaneous APs in the remaining (non-H-responder) DAMGO-exposed neurons, which appeared to show some neurons had reductions in spontaneous APs, and others had not. (c) We therefore graphed the fold change in spontaneous APs after DAMGO or vehicle saline. We established a 50% reduction (vertical dotted line) in spontaneous APs as a somewhat arbitrary cutoff to distinguish nonresponders from S- responders. (d) Rate of spontaneous APs in S-responders (N = 10) which had fell below the cutoff, and therefore were classified as S-responders. (e) Rate of spontaneous APs in the remaining pool of nonresponders (N = 12), which had not achieved that 50% cutoff. Figures (d) and (e) are therefore subsets of (b). Figure 17 depicts the fold change in number of spontaneous APs (slot 2/slot 1) in the saline group (N = 11) and the DAMGO-exposed group (N = 22). Because a saline control dropped spontaneous APs to zero, we could not use a “real datapoint” cutoff as we did with the H-responders. We therefore established a 50% reduction (0.50; at the vertical dotted line in Figure 17c) as a cutoff point for S-responders and nonresponders.

All DAMGO-exposed neurons to the left of that cutoff were considered to be “S- 132 responders” (N = 10). We did not analyze the effect of DAMGO on spontaneous APs beyond counting them, because 8 out of 10 S-responders had simply stopped firing spontaneous APs after DAMGO. Collectively, 21 out of 55 (38.2%) neurons fell into at least one of the two responder categories. Hyperpolarizing responders had an average

RMP of -54.0 mV (range -50.1 to -62.9 mV). The H-responders had, on average, hyperpolarized by -5.6 mV (range -3.5 to -8.6 mV) during the 80s period from slot 1

(saline) to slot 2 (DAMGO). Most H-responders were silent at rest, with only three firing spontaneous APs. On average the S-responders fired around 2.8 APs/s before DAMGO when unstimulated. This ranged from 0.17 APs/s to 12.1 APs/s. After DAMGO, 8 out of

10 S-responders simply stopped firing spontaneous APs, however 2 continued to fire spontaneous APs at lower rates after DAMGO. A neuron that was firing spontaneous APs at 12.1 APs/s reduced its spiking to 5.75 APs/s after DAMGO, whereas the remaining the remaining neuron underwent a nearly four-fold reduction in spontaneous APs (Figure

17d).

To further explore whether the H and S-responder classes of DAMGO-responsive cells represent distinct cell populations, we determined the pre-drug spiking pattern based on the Petilla Interneuron Nomenclature Group’s suggested categorization scheme (The

Petilla Interneuron Nomenclature Group, 2008). Categorization of H and S-responders was determined for based on their firing pattern during the 1s suprathreshold current application (Figure 18). Although we characterized their spiking pattern around their threshold Vm for the first AP, we observed that discharge patterns could change at higher current applications. For instance, fastspiking neurons typically had low firing rates

133 around threshold, but increased at more-depolarizing current steps, which has been reported before (Golomb et al., 2007).

Figure 18. Representative Examples of Firing Patterns. We classified the spiking characteristics of interneurons by observing the pattern of AP discharges that occurred around their Vm threshold for an AP. Note that there is some variability here in their threshold and RMP, which manifests as threshold Vm sometimes occurring at different episodes for each neuron. Firing patterns are arranged by row, and their episodes by column. (Top row) Adapting neurons tended to fire more frequently at the beginning of the current pulses and had steadily increasing interspike intervals through the current application. This pattern stayed stable even at high current applications (E-15 and E-20). (Second Row) Irregular spiking neurons fired irregular APs, or irregular bursts of APs, at suprathreshold current application. However, this firing pattern often transitioned to more evenly spaced APs at higher current applications. (Third row) Fast- spiking neurons tended to have more distance between their RMP and AP threshold. Most of their APs were evenly spaced throughout the current application. At higher current applications (E-15 and E-20) this firing pattern tended to fire more frequently, and correlate with the amplitude of the current injection. (Fourth Row) Nonadapting, nonfastspiking neurons had evenly spaced spikes throughout the current application. This firing pattern tended to have a stable discharge pattern regardless of the current application; unlike the neurons we classified as fastspiking which tended to increase their AP frequency when more-depolarizing currents were applied. We found that S-responders and non-responders were roughly equivalent in their distribution of firing patterns. H-responders, in contrast, were notable in having a

134 relatively high frequency of Adapting neurons, a relatively low frequency of nonadapting neurons and a lack of fast-spiking neurons. Collectively, these data suggest that the µOR effects might be heterogeneous and could vary between neuron populations.

Pattern H-responders (N = 11) S-responders (N = 10) Nonresponders Count (% of column) Count (% of column) (N = 34) Count (% of column) Fastspiking 0 (0%) 2 (20%) 8 (23.5%) Nonadapting, 3 (27.2%) 6 (60%) 16 (47.1%) nonfastspiking Adapting 6 (54.5%) 1 (10%) 4 (11.8%) Irregular Spiking 2 (18.2%) 1 (10%) 6 (17.6%) Table 5. Responder groups demonstrate different proportions of spike patterns. We used the PING classification system to categorize each group (by hand) based on their spiking pattern in response to a 1s depolarizing current at 2-3 steps above threshold, where the trace stabilizes, and multiple APs occur. While the proportions of spiking patterns in each pool (in parentheses) are relatively similar, it appears that adapting firing patterns are roughly 5 times as common in the H-responder group as S-responders and nonresponders. We did not find fastspiking neurons in the H-responder group, though they were found sometimes in S-responders and nonresponders. The PING nomenclature also designates “Intrinsic Burst Spiking” and “Accelerating” spike patterns; however, we did not observe these patterns in the sample (N = 55).

3.3.2 DAMGO effects on action potential activity and kinetics

Modulation of AP kinetics can affect neurotransmitter release by influencing the kinetics and amplitude of calcium influx (Y.-M. Yang & Wang, 2006). Features of AP kinetics are shaped by multiple types of ion channels (B Rudy et al., 2009). We therefore evoked

APs before and after DAMGO and αDTX to assess a potential role of µOR in regulating

αDTX-sensitive ion channels as a means of governing AP kinetics.

We analyzed 54 AP parameters, ratios, and membrane properties (Tables 7 and

9). Of those, seven have been reported to be affected by activation of µORs or αDTX in other studies, though not necessarily both. Hyperpolarization was the first measure, which is often found with DAMGO stimulation (Loose et al., 1990). Secondly, αDTX- 135 sensitive channels are shown to modulate Vm threshold for an AP (AP threshold)

(Bekkers & Delaney, 2001; Glazebrook et al., 2002; Kirchheim, Tinnes, Haas, Stegen, &

Wolfart, 2013; Pathak et al., 2016), though the µOR may not change it (Bekkers &

Delaney, 2001; Faber & Sah, 2004; Glazebrook et al., 2002; Kirchheim et al., 2013;

Pathak et al., 2016). Both the µOR and αDTX-sensitive channels also suppresses evoked

AP discharges, and so we investigated whether it decreases the number of evoked APs and increases their temporal spacing (interspike interval; ISI) (Faber & Sah, 2004; Mo,

Adamson, & Davis, 2002). The AP halfwidth (duration) and the AP maximum repolarization rate were also measured because there is some evidence to support their modulation by αDTX-sensitive channels (W. Wang, Kim, Lv, Tempel, & Yamoah,

2013). Finally, since many voltage-sensitive K channels contribute to the afterhyperpolarization (AHP), we measured the magnitude of afterhyperpolarizations as well (W. Wang et al., 2013).

3.3.3 DAMGO alters AP waveform and pattern of evoked APs in H-responders

To reduce family-wise errors from running sets of 7 analyses simultaneously, we first tested the combined 7 dependent variables (RMP, AP threshold, interspike interval, number of evoked APs, AP halfwidth, maximum repolarization rate, and afterhyperpolarization amplitude) in the H-responders (N = 11) versus the saline controls

(N = 21) for a significant Slot*Group interaction (see Table 4) in a mixed-model repeated measures MANOVA. When the Slot*Group interaction was significant, we tested the changes in the dependent variables individually.

136

We found that the H-responders had statistically significant (F(7,20) = 19.32, p <

0.001, Wilks’ λ = 0.871) Slot*Group interaction in the 7 combined variables that we were tested, which suggested a possible DAMGO effect in those parameters. We therefore carried out specific analyses of RMP, AP threshold, number of evoked APs, interspike interval, AP halfwidth, max repolarization rate, and afterhyperpolarization magnitude in

H-responders.

The αDTX-sensitive channels can modulate Vm threshold for an AP, however the

µOR does not necessarily change this property (Faber & Sah, 2004; Finnegan et al.,

2006). We analyzed their change in Vm threshold for the first AP (Figure 19a,b,c). We found no significant difference (t(20) = 0.42, p = 0.338) in the change in AP threshold for the H-responders (M = 0.99, SD = 0.07) versus the saline controls (푥̅ = 0.98, s = 0.07) after DAMGO. This shows that neocortical µORs do not modulate AP threshold in the H- responders.

We then investigated the patterns of AP discharges to determine whether how

DAMGO had influenced the number of evoked APs and their temporal spacing between

APs (Figure 19). These H-responders (N = 10, M = 1.35, SD = 0.36) had significantly increased interspike intervals (t(18) = 3.01, p = 0.004) when compared to the saline-only controls (n = 10, 푥̅ = 0.98, s = 0.12). The H-responders (N = 11, Mdn = 0.98) also had significantly reduced numbers of evoked action potentials (U = 12, p < 0.001) after application of DAMGO when compared to saline controls (n = 11, Mdn = 0.98). These data show that DAMGO reduced the number of APs evoked by suprathreshold currents.

137

Vm Threshold for an AP Vm Threshold for an AP a in Saline Controls b in H-responders c AP Threshold 0 0

-20 -20

DAMGO V

V -40 -40 m

m ns

-60 -60 Saline

-80 -80

.7 .8 .9 .0 .1 .2 Pre Post Pre Post 0 0 0 1 1 1 Fold Change From Initial Higher Threshold

Number of Evoked APs Number of Evoked APs d in Saline Controls e in H-responders f Evoked APs

30 30

s s P

P DAMGO A

A 20 20

f f

o o

r ***

e e b

b 10 10 m

m Saline

u u

N N

0 0 .0 .5 .0 .5 .0 Pre Post Pre Post 0 0 1 1 2 Fold Change From Initial

g Mean Interspike Interval h Mean Interspike Interval i in Saline Controls in H-responders Mean ISI

0.20 0.20 )

) 0.15 0.15 s

s DAMGO

( (

s s d

d 0.10 0.10 n

n **

o o

c c e

e Saline S S 0.05 0.05

0.00 0.00

.0 .5 .0 .5 .0 .5 Pre Post Pre Post 0 0 1 1 2 2 Fold Change From Initial

Figure 19. Spike frequency altered in H-responders. We tested whether H-responders had discernible changes after (3 µM) DAMGO AP threshold and frequency by comparing them to saline controls. (Left column) The saline controls (N = 21) show little systematic changes in their (a) AP threshold, (d) number of evoked APs, (g) and mean interspike interval during exposure to vehicle saline. (Middle column) Meanwhile the H-responders (N = 11) show little change after (3 µM) DAMGO in (b) AP threshold, however they generally show reduced numbers of (e) evoked APs, and (h) larger interspike intervals after DAMGO. (Right column) To illustrate and test these changes more precisely, we measured the fold change in those parameters against a random group of saline controls (n = 11). We found that (c) AP threshold change was unaffected by the DAMGO compared to the change in saline controls, but we found significantly decreased (p < 0.05) number of (f) evoked APs in H-responders after DAMGO and significantly increased (i) mean interspike interval. One H-responder is excluded because it became single- spiking after DAMGO (Left and Middle columns) connecting lines bridge the same neuron after an 80s period of vehicle or DAMGO incubation. Y axes are true values for those particular measures (Right column) isolates the fold change (slot 2/slot1; postdrug/initial) after vehicle saline or DAMGO. Edges of the boxes are drawn at the 1st and 3rd quartiles while the “whiskers” connect the largest and smallest values 138 for that group. We performed statistical testing of ISI and AP threshold with unpaired t-tests (one-tailed) after first testing for skewness and kurtosis with D’Agostino-Pearson omnibus K2 test (p > 0.05). Only number of evoked APs was significantly different from a normal distribution (p < 0.05) and therefore tested with a one-tailed Mann-Whitney U test.

We next analyzed the H-responder group to determine whether AP waveform was being altered by DAMGO (Figure 20). Among hyperpolarizing DAMGO-responders, we found several significant changes in action potential waveforms against the saline-only controls; we found significant reductions in AP halfwidth (U = 11, p < 0.001) in the

DAMGO H-responders (Mdn = 0.91) versus the saline controls, which trended slightly towards widening APs over that time period (Mdn = 1.05). The 11 DAMGO H- responders had increased their maximum repolarization rates (M = 1.29, SD = 0.44) significantly (t(20) = 2.91, p = 0.004) compared to the saline controls (푥̅ = 0.89, s = 0.11) which trended towards slower repolarizations by slot 2. These results show that DAMGO was hastening the maximum rates of AP repolarization, and shortening the duration of the

APs. Lastly, we compared the magnitude of afterhyperpolarizations (AHPs) in the saline controls (푥̅ = 0.94, s = 0.07) versus the H-responders (M = 1.10, SD = 0.30) and found that DAMGO had significantly increased (t(20) = 1.78, p = 0.045) the AHPs in the H-

139 responders.

AP Halfwidth AP Halfwidth in H-responders AP Halfwidth a in Saline Controls b c 10 10 8 8 DAMGO 6

6 s

s m

m *** 4 4 Saline 2 2

0 0

.0 .5 .0 .5 Pre Post Pre Post 0 0 1 1 Fold Change From Initial d e f AP Max Repolarization Rate AP Max Repolarization Rate AP Max Repolarization Rate H-responders in Saline Controls 100 DAMGO 100 80

80 ** s

s 60 60

m /

m Saline

/ V V 40

40 m m

20 20 .0 .5 .0 .5 .0 .5 0 0 1 1 2 2 0 0 Fold Change From Initial Pre Post Pre Post g h AHP Amplitude i AHP Amplitude in H-responders AHP Amplitude Saline Controls 25 25 20 20 DAMGO 15

15 V

m * 10 10 Saline 5 5

0 0

.0 .5 .0 .5 .0 Pre Post Pre Post 0 0 1 1 2 Fold Change From Initial

Figure 20. AP kinetics are altered in H-responders. We analyzed AP kinetics and afterhyperpolarization (AHP) amplitude before and after DAMGO or vehicle saline to determine whether DAMGO altered these properties. (Left column) Saline controls (N = 21) before and after an 80s period of exposure to the vehicle saline. In the saline group, there was little trends in (a) AP halfwidth, and (d) maximum repolarization rates (Middle column) DAMGO-exposed H-responders (N = 11) generally show reduced (b) halfwidths and (e) larger maximum repolarization rates after DAMGO, but (h) AHP amplitudes are generally stable. (Right column) To isolate and statistically test the change after DAMGO, we plotted the fold change after DAMGO or vehicle saline. We found that DAMGO exposure in H-responders resulted in significantly reduced (c) halfwidths compared to random saline controls (n = 11). This corresponded with significantly increased (f) maximum AP repolarization rates. We also found significantly larger (i) AHP amplitudes in H-responders relative to controls. (Left and Middle columns) connecting lines bridge the same neuron after an 80s period of vehicle or DAMGO incubation. Right column isolates the fold change (slot 2/slot1; 140 postdrug/initial) after vehicle saline or DAMGO. Edges of the boxes are drawn at the 1st and 3rd quartiles while the “whiskers” connect the largest and smallest values for that group. We performed statistical testing for maximum repolarization rate and AHP amplitude with unpaired t-tests (one-tailed) after testing for skewness and kurtosis with D’Agostino-Pearson omnibus K2 test. AP halfwidth change (for saline group) was nonnormally distributed (p < 0.05) and therefore tested with Mann-Whitney U tests.

PARAMETER Mean Difference P value 95% Confidence Interval for (DAMGO- Difference Saline) Lower Upper RMP 0.12 <0.001 0.09 0.16 AP Halfwidth -0.21 < 0.001 -0.32 -0.10 AP Max Repolarization 0.40 0.004 0.11 0.68 Rate Interspike Interval 0.37 0.004 0.11 0.62 Number of Evoked APs -0.35 < 0.001 -0.56 0.14 AP Threshold 0.01 0.338 -0.05 0.08 Afterhyperpolarization 0.16 0.045 -0.03 0.35 Amplitude Table 6. Summary of DAMGO effects in H-responders. We compiled a statistical summary and effect sizes for H-responders in all 7 measures that we were primarily investigating. The change in each property was derived by dividing post-DAMGO/pre-DAMGO values (i.e., slot 2/slot 1) to calculate the fold change. Here, these fold changes in the saline group (expected to represent the change due to time) was subtracted from the fold change in the DAMGO group to derive the effect of DAMGO. Each mean difference is a proportion of one (e.g., a mean difference of -0.21 is a reduction of 21% in the DAMGO condition over the saline control). Comparisons of the change in saline and DAMGO group were tested first for normality with the D’Agostino-Pearson omnibus K2 test. Normally distributed data were tested for significance (α = 0.05) with one-tailed unpaired t-tests and nonnormally-distributed data were tested with one-tailed Mann- Whitney U tests.

We classified the H-responders by spiking pattern after DAMGO to determine if

µORs could change the pattern of discharges, since there were changes in evoked AP number and ISI in these neurons (Table 7). Although most H-responders had a stable spiking pattern after DAMGO, we did notice a few changes. Three H-responders had shifted spiking categories, however the majority (8 out of the 11) H-responders had remained in the same spiking pattern after DAMGO. In 2 out of 3 cases where the H- responder shifted pattern, the neurons transitioned into an Adapting patten, which was consistently the prevailing spiking pattern in H-responders.

141

Pre-DAMGO Pattern Post-DAMGO Pattern Adapting Adapting Adapting Adapting Adapting Adapting Adapting Adapting Adapting Adapting Adapting Irregular Spiking Irregular Spiking Irregular Spiking Irregular Spiking Adapting Non-adapting, Non-fastspiking Non-adapting, Non-fastspiking Non-adapting, Non-fastspiking Non-adapting, Non-fastspiking Non-adapting, Non-fastspiking Adapting Table 7. Few H-responders shift spiking patterns after DAMGO. We identified the spiking patterns of H- responders based on PING categories before and after DAMGO. Neurons are ordered by row arbitrarily. Boldtype rows are neurons that change firing type. The majority of the neurons remained the same spiking patterns, though 3 had changed after DAMGO.

3.3.4 αDTX counteracts µOR-agonism in H-responders

Having found that DAMGO was altering AP waveform and pattern of discharge, particularly in hyperpolarizing DAMGO responders, we investigated whether αDTX- sensitive channels were involved in this effect; µOR elsewhere in the brain has previously been found to modulate an α-Dendrotoxin-sensitive channels (Faber & Sah,

2004; Finnegan et al., 2006). We therefore exposed the neurons to a combination of

DAMGO + αDTX (3 µM and 100 nM, respectively) after their prior exposure to (3 µM)

DAMGO to determine if the addition of αDTX could reverse the effects. We once again compared this drug group against saline-only recordings to account for stochastic effects and the effects of repeated stimulation, and thereby detect the true impact of αDTX in a background of DAMGO. For the following comparisons we also isolated the change by dividing the post αDTX+DAMGO/pre αDTX-DAMGO; slot 3/slot 2. In this case, the post-DAMGO (slot 2) measurements that had been used as the “after-DAMGO” recording were used here as the “pre-αDTX+DAMGO” recording (see Figure 15). 142

We tested for a significant Slot*Group interaction in the H-responders versus the controls in timeslots 2 to 3 (DAMGO → αDTX+DAMGO). This interaction was significant indicating a potential αDTX effect on the 7 variables (F(7,19)= 2.57, p =

0.048, Wilks’ λ = 0.486). We therefore conducted further analysis on each measure. We had previously found significant DAMGO effects in all the measures for H-responders, except for AP threshold and afterhyperpolarization amplitude (see Table 6), but we expected that αDTX’s effects may not be as extensive as DAMGO’s effect, since this channel may only mediate some of the µOR’s inhibitory mechanisms.

143

a RMP RMP Saline Controls DTX in H-responders RMP -40 -50

-45 -55 (DTX) H-Responders -50

V -60 V

m ns -55 m -65 Saline -60 -70

-65 -75 Pre Post 5 0 5 0 5 .8 .9 .9 .0 .0 Pre Post 0 0 0 1 1 Fold Change from pre-DTX Depolarized AP Threshold AP Threshold b in Saline Controls DTX in H-responders AP Threshold 0 -35

-40 (DTX) H-Responders -20

-45 V

V * m m -50 -40 Saline -55

-60 -60 .7 .8 .9 .0 .1 .2 .3 0 0 0 1 1 1 1 Pre Post Pre Post Fold Change from pre-DTX Lower Threshold c Number of APs Number of APs in Saline Controls DTX in H-responders Number of Evoked APs

30 40 s

s 30 (DTX) H-Responders P

P 20

A A

d ** e

e 20

k k

o o v

v 10 Saline

E E

0 0 .0 .5 .0 .5 .0 .5 Pre Post Pre Post 0 0 1 1 2 2 Fold Change from pre-DTX

ISI ISI d in Saline Controls DTX in H-responders AP Interspike Interval

0.15 0.20 )

) 0.15 (DTX) H-Responders

s s (

0.10 (

l l a

a 0.10 v

v **

r r

e e t

0.05 t n

n Saline I I 0.05

0.00 0.00 .5 .6 .7 .8 .9 .0 .1 Pre Post Pre Post 0 0 0 0 0 1 1 Fold Change from pre-DTX

144

AP Halfwidth AP Halfwidth e in Saline Controls DTX in H-responders AP Halfwidth

10 5

)

) s

8 s 4

(

(DTX) H-Responders

h t

6 t 3

i w

w ns

f l

4 l 2

a H

H Saline

2 1

A A 0 0 .0 .5 .0 .5 .0 Pre Post Pre Post 0 0 1 1 2 Fold Change from pre-DTX

AP Max Repolarization Rate AP Max Repolarization Rate

f in Saline Controls DTX in H-responders AP Max Repolarization Rate )

60 ) 100

s s

m m

/ / V

V 80

m m (

( (DTX) H-Responders

40 e

e 60

t t

a a

R R

ns l

l 40 o

20 o p

p Saline e

e 20

R R

x x a

0 a 0

M M .0 .5 .0 .5 Pre Post Pre Post 0 0 1 1 Fold Change from pre-DTX

g AHP Amplitude AHP Amplitude in Saline Controls DTX in H-responders AHP Magnitude

25 25

) )

V V m

m 20 20

( (

(DTX) H-Responders

e e

d d u

u 15 15

t t

i i

l l p p *

10 10

m m A

A Saline

P 5 5

H H A A 0 0 .0 .5 .0 .5 Pre Post Pre Post 0 0 1 1 Fold Change from pre-DTX

Figure 21. αDTX reverses DAMGO effects on ISI and AP number. We tested for αDTX effects in the H- responders by applying αDTX in a background of DAMGO and examined these parameters for changes that were in opposite polarity as DAMGO. These change ratios were calculated by dividing postDTX/preDTX (i.e., slot 3/slot 2). Thus, αDTX in a background of DAMGO. (Left column) Real values for saline controls during this period which show stochastic changes in each measure. (Middle column) Real values for H-responders before and after exposure to αDTX. (Right column) isolates the fold change in a subset (n = 11) of saline controls and the (N = 11) H-responders from their pre-DTX condition. Statistical testing was performed by comparing the changes in saline controls (n = 11) with the changes in H-responders (N = 11) after αDTX. We found no significant effects of αDTX on (a) RMP, (e,f) AP kinetics (halfwidth and max repolarization rate). However, we did find significant effects of αDTX on H-responders in a background of DAMGO in (d) interspike interval (c) number of evoked APs, as well as (g) afterhyperpolarization amplitude. We also found that αDTX lowered (b) AP threshold, which had not been modulated by DAMGO. One neuron was excluded from graph (c), but not statistical comparison, due to a 145

14-fold increase in evoked APs. Edges of the boxes are drawn at the 1st and 3rd quartiles while the “whiskers” connect the largest and smallest values for that group. We performed statistical testing with unpaired t-tests (one-tailed) after testing for skewness and kurtosis with D’Agostino-Pearson omnibus K2 test (p > 0.05).

We found no significant effect (t(20) = 0.33, p = 0.374) of αDTX on the RMP of

H-responders (M = 0.98, SD = 0.03) when compared to saline-controls (푥̅ = 0.98, SD =

0.04). This indicates that although DAMGO had hyperpolarized these neurons, αDTX- sensitive channels were not responsible for this effect. However, we did find a significant

(U = 32, p = 0.033) effect of αDTX on the AP Vm threshold of H-responders (Mdn =

1.04) when compared to the saline controls (Mdn = 1.00). This was a particularly interesting finding because DAMGO did not modulate AP threshold (Figure 19 and Table

6).

Additionally, αDTX significantly reduced ISI (t(20) = 2.70, p = 0.007) in H- responders (M = 0.86, SD = 0.13)) versus the saline controls (푥̅ = 0.97, s = 0.04)).

Similarly, αDTX significantly increased (U = 22, p = 0.005) the number of APs in the H- responders (Mdn = 1.22) versus the saline controls (Mdn = 1.00). Therefore, αDTX has an opposite effect to DAMGO, which raises ISI and decreases evoked APs in these H- responders. This suggests that αDTX-sensitive currents do indeed mediate some of the

DAMGO effect on suppressing the initiation of APs at suprathreshold depolarizations. In a later analysis, we analyzed whether αDTX fully or partially reverted these properties on a cell-by-cell basis (Figure 22).

We next investigated whether αDTX reversed the DAMGO effects on AP kinetics through halfwidth and maximum rates of repolarization of an AP. αDendrotoxin failed to

146 reverse (U = 48, p = 0.219) DAMGO effects in AP halfwidth in the H-responders (Mdn =

1.01) versus the saline controls (Mdn = 1.06). Similarly, we found that αDTX did not significantly (U = 57, p = 0.423) alter the maximum rate of AP repolarization (Mdn =

0.91) when the H-responders were compared to the saline controls (Mdn = 0.96). Thus, it appeared that the changes in AP waveform induced by DAMGO are mediated by channels other than the αDTX-sensitive channels.

Finally, we investigated whether αDTX altered AHPs. We found that αDTX produced a significant reduction (U = 29, p = 0.020) in AHP amplitude when H- responders (Mdn = 0.84) were compared to the saline controls (Mdn = 0.98). Therefore,

DAMGO increased AHP amplitude, while αDTX reduced it.

PARAMETER Mean Difference P value 95% Confidence Interval for Δ(αDTX+DAMGO)- Difference Δ(Saline) Lower Upper RMP -0.01 0.374 -0.04 0.03 AP Halfwidth 0.03 0.219 -0.13 0.19 AP Max Repolarization 0.00 0.423 -0.17 0.17 Rate Interspike Interval -0.11 0.007 -0.20 -0.03 Number of Evoked 1.51 0.011 -0.91 3.92 APs AP Threshold 0.07 0.033 0.00 0.15 AHP Magnitude -0.14 0.020 -0.27 0.00 Table 8. αDTX-sensitive channels reverse some DAMGO effects. Mean values in this summary were derived by subtracting the mean fold change in the saline group from the mean fold change in the DAMGO group. The fold changes for each of the groups were compared with unpaired t tests (one-tailed) for significance or Mann-Whitney U-tests.

3.3.5 αDTX and magnitude of reversal

In H-responders, αDTX affected ISI, number of evoked APs, AHP amplitude, and AP threshold (However DAMGO had not changed AP threshold). We next wanted to

147 determine whether αDTX was fully reversing the effects of DAMGO in these parameters, or only partially reversing them.

148 a RMP b AP Threshold of Individual H-Responders of Individual H-Responders 1.20 *** ns 1.2 ns * (Hyperpolarizes) (Threshold Hyperpolarizes)

e 1.15

e l

g 1.1

n n

a 1.10

R 1.0

1.05 r

T l

1.00

F o

P 0.9 F 0.95 A Initial DAMGO DTX+DAMGO 0.8 Initial DAMGO DTX+DAMGO

Number of Evoked APs ISI c in Individual H-Responders d of Individual H-Responders

1.5 *** ** 2.5 ** ** e

e 2.0

g n

n 1.0

a 1.5

1.0

o d

d 0.5

o o

E 0.5

F F

0.0 0.0 Initial DAMGO DTX+DAMGO Initial DAMGO DTX+DAMGO

AP Repolarization Rate AP HWs e of Individual H-Responders f of Individual H-Responders

2.5 ** ns 1.2 *** ns

n e

e 2.0

t n

n 1.0

a a

a 1.5

l 1.0

d d

o 0.8

o o

e 0.5

F R 0.0 0.6 Initial DAMGO DTX+DAMGO Initial DAMGO DTX+DAMGO

AHP Amplitude of Individual H-Responders *

g 2.0 *

d e

u 1.5

a p

h 1.0

l P

o 0.5

H A 0.0 Initial DAMGO DTX+DAMGO

149

Figure 22. αDTX partially reverses some effects of DAMGO in H-responders. To illustrate the fold changes in all 7 parameters across the 3 timepoints, we baselined value for these measures in the (N = 11) H-responders by their initial value; therefore, all neurons emanate from “1.00” which marks their original values. The dotted line at 1.00 represents baselined starting value for all neurons. These graphs show the fold changes over the course of their treatment with DAMGO, and then αDTX+DAMGO. (a) Change in RMP, which was the basis of selection for (N = 11) H-responders, and not significantly reversed by αDTX. (b) Change in AP threshold, which was not significantly altered by DAMGO, shifted towards hyperpolarized Vm after αDTX. (c) Evoked APs in H-responders and (b) ISI changes in H-responders (n = 10; one neuron was not included because it became single-spiking after DAMGO). Changes in (c,d) had before been found to be affected by both DAMGO and αDTX in H-responders. (e) Maximum repolarization rate and (f) AP halfwidth, which were both significantly changed by DAMGO, but not αDTX. (g) Afterhyperpolarizations were significantly enhanced by DAMGO and reduced by subsequent addition of αDTX. Lines on all graphs connect the same neuron’s values across all 3 timepoints. Statistical comparisons were made by comparing fold change in H-responders to fold changes in saline controls. Changes were compared for significance with unpaired t-tests or Mann-Whitney U tests.

To visualize this variability effects of αDTX in a background of DAMGO, we baselined each of the H-responders to their initial starting point and graphed all 7 changes over the course of slots 1-3 (Figure 22). These analyses provided insight into the direction of the αDTX effect, while also charting each neuron’s course individually.

From these graphs it is evident that not only did the effect of DAMGO and αDTX varies between the interneurons, which can be expected of a heterogenous population of neocortical interneurons.

150

Figure 23. Example of αDTX reversal of DAMGO. A trace representing αDTX reversing effects of DAMGO in AP number and interspike interval. Respectively, each row represents the initial, post- DAMGO, post-αDTX+DAMGO timeslots. Each column represents one of the 20 sequential episodes as the current applications progressively become more depolarizing. For example, E4 is the 4th episode of stimulation, which generally was around AP threshold, while E20 is the final and most-depolarizing of the 20 episodes of the stimulation. The blue horizontal bars below the APs represent the period of 1 second where the depolarizing current was being applied. (Top Row) This neuron began firing APs in the early, less-depolarizing current applications. Application of DAMGO (Middle Row) resulted in the neuron firing APs only at later episodes. (Bottom Row) After addition of αDTX, the neuron once again began firing at earlier episodes and more frequently than it had after exposure to only DAMGO. A slight RMP hyperpolarization after DAMGO is evident here, which was not reversed by αDTX.

3.3.6 DAMGO alters RMP and AP waveform in S-responders

We began testing for DAMGO effects in S-responders by testing the 7 combined variables in a mixed-model repeated measures MANOVA for a significant Slot*Group interaction. We found a significant effect of DAMGO in S-responders (F(7,20) = 2.57, p

= 0.046, Wilk’s λ = 0.473) and we therefore investigated them further.

151

a RMP b in S-Responders RMP in S-Responders

-40

-45 DAMGO

V -50

m *

-55 Saline

-60 0 5 0 5 0 Pre Post .9 .9 .0 .0 .1 0 0 1 1 1 Fold Change From Initial Hyperpolarized

Figure 24. S-responders also hyperpolarize. We compared the fold change in RMP in the (N = 10) S- responders versus random saline to determine if hyperpolarization is found in this group as well. (a) Real values of resting membrane potential in S-responders before and after (3 µM) DAMGO. (b) Fold changes in resting membrane potential in S-responders versus random saline controls (n = 10). We found that the S- responders too underwent a significant hyperpolarization after exposure to DAMGO (p < 0.05) compared to the polarization change in saline controls during the same period. Edges of the boxes are drawn at the 1st and 3rd quartiles while the “whiskers” connect the largest and smallest values for that group. We performed statistical testing by confirming distribution was non-significantly different from a normal distribution with the D’Agostino-Pearson omnibus K2 test, and then an unpaired t-test against the fold change of the saline controls.

We first compared their change in RMP (M = 1.02, SD = 0.04) against saline controls (푥̅ = 0.99, s = 0.02). Apparently, although this group fell short of the threshold for the H-responder group compared to the saline group (Figure 16) the S-responders too had undergone a significant hyperpolarization (t(18) = 2.53, p = 0.011) compared to the saline controls. Although these S-responders underwent statistically significant hyperpolarization after DAMGO, we classify them as S-responders because they were not selected by that particular criterion, as the H-responders had been.

152

a AP Threshold in S-Responders b Evoked APs in S-Responders c Mean ISI in S-Responders

DAMGO DAMGO DAMGO ns ns ns Saline Saline Saline

.7 .8 .9 .0 .1 .2 .0 .5 .0 .5 .6 .8 .0 .2 .4 0 0 0 1 1 1 0 0 1 1 0 0 1 1 1 Fold Change From Initial Fold Change From Initial Fold Change From Initial Higher Threshold

Vm Threshold for an AP Number of Evoked APs Mean Interspike Interval in S-Responders in S-Responders in S-Responders

0 25 0.15

s 20 )

-20 P

(

0.10

15 s

-40 r

b 10

c 0.05

e u

-60 S

N 5

-80 0 0.00 Pre Post Pre Post Pre Post

d AP Max Repolarization Rate Afterhypolarization Magnitude AP Halfwidth in S-Responders e in S-Responders f in S-responders

DAMGO DAMGO DAMGO ** * ns Saline Saline Saline

0 5 0 5 0 5 0 5 0 5 0 .6 .8 .0 .2 .4 ...... 0 0 1 1 1 0 0 1 1 2 2 0 0 1 1 2 Fold Change From Initial Fold Change From Initial Fold Change From Initial

AP Halfwidth AP Max Repolarization Rate AHP Amplitude in S-Responders S-Responders in S-Responders 10 60 15 8

6 40 10

/ m

4 V

m 20 5 2

0 0 0 Pre Post Pre Post Pre Post

Figure 25. S-responders have changes in AP waveform. We compared the change in each parameter in a sample (n = 10) saline group versus the (N = 10) S-responders. Boxes are drawn around each set of real values for each measure (bottom of each box), and the fold change (top of each box). We found that these S-responders had significant reductions after DAMGO in (d) halfwidths and significant increases in (e) maximum rates of repolarization compared to saline controls (p < 0.05). However, we did not find significant differences in the changes in (a) AP threshold, (b) number of evoked APs, (c) mean interspike interval, or (f) afterhyperpolarization magnitude (p > 0.05). Edges are drawn at the 1st and 3rd quartiles while the “whiskers” connect the largest and smallest values for that group. Prior to statistical testing we tested for normality with D’Agostino-Pearson omnibus K2 test (p > 0.05). In most instances, we performed statistical testing with unpaired t-tests (one-tailed). However, number of evoked APs and maximum repolarization rates were nonnormally distributed (p < 0.05) and we therefore tested those with Mann- Whitney U tests (one-tailed). 153

We tested their fold change in AP threshold to find a nonsignificant effect of

DAMGO (t(18) = 1.11, p = 0.141) when S-responders (M = 0.96, SD = 0.08) versus the saline controls (푥̅ = 0.96, s = 0.07); there was no effect of DAMGO on Vm threshold for an AP in the S or H-responders.

We did not find a significant difference in their changes in interspike interval

(t(18) = 1.42, p = 0.086) of these S-responders (M = 1.09, SD = 0.12) relative to the saline controls (푥̅ = 1.00, s = 0.14). Nor did we find a significant (U = 30, p = 0.072) decrease in number of evoked action potentials (Mdn = 0.90) relative to the saline-only controls (Mdn = 1.01). Interestingly, we found effects in these 2 properties in the H- responder group but apparently these effects are absent in the S-responder category; although these neurons were selected for on the basis of their decreased spontaneous APs, it appears that somatically-evoked APs were not similarly decreased or reduced in frequency.

We next analyzed the AP kinetics of the S-responders to determine whether they too had altered AP shapes after DAMGO as the H-responders did. We found that the S- responders had significantly (t(18) = 2.42, p = 0.013) reduced their AP halfwidths (M =

0.95, SD = 0.13) after DAMGO compared to saline controls (푥̅ =1.07, s = 0.10). These S- responders (Mdn = 1.07) also had a significant increase (U = 19, p = 0.009) in their maximum rate of repolarization when compared with saline controls (Mdn = 0.87).

154

Finally, we analyzed the fold change in afterhyperpolarizations in S-responders

(M = 1.07, SD = 0.24) versus changes in the saline control (푥̅ = 1.00, s = 0.09). However, there were no significant differences in the changes of their AHP amplitude (t(18) = 1.72, p = 0.051).

PARAMETER Mean Difference P value 95% Confidence Interval for (DAMGO-Saline) Difference Lower Upper RMP 0.03 0.003 0.00 0.06 AP Halfwidth -0.12 0.013 -0.23 -0.02 AP Max Repolarization 0.25 0.009 0.02 0.49 Rate Interspike Interval 0.08 0.086 -0.04 0.21 Number of Evoked APs -0.03 0.072 -0.03 0.07 AP Threshold 0.04 0.141 -0.03 0.10 Afterhyperpolarization 0.23 0.051 -0.05 0.52 Amplitude Table 9. DAMGO Effect in S-responders. A statistical summary and estimates of effect sizes for the spontaneous-AP responders (S-responders). The change in each property was derived by dividing the post- DAMGO/pre-DAMGO values (i.e., slot 2/slot 1) to calculate the proportion of the change. For instance, a change of 0.28 in AP maximum repolarization rate is an average increase of 28% in the DAMGO S- responders. Comparisons of the change in saline and DAMGO group were performed by unpaired t-tests (one tailed). Values for saline were subtracted from DAMGO values to calculate the true effect of the drug (and subtract the effect of time). Prior to statistical testing, we tested for normality with D’Agostino- Pearson omnibus K2 test. In most instances, we performed statistical testing with unpaired t-tests (one- tailed) (p > 0.05). However, number of evoked APs and maximum repolarization rates were nonnormally distributed (p < 0.05) and we therefore tested those with Mann-Whitney U tests (one-tailed) instead.

To summarize, spontaneous AP responders had relatively small hyperpolarizations, and changes in AP kinetics compared to H-responders (Table 10). S- responders lacked the DAMGO-induced ISI increase and reduced evoked AP number, which were observed in H-responders (Table 6). Thus, it appears that DAMGO influenced their RMP, spontaneous APs frequency, and AP repolarization kinetics, but it did not change the somatically-evoked AP frequency, as DAMGO had done in the H- responders. The effects of DAMGO in S-responders appeared to be smaller in magnitude and in the range of affected parameters (Table 10).

155

PARAMETER Mean Difference H-responder Mean Difference S- (DAMGO-Saline) responder (DAMGO- Saline) RMP 0.12 0.03 AP Halfwidth -0.21 -0.12 AP Max Repolarization Rate 0.40 0.25 Interspike Interval 0.37 - Number of Evoked APs -0.35 - AP Threshold - - Afterhyperpolarization Amplitude 0.16 - Table 10. Summary of effects for H-responders vs S-responders. To provide a side-by-side comparison of H and S-responders and compare their response to DAMGO, we complied data from H-responders (Table 5) and data from S-responders (Table 6) and re-display them in this table. Numerical values reflect the mean fold change in that particular responder group (see top row) for each parameter (left column). Mean differences were derived by subtracting fold change in the saline group from the DAMGO responder group to isolate the effect of DAMGO from the effect of time. For example, a value of 0.40 for H-responder AP max repolarization rate reflects a 40% increase for maximum rate of AP repolarization in H-responders compared to the control, which trended larger than the mean 28% change in the S-responders. Only significant (p < 0.05) values are shown in the table; cells are intentionally left blank when effects were nonsignificant (Unpaired t-test or Mann-Whitney U test, as described in Tables 3 and 6). We found that the H-responders had a wider range of detected effects to DAMGO, as well as having larger effect sizes.

3.3.7 αDTX effects absent in S-responders

We also investigated whether there were αDTX effects in the S-responders, although we did not expect a αDTX effect here since data from the H-responder suggested that the effects of αDTX were restricted to AP threshold, ISI and number of evoked APs, none of which were affected by DAMGO in the S-responders. We found a non-significant

Slot*Group interaction (F(7,19) = 0.81, p = 0.592, Wilks’ λ = 0.229) in this group for the combined variables, indicating that there was not an effect of αDTX in the S-responders.

We also wanted to determine whether αDTX was reversing the measurement that was used to initially identify the S-responders, their spontaneous APs. However, the number of spontaneous APs was not significantly changed when αDTX was applied in a background of DAMGO (U = 32, p = 0.069). Indeed, after DAMGO exposure, only 3 out of 10 S-responders were firing spontaneously, and only 1 out of 7 had resumed firing

156 spontaneous APs after αDTX. Thus, it appears that DAMGO-induced downregulation of spontaneous APs was not detectably related to αDTX-sensitive channels.

3.3.8 Investigating the influence of RMP polarization

The identification of H-responders to DAMGO involved the selection of neurons with the largest hyperpolarization in during those 80s of DAMGO incubation. Under that procedure, it was possible that any subsequent differences we observed was simply a consequence of that difference in polarity change between the sampling periods. To investigate the confound of polarization, we selected the neurons that had hyperpolarized the most in the saline control, and we also selected a comparison group of neurons that had depolarized the most in the DAMGO group. Although we expected that this depolarizing change was merely due to chance (and not a DAMGO effect) it enabled us to “flip the sign” so that the saline-control neurons were relatively more hyperpolarized than the DAMGO-exposed neurons. If all of these changes were actually a result of hyperpolarization, the most-hyperpolarizing saline neurons should actually undergo analogous changes as the H-responders. If they fail to do this, significant effects in AP parameters are more likely to be caused by differential ion channel regulation by µORs, rather than a polarizing shift in the Vm without ion channel regulation by µORs.

We found a significant Slot*Group interaction (F(7,12) = 2.57, p = 0.048, Wilks’

λ = 0.633) in these groups for the combined dependent variables, which included the

RMP change that they had been selected on. We therefore investigated them further for changes in other parameters.

157

a RMP b AP Threshold c Number of Evoked APs

DAMGO DAMGO DAMGO *** ns ns Saline Saline Saline

0 5 0 5 0 .9 .9 .0 .0 .1 0 0 1 1 1 8 9 0 1 2 .0 .5 .0 .5 0. 0. 1. 1. 1. 0 0 1 1 Fold Change From Initial Hyperpolarization Fold Change From Initial Fold Change From Initial Higher Threshold

Max Repolarization Rate d Mean ISI e f AP Halfwidth

DAMGO DAMGO DAMGO ns ns ns Saline Saline Saline

.6 .7 .8 .9 .0 .1 .2 0 5 0 5 .8 .9 .0 .1 .2 .3 0 0 0 0 1 1 1 0. 0. 1. 1. 0 0 1 1 1 1 Fold Change From Initial Fold Change From Initial Fold Change From Initial

AHP Magnitude g DAMGO ns Saline

6 8 0 2 4 0. 0. 1. 1. 1. Fold Change From Initial

Figure 26. Polarity shift does not account for other changes. In these analyses, we compared neurons that depolarized in the (n = 11) DAMGO condition and neurons that hyperpolarized in the (n = 11) saline control. These changes were expected to be due to stochastic fluctuations in their RMP during this period, and we therefore expected null results. We selected these neurons based on their changes in (a) RMP which, predictably, had significantly depolarized in the DAMGO condition relative to the change in saline controls (p < 0.05). However, none of the other changes we tested were significantly different from controls (p > 0.05); we did not find significant differences in their changes of (b) AP threshold, (c) number of evoked APs, (d) interspike interval, (e) maximum repolarization rates, (f) AP halfwidth, nor (g) magnitude of their hyperpolarization when the groups were compared to controls in the same manner as the H and S- responders were. Comparisons of the DAMGO-depolarizers versus saline-hyperpolarizers were made with unpaired t-tests (one-tailed) and summarized on Table 11. Edges of the boxes are drawn at the 1st and 3rd quartiles while the “whiskers” connect the largest and smallest values for that group. We performed statistical testing with unpaired t-tests (one-tailed) after testing for skewness and kurtosis with D’Agostino- Pearson omnibus K2 test (p > 0.05). Only halfwidth and maximum repolarization rates were significantly different (p < 0.05) from a normal distribution, and therefore tested with Mann-Whitney U tests.

158

PARAMETER Mean Difference P value 95% Confidence (DAMGO-Saline) Interval for Difference Lower Upper RMP -0.05 <0.001 -0.07 0.03 AP Halfwidth -0.02 0.405 -0.08 0.09 AP Max Repolarization -0.04 0.245 -0.15 0.08 Rate Interspike Interval -0.04 0.135 -0.04 0.05 Number of Evoked APs 0.11 0.154 -0.11 0.32 AP Threshold -0.03 0.104 -0.08 0.02 AHP Amplitude -0.01 0.219 -0.12 0.09 Table 11. Comparisons for the effect of polarity shift. Neurons that depolarized in the DAMGO condition were compared against the most hyperpolarizing neurons in the saline condition. This table compares the mean changes in each measure for these groups. Statistical testing revealed no significant changes in any parameter we compared, other than the RMP which was the basis for their selection to these groups. Change in RMP was significantly different from the controls (t(20) = 4.61, p < 0.001), however changes in AP halfwidth (t(20) = 0.24, p = 0.405), max repolarization rate (t(20) =0.25, p = 0.245), interspike interval (U = 43, p = 0.135), number of evoked APs (t(20) = 1.05, p = 0.154), AP threshold (t(20) = 0.70, p = 0.245), and AHP (U = 48, p = 0.219) were nonsignificantly (p < 0.05) different from each other. Prior to statistical comparisons, distributions were tested with D’Agostino-Pearson omnibus K2 test. Normally- distributed data (p > 0.05) were tested with unpaired t-tests (one-tailed) while nonnormally-distributed data were tested with Mann-Whitney U tests (one-tailed).

Other than the change in RMP that they had been selected on, we did not find significant changes in any of the 6 other parameters. These analyses suggest that parameters following RMP were not being influenced by the change in the magnitude of their polarity that served as the maker for H-responders, but instead were a result of differential ion channel regulation that occurs during a DAMGO-response.

3.3.9 Investigating false negatives

We next wanted to determine whether there were DAMGO-induced changes in the nonresponder group, which would have suggested that these groupings were too exclusionary. This could contribute false negatives, i.e., true responders that were excluded from the responder groups and passed unanalyzed into the pool of

“nonresponders.” Out of the 55 neurons that we recorded and exposed to DAMGO, there remained 34 recordings that were excluded from the S&H-responder categories. We 159 wanted to investigate this nonresponder pool for the same changes in AP parameters that we sought in the S&H-responder categories. We decided to begin to address this issue by seeking changes in the nonresponders for the 7 measures we have been testing for.

We individually graphed the change in polarity (x axis) versus all 7 measures tested (y axis). If neurons had undergone changes in those measures without hyperpolarization, they can be visually identified by appearing vertically displaced, but horizontally close to the graph’s origin (Figure 26).

160

H-responders only S&H-responders

Fold Change in Polarization Fold Change in Polarization a vs AP Threshold b vs AP Threshold

1.2 1.2 Saline Saline 1.1 H-Responders 1.1

d Both S&H Responders

d l

o Nonresponders o

h 1.0 Nonresponders h

s 1.0

e h

0.9 h T

T 0.9

P P

A 0.8 A 0.8

0.7 0.7 0.8 0.9 1.0 1.1 1.2 0.8 0.9 1.0 1.1 1.2 Polarity Polarity Hyperpolarized Hyperpolarized

Fold Change in Polarization Fold Change in Polarization c vs Number of Evoked APs d vs Number of Evoked APs

2.0

s 2.0 s

P Saline

A f

H-Responders f

o 1.5 Both S&H Responders

o 1.5

Nonresponders r

e Nonresponders

b b

m 1.0

0.5 i

e 0.5

a a

h 0.0

h 0.0 C 0.8 0.9 1.0 1.1 1.2 C 0.8 0.9 1.0 1.1 1.2 Fold Change in Polarity Polarity Hyperpolarized Hyperpolarized

Fold Change in Polarization Fold Change in Polarization e vs Mean Interspike Interval vs Mean Interspike Interval f 2.5 2.5

Saline Saline

l a

2.0 H-Responders a 2.0 Both S&H Responders

r e

Nonresponders e Nonresponders t

1.5 t 1.5

k i

1.0 i 1.0

e t

0.5 t 0.5

I I 0.0 0.0 0.9 1.0 1.1 1.2 0.9 1.0 1.1 1.2 Polarity Polarity Hyperpolarized Hyperpolarized

161

Fold Change in Polarization h Fold Change in Polarization g vs AP Halfwidth vs AP Halfwidth 1.5 1.5 Saline Saline

H-Responders Both

h t

t S&H-Responders

d d

1.0 i 1.0

i Nonresponders

w f

f Nonresponders

0.5 0.5

A A

0.0 0.0 0.8 0.9 1.0 1.1 1.2 0.8 0.9 1.0 1.1 1.2 Polarity Polarity Hyperpolarized Hyperpolarized

Fold Change in Polarization j Fold Change in Polarization

i vs AP Max Repolarization Rate vs AP Max Repolarization Rate e

e 2.5 2.5

t a

a Saline Saline

2.0 H-Responders 2.0 Both S&H

n o

o Responders i

i Nonresponders

t a

a 1.5 1.5 Nonresponders

a l

l 1.0 1.0

p e

e 0.5 0.5

x a

a 0.0 0.0 M M 0.8 0.9 1.0 1.1 1.2 0.8 0.9 1.0 1.1 1.2 Polarity Polarity Hyperpolarized Hyperpolarized

k l Fold Change in Polarization Fold Change in Polarization vs AHP Amplitude vs AHP Amplitude

2.0 2.0

Saline Saline e

e H-Responders Both S&H Responders d

d 1.5 1.5

t t

Nonresponders i Nonresponders

l p

p 1.0 1.0

P P

H 0.5

0.5 H

A A

0.0 0.0 0.8 0.9 1.0 1.1 1.2 0.8 0.9 1.0 1.1 1.2 Polarity Polarity Hyperpolarized Hyperpolarized

Figure 27. Responder categories captured most large changes. To determine whether the changes we were attempting to detect had successfully been predicted by the S and H responder groups, we compared polarization (x ais) with the changes in AP parameters we were measuring (y axis).Vertical displacement on the y axis therefore indicates that the neuron had large changes in the measure being compared and can be compared with its polarization (horizontal displacement). Left column only includes the H-responders (blue squares), while the right column expands the responder category to both S&H-responders (blue squares). Saline-controls (black) are included in these images to illustrate the pattern of stochastic or entropic changes that should be considered in interpretations of these graphs. (a,b) AP Threshold versus polarization. We previously found no DAMGO effect in this measure, and the distribution shows scattering around the origin. (c,d) Number of evoked APs show a downward and right skew, as we expected the DAMGO responders were firing fewer evoked APs. (e,f) Interspike interval showed and upwards and right skew in the responders, as we expected, since a DAMGO response should reduce spike frequency (and

162 increase temporal spacing). (g,h) AP halfwidth showed a downwards and right skew, particularly in DAMGO responders. (i,j) AP maximum repolarization rates show and upwards and right skew in the DAMGO responders. (k,l) Afterhyperpolarization amplitude versus polarization shows an upward spread of S and H-responders. This comparison revealed that these 2 categories of responders had captured most of the large changes in those parameters, and yet there seemed to be a small cluster of

“nonresponders” that seemingly had hyperpolarized and underwent very small reductions in halfwidth and small increases in AP repolarization rates. For instance, the saline- controls never fall in the hyperpolarized and quickly-repolarizing top-right quadrant

(Figure 26j) and yet 7 nonresponders do. Therefore, examining for hyperpolarization seemed to predict the larger changes in these measures, but it left open the possibility that weaker responders were being misclassified as DAMGO-nonresponders.

To investigate false negatives that were overlooked, we tested for a significant

Slot*Group interaction in the nonresponders with a mixed model repeated measures

MANOVA (N = 34) versus the saline controls (N = 21). However, the result for the combined variables was nonsignificant (F(7,44)= 1.17, p = 0.342, Wilks’ λ = 0.156) despite the higher sample size when compared to S and H-responders. While it’s possible that some responders with weak effects (Figure 26) may have been misclassified as nonresponders, the neurons with the largest changes in measures that we were testing for seem to have been identified as H or S-responders and thus were previously analyzed here.

3.3.10 Post-hoc analyses on DAMGO, αDTX and AP kinetics

Finally, we conducted a post-hoc analysis of 47 other parameters and ratios that have not been previously addressed here. This included individually analyzing sequential APs to 163 determine if the drugs had effects only one specific APs in their spike train, as well as examining ratios for changes between sequential APs. For these post-hoc analyses, we only considered H-responders for the DAMGO and αDTX effects, because only the H- responders had systematic responses to αDTX and had stronger responses to DAMGO

(Table 10).

164

PARAMETER DAMGO DAMGO αDTX αDTX p value Full MANOVA p Full Results value MANOVA Results Membrane and Input Resistance 0.064 F(1,28) = 3.71, 0.512 F(1,27) = 2 2 Vm Threshold ηp = 0.117 0.44, ηp = Parameters 0.016 AP Number at Threshold 0.321 F(1,28) = 1.02, 0.581 F(1,27) = 2 2 ηp = 0.035 0.31, ηp = 0.011 AP Halfwidth at <0.001 F(1,28) = 0.536 F(1,27) = 2 2 Threshold 23.13, ηp = 0.39, ηp = 0.452 0.014 First AP Maximum Depolarization <0.001 F(1,30) = 0.917 F(1,29) = 2 2 Parameters Rate 22.47, ηp = 0.78, ηp = 0.428 0.011 Average Depolarization 0.001 F(1,30) = 0.472 F(1,27) = 2 2 Rate 13.33, ηp = 0.53, ηp = 0.308 0.018 Time to Max 0.001 F(1,30) = 0.612 F(1,27) = 2 2 Depolarization Rate 12.32, ηp = 0.26, ηp = 0.291 0.009 Time to peak 0.003 F(1,30) = 0.785 F(1,27) = 2 2 10.60, ηp = 0.08, ηp = 0.251 0.003 Peak Amplitude 0.084 F(1,30) = 3.19, 0.172 F(1,27) = 2 2 ηp = 0.096 1.96, ηp = 0.063 Maximum Repolarization 0.001 F(1,30) = 0.917 F(1,27) = 2 2 Rate 14.32, ηp = 0.01, ηp < 0.323 0.001 Average Repolarization 0.001 F(1,30) = 0.716 F(1,27) = 2 2 Rate 13.65, ηp = 0.14, ηp = 0.313 0.005 Time to max 0.014 F(1,30) = 6.75, 0.949 F(1,27) < 2 2 Repolarization Rate ηp = 0.184 0.01, ηp < 0.001 AP1 Halfwidth <0.001 F(1,30) = 0.341 F(1,27) = 2 2 21.79, ηp = 0.94, ηp = 0.421 0.031 AP1 90% Width <0.001 F(1,30) = 0.949 F(1,27) < 2 2 21.79, ηp = 0.01, ηp < 0.421 0.001 Afterhyperpolarization 0.042 F(1,30) = 4.53, 0.043 F(1,27) = 2 2 Amplitude ηp = 0.131 4.48, ηp = 0.134 Second AP Maximum Depolarization 0.001 F(1,29) = 0.462 F(1,28) = 2 2 Parameters Rate 14.02, ηp = 0.56, ηp = 0.326 0.019 Average Depolarization 0.001 F(1,29) = 0.501 F(1,28) = 2 2 Rate 15.56, ηp = 0.47, ηp = 0.334 0.016 Time to Max 0.001 F(1,29) = 0.749 F(1,28) = 2 2 Depolarization Rate 13.80, ηp = 0.11, ηp = 0.322 0.004

165

Time to peak 0.961 F(1,29) < 0.01, 0.265 F(1,28) = 2 2 ηp = 0.000 1.30, ηp = 0.044 Peak Amplitude 0.028 F(1,29) = 5.34, 0.110 F(1,28) = 2 2 ηp = 0.156 2.73, ηp = 0.089 Maximum Repolarization <0.001 F(1,29) = 0.970 F(1,28) < 2 2 Rate 16.45, ηp = 0.01, ηp < 0.362 0.001 Average Repolarization <0.001 F(1,29) = 0.748 F(1,28) = 2 2 Rate 15.42, ηp = 0.11, ηp = 0.347 0.004 Time to max <0.001 F(1,29) = 0.438 F(1,28) = 2 2 Repolarization Rate 17.35, ηp = 0.62, ηp = 0.374 0.022 AP2 Halfwidth <0.001 F(1,29) = 0.915 F(1,28) = 2 2 18.04, ηp = 0.01, ηp < 0.384 0.001 AP2 90% Width 0.229 F(1,29) = 1.51, 0.868 F(1,28) = 2 2 ηp = 0.049 0.03, ηp = 0.001 Afterhyperpolarization 0.001 F(1,29) = 0.059 F(1,28) = 2 2 13.42, ηp = 1.91, ηp = 0.316 0.122 Average of Maximum Depolarization 0.001 F(1,29) = 0.371 F(1,28) = 2 2 AP3+ Rate 14.11, ηp = 0.83, ηp = 0.327 0.029 Average Depolarization 0.001 F(1,29) = 0.445 F(1,28) = 2 2 Rate 15.08, ηp = 0.60, ηp = 0.342 0.021 Time to Max <0.001 F(1,29) = 0.716 F(1,28) = 2 2 Depolarization Rate 27.15, ηp = 0.14, ηp = 0.484 0.005 Time to peak 0.332 F(1,29) = 0.440 F(1,28) = 2 2 16.21, ηp = 0.61, ηp = 0.033 0.021 Peak Amplitude 0.023 F(1,29) = 5.76, 0.098 F(1,28) = 2 2 ηp = 0.166 2.93, ηp = 0.095 Average Repolarization 0.001 F(1,29) = 0.607 F(1,28) = 2 2 Rate 14.89, ηp = 0.27, ηp = 0.339 0.010 Time to max 0.002 F(1,29) = 0.453 F(1,28) = 2 2 Repolarization Rate 12.02, ηp = 0.58, ηp = 0.293 0.020 AP Multi 90% Width 0.001 F(1,29) = 0.759 F(1,28) = 2 2 12.34, ηp = 0.10, ηp = 0.298 0.003 Interspike ISI Slope 0.016 F(1,26) = 6.71, 0.184 F(,25) = 2 2 Properties ηp = 0.205 1.86, ηp = 0.069 ISI Standard Deviation 0.029 F(1,26) = 5.33, 0.934 F(1,25) = 2 2 ηp = 0.170 0.01, ηp < 0.001

166

ISI min 0.001 F(1,26) = 0.001 F(1,25) = 2 2 15.36, ηp = 10.97, ηp 0.371 = 0.305 ISI max 0.014 F(1,26) = 6.93, 0.625 F(1,25) = 2 2 ηp = 0.210 0.25, ηp = 0.010 ISI covariance 0.114 F(1,26_ = 2.67, 0.417 F(1,25) = 2 2 ηp = 0.093 0.68, ηp = 0.027 Instantaneous Frequency <0.001 F(1,26) = 0.002 F(1,25) = 2 2 Median 18.77, ηp = 11.41, ηp 0.419 = 0.313 Instantaneous Frequency 0.002 F(1,26) = 0.003 F(1,25) = 2 2 Mean 12.30, ηp = 9.07, ηp = 0.321 0.266 Instantaneous Frequency <0.001 F(1,26) = <0.001 F(1,25) = 2 2 Max 18.95, ηp = 21.82, ηp 0.422 = 0.466 Instantaneous Frequency 0.002 F(1,26) – 0.335 F(1,25) = 2 2 Min 11.28, ηp = 0.97, ηp = 0.304 0.037 Ratios AP2/AP1 Halfwidth 0.237 F(1,26) = 1.47, 0.492 F(1,25) = 2 2 ηp = 0.053 0.49, ηp = 0.019 AP3+/AP1 Halfwidth 0.342 F(1,26) = 0.94, 0.668 F(1,25) = 2 2 ηp = 0.035 0.19, ηp = 0.007 AP2/AP1 AHP Amplitude 0.463 F(1,26) = 0.58, 0.916 F(1,25) = 2 2 ηp = 0.022 0.01, ηp < 0.001 APmulti/AP1 AHP 0.669 F(1,26) = 0.19, 0.387 F(1,25) = 2 2 Amplitude ηp = 0.007 0.19, ηp = 0.030 ISI max/min 0.453 F(1,26) = 0.58, 0.113 F(1,25) = 2 2 ηp = 0.022 2.70, ηp = 0.097 Table 12. Posthoc summary of electrophysiological changes. Our Python script analyzed 44 measures and 5 ratios, and only 7 have been discussed previously here. We therefore analyzed the 44 remaining measures and 5 ratios not presented yet to determine whether DAMGO or αDTX were have significant effects on parameters that have not yet been discussed here. Several mixed-model repeated measures MANOVAs were repeated for every subcategory (left column) for both DAMGO and αDTX to identify Slot*Drug interaction effects which may indicate that those drugs were modulating the properties listed on the left- side column H-responders (N = 11) were compared with saline controls (N = 21) on various parameters for a Slot*Group interaction that would suggest a drug (DAMGO or αDTX) that was independent of degradation reflected in the saline control group. Timeslots analyzed were Initial → DAMGO for DAMGO effects, and DAMGO → αDTX+DAMGO for αDTX effects in a background of DAMGO. In order to determine whether these drugs acted only specifically on certain APs (e.gs., only the first AP, second AP, or only APs after that), we analyzed these APs separately for Slot*Group interactions. Boldtype are significant effects.

We found that DAMGO was affecting a very wide range of AP kinetic parameters across many sequential APs, with the notable exception of spike amplitude. However,

167

αDTX’s effects were not seen in these parameters (Table 11). Broadly speaking, it appears that DAMGO consistently alters many different features, from individual AP kinetics, to the probability of firing an AP. Meanwhile αDTX-sensitive channels do not seem to contribute the shape of APs in neocortical interneurons, and its effects appear to be restricted to the probability of firing an AP and their firing frequency.

3.4 Discussion

Previous research indicates that the µOR exerts an excitatory effect on cortical networks by suppressing inhibitory neurons (Homayoun & Moghaddam, 2007; C. Jiang et al.,

2019; H. K. Lee et al., 1980; Madison & Nicoll, 1988; Pang & Rose, 1989;

Zieglgansberger et al., 1979). The experiments here likewise demonstrate that the µOR exerts a strong inhibition on many cortical interneurons. We found that over a third

(21/55; 38.1%) of the sample of interneurons either hyperpolarized (H-responder) or had reduced spontaneous action potentials (S-responder). The influence of DAMGO in these two populations of interneurons had overlapping characteristics but were distinct. For instance, in H-responders DAMGO elicited clear hyperpolarization and affected AP kinetics and firing frequency. In contrast, the S-responders exhibited relatively smaller hyperpolarizations and small changes in AP kinetics, and no alterations in the probability of somatically evoked APs.

The high degree of variability of neocortical interneurons between and within their classes makes identification of interneurons in our system difficult. We did not attempt cell-marker identification of the neocortical interneurons since identification would have required a thorough multimodal analysis of their characteristics (The Petilla

168

Interneuron Nomenclature Group, 2008). Particularly because our system enabled more expeditious data collection at the expense of laminar arrangement - itself an identifying characteristic for interneurons.

We found that most H-responders displayed an Adapting firing pattern, characterized by steadily-increasing interspike intervals. This is noteworthy because this spiking pattern was found at much lower rates in the non-responders and S-responders

(Table 5). Other researchers have found interneurons with this firing pattern amongst neuropeptide Y-expressing interneurons (Bruno Cauli et al., 1997; B. Cauli et al., 2004;

Férézou et al., 2007; Toledo-Rodriguez, Goodman, Illic, Wu, & Markram, 2005; Yun

Wang, Gupta, Toledo-Rodriguez, Wu, & Markram, 2002; Y. Wang et al., 2004). This firing pattern in NPY+ interneurons is particularly common amongst Layer I interneurons of the neocortex (Karagiannis et al., 2009), which others have found to express high rates of µORs and hyperpolarize strongly to DAMGO exposure (Férézou et al., 2007), which was the distinguishing feature of H-responders in our study but found in both groups of responders.

Our data also suggest that some of the S or H-responders may have features consistent with VIPergic interneurons of the neocortex based on reductions in inhibitory postsynaptic potentials that we observed during recording. These experiments did not include the use of GABAA and GABAB inhibitors, and IPSPs could therefore persist through the spontaneous activity of the interneurons. Although we did not analyze these events, we noticed reductions in IPSPs occurring after addition of DAMGO in some interneurons, including some H-responders (data not shown, though partially visible off-

169 pulse in Figure 23). The VIPergic interneurons tend to inhibit other interneurons, which may have been manifesting our data (Fu et al., 2014; Jackson et al., 2016; Pi et al., 2013).

It is interesting to note that previous research in neocortical slices has found a high degree of overlap between VIP and µOR immunoreactivities (Taki et al., 2000).

Electrophysiological studies have also found high rates of VIP expression amongst interneurons that hyperpolarize in response to DAMGO, including neurons that would have been classified here as having Adapting spiking patterns (Férézou et al., 2007).

While speculative, this may hint at a relationship between the H and S responders and

VIPergic neurons of the neocortex.

It is also important to point out that H and S-responders may not represent discrete populations of interneurons despite the different methodologies to distinguish them from nonresponders; numerous effects we found in H-responders fell just short of significance in the S-responders (Table 9). For example, many S-responders had particularly robust changes in AHP amplitude, which was not captured in the group mean. Although the S-responders generally had weaker responses to DAMGO (Table

10), significant differences in this group may have been more difficult to resolve due to lower samples sizes (N = 10), versus the H-responders (N = 11).

We also found that neurons within the H-responder group possess a variety of spiking patterns (Table 5) and also a range of responses to both DAMGO and αDTX

(Figure 22). Within the hyperpolarizing DAMGO responders, we found a spectrum of changes in their AP kinetic parameters; some neurons had pronounced changes in their parameters, and some changed barely at all despite their greater hyperpolarization. The

170 assortment of ion channels in each particular neuron may have led to varied responses to

DAMGO, but variability may also have resulted from the complex intracellular signaling cascades of GPCRs; µORs are shown in non-cortical regions to modulate αDTX- sensitive currents through the Gαi subunit, phospholipase A2 and the lipoxygenase pathway (Faber & Sah, 2004; Vaughan et al., 1997), meanwhile upregulation of GIRK channels appears to be mediated through direct interaction by activated Gβγ subunits (C.-

L. Huang, Slesinger, Casey, Jan, & Jan, 1995; Inanobe et al., 1995; Logothetis, Kurachi,

Galper, Neer, & Clapham, 1987; Wickman et al., 1994). It is therefore conceivable that neurons have a variety of responses to DAMGO ranging from activating GIRK channels, to activating voltage-sensitive currents, based on diverging signaling cascades.

The neurons in the H and S-responder categories were identified using two different measures (RMP and spontaneous APs, respectively), which themselves can be manifestations of different target ion channels of µORs, or different localization of the

µORs among the neuron types. Studies in the cortex suggest that hyperpolarization is not always found in a DAMGO response, and that there is incomplete overlap between GIRK channels and µORs (S. B. Bausch et al., 1995; Wimpey & Chavkin, 1991). Therefore, larger hyperpolarization may not always be seen in µOR+ neurons if GIRK channels are not involved. However, the drop in spontaneous APs seen in the S-responders is another interesting feature. Studies of this receptor in neocortical interneurons using DAMGO have found reductions in non-elicited APs, which appear to originate in distal processes of some interneurons (Krook-Magnuson et al., 2011; Norimitsu Suzuki et al., 2014).

171

These mechanisms illustrate that the µOR has several known mechanisms to downregulate interneurons which can vary among the interneuronal types.

We found a range of AP kinetics as well as firing frequency properties (Table 10) are altered by the µOR. We found that the AP kinetics were not likely to be mediated by the αDTX-sensitive channels, though measures of AP probability (ISI, AP number, and

AP threshold) were related to the αDTX-sensitive current (Figure 21 and Table 8). We found that αDTX frequently reversed (Figure 22) DAMGO-induced changes in spiking frequency in the H-responders with a wide spectrum of responses, with neurons that underwent DAMGO-induced changes in ISI and evoked APs usually being partially or fully reversed to baseline. Interestingly αDTX did seem to have an effect in 2 or 3 neurons in RMP and AP halfwidth which was not captured by the mean; however, it is difficult to say whether this was a genuine reaction to the drug by certain subclasses of interneurons, or if they were simply outliers in the data (Figure 22).

It is interesting that αDTX reversed DAMGO effects in ISI and number of evoked

APs without reversing any of the AP kinetic features that we measured. The mechanism behind the αDTX-sensitive current suppression of follow-up APs is not immediately clear considering that changes in AHP magnitude was not significantly changed by DAMGO nor αDTX. In other parts of the brain, this current is responsible for modulating spike frequency adaptation – a feature that limits follow-up APs during prolonged depolarizations (Faber & Sah, 2004). It appears therefore possible that, based on

DAMGO’s effect and αDTX’s (usually) partial reversal of it (Figure 22) that other voltage-sensitive potassium currents are being modulated by µORs. For example, the

172

Kv3 and Kv4 family potassium channels are found in cortical neurons and are implicated in modulating these properties (Burkhalter et al., 2006; Carrasquillo et al., 2012; Martina et al., 1998; B Rudy et al., 2009). The relationship of AP threshold with the µOR and

αDTX-sensitive current was a particularly nuanced feature in hyperpolarizing DAMGO responders and appears to track with prior findings from the BLA where µORs modulate

αDTX-sensitive current without changing the Vm threshold for an AP (Faber & Sah,

2004). Here, αDTX produced a hyperpolarizing shift in AP threshold despite the µOR’s apparent noninfluence over AP threshold. Yet the µOR still appears to regulate αDTX current; µOR stimulation with DAMGO suppressed follow-up APs, which was partially reversed by αDTX. At face value, it appears that µOR+ interneurons have αDTX- sensitive currents at rest that influence AP threshold, yet these currents can be upregulated to suppress APs after µOR-stimulation. One explanation is that αDTX- sensitive channels are expressed in areas that the µOR is not localized to. For instance,

αDTX-sensitive channels in axon initial segments of cortical neurons can modulate AP threshold (Inda, DeFelipe, & Muñoz, 2006; Kole, Letzkus, & Stuart, 2007), but if the

µORs localize to soma and dendrites, they may very well modulate somatic channels, but not AIS-localized ion channels, where they may be already functioning without DAMGO stimulation.

One drawback of this study is that we did not test the inverse drug exposure of

αDTX before DAMGO and therefore it is difficult to assess the baseline function of the

αDTX-sensitive channels in DAMGO-responders to directly relate µOR-stimulation with the upregulation of αDTX-sensitive channels. When designing these experiments, we

173 believed it might have been possible that αDTX would block DAMGO-induced hyperpolarization (the primary indicator of a positive DAMGO response) and therefore prevent us from ultimately identifying H-responders. However, we found that αDTX usually failed to reverse DAMGO effects, including hyperpolarization. The inverse drug sequence may be the most ideal way to measure baseline αDTX activity in DAMGO responders. On the other hand, there were some neurons that did undergo noticeable depolarization and had pronounced increases in spontaneous APs, and we may have overlooked several H-responders and S-responders had we used inverse drug exposures

(Figure 22a).

It may be argued that the hyperpolarizing effect of DAMGO found in the hyperpolarizing responders and, to a lesser degree, in spontaneous DAMGO responders may have influenced action potential parameters. This is unlikely for several reasons.

First, most of the action potential analyses were done at several current steps above threshold and we used the average values for APs 3-5 in those trains, thereby mitigating the influence of hyperpolarized RMP on action potential parameters. In addition, most of the APs we observed were initiated from approximately the same Vm regardless of their change in RMP. Second, αDTX reversed DAMGO-induced changes in ISI and number of evoked APs in hyperpolarizing DAMGO responders despite αDTX’s failure to reverse

DAMGO-induced hyperpolarization (Table 6 and Figure 21a) and broad range of other features in the post-hoc analyses (Table 12). Third, we compared DAMGO depolarizers

(predicted to have had no µOR response) against neurons that had hyperpolarized slightly in the saline group; this comparison “flipped the sign” to the saline group being

174 significantly more hyperpolarized then the DAMGO group. However, in all measures resulting from this comparison, we found no significant differences in action potential parameters, ISI, and number of evoked action potentials (Figure 25). This analysis suggests that the parameters we were investigating during the current steps were not being influenced by the neuron’s RMP regardless of if the neuron had depolarized under the DAMGO challenge.

Our data demonstrates that the μOR modulates AP parameters, including the halfwidth and maximum rates of repolarization to suppress the excitability of interneurons. We observed halfwidth changes of 12% (S-responder mean) 21% (H- responder mean) and mean maximum repolarization rates of 25% or 40% (Table 10) - with also a high degree of variation between neurons (Figure 22). The consequences of

AP kinetics have been explored in models with large synapses by using real and pseudo

APs to shown that kinetics have consequences for voltage-gated calcium channel

(VGCC) activity, calcium entry, and neurotransmitter release (Augustine, 1990; Klein &

Kandel, 1980; Llinas, Steinberg, & Walton, 1981). Computer simulations and developmental data from synapses suggest that most VGCCs are activated by an action potential, but further broadening prolongs the kinetics of the VGCCs, and therefore contributes to longer duration of Ca2+ entry (Borst & Sakmann, 1998; Geiger & Jonas,

2000; Sabatini & Regehr, 1997), Additional data from the calyx of Held have added that depolarization phases affect the number of VGCCs recruited, while repolarization affects their kinetics to influence the amount or duration of Ca2+ entry. Paired recordings have shown that both features can modulate the release of neurotransmitter and the amplitude

175 of postsynaptic potentials, depending on stages of development and adaptations at the synapse (Chao & Yang, 2019; Y.-M. Yang & Wang, 2006)

The αDTX-sensitive channels Kv1.1, Kv1.2, and Kv1.6 have previously been linked to regulation of cortical interneurons. In the neocortex, they have been found to strongly influence firing through their localization at axon initial segments of interneurons and excitatory neurons (Bekkers & Delaney, 2001; Guan et al., 2006). The have previously been linked to the excitability of cortical interneurons by modulating AP threshold and near-threshold spike frequency (Goldberg et al., 2008). In VIPergic interneurons, these currents affect the frequency of output (Porter et al., 1998). Genetic knockout of Kcna1 (encodes Kv1.1) and Kcna2 (encodes Kv1.2) are linked to seizures and ataxia in both humans and rat models (Adelman, Bond, Pessia, & Mayliet, 1995;

Brew et al., 2007; Heeroma et al., 2009; Robbins & Tempel, 2012). Restoring regulated activity of these channels has been proposed as a mechanism underlying a treatment for

Fragile X syndrome (Y.-M. Yang et al., 2020). Here, our data suggest that the μOR may upregulate these currents, as evidenced by αDTX’s attenuation of the DAMGO effect on firing frequency (Figure 22). The endogenous opioids of the neocortex are deeply involved in reward-seeking and motivated behaviors. Studies in animal models have shown enhanced µOR activity correlates with, and causes, binge-eating and drug intake

(Blasio et al., 2014; Morganstern, Liang, Ye, Karatayev, & Leibowitz, 2012; Unterwald,

Rubenfeld, & Kreek, 1994). The µOR is believed to reshape neocortical activity by releasing Pyramidal Neurons from inhibition by cortical interneurons (Férézou et al.,

2007; Homayoun & Moghaddam, 2007; C. Jiang et al., 2019; H. K. Lee et al., 1980;

176

Madison & Nicoll, 1988; Pang & Rose, 1989; Taki et al., 2000; Zieglgansberger et al.,

1979). Yet despite the supporting evidence for µOR’s disinhibitory role in the neocortex, this receptor’s inhibitory influence has rarely been explored beyond its hyperpolarizing effects (Baldo, 2016; Férézou et al., 2007). Here, we found that the neocortical µOR has far-reaching inhibitory effects in neocortical interneurons that had not yet been described, which includes the potential upregulation of αDTX-sensitive channels to modulate interneuronal output through firing frequency.

CHAPTER 4: COMPREHENSIVE DISCUSSION

4.1 Overview

These chapters provide a comprehensive summary of the neurophysiological effects of

µOR agonism by documenting the electrophysiological effects within cortical interneurons, as well as the downstream hyperactivity effects by using spontaneous calcium oscillations s as an indicator for bursts of neuronal activity. These findings advance our knowledge of the µOR significantly by providing a robust statistical profile of DAMGO effects (Table 12) while also showing that most of the µOR’s net excitatory effect appears to be secondary to inhibition of cortical interneurons. This dissertation sheds light on the mechanisms and effects of the µOR, how opioid drugs dysregulate cortical activity, and how endogenous opioid signaling modulates networks and individual neurons.

177

The studies in this dissertation are laid on a foundation of confirmatory evidence for research that previous investigators have discovered, while furthering our knowledge of the µOR as well. For example, we detected µOR responses in the expected range of

1/3rd of cortical interneurons, and that hyperpolarization and reduction of spontaneous

APs is a typical indicator, which has been observed (Férézou et al., 2007). On top of that foundation, we show that the µOR has many other effects on cortical interneurons that have not been previously described. Additionally, the conclusions and data presented here provide numerous other possible research avenues that can be utilized to thoroughly understand the target channels of the µOR, as well as to understand the net effects of the receptor on neuronal circuits.

Our calcium imaging studies also corroborate previous reports that investigated the receptor – that it reduces GABA release onto both GABAA and GABAB receptors

(Olah et al., 2007; Tamás, Buhl, Lörincz, & Somogyi, 2000). But our experiments show that calcium-imaging and SCOs can be used to detect this relationship as well, and they lend a convenient model to studying the network dynamics by visualizing the network activity of neurons. Furthermore, this model provided an effective means to assess the respective roles of GABAA and GABAB receptors on the SCOs.

Data from electrophysiology corroborate other studies from neocortical interneurons. For instance, we found that DAMGO hyperpolarizes interneurons and reduces spontaneous firing. We found that the µOR employs several fascinating mechanisms for suppressing interneurons, including hyperpolarization, but it also affects

AP kinetics – which has not previously been described. Therefore, the µOR utilizes

178 several mechanisms to suppress interneurons, and many of these means have not been previously described before the detailed analyses presented here.

4.2 αDTX-sensitive/insensitive effects

Our electrophysiology experiments sought to determine whether the µOR positively regulated αDTX-sensitive currents, and our results suggest that they do; we observed a

αDTX-reversible change in AP frequency in H-responders. However, we did not narrow this down to specific subunits (e.gs: Kv1.1, Kv1.2, and Kv1.6). It is presently unclear which subunits are mediating this effect, though Kv1.1 or Kv1.2 are the mostly likely suspects, given their previously-established relationship with the µOR (Faber & Sah,

2004; Finnegan et al., 2006; Vaughan et al., 1997) and their known localization in cortical interneurons (J. Connor & Stevens, 1971; X. Li et al., 2012).

One of our main findings was that the µOR has far-reaching neurophysiological effects and the αDTX-sensitive current mostly modulates firing frequency, and not the

AP kinetics. Naturally, this does lead to the question of which ion channels are mediating these effects since it appears kinetics may be mediated by other channels. The most likely suspects are K channel subunits that are more classically known to mediate faster (A- type) current (B Rudy et al., 2009), even though αDTX-sensitive currents are themselves capable of faster kinetics (Casale et al., 2015; Geiger & Jonas, 2000; Pathak et al., 2016).

Considering the tendency for αDTX-sensitive channels to heterotetramerize when the two are coexpressed, it would also be interesting to determine whether this takes place in responder neurons (B Rudy et al., 2009). Further identification of subunits mediating this effect of DAMGO could therefore be warranted.

179

4.3 Predicting a DAMGO response

Future studies in this area may require higher response rates than achieved here. Data collected from the electrophysiological experiments mostly use random selection of mRuby2-positive interneurons to record from and we achieved an average response rate of 1/3rd between the H-responder and S-responder criteria. While this response rate may be adequate for future studies, there may also be other diagnostic criteria that can be used to predict DAMGO responses. Data collected and presented in this dissertation can also provide a basis for more targeted experiments directed at understanding the electrophysiology of the µOR.

One of our first attempts to predict DAMGO responses was the use of morphology. Unfortunately, we found that neurons in vitro had indiscernible shapes that could not be identified; nearly every neuron appears to have a spherical soma with indistinct neurites. Shapes such as pyramidal, bipolar/bitufted, or multipolar cannot easily be determined in these cultures. However, the use of the mRuby2 AAV improves this situation substantially, not only by selectively fluorescing the interneurons, but also allowing the viewer to resolve the precise and shape and neurites of each neuron (Figure

8 and Figure 14). While I did not attempt to use this technique to predict a DAMGO response, interneuronal shapes are much more resolvable with mRuby2. Anecdotally, it may not be an entirely useful property to predict a DAMGO response, however mRuby2 makes it exceptionally easy to visualize the shapes of interneurons and, importantly, it can be used effectively before committing to the slow step of the recording process: sealing and breaking in.

180

Resting membrane potential and membrane resistance tend to be characteristic features of neurons as well. However, use of these features to predict a DAMGO response would probably be unsuccessful. We found RMP tended to fall in a fairly tight range, and membrane resistance is sometimes difficult to consistently determine in real- time. In addition, these properties can vary over the course of seconds, and spontaneous spiking of the neurons also disrupts accurate reading. We typically calculated the RMP based on an average value that represented its RMP over a period of seconds, rather than being an instantaneous determination. Therefore, these basic properties of neurons were not very useful for predicting DAMGO responses when they are put into practice during the recording process. This made these properties more problematic to utilize to predict a

DAMGO response.

Our electrophysiological data suggest that most H-responders have Adapting firing patterns, and that the Adapting firing pattern is less prevalent in the nonresponder pool. Future investigations into the µOR should consider using this firing pattern to predict a response. However, half of H-responders did not display an Adapting firing pattern, and thus using this diagnostic feature may rule out potential H-responders. For example, these Adapting neurons likely have unique ion channel profiles that lend them their unique firing pattern. Including or excluding them can skew analyses. On the other hand, perhaps future investigations may choose to focus on neurons with that firing property. In summary, the most useful predictors for DAMGO responses are most likely the spiking pattern, which can only be determined after the slow steps of the recording

181 process (positioning and intrusion of the electrode), and also morphology – but only with the assistance of the mRuby2 AAV credited in the methods section of those experiments.

4.4 Identifying responders by subtype

Cortical interneurons are remarkably diverse and often identified through the use of protein-expressional markers (The Petilla Interneuron Nomenclature Group, 2008). To expedite the data collection process, we did not identify neurons using protein markers.

Unfortunately, equipment malfunctions prevented us from collecting images for most of the experiment as well. We instead utilized spike patterns to classify interneurons, which is also a well-recognized characterizing feature (The Petilla Interneuron Nomenclature

Group, 2008). However subtyping interneurons by cell marker could assist in making inferences about the roles of these interneurons in intact neocortex – in addition to perhaps explaining some of the variability that we observed (Figure 22).

Our experiments grouped neurons into H-responders and S-responders categories based on their response to DAMGO. These typings appeared to be necessary, because they had been identified based on their drug responses, which were different criteria from each other. However, it is possible that interneurons in the 2 groups are actually more similar than they are presented, and their neuronal categories and typing by interneuronal marker may reveal overlap between the populations. On the other hand, H-responders were considerably more likely to be Adapting firing patterns than S-responders or nonresponders, and therefore it is quite possible that the H-responder and S-responders are composed of different interneuronal populations.

182

4.5 Origin of spontaneous APs

We used a decreased rate of spontaneous APs as a diagnostic feature of S-responders in this study. One interesting area of interest raised by this observation and technique is the exact cause of spontaneous APs in this model, and the mechanism by which spontaneous

APs were produced and reduced. The simplest explanation is that these neurons RMP was very close to its AP threshold, and the frequency of APs was simply a matter of probability of APs at that Vm. While hyperpolarization was not the diagnostic criterion for this group, the S-responders did hyperpolarize on average (Figure 23). Therefore, this is a simple and plausible scenario for this observation.

Another more interesting and complex explanation may be lent by pioneering studies in PV+ and neurogliaform neurons, which suggest that APs in these neurons can be generated in distal processes, such as their axons and conduct back towards their soma. This phenomenon is called retroaxonal barrage firing. These APs may only be accompanied by small hyperpolarizations, which possibly indicates a larger hyperpolarization in its distant axon. DAMGO has already been shown to reduce these tonic APs, and thus this mechanism may apply here as well (Krook-Magnuson et al.,

2011; M. E. Sheffield et al., 2013; M. E. J. Sheffield et al., 2011; Norimitsu Suzuki et al.,

2014). In addition, these tonic APs can occur after depolarizing trains of current, due to axonal terminals becoming tonically active and conduction of APs through gap junctions which can be conducted into other interneurons; therefore, these are delayed evoked action potentials, rather than truly being unevoked and spontaneous APs. The involvement of gap junctions suggests that perhaps S-responders may not necessarily

183 indicate a µOR+ neuron, but instead a neuron junctionally-linked to an µOR+ neuron.

However, this phenomenon has never been demonstrated in neuronal culture models, and the frequency of this type of spiking seems to decrease from in vivo to brain slices

(Krook-Magnuson et al., 2011; M. E. Sheffield et al., 2013; M. E. J. Sheffield et al.,

2011; Norimitsu Suzuki et al., 2014). It’s therefore quite possible this phenomenon may not occur in neuronal cultures.

This phenomenon, however, can be experimentally tested for, as has already been done in vivo and brain slices (Krook-Magnuson et al., 2011; M. E. Sheffield et al., 2013;

M. E. J. Sheffield et al., 2011; Norimitsu Suzuki et al., 2014). In this case a long depolarizing current could be applied and the “spontaneous” APs in the minutes that follow this train could be counted. By comparing these tonic action potentials before and after the current application, it may be possible to test for retroaxonal barrage firing in vitro.

4.6 Drawbacks of whole-cell recording

Our approach during electrophysiology was establishing whole-cell configuration by breaking into neurons for recording. Unfortunately, this may lead to artifactual changes, as opposed to cell-attached which would have allowed us to record from neurons without breaking into them and thereby altering their intracellular environment. Cell-attached recordings were initially attempted during these experiments, but then aborted due to spontaneous break-ins during the recording process. It became clear that the sample size would have been prohibitively small to allow us to record from enough neurons to have a reasonable number of responders, and we switched to whole-cell recording. Therefore the

184

“real” values to AP kinetics may be subjected to artifactual challenges. However, we did record from many saline controls as well, and our main data analyses were performed by comparing fold changes in DAMGO-exposed neurons with those saline controls.

Therefore, artifacts were most likely evenly reflected in both groups and fold changes that are calculated should still be accurate, even if true values may be subject to some deviation. Utilizing this whole-cell recording approach therefore allowed us to collected substantial amounts of recordings in a relatively short period, and further enabled the data analysis step to identify changes in AP kinetic parameters occurring in responders. In many ways, this expedited approach facilitated the robust data analysis and statistical profiling shown here.

4.7 DAMGO and GABARs

Our experiments with SCOs are a unique approach in determining how DAMGO affects

GABAergic signaling and identifying the GABARs that are mediating that effect. This approach enabled us to record from many (20+) neurons in a single recording, in order to understand the dynamics of network activity and how DAMGO affects it. Using this model, we reproduced several important findings in µOR research while answering key questions about the receptor’s mechanisms. Firstly, this model successfully showed that

DAMGO does indeed lead to net excitation in neuronal networks, which was well- indicated by the SCO duration (halfwidth) enhancement by DAMGO. Secondly, our data corroborate the well-supported view that DAMGO increases net excitation in neuronal circuits by inhibiting GABAergic interneurons, and that net excitation is secondary to interneuronal suppression.

185

Our data also advance our knowledge of the field in several important ways; we show that not only does DAMGO suppress GABA release onto GABAA receptors, but it also suppresses GABA release onto GABAB receptors as well. On one hand, this result is well-aligned with the established view that µORs localize to neurogliaform neurons which activate postsynaptic GABAB receptors, but our data further suggest that the effect of GABAB may be influential for terminating bursts of activity that are indicated by each

SCO. While extrapolating from these data into in vivo brain structures is inherently speculative, these data may suggest that DAMGO enhances net excitation by promoting bursts of activity and increasing the lag time before GABAergic interneurons are capable of terminating that burst of activity – which clearly seems to be a role of GABAB receptors in this model. Whereas our electrophysiological recordings helped us determine the effects of DAMGO in single neurons, the SCO model illustrated how networks of neurons interact and suggest that there may be a delay in responsiveness to excitation within cortical interneurons. This aspect of time lag to inhibit discrete bursts of activity in cortical networks could be an interesting topic to explore.

4.8 DAMGO inhibition and GABA release

As explained, we observed far-reaching electrophysiological consequences for the 1/3rd of interneurons susceptible to DAMGO. But what about the consequences of this for synaptic GABA release? This question is largely unanswered by the current data – though the data in this dissertation certainly provide a necessary basis for such follow-up experiments; we identified AP kinetic changes, but it’s not clear how this relates to

186

GABA release. This topic could be explored by dual electrophysiological recordings to record inhibitory postsynaptic potential magnitudes.

There are likewise similar experiments that could be done to address the neuronal networks without the use of CNQX + AP5 at the neuronal network model. For example, the cultured neurons could be exposed to DAMGO and concentrations of GABA could be quantitatively measured before and after certain timepoints to determine how this manipulation affects GABA release. Though, the results from this experiment may be difficult to predict or draw conclusions from. For instance, DAMGO may counter- intuitively increase GABA concentrations due to the greater glutamatergic input on

GABAergic interneurons; GABA concentrations during the experiment would provide little information about whether bursts of GABA release into synapses was happening too slowly. Perhaps another important factor to consider, for reasons described later in this chapter, is proportionality to their glutamatergic input, i.e., standardizing the GABA concentration by glutamate concentration. This may help evaluate not only whether

GABAergic interneurons are increasing GABA release, but whether their response is proportional to glutamatergic input before and after DAMGO.

4.9 True role of µOR+ interneurons in cortical networks

Our SCO model for DAMGO responses showed a powerful way to visualize disinhibition and cortical hyperactivity. Yet, this was one of the biggest surprises for these experiments

– why was hyperactivity with DAMGO so common at the circuit level, and yet the receptor acted mainly through disinhibition (interneurons), and interneurons expressed the receptor uncommonly?

187

There were several striking features about the SCOs that we observed, which were surprising given the likely low expression of the receptor. In our early experiments with SCOs, we noticed how DAMGO’s enhancement of SCO duration (and sometimes amplitude) were particularly prevalent. There may be several explanations for this.

Firstly, we observed 3 - 11 red interneurons in each imaging field, and our electrophysiological experiments suggested that about 1/3rd were inhibited by DAMGO.

It is therefore possible that an average of 1-3 neurons in every frame were µOR+. We also observed that neurites at this stage spanned far distances. Taken together, this suggests that the interneurons were highly interlinked with many other neurons in an imaging field (Figure 8 and Figure 14). Therefore, despite their low frequency, even one

µOR+ interneuron in a field could potentially influence many other neurons in a field.

However, there is a counterargument to this point, which may further reduce the number of DAMGO responders that enable DAMGO-enhancement of SCOs; about a third of responders (8/21) did not have spontaneous firing, which reduces the 1/3rd responder proportion even lower, because many responders were not actively firing at rest, and therefore inhibition of those neurons would presumably produce very little effect on the network of neurons. To summarize the problem, about 17% of neurons were interneurons, 38.1% of them were responders, and 61.9% of them spiked spontaneously.

This leaves about 4% of neurons which have the µOR+ and spike spontaneously (and thus exert effects on the network at rest). Given that an imaging frame only contained 30-

60 neurons, it seems hard to imagine why the calcium imaging effects were so common and noticeable.

188

On the other hand, though, about 2/3rds of responders were firing at rest – which is a slight majority. Additionally, and importantly, if excitatory activity in the network were to increase, then the 1/3rd of responders (8/11 H-responders) may fail to be recruited into inhibiting that excitation – because they may have undergone a severe hyperpolarization that makes them less responsive (by requiring that they undergo more depolarization to close the gap between AP threshold and RMP), in addition to the reduced firing frequency and shortened APs that we observed (Figure 19). This model would suggest that not only does DAMGO inhibit the spontaneous activity of interneurons, but also prevents them from suppressing overexcitation in neuronal networks. In other words, not only do µORs provide constant and tonic inhibition, but perhaps the µOR+ interneurons also stop neuronal excitation from “snowballing.” Therefore, DAMGO might cause an uptick in neuronal excitation due to reduced inhibition, but it may also prevent H- responders from ameliorating this increased glutamatergic activity. Thus, it is a problem of cortical interneurons not being able to mobilize effectively, despite the greater glutamatergic stimulation being applied to them.

Another possible consideration is the interlinkage of interneurons, and the influence of neurogliaform interneurons, which are already known to express this receptor (Férézou et al., 2007; Krook-Magnuson et al., 2011). These interneurons are believed to be powerfully inhibitory in neuronal circuits; they create gap junctions to numerous neuronal types and synapse widely (Chittajallu, Pelkey, & McBain, 2013;

Niquille et al., 2018; Olah et al., 2007; Overstreet-Wadiche & McBain, 2015; A. Simon,

Olah, Molnar, Szabadics, & Tamas, 2005). Therefore, inhibiting neurogliaform neurons

189 may cause a very disproportionate uptick in excitatory activity relative to their frequency, which is very low. While we are not certain about their frequency in this neuronal culture model, their postsynaptic target receptors (GABAB receptors) are apparently found in this preparation, based on the substantial changes induced CGP55845 (Figure 12). Though given the in vitro nature of this model, we cannot be completely certain that this is an expressional aberration and perhaps an imperfect indicator of neurogliaform neurons in culture, in spite of its congruence to findings in brain slices. We never attempted to localize these neurons in this model due to their unfortunate lack of unique and distinct cell markers (Conde, Lund, Jacobowitz, Baimbridge, & Lewis, 1994; Niquille et al.,

2018; Overstreet-Wadiche & McBain, 2015). Neurogliaform neurons are often known to have late-spiking characteristics (Markwardt, Dieni, Wadiche, & Overstreet-Wadiche,

2011; Olah et al., 2007; Overstreet-Wadiche & McBain, 2015), and we did not observe this firing pattern in any recording.

While some of the assumptions here are speculative, they do lay some groundwork for future studies that explore the roles of µOR+ interneurons in cortical network. The experiments in this dissertation not only provide experimental evidence for the effects of the µOR in the neocortex, but they provide a substantial basis for this topic to be further explored in greater detail.

4.10 GABAA receptors counteract hypersynchrony

We observed another interesting drug effect, which may not be obvious in the quantitative data we collected. Picrotoxin (100 µM) always caused all neurons to produce synchronous SCOs, whereas pre-picrotoxin conditions had some out-of-sync neurons or

190 populations of neurons. It was always very clear in the images whether picrotoxin was applied due to the nearly perfect synchrony of neurons in the field. Also, very noticeably, picrotoxin dramatically increased amplitude and duration of SCOs (Figure 7). Together, this created a profound and bright “blinking” effect as all neurons flickered on and off collectively. But what does this tell us about interneurons, and their role in neuronal networks? And what does this tell us about the particular roles of GABAA and GABAB in neuronal networks? These basic questions have not been raised earlier due to the lack of quantitation, but the experiments conducted here do shed some light on the answers to these questions.

I posit that in pre-drug conditions, the interneurons maintained separate and discrete populations of neurons, and kept the neurons from achieving perfect synchrony with all their neighbors, though neurons started off fairly synchronous (Figure 2). This may allow groups of neurons to act as semi-independent units for cortical processing.

Upon application of picrotoxin, the interneurons lost their ability to segregate the neuronal networks, which produced universal SCOs and synchronous excitation in the whole network of neurons. This would predict that the fast inhibition of the GABAA receptor was primarily responsible for preventing excitation from simply reverberating through the culture. Upon applying GABAB receptor antagonist (CGP55845) in a background of picrotoxin, this universal synchrony did not seem to change, but instead enhanced the durations of the synchronous SCOs (Figure 12). While this model is speculative, the synchrony produced by picrotoxin should be noted because it was not reflected in the quantitation we did. In addition, it is difficult to extrapolate what this

191 means for in vivo brain circuits. Perhaps the beginning point for testing this would be to measure SCOs in brain slices and determining whether picrotoxin likewise causes complete synchrony in that model as well.

4.11 Unknown physiology of SCOs

One of the more obvious questions that the SCO study raises is the physiological events that contribute to SCO duration and amplitude. SCO amplitude may be modulated by

NMDA receptor ion traffic, which could function to both depolarize the neuron as well as conduct inward Ca2+ currents that induce an SCO (Canepari et al., 1997; Penn et al.,

2016; Przewlocki et al., 1999; T. Tanaka et al., 1996). Of the SCO phenomena, amplitude remains the most well-studied measurement. However, halfwidth (a measure of duration) is a much more complicated matter because few researchers have provided quantitation.

Of the studies that I cite in Chapter 2, (T. H. Murphy et al., 1992) is the only study I am aware that examined it. In that study, function of the NMDAR (modulated through

2+ manipulations of [Mg ]out) is shown to strongly modulate SCO duration. However, my own results repeatedly demonstrate that function of the GABAB receptor strongly modulates SCO halfwidth, and therefore I believe it may also be indicative of

GABAergic termination of bursts of APs. However, these explanations are non-mutually exclusive because both receptors may mediate this effect, and GABAB receptor activity inhibits PNs, which may then release glutamate onto NMDARs.

My speculation is that the proximal cause of SCO halfwidth is the duration of the

EPSPs, which may allow NMDARs to conduct inward Ca2+ currents for longer periods.

Evidence for this interpretation is supported by our well-replicated finding that GABAB

192 receptors appear to strongly modulate SCO halfwidth, and their expression on PNs which release glutamate (J. R. Chalifoux & Carter, 2010; Jason R. Chalifoux & Carter, 2011;

Corrêa, Munton, Nishimune, Fitzjohn, & Henley, 2004; Olah et al., 2007).

On the matter of SCO coupling to calcium-induced calcium release (CICR), we did often observe both smaller high-frequency SCOs, as well as much larger ones that often showed a refractory period associated with them (Figure 4). My interpretation of these data is that the peaks caused by VSCCs, NMDARs and calcium-conducting

AMPARs might trigger CICR when the peaks are of a certain threshold amplitude or duration. This may make tall peaks with a refractory period that tower over smaller calcium peaks that have failed to induce CICR. Coupling between extracellular sources of calcium and intracellular calcium release could potentially be an interesting study that can show how various drug interventions modulate the coupling efficiency between the two. However, there are several practical problems that can limit such a follow up.

Firstly, it may be difficult to identify peaks that are linked to CICR from those that are only from extracellular sources and did not induce CICR. Secondly, my experiments were unable to determine how the µOR itself couples to intracellular calcium release because the µOR is only expected to be expressed in about a third of the interneurons that

I recorded from. Thirdly, the SCO is a complicated model to study because it involves a web of interacting factors. For instance, applying DAMGO often intensifies SCO duration and amplitude, yet it happens in too many neurons to be a primary effect of the

µOR. Instead, the µOR likely enhances glutamatergic outflow which primes NMDARs and enables inward calcium current. NMDARs, in turn, may be in an ideal location for

193 coupling smaller SCOs with the larger CICR-induced SCOs. In this scenario the ultimate cause may have been µOR agonism, but the proximal cause of enhanced coupling between small SCOs and CICR SCOs are the NMDARs, and not any intracellular mechanism of the µOR. Although this exact scenario may not be shown to be true, follow-up studies should consider alternate possibilities.

4.12 SCOs: glutamatergic input or neuronal output?

Our experiments in this dissertation contributed to the field by resolving ambiguity about how the µOR exerted its effects by showing disinhibition and not direct excitation.

However, it also presents another compelling opportunity for future research; investigating whether SCOs correspond to (glutamatergic) input or neuronal output

(through prolonged opening of voltage-sensitive Ca2+ channels). Thus, whether SCOs in a given neuron corresponds to its incoming input or its own action potential output.

In the discussion of Chapter, I previously described how red and nonred neurons did not generally react to the various drug combinations differently from each other. I speculated that their lack of a difference, even as the various drugs had clear effects on all the neurons, may have been due to the origin of the calcium transients; most literature suggest that these calcium oscillations are actually a measure of the glutamatergic input being applied to the neurons by their neighbors, and therefore that calcium transients are not a direct measure of the output of the neurons, i.e., VS Ca2+ channels only constitute a small component of the calcium oscillations (Canepari et al., 1997; Dravid & Murray,

2004; Flint et al., 1999; Penn et al., 2016; Przewlocki et al., 1999; T. Tanaka et al., 1996).

However, this is not a settled issue, and some authors have proposed that VSCCs are

194 modest or substantial components of SCOs (Bacci et al., 1999; Dravid & Murray, 2004;

Inglefield & Shafer, 2000). It’s quite possible this is due to variance in neuronal types, and experimental conditions.

While glutamatergic input and output of the neuron are almost certainly positively correlated, they are not exactly the same thing. Specifically testing the relationship between input and output in the red vs nonred neurons would be an interesting follow-up to my experiments, because this would predict that a proportion (maybe a 1/3rd, a suggested by my electrophysiology data) may be µOR+. The consequences of a DAMGO response may cause them to become relatively suppressed in proportion to their input.

For instance, these µOR+ interneurons may fire fewer APs, and their APs may be hastened in response to DAMGO. In other words, their output given their input may become less intense. This output/input ratio may be difficult to measure, because it would require calculating an area-under-the curve estimate in the electrophysiological trace (for instance, calculating the area of an excitatory postsynaptic potential) and then also quantify their (indicator fluorescence) output in the imaging trace, perhaps also through an area-under-the curve estimate. Perhaps another way to quantify output is counting discrete APs, though bursts of APs appear to have various and irregular amplitudes and this method would also ignore the duration of APs. This output/input calculation may demonstrate that although excitatory drive increases in µOR+ neurons, their action potential output may not be proportional to their excitatory drive, as indicated by the area of the SCO. To synthesize the idea into a hypothesis: DAMGO reduces the output/input ratio in µOR+ interneurons, which then facilitates disinhibition. Affirming this

195 hypothesis would assist in explaining why interneurons are unable to prevent net excitation in spite of their uptick in glutamatergic input.

While most articles on the topic of SCOs support the view that they are visualizations of glutamatergic input, other explanations exist. Some data by other labs suggest that voltage-sensitive calcium channels contribute to this affect as well (Dravid &

Murray, 2004; Robinson et al., 1993). Indeed, while our data show that CNQX+AP5 combination prevented SCOs from occurring (Figure 3), I also observed that under this condition some rare flickering in a few neurons still seemed to occur. This may correspond to bursts of tonic activity and voltage-sensitive calcium channels, but it is unclear. In addition, application of depolarizing current into neurons could cause a calcium oscillation, though these calcium oscillations were frequently not as large as

SCOs, though, but sometimes they were. It’s possible that VSCCs may contribute to this effect as well. In these scenarios, the SCOs may indicate output of the neurons, and not necessarily their glutamatergic input.

4.13 SCOs: AMPARs, NMDRs or mGluRs?

In our experiments with SCOs, we blocked VS sodium channels with TTX, and we also blocked glutamate receptors with the combination of CNQX and dAP5. However, we did not separately block AMPARs and NMDARs. Although practically-speaking, these experiments would be trivial to conduct using our testing model, the main issue was data interpretation and understanding the conclusions that could be drawn from the data since these SCOs are spontaneous – not evoked by a constant stimulus. We anticipated that decreasing excitatory drive by separately blocking NMDARs or AMPARs may simply

196 reduce excitatory drive in this model, and therefore reducing the proximal factor mediating the SCO. For example, blocking AMPA receptors selectively may also lead to a substantial reduction in NMDAR activity by reducing the depolarizations produced by

EPSPs. Similarly blocking NMDARs or AMPARs would likely reduce excitatory drive being input into glutamatergic neurons that ultimately release glutamate onto Gq receptors.

4.14 Relationship between SCO halfwidth and interneuronal function

Using the SCO model to correlate neuronal activity enabled us to identify a particularly persistent feature of a DAMGO response: the prominent increase in halfwidth of SCOs.

This feature was penetrant in the presence or preincubation with picrotoxin and absent in the presence or preincubation with CGP55845 (Figure 7). We also found that CGP55845 could substantially and significantly increase SCO duration (Figure 12). Utilizing this model allowed us to effectively measure the effect of DAMGO in this environment and determine that it disinhibited cultures and acted through GABAA and GABAB receptors.

However, going beyond this; what does this feature tell us about interneuronal activity?

What does halfwidth indicate about GABAergic interneurons in this environment?

This question is largely unanswered by these data, yet the electrophysiological recordings could provide a reasonable starting point, because those data report the effects in single neurons. This starting point could be extrapolated from, in order to provide theoretical framework for understanding what halfwidth tells us about interneuronal function in this mode.

197

Recalling our electrophysiology protocol; we applied depolarizing current over a fixed 1s period to measure DAMGO’s effects on interneuronal action potentials. One clear metric involved is the latency to spike, which records how quickly a neuron can be induced into firing an action potential. However, this is unlikely the whole explanation here because it seems unlikely that interneurons firing their first AP can bring an end to glutamate release in the network. Instead, a critical amount and duration of GABA might be required for glutamate release from PNs to grind to a gradual halt and terminate the

SCO for that cycle. SCOs did not stop abruptly; they had slow decline phases. This suggests that sustained actions of GABAergic interneurons were required to stop SCOs.

Therefore, The AP halfwidth and spike frequency may modulate this critical amount of

GABA, and latency to spike is only one feature of this model.

SCO halfwidth is clearly a complex and multifactorial phenomenon, and given its consistent modulation by DAMGO, it warrants more attention. It suggests a complex interaction between glutamatergic and GABAergic signaling and it may assay the ability of interneurons to respond quickly to bursts of excitation from PNs.

4.15 Factors other than GABARs can modulate SCO halfwidth

In the picrotoxin condition, which induced long and synchronous SCOs, the long halfwidths likely reflected that only the GABAB receptor was functioning due to the block on GABAA by picrotoxin. The GABAB receptors appear to require longer durations of GABAergic stimulation to suppress the PNs. There were clear effects of CGP55845 in a background of picrotoxin, which supports this idea (Figure 12). Yet, in the combined picrotoxin + CGP 55845 condition, the neurons still flickered (albeit slowly, perhaps only

198

2-3 times per 80 seconds); but why did the neurons flicker at all – why did they not simply stay “on” since the GABAA and GABAB receptors were blocked? If GABA receptors were required to end an SCO for a cycle, the neurons would simply stay

“flickered on” for the entire duration of the recording, as the PNs release glutamate more- or-less indefinitely. On one hand, it may be possible that 100 uM picrotoxin and 10 uM

CGP55845 did not completely block all the GABA receptors. Or perhaps factors other than GABA receptors were in play in this condition. One possible explanation is that the

PNs have internal mechanisms that can terminate SCOs when GABAergic inhibition is absent. For example, delayed rectifier K channels, slow inactivation of synaptic terminal

VSCCs, or calcium-activated K channels. Or perhaps supply of synaptic glutamate becomes depleted, due to the seconds-long bursts of APs that PNs appear to undergo.

Another possibility is that this just a result of calcium stores becoming depleted from internal stores and undergoes a refractory period, and thus not necessarily an electrophysiological phenomenon.

It would be interesting to record from PNs in the presence of picrotoxin and

CGP55845 and observing the neurophysiological events that correspond to the end of an

SCO in this condition. As mentioned in this paragraph, it may simply be slow actions of

K channels that terminate bursts of activity from PNs, or depletion of synaptic glutamate.

If the termination of SCOs in picrotoxin + CGP55845 did not correspond to an abatement of electrophysiological phenomena (i.e., the Pyramidal Neuron kept firing even after the

SCO ends), then depletion of glutamate or calcium stores seems more likely. Though,

199 depletion of glutamate may cause a chain reaction through the culture that terminates the

SCO by diminishing excitatory drive.

To provide a more comprehensive summary of factors modulating SCO halfwidth; perhaps there are layers of factors that can modulate this feature. Perhaps

GABAA receptors provide quick modulation and blockading them makes the GABAB receptor the next-in-line mitigator. After blockade of both GABAA and GABAB, perhaps only the slow action of intracellularly-stored calcium can stop SCOs. It’s fairly clear from the data that a multitude of factors can modulate this and that they may come with their own temporal range of responses.

4.16 Sub-threshold calcium oscillations

We used an all-purpose general peakfinder setting that culled a wide variety of peak sizes with a lower cutoff, but virtually no upper cutoff. These settings were chosen after an extensive trial-and-error period of peakfinder adjustments, an ensuring that many peaks were culled and not overlooked. It also screened out minor fluctuations in baseline that could have been culled alongside larger peaks. This separation prevented us from conflating miniscule peaks from much larger ones.

However, some refinements or adjustments may also be made. For example, perhaps certain drugs only affect small peaks, and not larger ones. In this case, perhaps multiple peak profiles could have been used. Anecdotally, I rarely observed specific drug effects in certain peak profiles during the trial-and-error period; drug effects were usually consistent. Nevertheless, this topic could be explored further.

200

In addition, analyzing the RMS may be a useful measure, which could presumably measure some of the frequent, small-scale calcium fluctuations which were not explored here. This perhaps may be a way of analyzing calcium events that are too small to induce CICR and may represent scattered synaptic events that are too out-of- sync to be interpreted as an SCO.

4.17 Additional peakfinder measures

In a similar vein, other features of SCOs could have been analyzed as well, analogous to the action potential peakfinder. This calcium peak finder could include rate of rise, rate of decay, time to maximum rates, 90% width, and maximum rates. These may all be important and distinctive features that can inform us on the physiology of SCOs. This may lie outside the scope of this dissertation, since we were primarily focused on amplitude and duration to find DAMGO effects, but such analyses could provide grounds for future studies.

We sometimes noticed that neighboring neurons had different characteristics, including shape and size. One explanation for this is that neighboring neurons have different physiological characteristics, which create SCOs of different sizes. Drastic differences in peak amplitudes between (and within) neurons may indicate different physiological mechanism; for instance; large peaks may indicate CICR, whereas smaller peaks may be NMDARs/AMPARs. Perhaps shapes as well indicate different physiological mechanisms. For instance, CICR might correspond with sudden upward inflections, and slow decays. This may or may not be associated with Gq receptors

(because membrane calcium channels may also induce CICR). NMDARs, on the other

201 hand, may be associated with faster kinetics than CICR, but perhaps could also induce

CICR on top of their own peaks. AMPARs may be similar. Meanwhile, even smaller peaks may be simply voltage-sensitive calcium channels. RMS values may capture

VSCC activity, due to their fast kinetics and constant fluctuations. Perhaps neuronal variability and testing systems all contribute to some of the conflicts in the literature (as discussed earlier, whether SCO amplitude is due to VSCCs or glutamate receptors).

While speculative, these can beget testable hypotheses directed at understanding all the underpinnings of SCOs.

4.18 SCOs for drug discovery and assays of glutamatergic activity

Other labs have proposed the idea of developing a medium-throughput platform for measuring and analyzing SCOs as a means of drug discovery and mechanistic studies on epilepsy and its treatment (Cao et al., 2015; Pacico & Mingorance-Le Meur, 2014). Our lab performed a variety of experiments with other drugs by using SCOs, after we had developed the software to analyze them. The data presented here were deemed central to the hypotheses presented here, but we applied this testing system more broadly to test other drugs. These pilot experiments could have led to other projects and some data is therefore shown here.

As an example, I hypothesized that nicotinic agonism would have opposing effects of DAMGO, considering their antagonistic effects in µOR+ interneurons (Férézou et al., 2007). I therefore tested the effects of the nicotinic receptor agonist 1,1-Dimethyl-

4-phenylpiperazinium (100 µM; DMPP) on primary neuronal cultures using the same time course presented in Chapter 2 (Figure 5).

202

Change in Average Peak Halfwidth

DMPP+CTAP *** DMPP ns * Saline

0 1 2 3 Fold change from baseline

Figure 28. Nicotinic modulation of SCOs. We tested the effects of the nicotinic acetylcholine receptor agonist 1,1-Dimethyl-4-phenylpiperazinium (DMPP; 100 µM) using the same slot model (Figure 5). DMPP alone significantly enhances SCO halfwidth. But the combination of DMPP + the µOR antagonist D-Phe- Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP; 1 µM) blocked the effect of DMPP, which was significantly (p < 0.05) lower than DMPP alone, but not significantly different from the saline controls (p > 0.05).

Surprisingly, we found that DMPP actually enhanced SCO halfwidth compared the no- drug condition, in a manner similar to DAMGO. However, nicotinic agonism could also stimulate enkephalin release in cortical neurons (Dhatt et al., 1995), and therefore we tested DMPP in the presence of D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP) to block the µORs. This drug blocked the effect we observed with DMPP. These results do indeed support the notion that DMPP enhancement of SCO halfwidth was due to increased enkephalin release, which was blocked in the presence of CTAP. This line of evidence was not followed up on further, but these results do show that this testing system has broader implications for drug discovery and mechanistic studies. We attempted similar experiments with Cannabidiol, but we obtained nonsignificant results;

203 however, we did not troubleshoot potential problems with solubility because it was not a necessary part of this dissertation. In conclusion, while this experiment did not contribute to the central experiments of this dissertation, testing SCOs using our experimental design may provide a platform for future studies that are directed at determining how various drugs affect glutamate release.

4.19 Between SCOs and interneuronal suppression

The experiments in this dissertation found inhibition of about one third of interneurons, and the calcium imaging approach found detectable enhancement of duration of SCOs, and often amplitude as well. However, the question remains; how does suppression of interneurons produce an enhancement of SCOs? In other words, having found that

DAMGO inhibits a proportion of interneurons – how do we explain increased durations of bursts of activity (SCO halfwidth)?

Both Chapter 2 and Chapter 3 are congruent insomuch as they both suggest that the µOR suppresses cortical interneurons. Yet, these experiments set two points, and it is presently difficult to develop a model that relates these points. For example: how does reduced AP halfwidth and firing frequency in interneurons produce SCOs of longer durations in both interneurons and excitatory neurons? As mentioned earlier, these changes may result in interneurons that are simply slower to be recruited to stop bursts of excitatory activity. Perhaps the 1/3rd of cortical interneurons that are inhibited by

DAMGO are too slow to secrete critical amounts of GABA that halts glutamate release.

However, there may be practical experiments that can be done as well. This idea might be tested by tracking SCOs through inhibitory interneurons (using mRuby2) and PNs with a

204 much higher-speed camera. This may allow investigators to resolve the precise timing between the two populations of neurons – with glutamatergic neurons causing SCOs to move back and forth between PNs and interneurons, and then perhaps allowing researchers to determine whether DAMGO alters the temporal delay. However, SCOs likely only monitor glutamatergic input, and not neuronal output. Thus, SCOs in a given cell will indicate their glutamatergic excitation – not their delay to start spiking. Perhaps paired electrophysiological recordings may assist in this determination.

4.20 Summary

The calcium imaging studies highlight that DAMGO enhancement of cortical activity is likely secondary to inhibition of interneurons, and therefore provides some clarity in a field of ambiguities. Additionally, we substantiate findings that the µOR inhibits GABA release onto both GABAA and GABAB receptors, and therefore provides evidence that the µOR may be found on neurogliaform neurons. Going further than that however, these results provide fertile ground for follow-up studies directed at investigating the physiological correspondent phenomena that contribute to SCOs and may serve as an initial screening test for drug discovery assays.

Throughout the experiments presented here, I show that the µOR inhibits interneurons in numerous ways, and that a spectrum of responses may represent interneuronal type. I provide a detailed statistical analysis of these data, and a framework of categorizing DAMGO responses for future studies. Our data show that DAMGO not only inhibits interneurons through hyperpolarization, but also causes detectable changes

205 in their firing kinetics and AP frequency, which could have consequences for their control of network activity.

As described in this chapter, these studies are built upon a foundation of studies that have investigated the µOR. We successfully recapitulated many expected findings about the µOR and its excitatory effects on neuronal circuits, hyperpolarizing effects, and suppression of spontaneous APs. In doing this, we provide a framework for further investigations of the receptor. On top of that foundation, these studies profile many more complex effects of the µOR on APs and affirms that the receptor positively regulates

αDTX-sensitive currents to modulate some of these effects, as well as non-αDTX- sensitive currents to mediate other effects.

The experiments presented here significantly promotes our understanding of the

µOR in the brain. We show how the receptor produces dysregulation of cortical circuits by increasing excitatory outflow. These studies shed light on the mechanisms of the µOR, and its effects in cortical neurons.

BIBLIOGRAPHY

Abbadie, C., Pan, Y., Drake, C. T., & Pasternak, G. W. (2000). Comparative immunohistochemical distributions of carboxy terminus epitopes from the mu-opioid receptor splice variants MOR-1D, MOR-1 and MOR-1C in the mouse and rat CNS. Neuroscience, 100(1), 141-153. doi:10.1016/s0306-4522(00)00248-7 Abe, H., Keen, K. L., & Terasawa, E. (2008). Rapid Action of Estrogens on Intracellular Calcium Oscillations in Primate Luteinizing Hormone-Releasing Hormone-1 Neurons. Endocrinology, 149(3), 1155-1162. doi:10.1210/en.2007-0942 %J Endocrinology Adelman, J. P., Bond, C. T., Pessia, M., & Mayliet, J. (1995). Episodic ataxia results from voltage-dependent potassium channels with altered functions. Neuron, 15(6), 1449-1454.

206

Agmon, A., & Connors, B. (1991). Thalamocortical responses of mouse somatosensory (barrel) cortex in vitro. Neuroscience, 41(2-3), 365-379. Akil, H., Mayer, D., & Liebeskind, J. C. (1976). Antagonism of stimulation-produced analgesia by naloxone, a narcotic antagonist. Science. Al-Hasani, R., & Bruchas, M. R. (2011). Molecular Mechanisms of Opioid Receptor-Dependent Signaling and Behavior. Anesthesiology, 115(6), 1363-1381. doi:10.1097/ALN.0b013e318238bba6 Allouche, S., Noble, F., & Marie, N. (2014). Opioid receptor desensitization: mechanisms and its link to tolerance. Frontiers in pharmacology, 5, 280-280. doi:10.3389/fphar.2014.00280 Allouche, S., Polastron, J., Hasbi, A., Homburger, V., & Jauzac, P. (1999). Differential G-protein activation by alkaloid and peptide opioid agonists in the human neuroblastoma cell line SK-N-BE. Biochem J, 342 ( Pt 1)(Pt 1), 71-78. Alvarez‐Bolado, G., Fairén, A., Douglass, J., & Naranjo, J. R. (1990). Expression of the prodynorphin gene in the developing and adult cerebral cortex of the rat: an in situ hybridization study. Journal of Comparative Neurology, 300(3), 287-300. Amaral, D. G., Schumann, C. M., & Nordahl, C. W. (2008). Neuroanatomy of autism. Trends in Neurosciences, 31(3), 137-145. doi:https://doi.org/10.1016/j.tins.2007.12.005 American Psychiatric Association. (2013). Diagnostic statistical manual of mental disorders: DSM. 5. Amir, S., & Amit, Z. (1978). Endogenous opioid ligands may mediate stress-induced changes in the affective properties of pain related behavior in rats. Life Sciences, 23(11), 1143-1151. doi:https://doi.org/10.1016/0024-3205(78)90348-X An, X., Bandler, R., Öngür, D., & Price, J. (1998). Prefrontal cortical projections to longitudinal columns in the midbrain periaqueductal gray in macaque monkeys. Journal of Comparative Neurology, 401(4), 455-479. Anderson, S. A., Eisenstat, D. D., Shi, L., & Rubenstein, J. L. R. (1997). Interneuron Migration from Basal Forebrain to Neocortex: Dependence on Dlx Science, 278(5337), 474-476. doi:10.1126/science.278.5337.474 Anderson, S. A., Marín, O., Horn, C., Jennings, K., & Rubenstein, J. (2001). Distinct cortical migrations from the medial and lateral ganglionic eminences. Development, 128(3), 353- 363. Andoh, T., Yageta, Y., Konno, M., Yamaguchi-Miyamoto, T., Takahata, H., Nojima, H., . . . Kuraishi, Y. (2008). Evidence for separate involvement of different mu-opioid receptor subtypes in itch and analgesia induced by supraspinal action of opioids. J Pharmacol Sci, 106(4), 667-670. doi:10.1254/jphs.08004sc Arttamangkul, S., Torrecilla, M., Kobayashi, K., Okano, H., & Williams, J. T. (2006). Separation of μ-opioid receptor desensitization and internalization: endogenous receptors in primary neuronal cultures. Journal of Neuroscience, 26(15), 4118-4125. Arvidsson, U., Riedl, M., Chakrabarti, S., Lee, J.-H., Nakano, A. H., Dado, R., . . . Elde, R. (1995). Distribution and targeting of a mu-opioid receptor (MOR1) in brain and spinal cord. Journal of Neuroscience, 15(5), 3328-3341. Ashok, A. H., Myers, J., Reis Marques, T., Rabiner, E. A., & Howes, O. D. (2019). Reduced mu opioid receptor availability in schizophrenia revealed with [11C]-carfentanil positron emission tomographic Imaging. Nature Communications, 10(1), 4493. doi:10.1038/s41467-019-12366-4 Augustine, G. (1990). Regulation of transmitter release at the squid giant synapse by presynaptic delayed rectifier potassium current. The Journal of physiology, 431(1), 343-364.

207

Bacci, A., Verderio, C., Pravettoni, E., & Matteoli, M. (1999). Synaptic and intrinsic mechanisms shape synchronous oscillations in hippocampal neurons in culture. European Journal of Neuroscience, 11(2), 389-397. Bailly, J., Del Rossi, N., Runtz, L., Li, J.-J., Park, D., Scherrer, G., . . . Kieffer, B. L. (2020). Targeting Morphine-Responsive Neurons: Generation of a Knock-In Mouse Line Expressing Cre Recombinase from the Mu-Opioid Receptor Gene Locus. eneuro, 7(3), ENEURO.0433-0419.2020. doi:10.1523/ENEURO.0433-19.2020 Balaji, R., Bielmeier, C., Harz, H., Bates, J., Stadler, C., Hildebrand, A., & Classen, A.-K. (2017). Calcium spikes, waves and oscillations in a large, patterned epithelial tissue. Scientific reports, 7(1), 42786. doi:10.1038/srep42786 Baldo, B. A. (2016). Prefrontal Cortical Opioids and Dysregulated Motivation: A Network Hypothesis. Trends in Neurosciences, 39(6), 366-377. doi:https://doi.org/10.1016/j.tins.2016.03.004 Barbas, H. (2000). Connections underlying the synthesis of cognition, memory, and emotion in primate prefrontal cortices. Brain Research Bulletin, 52(5), 319-330. Barker, A. T., Jalinous, R., & Freeston, I. L. (1985). Non-invasive magnetic stimulation of human motor cortex. Lancet, 1(8437), 1106-1107. doi:10.1016/s0140-6736(85)92413-4 Bartos, M., & Elgueta, C. (2012). Functional characteristics of parvalbumin‐and cholecystokinin‐ expressing basket cells. The Journal of physiology, 590(4), 669-681. Bartus, R. T., Emerich, D. F., Hotz, J., Blaustein, M., Dean, R. L., Perdomo, B., & Basile, A. S. (2003). Vivitrex®, an Injectable, Extended-Release Formulation of Naltrexone, Provides Pharmacokinetic and Pharmacodynamic Evidence of Efficacy for 1 Month in Rats. Neuropsychopharmacology, 28(11), 1973-1982. doi:10.1038/sj.npp.1300274 Batista-Brito, R., Machold, R., Klein, C., & Fishell, G. (2008). Gene expression in cortical interneuron precursors is prescient of their mature function. Cereb Cortex, 18(10), 2306- 2317. doi:10.1093/cercor/bhm258 Batista‐Brito, R., & Fishell, G. (2009). The developmental integration of cortical interneurons into a functional network. Current topics in developmental biology, 87, 81-118. Bausch, S. B., & Chavkin, C. (1995). Colocalization of mu and delta opioid receptors with GABA, parvalbumin and a G-protein-coupled inwardly rectifying potassium channel (GIRK1) in the rodent brain. Analgesia, 1(4-5), 282-285. Bausch, S. B., Patterson, T. A., Ehrengruber, M. U., Lester, H. A., Davidson, N., & Chavkin, C. (1995). Colocalization of mu opioid receptors with GIRK1 potassium channels in the rat brain: an immunocytochemical study. Receptors Channels, 3(3), 221-241. Beaulieu, C. (1993). Numerical data on neocortical neurons in adult rat, with special reference to the GABA population. Brain Res, 609(1-2), 284-292. doi:10.1016/0006-8993(93)90884- p Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (1997). Deciding advantageously before knowing the advantageous strategy. Science, 275(5304), 1293-1295. Bekkers, J. M., & Delaney, A. J. (2001). Modulation of excitability by α-dendrotoxin-sensitive potassium channels in neocortical pyramidal neurons. Journal of Neuroscience, 21(17), 6553-6560. Bencherif, B., Guarda, A. S., Colantuoni, C., Ravert, H. T., Dannals, R. F., & Frost, J. J. (2005). Regional mu-opioid receptor binding in insular cortex is decreased in bulimia nervosa and correlates inversely with fasting behavior. J Nucl Med, 46(8), 1349-1351. Bencherif, B., Guarda, A. S., Colantuoni, C., Ravert, H. T., Dannals, R. F., & Frost, J. J. (2005). Regional μ-opioid receptor binding in insular cortex is decreased in bulimia nervosa and correlates inversely with fasting behavior. Journal of Nuclear Medicine, 46(8), 1349- 1351. 208

Benes, F. M., McSparren, J., Bird, E. D., SanGiovanni, J., & Vincent, S. L. (1991). Deficits in small interneurons in prefrontal and cingulate cortices of schizophrenic and schizoaffective patients. Archives of General Psychiatry, 48(11), 996-1001. doi:10.1001/archpsyc.1991.01810350036005 Berger, H. (1929). Über das Elektrenkephalogramm des Menschen. Archiv für Psychiatrie und Nervenkrankheiten, 87(1), 527-570. doi:10.1007/BF01797193 Berke, J. D. (2009). Fast oscillations in cortical-striatal networks switch frequency following rewarding events and stimulant drugs. European Journal of Neuroscience, 30(5), 848- 859. doi:10.1111/j.1460-9568.2009.06843.x Berridge, M. J. (1993). Inositol trisphosphate and calcium signalling. Nature, 361(6410), 315- 325. Biber, K., Laurie, D. J., Berthele, A., Sommer, B., Tölle, T. R., Gebicke‐Härter, P. J., . . . Boddeke, H. W. J. J. o. n. (1999). Expression and signaling of group I metabotropic glutamate receptors in astrocytes and microglia. 72(4), 1671-1680. Biedler, J. L., Helson, L., & Spengler, B. A. (1973). Morphology and Growth, Tumorigenicity, and Cytogenetics of Human Neuroblastoma Cells in Continuous Culture. Cancer Research, 33(11), 2643. Bilecki, W., & Przewlocki, R. (2000). Effect of opioids on Ca2+/cAMP responsive element binding protein. Acta Neurobiol. Exp, 60, 557-567. Bilkei-Gorzo, A., Michel, K., Noble, F., Roques, B. P., & Zimmer, A. (2007). Preproenkephalin Knockout Mice Show No Depression-Related Phenotype. Neuropsychopharmacology, 32(11), 2330-2337. doi:10.1038/sj.npp.1301370 Bingel, U., Wanigasekera, V., Wiech, K., Ni Mhuircheartaigh, R., Lee, M. C., Ploner, M., & Tracey, I. (2011). The Effect of Treatment Expectation on Drug Efficacy: Imaging the Analgesic Benefit of the Opioid Remifentanil. 3(70), 70ra14-70ra14. doi:10.1126/scitranslmed.3001244 %J Science Translational Medicine Blakemore, P. R., & White, J. D. (2002). Morphine, the Proteus of organic molecules. Chem Commun (Camb)(11), 1159-1168. doi:10.1039/b111551k Blanchet, C., & Lüscher, C. (2002). Desensitization of μ-opioid receptor-evoked potassium currents: Initiation at the receptor, expression at the effector. 99(7), 4674-4679. doi:10.1073/pnas.072075399 %J Proceedings of the National Academy of Sciences Blasio, A., Steardo, L., Sabino, V., & Cottone, P. (2014). Opioid system in the medial prefrontal cortex mediates binge‐like eating. Addiction biology, 19(4), 652-662. Blatow, M., Rozov, A., Katona, I., Hormuzdi, S. G., Meyer, A. H., Whittington, M. A., . . . Monyer, H. (2003). A Novel Network of Multipolar Bursting Interneurons Generates Theta Frequency Oscillations in Neocortex. Neuron, 38(5), 805-817. doi:https://doi.org/10.1016/S0896-6273(03)00300-3 Bohn, L., Dykstra, L., Lefkowitz, R. J., Caron, M., & Barak, L. (2004). Relative opioid efficacy is determined by the complements of the G protein-coupled receptor desensitization machinery. Mol Pharmacol, 66(1), 106-112. Borst, J. G. G., & Sakmann, B. (1998). Calcium current during a single action potential in a large presynaptic terminal of the rat brainstem. The Journal of physiology, 506(Pt 1), 143. Bozzi, Y., Provenzano, G., & Casarosa, S. (2018). Neurobiological bases of autism–epilepsy comorbidity: a focus on excitation/inhibition imbalance. European Journal of Neuroscience, 47(6), 534-548. Bragin, A., Jandó, G., Nádasdy, Z., Hetke, J., Wise, K., & Buzsáki, G. (1995). Gamma (40-100 Hz) oscillation in the hippocampus of the behaving rat. Journal of neuroscience, 15(1), 47-60.

209

Bremner, J. D., Vythilingam, M., Vermetten, E., Nazeer, A., Adil, J., Khan, S., . . . Charney, D. S. (2002). Reduced volume of orbitofrontal cortex in major depression. Biological Psychiatry, 51(4), 273-279. Brew, H. M., Gittelman, J. X., Silverstein, R. S., Hanks, T. D., Demas, V. P., Robinson, L. C., . . . Messing, A. J. J. o. n. (2007). Seizures and reduced life span in mice lacking the potassium channel subunit Kv1. 2, but hypoexcitability and enlarged Kv1 currents in auditory neurons. 98(3), 1501-1525. Brewer, G. J., Torricelli, J. R., Evege, E. K., & Price, P. J. (1993). Optimized survival of hippocampal neurons in B27-supplemented Neurobasal, a new serum-free medium combination. Journal of neuroscience research, 35(5), 567-576. doi:10.1002/jnr.490350513 Bright, D. P., Renzi, M., Bartram, J., McGee, T. P., MacKenzie, G., Hosie, A. M., . . . Brickley, S. G. J. J. o. N. (2011). Profound desensitization by ambient GABA limits activation of δ-containing GABAA receptors during spillover. 31(2), 753-763. Brown, A. M., & Birnbaumer, L. (1990). Ionic channels and their regulation by G protein subunits. Annual review of physiology, 52(1), 197-213. Brown, D. (1988). Ion channels. In: Plenum, New York. Brownstein, M. J. (1993). A brief history of opiates, opioid peptides, and opioid receptors. Proceedings of the National Academy of Sciences of the United States of America, 90(12), 5391-5393. doi:10.1073/pnas.90.12.5391 Brunton, J., & Charpak, S. (1998). μ-Opioid peptides inhibit thalamic neurons. Journal of Neuroscience, 18(5), 1671-1678. Burkhalter, A., Gonchar, Y., Mellor, R. L., & Nerbonne, J. M. (2006). Differential Expression of IA Channel Subunits Kv4.2 and Kv4.3 in Mouse Visual Cortical Neurons and Synapses. The Journal of Neuroscience, 26(47), 12274. doi:10.1523/JNEUROSCI.2599-06.2006 Bushell, T., Endoh, T., Simen, A., Ren, D., Bindokas, V., & Miller, R. (2002). Molecular components of tolerance to opiates in single hippocampal neurons. Mol Pharmacol, 61(1), 55-64. Butelman, E. R., & Kreek, M. J. (2015). Salvinorin A, a kappa-opioid receptor agonist hallucinogen: pharmacology and potential template for novel pharmacotherapeutic agents in neuropsychiatric disorders. Frontiers in pharmacology, 6, 190-190. doi:10.3389/fphar.2015.00190 Butler, R. K., Nilsson-Todd, L., Cleren, C., Léna, I., Garcia, R., & Finn, D. P. (2011). Molecular and electrophysiological changes in the prefrontal cortex–amygdala–dorsal periaqueductal grey pathway during persistent pain state and fear-conditioned analgesia. Physiology & behavior, 104(5), 1075-1081. Butt, S. J., Fuccillo, M., Nery, S., Noctor, S., Kriegstein, A., Corbin, J. G., & Fishell, G. (2005). The temporal and spatial origins of cortical interneurons predict their physiological subtype. Neuron, 48(4), 591-604. doi:10.1016/j.neuron.2005.09.034 Buzsáki, G., & Draguhn, A. (2004). Neuronal Oscillations in Cortical Networks. 304(5679), 1926-1929. doi:10.1126/science.1099745 %J Science Buzsáki, G., Geisler, C., Henze, D. A., & Wang, X.-J. (2004). Interneuron Diversity series: Circuit complexity and axon wiring economy of cortical interneurons. Trends in Neurosciences, 27(4), 186-193. doi:https://doi.org/10.1016/j.tins.2004.02.007 Cai, Q., Qiu, C.-Y., Qiu, F., Liu, T.-T., Qu, Z.-W., Liu, Y.-M., & Hu, W.-P. (2014). Morphine inhibits acid-sensing ion channel currents in rat dorsal root ganglion neurons. Brain Research Bulletin, 1554, 12-20. Campbell, K. (2005). Cortical Neuron Specification: It Has Its Time and Place. Neuron, 46(3), 373-376. doi:https://doi.org/10.1016/j.neuron.2005.04.014 210

Canepari, M., Bove, M., Maeda, E., Cappello, M., & Kawana, A. (1997). Experimental analysis of neuronal dynamics in cultured cortical networks and transitions between different patterns of activity. Biological Cybernetics, 77(2), 153-162. doi:10.1007/s004220050376 Cao, Z., Cui, Y., Busse, E., Mehrotra, S., Rainier, J. D., & Murray, T. F. (2014). Gambierol Inhibition of Voltage-Gated Potassium Channels Augments Spontaneous Ca2+ Oscillations in Cerebrocortical Neurons. 350(3), 615-623. doi:10.1124/jpet.114.215319 %J Journal of Pharmacology and Experimental Therapeutics Cao, Z., Zou, X., Cui, Y., Hulsizer, S., Lein, P. J., Wulff, H., & Pessah, I. N. (2015). Rapid Throughput Analysis Demonstrates that Chemicals with Distinct Seizurogenic Mechanisms Differentially Alter Ca2+ Dynamics in Networks Formed by Hippocampal Neurons in Culture. Mol Pharmacol, 87(4), 595. doi:10.1124/mol.114.096701 Capogna, M., Gähwiler, B. H., & Thompson, S. M. (1993). Mechanism of mu-opioid receptor- mediated presynaptic inhibition in the rat hippocampus in vitro. The Journal of physiology, 470(1), 539-558. doi:10.1113/jphysiol.1993.sp019874 Carper, R. A., & Courchesne, E. (2005). Localized enlargement of the frontal cortex in early autism. Biological Psychiatry, 57(2), 126-133. Carr, D. B., & Sesack, S. R. (2000). Projections from the Rat Prefrontal Cortex to the Ventral Tegmental Area: Target Specificity in the Synaptic Associations with Mesoaccumbens and Mesocortical Neurons. The Journal of Neuroscience, 20(10), 3864. doi:10.1523/JNEUROSCI.20-10-03864.2000 Carrasquillo, Y., Burkhalter, A., & Nerbonne, J. M. (2012). A‐type K+ channels encoded by Kv4. 2, Kv4. 3 and Kv1. 4 differentially regulate intrinsic excitability of cortical pyramidal neurons. The Journal of physiology, 590(16), 3877-3890. Carrero, J. P., Kaigler, K. F., Hartshorn, G. H., Fadel, J. R., & Wilson, M. A. (2019). Mu opioid receptor regulation of glutamate efflux in the central amygdala in response to predator odor. Neurobiology of Stress, 11, 100197. Casale, A. E., Foust, A. J., Bal, T., & McCormick, D. A. (2015). Cortical interneuron subtypes vary in their axonal action potential properties. Journal of Neuroscience, 35(47), 15555- 15567. Castro, D. C., & Berridge, K. C. (2014). Opioid Hedonic Hotspot in Nucleus Accumbens Shell: Mu, Delta, and Kappa Maps for Enhancement of Sweetness “Liking” and “Wanting”. The Journal of Neuroscience, 34(12), 4239-4250. doi:10.1523/jneurosci.4458-13.2014 Castro, D. C., & Berridge, K. C. (2017). Opioid and orexin hedonic hotspots in rat orbitofrontal cortex and insula. Proceedings of the National Academy of Sciences, 114(43), E9125. doi:10.1073/pnas.1705753114 Cauli, B., Audinat, E., Lambolez, B., Angulo, M. C., Ropert, N., Tsuzuki, K., . . . Rossier, J. (1997). Molecular and Physiological Diversity of Cortical Nonpyramidal Cells. The Journal of Neuroscience, 17(10), 3894-3906. Cauli, B., Tong, X. K., Rancillac, A., Serluca, N., Lambolez, B., Rossier, J., & Hamel, E. (2004). Cortical GABA interneurons in neurovascular coupling: relays for subcortical vasoactive pathways. J Neurosci, 24(41), 8940-8949. doi:10.1523/jneurosci.3065-04.2004 Chalifoux, J. R., & Carter, A. G. (2010). GABAB receptors modulate NMDA receptor calcium signals in dendritic spines. Neuron, 66(1), 101-113. doi:10.1016/j.neuron.2010.03.012 Chalifoux, J. R., & Carter, A. G. (2011). GABAB Receptor Modulation of Voltage-Sensitive Calcium Channels in Spines and Dendrites. The Journal of Neuroscience, 31(11), 4221- 4232. doi:10.1523/jneurosci.4561-10.2011 Chan, K. Y., Jang, M. J., Yoo, B. B., Greenbaum, A., Ravi, N., Wu, W. L., . . . Gradinaru, V. (2017). Engineered AAVs for efficient noninvasive gene delivery to the central and peripheral nervous systems. Nat Neurosci, 20(8), 1172-1179. doi:10.1038/nn.4593 211

Chao, O. Y., & Yang, Y.-M. (2019). Timing constraints of action potential evoked Ca2+ current and transmitter release at a central nerve terminal. Scientific reports, 9(1), 4448. doi:10.1038/s41598-019-41120-5 Chartoff, E. H., & Connery, H. S. (2014). It’s MORe exciting than mu: crosstalk between mu opioid receptors and glutamatergic transmission in the mesolimbic dopamine system. 5(116). doi:10.3389/fphar.2014.00116 Chavkin, C. (2011). The therapeutic potential of κ-opioids for treatment of pain and addiction. Neuropsychopharmacology, 36(1), 369-370. doi:10.1038/npp.2010.137 Chen, Y., & Yu, L. (1994). Differential regulation by cAMP-dependent protein kinase and protein kinase C of the mu opioid receptor coupling to a G protein-activated K+ channel. Journal of Biological Chemistry, 269(11), 7839-7842. Cheriyan, J., Kaushik, M. K., Ferreira, A. N., & Sheets, P. L. (2016). Specific targeting of the basolateral amygdala to projectionally defined pyramidal neurons in prelimbic and infralimbic cortex. eneuro, 3(2). Cherny, N., Ripamonti, C., Pereira, J., Davis, C., Fallon, M., McQuay, H., . . . Ventafridda, V. (2001). Strategies to manage the adverse effects of oral morphine: an evidence-based report. J Clin Oncol, 19(9), 2542-2554. doi:10.1200/jco.2001.19.9.2542 Chiappalone, M., Bove, M., Vato, A., Tedesco, M., & Martinoia, S. (2006). Dissociated cortical networks show spontaneously correlated activity patterns during in vitro development. Brain Res, 1093(1), 41-53. doi:https://doi.org/10.1016/j.brainres.2006.03.049 Chieng, B., Azriel, Y., Mohammadi, S., & Christie, M. J. (2011). Distinct cellular properties of identified dopaminergic and GABAergic neurons in the mouse ventral tegmental area. The Journal of physiology, 589(15), 3775-3787. Chieng, B., & Christie, M. (1994). Hyperpolarization by opioids acting on μ‐receptors of a sub‐ population of rat periaqueductal gray neurones in vitro. British journal of pharmacology, 113(1), 121-128. Chieng, B., Connor, M., & Christie, M. J. (1996). The mu-opioid receptor antagonist D-Phe-Cys- Tyr-D-Trp-Orn-Thr-Pen-Thr-NH2 (CTOP) [but not D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen- Thr-NH2 (CTAP)] produces a nonopioid receptor-mediated increase in K+ conductance of rat locus ceruleus neurons. Mol Pharmacol, 50(3), 650-655. Childers, S. R. (1991). Opioid receptor-coupled second messenger systems. Life Sciences, 48(21), 1991-2003. doi:https://doi.org/10.1016/0024-3205(91)90154-4 Chiou, L. C., & Huang, L. Y. (1999). Mechanism underlying increased neuronal activity in the rat ventrolateral periaqueductal grey by a mu-opioid. The Journal of physiology, 518 ( Pt 2)(Pt 2), 551-559. doi:10.1111/j.1469-7793.1999.0551p.x Chittajallu, R., Pelkey, K. A., & McBain, C. J. (2013). Neurogliaform cells dynamically regulate somatosensory integration via synapse-specific modulation. Nat Neurosci, 16(1), 13-15. doi:10.1038/nn.3284 Cho, R., Konecky, R., & Carter, C. S. (2006). Impairments in frontal cortical γ synchrony and cognitive control in schizophrenia. Proceedings of the National Academy of Sciences, 103(52), 19878-19883. Choi, J., Kim, J., Kang, E., Lee, J. M., Cha, J., Kim, Y., . . . Yi, D. J. (2016). Brain mechanisms of pain relief by transcutaneous electrical nerve stimulation: a functional magnetic resonance imaging study. European Journal of Pain, 20(1), 92-105. Christie, M. J., Connor, M., Vaughan, C. W., Ingram, S. L., & Bagley, E. E. (2000). Cellular Actions Of Opioids And Other Analgesics: Implications For Synergism In Pain Relief. 27(7), 520-523. doi:10.1046/j.1440-1681.2000.03291.x

212

Christie, M. J., Vaughan, C. W., & Ingram, S. L. (1999). Opioids, NSAIDs and 5-lipoxygenase inhibitors act synergistically in brain via arachidonic acid metabolism. Inflammation Research, 48(1), 1-4. doi:10.1007/s000110050367 Chu, Z., Galarreta, M., & Hestrin, S. (2003). Synaptic interactions of late-spiking neocortical neurons in layer 1. J Neurosci, 23(1), 96-102. doi:10.1523/jneurosci.23-01-00096.2003 Classification, I. U. o. B. a. C. P. C. o. R. N. a. D. (2020). The IUPHAR/BPS Guide to PHARMACOLOGY. https://www.guidetopharmacology.org/GRAC/GPCRListForward?class=A Clendeninn, N. J., Petraitis, M., & Simon, E. J. (1976). Ontological development of opiate receptors in rodent brain. Brain Res, 118(1), 157-160. doi:10.1016/0006-8993(76)90852- 0 Cobb, S., Buhl, E., Halasy, K., Paulsen, O., & Somogyi, P. (1995). Synchronization of neuronal activity in hippocampus by individual GABAergic interneurons. Nature, 378(6552), 75- 78. Colasanti, A., Searle, G. E., Long, C. J., Hill, S. P., Reiley, R. R., Quelch, D., . . . Lingford- Hughes, A. R. (2012). Endogenous opioid release in the human brain reward system induced by acute amphetamine administration. Biological Psychiatry, 72(5), 371-377. Conde, F., Lund, J. S., Jacobowitz, D. M., Baimbridge, K. G., & Lewis, D. A. (1994). Local circuit neurons immunoreactive for calretinin, calbindin D-28k or parvalbumin in monkey prefrontal cortex: distribution and morphology. J Comp Neurol, 341(1), 95-116. doi:10.1002/cne.903410109 Connor, J., & Stevens, C. (1971). Prediction of repetitive firing behaviour from voltage clamp data on an isolated neurone soma. The Journal of physiology, 213(1), 31. Connor, M., Schuller, A., Pintar, J. E., & Christie, M. J. (1999). μ‐opioid receptor modulation of calcium channel current in periaqueductal grey neurons from C57B16/J mice and mutant mice lacking MOR‐1. British journal of pharmacology, 126(7), 1553-1558. Corrêa, S. A. L., Munton, R., Nishimune, A., Fitzjohn, S., & Henley, J. M. (2004). Development of GABAB subunits and functional GABAB receptors in rat cultured hippocampal neurons. Neuropharmacology, 47(4), 475-484. doi:10.1016/j.neuropharm.2004.04.021 Courchesne, E., & Pierce, K. (2005). Why the frontal cortex in autism might be talking only to itself: local over-connectivity but long-distance disconnection. Current opinion in neurobiology, 15(2), 225-230. doi:https://doi.org/10.1016/j.conb.2005.03.001 Coyle, J. T., & Pert, C. B. (1976). Ontogenetic development of [3H]naloxone binding in rat brain. Neuropharmacology, 15(9), 555-560. doi:10.1016/0028-3908(76)90107-6 Craig, M. T., & McBain, C. J. (2014). The emerging role of GABAB receptors as regulators of network dynamics: fast actions from a 'slow' receptor? Current opinion in neurobiology, 26, 15-21. doi:10.1016/j.conb.2013.10.002 Crain, S., Shen, K.-F., & Chalazonitis, A. (1988). Opioids excite rather than inhibit sensory neurons after chronic opioid exposure of spinal cord-ganglion cultures. Brain Res, 455(1), 99-109. Crawley, J. N. (1985). Comparative distribution of cholecystokinin and other neuropeptides. Why is this peptide different from all other peptides? Ann N Y Acad Sci, 448, 1-8. doi:10.1111/j.1749-6632.1985.tb29900.x Csicsvari, J., Hirase, H., Czurkó, A., Mamiya, A., & Buzsáki, G. (1999). Oscillatory Coupling of Hippocampal Pyramidal Cells and Interneurons in the Behaving Rat. The Journal of Neuroscience, 19(1), 274. doi:10.1523/JNEUROSCI.19-01-00274.1999 Csicsvari, J., Jamieson, B., Wise, K. D., & Buzsáki, G. (2003). Mechanisms of gamma oscillations in the hippocampus of the behaving rat. Neuron, 37(2), 311-322.

213

Cvejic, S., & Devi, L. A. (1997). Dimerization of the δ opioid receptor: implication for a role in receptor internalization. Journal of Biological Chemistry, 272(43), 26959-26964. Dani, J. W., Chernjavsky, A., & Smith, S. J. (1992). Neuronal activity triggers calcium waves in hippocampal astrocyte networks. Neuron, 8(3), 429-440. doi:https://doi.org/10.1016/0896-6273(92)90271-E Dani, V. S., Chang, Q., Maffei, A., Turrigiano, G. G., Jaenisch, R., & Nelson, S. B. (2005). Reduced cortical activity due to a shift in the balance between excitation and inhibition in a mouse model of Rett syndrome. Proceedings of the National Academy of Sciences of the United States of America, 102(35), 12560-12565. doi:10.1073/pnas.0506071102 de la Fuente-Sandoval, C., Favila, R., Gómez-Martín, D., León-Ortiz, P., & Graff-Guerrero, A. (2012). Neural response to experimental heat pain in stable patients with schizophrenia. Journal of Psychiatric Research, 46(1), 128-134. doi:https://doi.org/10.1016/j.jpsychires.2011.09.008 de Lombares, C., Heude, E., Alfama, G., Fontaine, A., Hassouna, R., Vernochet, C., . . . Narboux- Nême, N. (2019). Dlx5 and Dlx6 expression in GABAergic neurons controls behavior, metabolism, healthy aging and lifespan. Aging, 11(17), 6638-6656. doi:10.18632/aging.102141 DeFelipe, J., & Fariñas, I. (1992). The pyramidal neuron of the cerebral cortex: Morphological and chemical characteristics of the synaptic inputs. Progress in neurobiology, 39(6), 563- 607. doi:http://dx.doi.org/10.1016/0301-0082(92)90015-7 Der-Avakian, A., & Markou, A. (2012). The neurobiology of anhedonia and other reward-related deficits. Trends in Neurosciences, 35(1), 68-77. doi:10.1016/j.tins.2011.11.005 Devinsky, O., Morrell, M. J., & Vogt, B. A. (1995). Contributions of anterior cingulate cortex to behaviour. Brain, 118 ( Pt 1), 279-306. doi:10.1093/brain/118.1.279 Dhatt, R. K., Gudehithlu, K. P., Wemlinger, T. A., Tejwani, G. A., Neff, N. H., & Hadjiconstantinou, M. J. J. o. n. (1995). Preproenkephalin mRNA and methionine‐ enkephalin content are increased in mouse striatum after treatment with nicotine. 64(4), 1878-1883. Dichter, M. A. (1978). Rat cortical neurons in cell culture: Culture methods, cell morphology, electrophysiology, and synapse formation. Brain Res, 149(2), 279-293. doi:https://doi.org/10.1016/0006-8993(78)90476-6 Dolmetsch, R. E., Xu, K., & Lewis, R. S. (1998). Calcium oscillations increase the efficiency and specificity of gene expression. Nature, 392(6679), 933-936. Drake, C. T., & Milner, T. A. (1999). Mu opioid receptors are in somatodendritic and axonal compartments of GABAergic neurons in rat hippocampal formation. Brain Res, 849(1– 2), 203-215. doi:http://dx.doi.org/10.1016/S0006-8993(99)01910-1 Drake, C. T., & Milner, T. A. (2002). Mu opioid receptors are in discrete hippocampal interneuron subpopulations. Hippocampus, 12(2), 119-136. doi:10.1002/hipo.1107 Drake, C. T., & Milner, T. A. (2006). Mu opioid receptors are extensively co-localized with parvalbumin, but not somatostatin, in the dentate gyrus. Neuroscience Letters, 403(1–2), 176-180. doi:http://dx.doi.org/10.1016/j.neulet.2006.04.047 Dravid, S. M., & Murray, T. F. (2004). Spontaneous synchronized calcium oscillations in neocortical neurons in the presence of physiological [Mg(2+)]: involvement of AMPA/kainate and metabotropic glutamate receptors. Brain Res, 1006(1), 8-17. doi:10.1016/j.brainres.2004.01.059 Drevets, W. C. (2007). Orbitofrontal cortex function and structure in depression. Ann N Y Acad Sci, 1121(1), 499-527.

214

Dutkiewicz, A. P., & Morielli, A. D. (2020). μ Opioid Receptors Modulate Action Potential Kinetics and Firing Frequency in Neocortical Interneurons. 2020.2011.2020.391508. doi:10.1101/2020.11.20.391508 %J bioRxiv Ebrahimian, M., Razeghi, M., Zamani, A., Bagheri, Z., Rastegar, K., & Motealleh, A. (2018). Does High Frequency Transcutaneous Electrical Nerve Stimulation (TENS) Affect EEG Gamma Band Activity? Journal of biomedical physics & engineering, 8(3), 271-280. Edelstein, A. D., Tsuchida, M. A., Amodaj, N., Pinkard, H., Vale, R. D., & Stuurman, N. (2014). Advanced methods of microscope control using μManager software. 2014. doi:10.14440/jbm.2014.36
Fatal error: Call to undefined method PublishedArticle::getStartingPage() in /home/jbm/public_html/cache/t_compile/%%5F^5F4^5F4CE8DE%%citation.tpl.php on line 47
Eisch, A. J., Barrot, M., Schad, C. A., Self, D. W., & Nestler, E. J. (2000). Opiates inhibit neurogenesis in the adult rat hippocampus. Proceedings of the National Academy of Sciences, 97(13), 7579. doi:10.1073/pnas.120552597 Elghaba, R., & Bracci, E. (2017). Dichotomous Effects of Mu Opioid Receptor Activation on Striatal Low-Threshold Spike Interneurons. 11(385). doi:10.3389/fncel.2017.00385 Elston, T. W., Croy, E., & Bilkey, D. K. (2019). Communication between the anterior cingulate cortex and ventral tegmental area during a cost-benefit reversal task. Cell reports, 26(9), 2353-2361. e2353. Endres-Becker, J., Heppenstall, P. A., Mousa, S. A., Labuz, D., Oksche, A., Schäfer, M., . . . Zöllner, C. (2007). μ-Opioid receptor activation modulates transient receptor potential vanilloid 1 (TRPV1) currents in sensory neurons in a model of inflammatory pain. Mol Pharmacol, 71(1), 12-18. Engel, J. J. C. J. o. N. S. (1996). Excitation and inhibition in epilepsy. 23(3), 167-174. Erbs, E., Faget, L., Scherrer, G., Matifas, A., Filliol, D., Vonesch, J.-L., . . . Function. (2015). A mu–delta opioid receptor brain atlas reveals neuronal co-occurrence in subcortical networks. Brain Structure, 220(2), 677-702. F, M. J., van der Kooy, D., & Bloom, F. E. J. J. o. N. (1984). The distribution and morphology of opioid peptide immunoreactive neurons in the cerebral cortex of rats. 4(4), 1104-1117. Faber, E. S. L., & Sah, P. (2004). Opioids Inhibit Lateral Amygdala Pyramidal Neurons by Enhancing a Dendritic Potassium Current. The Journal of Neuroscience, 24(12), 3031- 3039. doi:10.1523/jneurosci.4496-03.2004 Falk, T., Bartlett, M., So, L., Szabò, L., Skinner, D., Parent, K., . . . Sherman, S. (2020). Highly- selective μ-opioid receptor antagonism does not block L-DOPA-induced dyskinesia in a rodent model. Fallon, J. H., & Leslie, F. M. (1986). Distribution of dynorphin and enkephalin peptides in the rat brain. Journal of Comparative Neurology, 249(3), 293-336. doi:10.1002/cne.902490302 Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41(4), 1149-1160. doi:10.3758/BRM.41.4.1149 Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175-191. doi:10.3758/BF03193146 Fazzari, P., Paternain, A. V., Valiente, M., Pla, R., Lujan, R., Lloyd, K., . . . Rico, B. (2010). Control of cortical GABA circuitry development by Nrg1 and ErbB4 signalling. Nature, 464(7293), 1376-1380. doi:10.1038/nature08928

215

Fell, J., Klaver, P., Lehnertz, K., Grunwald, T., Schaller, C., Elger, C. E., & Fernández, G. (2001). Human memory formation is accompanied by rhinal–hippocampal coupling and decoupling. Nature Neuroscience, 4(12), 1259-1264. Férézou, I., Hill, E. L., Cauli, B., Gibelin, N., Kaneko, T., Rossier, J., & Lambolez, B. (2007). Extensive Overlap of Mu-Opioid and Nicotinic Sensitivity in Cortical Interneurons. Cerebral Cortex, 17(8), 1948-1957. doi:10.1093/cercor/bhl104 Ferguson, B. R., & Gao, W.-J. (2018). PV Interneurons: Critical Regulators of E/I Balance for Prefrontal Cortex-Dependent Behavior and Psychiatric Disorders. 12(37). doi:10.3389/fncir.2018.00037 Ferguson, S. S., Downey, W. E., Colapietro, A.-M., Barak, L. S., Ménard, L., & Caron, M. G. (1996). Role of β-arrestin in mediating agonist-promoted G protein-coupled receptor internalization. Science, 271(5247), 363-366. Fields, H. L., & Margolis, E. B. (2015). Understanding opioid reward. Trends in Neurosciences, 38(4), 217-225. doi:https://doi.org/10.1016/j.tins.2015.01.002 Filliol, D., Ghozland, S., Chluba, J., Martin, M., Matthes, H. W., Simonin, F., . . . Kieffer, B. L. (2000). Mice deficient for delta- and mu-opioid receptors exhibit opposing alterations of emotional responses. Nat Genet, 25(2), 195-200. doi:10.1038/76061 Finn, A. K., & Whistler, J. L. (2001). Endocytosis of the Mu Opioid Receptor Reduces Tolerance and a Cellular Hallmark of Opiate Withdrawal. Neuron, 32(5), 829-839. doi:10.1016/S0896-6273(01)00517-7 Finnegan, T. F., Chen, S. R., & Pan, H. L. (2006). Mu opioid receptor activation inhibits GABAergic inputs to basolateral amygdala neurons through Kv1.1/1.2 channels. J Neurophysiol, 95(4), 2032-2041. doi:10.1152/jn.01004.2005 Flames, N., & Marin, O. (2005). Developmental mechanisms underlying the generation of cortical interneuron diversity. Neuron, 46(3), 377-381. doi:10.1016/j.neuron.2005.04.020 Flint, A. C., Dammerman, R. S., & Kriegstein, A. R. (1999). Endogenous activation of metabotropic glutamate receptors in neocortical development causes neuronal calcium oscillations. Proceedings of the National Academy of Sciences, 96(21), 12144. doi:10.1073/pnas.96.21.12144 Fries, P., Reynolds, J. H., Rorie, A. E., & Desimone, R. (2001). Modulation of oscillatory neuronal synchronization by selective visual attention. Science, 291(5508), 1560-1563. Fu, Y., Tucciarone, Jason M., Espinosa, J. S., Sheng, N., Darcy, Daniel P., Nicoll, Roger A., . . . Stryker, Michael P. (2014). A Cortical Circuit for Gain Control by Behavioral State. Cell, 156(6), 1139-1152. doi:https://doi.org/10.1016/j.cell.2014.01.050 Fukuda, T., Kosaka, T., Singer, W., & Galuske, R. A. (2006). Gap junctions among dendrites of cortical GABAergic neurons establish a dense and widespread intercolumnar network. Journal of Neuroscience, 26(13), 3434-3443. Funahashi, S. (2013). Thalamic mediodorsal nucleus and its participation in spatial working memory processes: comparison with the prefrontal cortex. Frontiers in Systems Neuroscience, 7, 36. Galarreta, M., & Hestrin, S. (1999). A network of fast-spiking cells in the neocortex connected by electrical synapses. Nature, 402(6757), 72-75. Galarreta, M., & Hestrin, S. (2002). Electrical and chemical synapses among parvalbumin fastspiking GABAergic interneurons in adult mouse neocortex. Proceedings of the National Academy of Sciences, 99(19), 12438-12443. Garcin, F., & Coyle, J. T. (1977). Effects of perinatal 6-hydroxydopamine treatment on opiate receptor distribution in adult brain. Commun Psychopharmacol, 1(3), 283-290.

216

Gardon, O., Faget, L., Chung, P. C. S., Matifas, A., Massotte, D., & Kieffer, B. (2014). Expression of mu opioid receptor in dorsal diencephalic conduction system: new insights for the medial habenula. Neuroscience, 277, 595-609. Gavériaux-Ruff, C., & Kieffer, B. L. (2011). Delta opioid receptor analgesia: recent contributions from pharmacology and molecular approaches. Behavioural pharmacology, 22(5-6), 405- 414. doi:10.1097/FBP.0b013e32834a1f2c Geddes-Klein, D. M., Serbest, G., Mesfin, M. N., Cohen, A. S., & Meaney, D. F. (2006). Pharmacologically induced calcium oscillations protect neurons from increases in cytosolic calcium after trauma. 97(2), 462-474. doi:10.1111/j.1471-4159.2006.03761.x Gee, C. E., Chen, C.-L. C., Roberts, J. L., Thompson, R., & Watson, S. J. (1983). Identification of proopiomelanocortin neurones in rat hypothalamus by in situ cDNA-mRNA hybridization. Nature, 306(5941), 374-376. Geiger, J. R., & Jonas, P. (2000). Dynamic control of presynaptic Ca2+ inflow by fast- inactivating K+ channels in hippocampal mossy fiber boutons. Neuron, 28(3), 927-939. George, M. S., Nahas, Z., Molloy, M., Speer, A. M., Oliver, N. C., Li, X.-B., . . . Ballenger, J. C. (2000). A controlled trial of daily left prefrontal cortex TMS for treating depression. Biological Psychiatry, 48(10), 962-970. Gibson, J. R., Beierlein, M., & Connors, B. W. (1999). Two networks of electrically coupled inhibitory neurons in neocortex. Nature, 402(6757), 75-79. Giuliano, C., & Cottone, P. (2015). The role of the opioid system in binge eating disorder. CNS spectrums, 20(6), 537-545. doi:10.1017/S1092852915000668 Glaum, S. R., Holzwarth, J. A., & Miller, R. J. J. P. o. t. N. A. o. S. (1990). Glutamate receptors activate Ca2+ mobilization and Ca2+ influx into astrocytes. 87(9), 3454-3458. Glausier, J. R., Fish, K. N., & Lewis, D. A. (2014). Altered parvalbumin basket cell inputs in the dorsolateral prefrontal cortex of schizophrenia subjects. Mol Psychiatry, 19(1), 30-36. doi:10.1038/mp.2013.152 Glazebrook, P. A., Ramirez, A. N., Schild, J. H., Shieh, C.-C., Doan, T., Wible, B. A., & Kunze, D. L. (2002). Potassium channels Kv1.1, Kv1.2 and Kv1.6 influence excitability of rat visceral sensory neurons. The Journal of physiology, 541(2), 467-482. doi:10.1113/jphysiol.2001.018333 Glickfeld, L. L., Atallah, B. V., & Scanziani, M. (2008). Complementary modulation of somatic inhibition by opioids and cannabinoids. Journal of Neuroscience, 28(8), 1824-1832. Goldberg, E. M., Clark, B. D., Zagha, E., Nahmani, M., Erisir, A., & Rudy, B. (2008). K+ Channels at the Axon Initial Segment Dampen Near-Threshold Excitability of Neocortical Fast-Spiking GABAergic Interneurons. Neuron, 58(3), 387-400. doi:https://doi.org/10.1016/j.neuron.2008.03.003 Golding, N. L., Jung, H.-y., Mickus, T., & Spruston, N. (1999). Dendritic Calcium Spike Initiation and Repolarization Are Controlled by Distinct Potassium Channel Subtypes in CA1 Pyramidal Neurons. The Journal of Neuroscience, 19(20), 8789. doi:10.1523/JNEUROSCI.19-20-08789.1999 Goldstein, A., Lowney, L. I., & Pal, B. (1971). Stereospecific and nonspecific interactions of the morphine congener levorphanol in subcellular fractions of mouse brain. Proceedings of the National Academy of Sciences, 68(8), 1742-1747. Golomb, D., Donner, K., Shacham, L., Shlosberg, D., Amitai, Y., & Hansel, D. (2007). Mechanisms of Firing Patterns in Fast-Spiking Cortical Interneurons. PLOS Computational Biology, 3(8), e156. doi:10.1371/journal.pcbi.0030156 Goodman, O. B., Krupnick, J. G., Santini, F., Gurevich, V. V., Penn, R. B., Gagnon, A. W., . . . Benovic, J. L. (1996). β-Arrestin acts as a clathrin adaptor in endocytosis of the β 2- adrenergic receptor. Nature, 383(6599), 447-450. 217

Gordon, J., Amini, S., & White, M. K. (2013). General overview of neuronal cell culture. Methods in molecular biology (Clifton, N.J.), 1078, 1-8. doi:10.1007/978-1-62703-640- 5_1 Gorski, J. A., Talley, T., Qiu, M., Puelles, L., Rubenstein, J. L. R., & Jones, K. R. (2002). Cortical Excitatory Neurons and Glia, But Not GABAergic Neurons, Are Produced in the Emx1-Expressing Lineage. The Journal of Neuroscience, 22(15), 6309. doi:10.1523/JNEUROSCI.22-15-06309.2002 Gray, C. M. (1994). Synchronous oscillations in neuronal systems: mechanisms and functions. Journal of Computational Neuroscience, 1(1-2), 11-38. Grecksch, G., Just, S., Pierstorff, C., Imhof, A. K., Glück, L., Doll, C., . . . Schulz, S. (2011). Analgesic tolerance to high-efficacy agonists but not to morphine is diminished in phosphorylation-deficient S375A μ-opioid receptor knock-in mice. J Neurosci, 31(39), 13890-13896. doi:10.1523/jneurosci.2304-11.2011 Green, M. F., & Nuechterlein, K. H. (1999). Cortical Oscillations and Schizophrenia: Timing Is of the Essence. Archives of General Psychiatry, 56(11), 1007-1008. doi:10.1001/archpsyc.56.11.1007 %J Archives of General Psychiatry Groer, C. E., Tidgewell, K., Moyer, R. A., Harding, W. W., Rothman, R. B., Prisinzano, T. E., & Bohn, L. M. (2007). An opioid agonist that does not induce mu-opioid receptor--arrestin interactions or receptor internalization. Mol Pharmacol, 71(2), 549-557. doi:10.1124/mol.106.028258 Grudt, T. J., & Williams, J. T. (1993). kappa-Opioid receptors also increase potassium conductance. Proceedings of the National Academy of Sciences, 90(23), 11429-11432. Gu, Z. H., Wang, B., Kou, Z. Z., Bai, Y., Chen, T., Dong, Y. L., . . . Li, Y. Q. (2017). Endomorphins: Promising Endogenous Opioid Peptides for the Development of Novel Analgesics. Neurosignals, 25(1), 98-116. doi:10.1159/000484909 Guan, D., Lee, J. C. F., Tkatch, T., Surmeier, D. J., Armstrong, W. E., & Foehring, R. C. (2006). Expression and biophysical properties of Kv1 channels in supragranular neocortical pyramidal neurones. The Journal of physiology, 571(Pt 2), 371-389. doi:10.1113/jphysiol.2005.097006 Gulyás, A. I., Szabó, G. G., Ulbert, I., Holderith, N., Monyer, H., Erdélyi, F., . . . Hájos, N. (2010). Parvalbumin-Containing Fast-Spiking Basket Cells Generate the Field Potential Oscillations Induced by Cholinergic Receptor Activation in the Hippocampus. The Journal of Neuroscience, 30(45), 15134. doi:10.1523/JNEUROSCI.4104-10.2010 Hackler, L., Zadina, J. E., Ge, L.-J., & Kastin, A. J. (1997). Isolation of Relatively Large Amounts of Endomorphin-1 and Endomorphin-2 From Human Brain Cortex. Peptides, 18(10), 1635-1639. doi:https://doi.org/10.1016/S0196-9781(97)00259-3 Haider, B., Duque, A., Hasenstaub, A. R., & McCormick, D. A. (2006). Neocortical network activity in vivo is generated through a dynamic balance of excitation and inhibition. Journal of Neuroscience, 26(17), 4535-4545. Handa, B. K., Lane, A. C., Lord, J. A. H., Morgan, B. A., Rance, M. J., & Smith, C. F. C. (1981). Analogues of β-LPH61–64 posessing selective agonist activity at μ-opiate receptors. European Journal of Pharmacology, 70(4), 531-540. doi:https://doi.org/10.1016/0014- 2999(81)90364-2 Hardy, S., & Leichnetz, G. (1981). Cortical projections to the periaqueductal gray in the monkey: a retrograde and orthograde horseradish peroxidase study. Neuroscience Letters, 22(2), 97-101. Harris, G., & Williams, J. (1991). Transient homologous mu-opioid receptor desensitization in rat locus coeruleus neurons. 11(8), 2574-2581.

218

Harrison, R. G. (1910). The outgrowth of the nerve fiber as a mode of protoplasmic movement. 9(4), 787-846. doi:10.1002/jez.1400090405 Hartwich, K., Pollak, T., & Klausberger, T. (2009). Distinct Firing Patterns of Identified Basket and Dendrite-Targeting Interneurons in the Prefrontal Cortex during Hippocampal Theta and Local Spindle Oscillations. The Journal of Neuroscience, 29(30), 9563. doi:10.1523/JNEUROSCI.1397-09.2009 Hashimoto, T., Arion, D., Unger, T., Maldonado-Aviles, J. G., Morris, H. M., Volk, D. W., . . . Lewis, D. A. (2008). Alterations in GABA-related transcriptome in the dorsolateral prefrontal cortex of subjects with schizophrenia. Mol Psychiatry, 13(2), 147-161. doi:10.1038/sj.mp.4002011 Hawkins, J., Ahmad, S., & Cui, Y. (2017). A Theory of How Columns in the Neocortex Enable Learning the Structure of the World. 11(81). doi:10.3389/fncir.2017.00081 He, L., Fong, J., Von Zastrow, M., & Whistler, J. L. (2002). Regulation of opioid receptor trafficking and morphine tolerance by receptor oligomerization. Cell, 108(2), 271-282. He, S. Q., Zhang, Z. N., Guan, J. S., Liu, H. R., Zhao, B., Wang, H. B., . . . Zhang, X. (2011). Facilitation of μ-opioid receptor activity by preventing δ-opioid receptor-mediated codegradation. Neuron, 69(1), 120-131. doi:10.1016/j.neuron.2010.12.001 Heeroma, J. H., Henneberger, C., Rajakulendran, S., Hanna, M. G., Schorge, S., & Kullmann, D. M. (2009). Episodic ataxia type 1 mutations differentially affect neuronal excitability and transmitter release. Disease models & mechanisms, 2(11-12), 612-619. doi:10.1242/dmm.003582 Heilbronner, S. R., Rodriguez-Romaguera, J., Quirk, G. J., Groenewegen, H. J., & Haber, S. N. (2016). Circuit-Based Corticostriatal Homologies Between Rat and Primate. Biological Psychiatry, 80(7), 509-521. doi:https://doi.org/10.1016/j.biopsych.2016.05.012 Helmstetter, F. J., Tershner, S. A., Poore, L. H., & Bellgowan, P. S. F. (1998). Antinociception following opioid stimulation of the basolateral amygdala is expressed through the periaqueductal gray and rostral ventromedial medulla. Brain Res, 779(1), 104-118. doi:https://doi.org/10.1016/S0006-8993(97)01104-9 Hemstapat, K., Smith, M. T., & Monteith, G. R. (2004). Measurement of intracellular Ca2+ in cultured rat embryonic hippocampal neurons using a fluorescence microplate reader: potential application to biomolecular screening. Journal of Pharmacological and Toxicological Methods, 49(2), 81-87. doi:https://doi.org/10.1016/j.vascn.2003.10.002 Henry, D. J., Grandy, D. K., Lester, H. A., Davidson, N., & Chavkin, C. (1995). Kappa-opioid receptors couple to inwardly rectifying potassium channels when coexpressed by Xenopus oocytes. Mol Pharmacol, 47(3), 551-557. Hestrin, S., & Galarreta, M. (2005). Electrical synapses define networks of neocortical GABAergic neurons. Trends Neurosci, 28(6), 304-309. doi:10.1016/j.tins.2005.04.001 Hiller, J. M., & Fan, L. Q. (1996). Laminar distribution of the multiple opioid receptors in the human cerebral cortex. Neurochem Res, 21(11), 1333-1345. doi:10.1007/bf02532374 Hirano, Y., Oribe, N., Kanba, S., Onitsuka, T., Nestor, P. G., & Spencer, K. M. (2015). Spontaneous Gamma Activity in Schizophrenia. JAMA psychiatry, 72(8), 813-821. doi:10.1001/jamapsychiatry.2014.2642 %J JAMA Psychiatry Homayoun, H., & Moghaddam, B. (2007). NMDA Receptor Hypofunction Produces Opposite Effects on Prefrontal Cortex Interneurons and Pyramidal Neurons. The Journal of Neuroscience, 27(43), 11496-11500. doi:10.1523/jneurosci.2213-07.2007 Hoover, W. B., & Vertes, R. P. (2007). Anatomical analysis of afferent projections to the medial prefrontal cortex in the rat. Brain Structure and Function, 212(2), 149-179. doi:10.1007/s00429-007-0150-4 Hosztafi, S. (2001). The History of Heroin. Acta Pharm Hung, 71(2), 233-242. 219

Hu, H., Vervaeke, K., & Storm, J. F. (2007). M-Channels (Kv7/KCNQ Channels) That Regulate Synaptic Integration, Excitability, and Spike Pattern of CA1 Pyramidal Cells Are Located in the Perisomatic Region. The Journal of Neuroscience, 27(8), 1853. doi:10.1523/JNEUROSCI.4463-06.2007 Huang, C.-L., Slesinger, P. A., Casey, P. J., Jan, Y. N., & Jan, L. Y. (1995). Evidence that direct binding of Gβγ to the GIRK1 G protein-gated inwardly rectifying K+ channel is important for channel activation. Neuron, 15(5), 1133-1143. Huang, C. L., Feng, S., & Hilgemann, D. W. (1998). Direct activation of inward rectifier potassium channels by PIP2 and its stabilization by Gbetagamma. Nature, 391(6669), 803-806. doi:10.1038/35882 Huang, X., Tang, J.-S., Yuan, B., & Jia, H. (2001). Morphine applied to the ventrolateral orbital cortex produces a naloxone-reversible antinociception in the rat. Neuroscience Letters, 299(3), 189-192. Huo, F.-Q., Wang, J., Li, Y.-Q., Chen, T., Han, F., & Tang, J.-S. (2005). GABAergic neurons express μ-opioid receptors in the ventrolateral orbital cortex of the rat. Neuroscience Letters, 382(3), 265-268. doi:http://dx.doi.org/10.1016/j.neulet.2005.03.070 Hyman, S. E., Malenka, R. C., & Nestler, E. J. (2006). Neural Mechanisms of Addiction: The Role of Reward-Related Learning and Memory. 29(1), 565-598. doi:10.1146/annurev.neuro.29.051605.113009 Ikeda, K., Kobayashi, T., Kumanishi, T., Niki, H., & Yano, R. (2000). Involvement of G-protein- activated inwardly rectifying K+ (GIRK) channels in opioid-induced analgesia. Neuroscience Research, 38(1), 113-116. Ikeda, K., Yoshii, M., Sora, I., & Kobayashi, T. (2003). Opioid Receptor Coupling to GIRK Channels. In Z. Z. Pan (Ed.), Opioid Research: Methods and Protocols (pp. 53-64). Totowa, NJ: Humana Press. Inanobe, A., Morishige, K.-I., Takahashi, N., Ito, H., Yamada, M., Takumi, T., . . . Katada, T. (1995). Gβγ directly binds to the carboxyl terminus of the G protein-gated muscarinic K+ channel, GIRK1. Biochemical and biophysical research communications, 212(3), 1022- 1028. Inda, M. C., DeFelipe, J., & Muñoz, A. (2006). Voltage-gated ion channels in the axon initial segment of human cortical pyramidal cells and their relationship with chandelier cells. Proceedings of the National Academy of Sciences of the United States of America, 103(8), 2920-2925. doi:10.1073/pnas.0511197103 Inglefield, J. R., & Shafer, T. J. (2000). Polychlorinated Biphenyl-Stimulation of Ca2+ Oscillations in Developing Neocortical Cells: A Role for Excitatory Transmitters and L- Type Voltage-Sensitive Ca2+ Channels. 295(1), 105-113. Ingram, S. L., & Williams, J. T. (1994). Opioid inhibition of Ih via adenylyl cyclase. Neuron, 13(1), 179-186. Isaacson, Jeffry S., & Scanziani, M. (2011). How Inhibition Shapes Cortical Activity. Neuron, 72(2), 231-243. doi:https://doi.org/10.1016/j.neuron.2011.09.027 Jackson, J., Ayzenshtat, I., Karnani, M. M., & Yuste, R. (2016). VIP+ interneurons control neocortical activity across brain states. J Neurophysiol, 115(6), 3008-3017. doi:10.1152/jn.01124.2015 Janak, P. H., & Tye, K. M. (2015). From circuits to behaviour in the amygdala. Nature, 517(7534), 284-292. doi:10.1038/nature14188 Jasper, H. H. (2012). Jasper's basic mechanisms of the epilepsies (Vol. 80): OUP USA. Jentsch, J. D., & Pennington, Z. T. (2014). Reward, interrupted: Inhibitory control and its relevance to addictions. Neuropharmacology, 76, 479-486.

220

Jeong, S. W., & Ikeda, S. R. (1998). G protein alpha subunit G alpha z couples neurotransmitter receptors to ion channels in sympathetic neurons. Neuron, 21(5), 1201-1212. doi:10.1016/s0896-6273(00)80636-4 Jiang, C., Wang, X., Le, Q., Liu, P., Liu, C., Wang, Z., . . . Ma, L. (2019). Morphine coordinates SST and PV interneurons in the prelimbic cortex to disinhibit pyramidal neurons and enhance reward. Molecular Psychiatry. doi:10.1038/s41380-019-0480-7 Jiang, X., Shen, S., Cadwell, C. R., Berens, P., Sinz, F., Ecker, A. S., . . . Tolias, A. S. (2015). Principles of connectivity among morphologically defined cell types in adult neocortex. Science, 350(6264), aac9462. doi:10.1126/science.aac9462 Jo, M., & Jung, S. T. (2016). Engineering therapeutic antibodies targeting G-protein-coupled receptors. Experimental & molecular medicine, 48(2), e207-e207. doi:10.1038/emm.2015.105 Johns, P. (2014). Chapter 5 - Neurons and glial cells. In P. Johns (Ed.), Clinical Neuroscience (pp. 61-69): Churchill Livingstone. Johnson, S. W., & North, R. A. (1992). Opioids excite dopamine neurons by hyperpolarization of local interneurons. The Journal of Neuroscience, 12(2), 483. doi:10.1523/JNEUROSCI.12-02-00483.1992 Jones, E. (1993). GABAergic neurons and their role in cortical plasticity in primates. Cerebral Cortex, 3(5), 361-372. Joutsa, J., Karlsson, H. K., Majuri, J., Nuutila, P., Helin, S., Kaasinen, V., & Nummenmaa, L. (2018). Binge eating disorder and morbid obesity are associated with lowered mu-opioid receptor availability in the brain. Psychiatry Research: Neuroimaging, 276, 41-45. Juszczak, G. R., & Swiergiel, A. H. (2009). Properties of gap junction blockers and their behavioural, cognitive and electrophysiological effects: Animal and human studies. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 33(2), 181-198. doi:http://dx.doi.org/10.1016/j.pnpbp.2008.12.014 Karagiannis, A., Gallopin, T., Dávid, C., Battaglia, D., Geoffroy, H., Rossier, J., . . . Cauli, B. (2009). Classification of NPY-Expressing Neocortical Interneurons. J Neurosci, 29, 3642-3659. doi:10.1523/JNEUROSCI.0058-09.2009 Karayannis, T., Elfant, D., Huerta-Ocampo, I., Teki, S., Scott, R. S., Rusakov, D. A., . . . Capogna, M. (2010). Slow GABA transient and receptor desensitization shape synaptic responses evoked by hippocampal neurogliaform cells. J Neurosci, 30(29), 9898-9909. doi:10.1523/jneurosci.5883-09.2010 Karschin, C., Dißmann, E., Stühmer, W., & Karschin, A. (1996). IRK (1–3) and GIRK (1–4) inwardly rectifying K+ channel mRNAs are differentially expressed in the adult rat brain. Journal of Neuroscience, 16(11), 3559-3570. Kas, M. J., Van Den Bos, R., Baars, A. M., Lubbers, M., Lesscher, H. M., Hillebrand, J. J., . . . Spruijt, B. M. (2004). Mu‐opioid receptor knockout mice show diminished food‐ anticipatory activity. European Journal of Neuroscience, 20(6), 1624-1632. Katritch, V., Cherezov, V., & Stevens, R. C. (2013). Structure-Function of the G Protein– Coupled Receptor Superfamily. Annual Review of Pharmacology and Toxicology, 53(1), 531-556. doi:10.1146/annurev-pharmtox-032112-135923 Kawabata, S., Tsutsumi, R., Kohara, A., Yamaguchi, T., Nakanishi, S., & Okada, M. (1996). Control of calcium oscillations by phosphorylation of metabotropic glutamate receptors. Nature, 383(6595), 89-92. Kawaguchi, Y., & Kubota, Y. (1997). GABAergic cell subtypes and their synaptic connections in rat frontal cortex. Cerebral cortex (New York, NY: 1991), 7(6), 476-486. Kaye, W. H., Wierenga, C. E., Bailer, U. F., Simmons, A. N., Wagner, A., & Bischoff-Grethe, A. (2013). Does a shared neurobiology for foods and drugs of abuse contribute to extremes 221

of food ingestion in anorexia and bulimia nervosa? Biological Psychiatry, 73(9), 836- 842. Keith, D. E., Anton, B., Murray, S. R., Zaki, P. A., Chu, P. C., Lissin, D. V., . . . von Zastrow, M. (1998). μ-Opioid receptor internalization: opiate drugs have differential effects on a conserved endocytic mechanism in vitro and in the mammalian brain. Mol Pharmacol, 53(3), 377-384. Keith, D. E., Murray, S. R., Zaki, P. A., Chu, P. C., Lissin, D. V., Kang, L., . . . von Zastrow, M. (1996). Morphine activates opioid receptors without causing their rapid internalization. Journal of Biological Chemistry, 271(32), 19021-19024. Kekesi, O., Tuboly, G., Szucs, M., Birkas, E., Morvay, Z., Benedek, G., & Horvath, G. (2011). Long-lasting, distinct changes in central opioid receptor and urinary bladder functions in models of schizophrenia in rats. European Journal of Pharmacology, 661(1), 35-41. doi:https://doi.org/10.1016/j.ejphar.2011.04.022 Kelamangalath, L., Dravid, S. M., George, J., Aldrich, J. V., & Murray, T. F. (2011). κ-Opioid Receptor Inhibition of Calcium Oscillations in Spinal Cord Neurons. 79(6), 1061-1071. doi:10.1124/mol.111.071456 %J Molecular Pharmacology Kelly, M. J., Loose, M. D., & Ronnekleiv, O. K. (1990). Opioids hyperpolarize β-endorphin neurons via μ-receptor activation of a potassium conductance. Neuroendocrinology, 52(3), 268-275. Kim, F. J., Kovalyshyn, I., Burgman, M., Neilan, C., Chien, C. C., & Pasternak, G. W. (2010). Sigma 1 receptor modulation of G-protein-coupled receptor signaling: potentiation of opioid transduction independent from receptor binding. Mol Pharmacol, 77(4), 695-703. doi:10.1124/mol.109.057083 Kim, K.-W., Woo, R.-S., Kim, C.-J., Cheong, Y.-P., Kim, J.-K., Kwun, J., & Cho, K.-P. (2000). Receptor selectivity of Met-enkephalin-Arg6-Phe7, an endogenous opioid peptide, in cerebral cortex of human and rat. Life Sciences, 67(1), 61-71. doi:https://doi.org/10.1016/S0024-3205(00)00606-8 Kirchheim, F., Tinnes, S., Haas, C. A., Stegen, M., & Wolfart, J. (2013). Regulation of action potential delays via voltage-gated potassium Kv1.1 channels in dentate granule cells during hippocampal epilepsy. Front Cell Neurosci, 7, 248-248. doi:10.3389/fncel.2013.00248 Klausberger, T., Magill, P. J., Márton, L. F., Roberts, J. D. B., Cobden, P. M., Buzsáki, G., & Somogyi, P. (2003). Brain-state-and cell-type-specific firing of hippocampal interneurons in vivo. Nature, 421(6925), 844-848. Klein, M., & Kandel, E. R. (1980). Mechanism of calcium current modulation underlying presynaptic facilitation and behavioral sensitization in Aplysia. 77(11), 6912-6916. Knoll, A. T., & Carlezon, W. A. (2010). Dynorphin, stress, and depression. Brain Res, 1314, 56- 73. doi:https://doi.org/10.1016/j.brainres.2009.09.074 Kobayashi, T., Ikeda, K., Ichikawa, T., Togashi, S., & Kumanishi, T. (1996). Effects of sigma ligands on the cloned μ-, δ- and k-opioid receptors co-expressed with G-protein-activated K+ (GIRK) channel in Xenopus oocytes. 119(1), 73-80. doi:10.1111/j.1476- 5381.1996.tb15679.x Koch, T., Widera, A., Bartzsch, K., Schulz, S., Brandenburg, L.-O., Wundrack, N., . . . Höllt, V. (2005). Receptor Endocytosis Counteracts the Development of Opioid Tolerance. Mol Pharmacol, 67(1), 280. doi:10.1124/mol.104.004994 Kole, M. H., Letzkus, J. J., & Stuart, G. J. (2007). Axon initial segment Kv1 channels control axonal action potential waveform and synaptic efficacy. Neuron, 55(4), 633-647.

222

König, M., Zimmer, A. M., Steiner, H., Holmes, P. V., Crawley, J. N., Brownstein, M. J., & Zimmer, A. (1996). Pain responses, anxiety and aggression in mice deficient in preproenkephalin. Nature, 383(6600), 535-538. doi:10.1038/383535a0 Konradi, C., Cole, R. L., Heckers, S., & Hyman, S. E. (1994). Amphetamine regulates gene expression in rat striatum via transcription factor CREB. The Journal of Neuroscience, 14(9), 5623. doi:10.1523/JNEUROSCI.14-09-05623.1994 Kovoor, A., Celver, J., Abdryashitov, R. I., Chavkin, C., & Gurevich, V. (1999). Targeted construction of phosphorylation-independent β-arrestin mutants with constitutive activity in cells. Journal of Biological Chemistry, 274(11), 6831-6834. Kovoor, A., Henry, D. J., & Chavkin, C. (1995). Agonist-induced Desensitization of the Mu Opioid Receptor-coupled Potassium Channel (GIRK1). Journal of Biological Chemistry, 270(2), 589-595. doi:10.1074/jbc.270.2.589 Krook-Magnuson, E., Luu, L., Lee, S.-H., Varga, C., & Soltesz, I. (2011). Ivy and neurogliaform interneurons are a major target of μ-opioid receptor modulation. The Journal of Neuroscience, 31(42), 14861-14870. Kubota, Y., & Kawaguchi, Y. (1997). Two distinct subgroups of cholecystokinin-immunoreactive cortical interneurons. Brain Res, 752(1–2), 175-183. doi:http://dx.doi.org/10.1016/S0006- 8993(96)01446-1 Kuhar, M. J., Pert, C. B., & Snyder, S. H. (1973). Regional distribution of opiate receptor binding in monkey and human brain. Nature. Labro, A. J., Priest, M. F., Lacroix, J. J., Snyders, D. J., & Bezanilla, F. (2015). Kv3.1 uses a timely resurgent K+ current to secure action potential repolarization. Nature Communications, 6(1), 10173. doi:10.1038/ncomms10173 Lafourcade, C. A., & Alger, B. E. (2008). Distinctions among GABAA and GABAB responses revealed by calcium channel antagonists, cannabinoids, opioids, and synaptic plasticity in rat hippocampus. Psychopharmacology (Berl), 198(4), 539-549. doi:10.1007/s00213- 007-1040-4 Lagrange, A. H., Ronnekleiv, O. K., & Kelly, M. J. (1994). The potency of mu-opioid hyperpolarization of hypothalamic arcuate neurons is rapidly attenuated by 17 beta- estradiol. The Journal of Neuroscience, 14(10), 6196. doi:10.1523/JNEUROSCI.14-10- 06196.1994 Lagrange, A. H., Rønnekleiv, O. K., & Kelly, M. J. (1995). Estradiol-17 beta and mu-opioid peptides rapidly hyperpolarize GnRH neurons: a cellular mechanism of negative feedback? Endocrinology, 136(5), 2341-2344. doi:10.1210/endo.136.5.7720682 Lambe, E. K., & Aghajanian, G. K. (2001). The Role of Kv1.2-Containing Potassium Channels in Serotonin-Induced Glutamate Release from Thalamocortical Terminals in Rat Frontal Cortex. The Journal of Neuroscience, 21(24), 9955. doi:10.1523/JNEUROSCI.21-24- 09955.2001 Langlois, L. D., & Nugent, F. S. (2017). Opiates and Plasticity in the Ventral Tegmental Area. ACS chemical neuroscience, 8(9), 1830-1838. doi:10.1021/acschemneuro.7b00281 Lasagna, L. (1964). The Clinical Evaluation of Morphine and its Substitutes as Analgesics. Pharmacol Rev, 16, 47-83. Lau, B. K., Ambrose, B. P., Thomas, C. S., Qiao, M., & Borgland, S. L. (2020). Mu-opioids suppress GABAergic synaptic transmission onto orbitofrontal cortex pyramidal neurons with subregional selectivity. bioRxiv, 2020.2005.2012.091678. doi:10.1101/2020.05.12.091678 Law, P.-Y., Wong, Y. H., & Loh, H. H. (2000). Molecular Mechanisms and Regulation of Opioid Receptor Signaling. Annual Review of Pharmacology and Toxicology, 40(1), 389-430. doi:10.1146/annurev.pharmtox.40.1.389 223

Lawrence, J. J., Saraga, F., Churchill, J. F., Statland, J. M., Travis, K. E., Skinner, F. K., & McBain, C. J. (2006). Somatodendritic Kv7/KCNQ/M channels control interspike interval in hippocampal interneurons. Journal of Neuroscience, 26(47), 12325-12338. Lawrie, A. M., Graham, M. E., Thorn, P., Gallacher, D. V., & Burgoyne, R. D. (1993). Synchronous calcium oscillations in cerebellar granule cells in culture mediated by NMDA receptors. Neuroreport, 4(5), 539-542. doi:10.1097/00001756-199305000-00019 Leão, R. N., Tan, H. M., & Fisahn, A. (2009). Kv7/KCNQ Channels Control Action Potential Phasing of Pyramidal Neurons during Hippocampal Gamma Oscillations In Vitro. The Journal of Neuroscience, 29(42), 13353. doi:10.1523/JNEUROSCI.1463-09.2009 Lee, H. K., Dunwiddie, T., & Hoffer, B. (1980). Electrophysiological interactions of enkephalins with neuronal circuitry in the rat hippocampus. II. Effects on interneuron excitability. Brain Res, 184(2), 331-342. doi:https://doi.org/10.1016/0006-8993(80)90802-1 Lee, M. C., Cahill, C. M., Vincent, J. P., & Beaudet, A. (2002). Internalization and trafficking of opioid receptor ligands in rat cortical neurons. Synapse, 43(2), 102-111. Lee, S., Hjerling-Leffler, J., Zagha, E., Fishell, G., & Rudy, B. (2010). The largest group of superficial neocortical GABAergic interneurons expresses ionotropic serotonin receptors. Journal of Neuroscience, 30(50), 16796-16808. Lemogne, C., Delaveau, P., Freton, M., Guionnet, S., & Fossati, P. (2012). Medial prefrontal cortex and the self in major depression. Journal of affective disorders, 136(1-2), e1-e11. Lewis, D. A. (2014). Inhibitory neurons in human cortical circuits: substrate for cognitive dysfunction in schizophrenia. Cerebral Cortex, 26, 22-26. Lewis, D. A., Curley, A. A., Glausier, J. R., & Volk, D. W. (2012). Cortical parvalbumin interneurons and cognitive dysfunction in schizophrenia. Trends in Neurosciences, 35(1), 57-67. doi:10.1016/j.tins.2011.10.004 Lewis, D. A., Hashimoto, T., & Volk, D. W. (2005). Cortical inhibitory neurons and schizophrenia. Nature reviews. Neuroscience, 6(4), 312-324. doi:10.1038/nrn1648 Leybaert, L., & Sanderson, M. J. (2012). Intercellular Ca(2+) waves: mechanisms and function. Physiological reviews, 92(3), 1359-1392. doi:10.1152/physrev.00029.2011 Li, J.-G., Haines, D. S., & Liu-Chen, L.-Y. (2008). Agonist-promoted Lys63-linked polyubiquitination of the human κ-opioid receptor is involved in receptor downregulation. Mol Pharmacol, 73(4), 1319-1330. Li, T., Tian, C., Scalmani, P., Frassoni, C., Mantegazza, M., Wang, Y., . . . Shu, Y. (2014). Action Potential Initiation in Neocortical Inhibitory Interneurons. PLOS Biology, 12(9), e1001944. doi:10.1371/journal.pbio.1001944 Li, X., Surguchev, A., Bian, S., Navaratnam, D., & Santos-Sacchi, J. (2012). Extracellular chloride regulation of Kv2.1, contributor to the major outward Kv current in mammalian outer hair cells. Am J Physiol Cell Physiol, 302(1), C296-306. doi:10.1152/ajpcell.00177.2011 Lindberg, I., Smythe, S. J., & Dahl, J. L. (1979). Regional distribution of enkephalin in bovine brain. Brain Research Bulletin, 168(1), 200-204. Linnman, C., Coombs, G., Goff, D. C., & Holt, D. J. (2013). Lack of insula reactivity to aversive stimuli in schizophrenia. Schizophrenia Research, 143(1), 150-157. doi:https://doi.org/10.1016/j.schres.2012.10.038 Liu, P. W., & Bean, B. P. (2014). Kv2 channel regulation of action potential repolarization and firing patterns in superior cervical ganglion neurons and hippocampal CA1 pyramidal neurons. J Neurosci, 34(14), 4991-5002. doi:10.1523/JNEUROSCI.1925-13.2014 Llinás, R., Ribary, U., Contreras, D., & Pedroarena, C. (1998). The neuronal basis for consciousness. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, 353(1377), 1841-1849. 224

Llinas, R., Steinberg, I., & Walton, K. (1981). Relationship between presynaptic calcium current and postsynaptic potential in squid giant synapse. Biophysical journal, 33(3), 323-351. Lo, T.-M., Fallert, C. J., Piser, T. M., & Thayer, S. A. (1992). HIV-1 envelope protein evokes intracellular calcium oscillations in rat hippocampal neurons. Brain Res, 594(2), 189-196. doi:https://doi.org/10.1016/0006-8993(92)91125-X Lodge, D. J., Behrens, M. M., & Grace, A. A. (2009). A loss of parvalbumin-containing interneurons is associated with diminished oscillatory activity in an animal model of schizophrenia. Journal of Neuroscience, 29(8), 2344-2354. Logothetis, D. E., Kurachi, Y., Galper, J., Neer, E. J., & Clapham, D. E. (1987). The βγ subunits of GTP-binding proteins activate the muscarinic K+ channel in heart. Nature, 325(6102), 321-326. Loose, M. D., & Kelly, M. J. (1990). Opioids act at mu-receptors to hyperpolarize arcuate neurons via an inwardly rectifying potassium conductance. Brain Res, 513(1), 15-23. doi:10.1016/0006-8993(90)91084-t Loose, M. D., Ronnekleiv, O. K., & Kelly, M. J. (1990). Membrane properties and response to opioids of identified dopamine neurons in the guinea pig hypothalamus. The Journal of Neuroscience, 10(11), 3627-3634. doi:10.1523/jneurosci.10-11-03627.1990 Lui, J. H., Hansen, D. V., & Kriegstein, A. R. (2011). Development and evolution of the human neocortex. Cell, 146(1), 18-36. doi:10.1016/j.cell.2011.06.030 Lüscher, C., & Slesinger, P. A. (2010). Emerging roles for G protein-gated inwardly rectifying potassium (GIRK) channels in health and disease. Nature Reviews Neuroscience, 11(5), 301-315. Lutz, P.-E., & Kieffer, B. L. (2013). Opioid receptors: distinct roles in mood disorders. Trends in Neurosciences, 36(3), 195-206. doi:https://doi.org/10.1016/j.tins.2012.11.002 Lytton, W. W., & Sejnowski, T. J. (1991). Simulations of cortical pyramidal neurons synchronized by inhibitory interneurons. J Neurophysiol, 66(3), 1059-1079. Ma, X., Chen, R., Huang, M., Wang, W., Luo, L., Kim, D. K., . . . Xu, T. (2020). DAMGO- induced μ opioid receptor internalization and recycling restore morphine sensitivity in tolerant rat. European Journal of Pharmacology, 173118. Madison, D., & Nicoll, R. (1988). Enkephalin hyperpolarizes interneurones in the rat hippocampus. The Journal of physiology, 398(1), 123-130. Mains, R. E., & Patterson, P. H. (1973). Primary cultures of dissociated sympathetic neurons: I. Establishment of long-term growth in culture and studies of differentiated properties. The Journal of Cell Biology, 59(2), 329-345. Mandillo, S., & Kanarek, R. B. (2001). Chronic sucrose intake enhances nicotine-induced antinociception in female but not male Long–Evans rats. Pharmacology Biochemistry and Behavior, 68(2), 211-219. doi:https://doi.org/10.1016/S0091-3057(00)00464-0 Margolis, E. B., Toy, B., Himmels, P., Morales, M., & Fields, H. L. (2012). Identification of rat ventral tegmental area GABAergic neurons. PloS one, 7(7), e42365. Marker, C. L., Lujan, R., Loh, H. H., & Wickman, K. (2005). Spinal G-protein-gated potassium channels contribute in a dose-dependent manner to the analgesic effect of mu- and delta- but not kappa-opioids. J Neurosci, 25(14), 3551-3559. doi:10.1523/jneurosci.4899- 04.2005 Markram, H., Toledo-Rodriguez, M., Wang, Y., Gupta, A., Silberberg, G., & Wu, C. (2004). Interneurons of the neocortical inhibitory system. Nature reviews. Neuroscience, 5(10), 793-807. doi:http://www.nature.com/nrn/journal/v5/n10/suppinfo/nrn1519_S1.html Markwardt, S. J., Dieni, C. V., Wadiche, J. I., & Overstreet-Wadiche, L. (2011). Ivy/neurogliaform interneurons coordinate activity in the neurogenic niche. Nature Neuroscience, 14(11), 1407-1409. doi:10.1038/nn.2935 225

Martin-Schild, S., Gerall, A. A., Kastin, A. J., & Zadina, J. E. (1999). Differential distribution of endomorphin 1- and endomorphin 2-like immunoreactivities in the CNS of the rodent. J Comp Neurol, 405(4), 450-471. Martin, M., Matifas, A., Maldonado, R., & Kieffer, B. L. (2003). Acute antinociceptive responses in single and combinatorial opioid receptor knockout mice: distinct mu, delta and kappa tones. Eur J Neurosci, 17(4), 701-708. doi:10.1046/j.1460-9568.2003.02482.x Martina, M., Schultz, J. H., Ehmke, H., Monyer, H., & Jonas, P. (1998). Functional and Molecular Differences between Voltage-Gated K+ Channels of Fast-Spiking Interneurons and Pyramidal Neurons of Rat Hippocampus. The Journal of Neuroscience, 18(20), 8111. doi:10.1523/JNEUROSCI.18-20-08111.1998 Massotte, D., Brillet, K., Kieffer, B., & Milligan, G. (2002). Agonists activate Gi1α or Gi2α fused to the human mu opioid receptor differently. 81(6), 1372-1382. doi:10.1046/j.1471- 4159.2002.00946.x Mathieu-Kia, A., Kellogg, S., Butelman, E., & Kreek, M. (2002). Nicotine addiction: insights from recent animal studies. Psychopharmacology, 162(2), 102-118. doi:10.1007/s00213- 002-1096-0 Matthes, H. W. D., Maldonado, R., Simonin, F., Valverde, O., Slowe, S., Kitchen, I., . . . Kieffer, B. L. (1996). Loss of morphine-induced analgesia, reward effect and withdrawal symptoms in mice lacking the µ-opioid-receptor gene. Nature, 383(6603), 819-823. doi:10.1038/383819a0 Matthias, K., Kirchhoff, F., Seifert, G., Hüttmann, K., Matyash, M., Kettenmann, H., & Steinhäuser, C. J. J. o. N. (2003). Segregated expression of AMPA-type glutamate receptors and glutamate transporters defines distinct astrocyte populations in the mouse hippocampus. 23(5), 1750-1758. McCormick, D. A., Connors, B. W., Lighthall, J. W., & Prince, D. A. (1985). Comparative electrophysiology of pyramidal and sparsely spiny stellate neurons of the neocortex. J Neurophysiol, 54(4), 782-806. McLaughlin, P. J. (2013). Chapter 219 - Proenkephalin-Derived Peptides. In A. J. Kastin (Ed.), Handbook of Biologically Active Peptides (Second Edition) (pp. 1602-1609). Boston: Academic Press. McLaughlin, P. J., & Zagon, I. S. (2013). Chapter 217 - POMC-Derived Opioid Peptides. In A. J. Kastin (Ed.), Handbook of Biologically Active Peptides (Second Edition) (pp. 1592- 1595). Boston: Academic Press. McNally, J. M., & McCarley, R. W. J. (2016). Gamma band oscillations: a key to understanding schizophrenia symptoms and neural circuit abnormalities. Current opinion in psychiatry, 29(3), 202. McPherson, J., Rivero, G., Baptist, M., Llorente, J., Al-Sabah, S., Krasel, C., . . . Kelly, E. (2010). μ-opioid receptors: correlation of agonist efficacy for signalling with ability to activate internalization. Mol Pharmacol, 78(4), 756-766. doi:10.1124/mol.110.066613 McQuiston, A. R. (2008). Layer selective presynaptic modulation of excitatory inputs to hippocampal cornu Ammon 1 by mu-opioid receptor activation. Neuroscience, 151(1), 209-221. doi:10.1016/j.neuroscience.2007.09.077 Meilandt, W. J., Barea-Rodriguez, E., Harvey, S. A. K., & Martinez, J. L. (2004). Role of Hippocampal CA3 μ-Opioid Receptors in Spatial Learning and Memory. 24(12), 2953- 2962. doi:10.1523/JNEUROSCI.5569-03.2004 %J The Journal of Neuroscience Melo, I., Drews, E., Zimmer, A., & Bilkei-Gorzo, A. (2014). Enkephalin knockout male mice are resistant to chronic mild stress. 13(6), 550-558. doi:10.1111/gbb.12139

226

Mena, J. D., Sadeghian, K., & Baldo, B. A. (2011). Induction of hyperphagia and carbohydrate intake by μ-opioid receptor stimulation in circumscribed regions of frontal cortex. Journal of Neuroscience, 31(9), 3249-3260. Mena, J. D., Selleck, R. A., & Baldo, B. A. (2013). Mu-opioid stimulation in rat prefrontal cortex engages hypothalamic orexin/hypocretin-containing neurons, and reveals dissociable roles of nucleus accumbens and hypothalamus in cortically driven feeding. Journal of Neuroscience, 33(47), 18540-18552. Merrer, J. L., Becker, J. A. J., Befort, K., & Kieffer, B. L. (2009). Reward Processing by the Opioid System in the Brain. 89(4), 1379-1412. doi:10.1152/physrev.00005.2009 Merriam, E. P., Thase, M. E., Haas, G. L., Keshavan, M. S., & Sweeney, J. A. (1999). Prefrontal cortical dysfunction in depression determined by Wisconsin Card Sorting Test performance. American Journal of Psychiatry, 156(5), 780-782. Metin, C., Baudoin, J. P., Rakic, S., & Parnavelas, J. G. (2006). Cell and molecular mechanisms involved in the migration of cortical interneurons. Eur J Neurosci, 23(4), 894-900. doi:10.1111/j.1460-9568.2006.04630.x Meyer, H. S., Schwarz, D., Wimmer, V. C., Schmitt, A. C., Kerr, J. N. D., Sakmann, B., & Helmstaedter, M. (2011). Inhibitory interneurons in a cortical column form hot zones of inhibition in layers 2 and 5A. Proceedings of the National Academy of Sciences, 108(40), 16807. doi:10.1073/pnas.1113648108 Mick, I., Myers, J., Ramos, A. C., Stokes, P. R., Erritzoe, D., Colasanti, A., . . . Waldman, A. D. (2016). Blunted endogenous opioid release following an oral amphetamine challenge in pathological gamblers. Neuropsychopharmacology, 41(7), 1742. Mick, I., Myers, J., Stokes, P. R., Erritzoe, D., Colasanti, A., Bowden-Jones, H., . . . Searle, G. E. (2014). Amphetamine induced endogenous opioid release in the human brain detected with [11C] carfentanil PET: replication in an independent cohort. International Journal of Neuropsychopharmacology, 17(12), 2069-2074. Miles, R. (2000). Diversity in Inhibition. Science, 287(5451), 244-246. doi:10.1126/science.287.5451.244 Minzenberg, M. J., Firl, A. J., Yoon, J. H., Gomes, G. C., Reinking, C., & Carter, C. S. (2010). Gamma oscillatory power is impaired during cognitive control independent of medication status in first-episode schizophrenia. Neuropsychopharmacology, 35(13), 2590-2599. Mitchell, J. M., O’Neil, J. P., Janabi, M., Marks, S. M., Jagust, W. J., & Fields, H. L. (2012). Alcohol consumption induces endogenous opioid release in the human orbitofrontal cortex and nucleus accumbens. Science translational medicine, 4(116), 116ra116- 116ra116. Mitrovic, I., & Celeste Napier, T. (1998). Substance P attenuates and DAMGO potentiates amygdala glutamatergic neurotransmission within the ventral pallidum. Brain Res, 792(2), 193-206. doi:https://doi.org/10.1016/S0006-8993(98)00130-9 Miyoshi, G., Hjerling-Leffler, J., Karayannis, T., Sousa, V. H., Butt, S. J. B., Battiste, J., . . . Fishell, G. (2010). Genetic Fate Mapping Reveals That the Caudal Ganglionic Eminence Produces a Large and Diverse Population of Superficial Cortical Interneurons. The Journal of Neuroscience, 30(5), 1582. Mizoguchi, H., Sakurada, T., & Sakurada, S. (2013). Chapter 212 - Endomorphins. In A. J. Kastin (Ed.), Handbook of Biologically Active Peptides (Second Edition) (pp. 1556- 1561). Boston: Academic Press. Mo, Z. L., Adamson, C. L., & Davis, R. L. (2002). Dendrotoxin-sensitive K(+) currents contribute to accommodation in murine spiral ganglion neurons. J Physiol, 542(Pt 3), 763-778.

227

Mogil, J. S. (1999). The genetic mediation of individual differences in sensitivity to pain and its inhibition. Proceedings of the National Academy of Sciences of the United States of America, 96(14), 7744-7751. doi:10.1073/pnas.96.14.7744 Molnár, Z., & Pollen, A. (2014). How unique is the human neocortex? Development, 141(1), 11. doi:10.1242/dev.101279 Morganstern, I., Liang, S., Ye, Z., Karatayev, O., & Leibowitz, S. F. (2012). Disturbances in behavior and cortical enkephalin gene expression during the anticipation of ethanol in rats characterized as high drinkers. Alcohol, 46(6), 559-568. Morrison, S. E., & Salzman, C. D. (2010). Re-valuing the amygdala. Current opinion in neurobiology, 20(2), 221-230. doi:https://doi.org/10.1016/j.conb.2010.02.007 Morse, M., Sun, H., Tran, E., Levenson, R., & Fang, Y. (2013). Label-free integrative pharmacology on-target of opioid ligands at the opioid receptor family. BMC Pharmacol Toxicol, 14, 17. doi:10.1186/2050-6511-14-17 Moulin, D. E., Ling, G. S., & Pasternak, G. W. (1988). Unidirectional analgesic cross-tolerance between morphine and levorphanol in the rat. Pain, 33(2), 233-239. doi:10.1016/0304- 3959(88)90095-4 Murakoshi, H., & Trimmer, J. S. (1999). Identification of the Kv2.1 K+ Channel as a Major Component of the Delayed Rectifier K+ Current in Rat Hippocampal Neurons. The Journal of Neuroscience, 19(5), 1728. doi:10.1523/JNEUROSCI.19-05-01728.1999 Muramoto, K., Ichikawa, M., Kawahara, M., Kobayashi, K., & Kuroda, Y. (1993). Frequency of synchronous oscillations of neuronal activity increases during development and is correlated to the number of synapses in cultured cortical neuron networks. Neurosci Lett, 163(2), 163-165. doi:10.1016/0304-3940(93)90372-r Murer, G., Adelbrecht, C., Lauritzen, I., Lesage, F., Lazdunski, M., Agid, Y., & Raisman-Vozari, R. (1997). An immunocytochemical study on the distribution of two G-protein-gated inward rectifier potassium channels (GIRK2 and GIRK4) in the adult rat brain. Neuroscience, 80(2), 345-357. doi:10.1016/s0306-4522(97)00001-8 Murphy, M. J. M., & Deutch, A. Y. (2018). Organization of afferents to the orbitofrontal cortex in the rat. Journal of Comparative Neurology, 526(9), 1498-1526. doi:10.1002/cne.24424 Murphy, T. H., Blatter, L. A., Wier, W. G., & Baraban, J. M. (1992). Spontaneous synchronous synaptic calcium transients in cultured cortical neurons. J Neurosci, 12(12), 4834-4845. Nadal, X., Baños, J. E., Kieffer, B. L., & Maldonado, R. (2006). Neuropathic pain is enhanced in δ‐opioid receptor knockout mice. European Journal of Neuroscience, 23(3), 830-834. Nakanishi, K., & Kukita, F. (1998). Functional synapses in synchronized bursting of neocortical neurons in culture. Brain Res, 795(1), 137-146. doi:https://doi.org/10.1016/S0006- 8993(98)00283-2 Narita, M., Matsushima, Y., Niikura, K., Narita, M., Takagi, S., Nakahara, K., . . . Asato, M. (2010). Implication of dopaminergic projection from the ventral tegmental area to the anterior cingulate cortex in μ‐opioid‐induced place preference. Addiction biology, 15(4), 434-447. Navratilova, E., Xie, J. Y., Meske, D., Qu, C., Morimura, K., Okun, A., . . . Porreca, F. (2015). Endogenous opioid activity in the anterior cingulate cortex is required for relief of pain. Journal of Neuroscience, 35(18), 7264-7271. Nelson, S. B., & Valakh, V. (2015). Excitatory/inhibitory balance and circuit homeostasis in autism spectrum disorders. Neuron, 87(4), 684-698. Newman, S., Pascal, L., Sadeghian, K., & Baldo, B. A. (2013). Sweetened-fat intake sensitizes gamma-aminobutyric acid-mediated feeding responses elicited from the nucleus accumbens shell. Biol Psychiatry, 73(9), 843-850. doi:10.1016/j.biopsych.2012.11.027

228

Niquille, M., Limoni, G., Markopoulos, F., Cadilhac, C., Prados, J., Holtmaat, A., & Dayer, A. (2018). Neurogliaform cortical interneurons derive from cells in the preoptic area. Elife, 7, e32017. Noback, C. R., Strominger, N. L., Demarest, R. J., & Ruggiero, D. A. (2005). The human nervous system: structure and function: Springer Science & Business Media. Noble, F., & Cox, B. M. (1996). Differential desensitization of mu- and delta- opioid receptors in selected neural pathways following chronic morphine treatment. British journal of pharmacology, 117(1), 161-169. doi:10.1111/j.1476-5381.1996.tb15169.x Nockemann, D., Rouault, M., Labuz, D., Hublitz, P., McKnelly, K., Reis, F. C., . . . Heppenstall, P. A. (2013). The K+ channel GIRK2 is both necessary and sufficient for peripheral opioid‐mediated analgesia. EMBO molecular medicine, 5(8), 1263-1277. Nonnemacher, M., Banerjee, A., Li, L., Pirrone, V., Krebs, F., & Wigdahl, B. (2017). cAMP Signaling Enhances HIV-1 Long Terminal Repeat (LTR)-directed Transcription and Viral Replication in Bone Marrow Progenitor Cells. Clinical Medicine Insights: Pathology, 10. doi:10.1177/1179555717694535 North, R. A. (1989). Twelfth Gaddum memorial lecture. Drug receptors and the inhibition of nerve cells. British journal of pharmacology, 98(1), 13. North, R. A., Williams, J. T., Surprenant, A., & Christie, M. J. (1987). Mu and delta receptors belong to a family of receptors that are coupled to potassium channels. Proceedings of the National Academy of Sciences, 84(15), 5487. doi:10.1073/pnas.84.15.5487 Nuñez, L., Sanchez, A., Fonteriz, R. I., & Garcia‐Sancho, J. (1996). Mechanisms for synchronous calcium oscillations in cultured rat cerebellar neurons. European Journal of Neuroscience, 8(1), 192-201. Oertel, B., Preibisch, C., Wallenhorst, T., Hummel, T., Geisslinger, G., Lanfermann, H., & Lötsch, J. (2008). Differential Opioid Action on Sensory and Affective Cerebral Pain Processing. 83(4), 577-588. doi:10.1038/sj.clpt.6100441 Olah, S., Fule, M., Komlosi, G., Varga, C., Baldi, R., Barzo, P., & Tamas, G. (2009). Regulation of cortical microcircuits by unitary GABA-mediated volume transmission. Nature, 461(7268), 1278-1281. doi:10.1038/nature08503 Olah, S., Komlosi, G., Szabadics, J., Varga, C., Toth, E., Barzo, P., & Tamas, G. (2007). Output of neurogliaform cells to various neuron types in the human and rat cerebral cortex. Frontiers in neural circuits, 1, 4. doi:10.3389/neuro.04.004.2007 Olenik, C., & Meyer, D. K. (1997). Development of proenkephalin gene expression in rat neocortex: a non-radioactive in situ hybridization study. Molecular brain research, 44(1), 83-91. Ong, W.-Y., Stohler, C. S., & Herr, D. R. (2019). Role of the Prefrontal Cortex in Pain Processing. Molecular Neurobiology, 56(2), 1137-1166. doi:10.1007/s12035-018-1130-9 Onozawa, K., Yagasaki, Y., Izawa, Y., Abe, H., & Kawakami, Y. (2011). Amygdala-prefrontal pathways and the dopamine system affect nociceptive responses in the prefrontal cortex. BMC Neurosci, 12, 115. doi:10.1186/1471-2202-12-115 Orekhova, E. V., Stroganova, T. A., Nygren, G., Tsetlin, M. M., Posikera, I. N., Gillberg, C., & Elam, M. (2007). Excess of High Frequency Electroencephalogram Oscillations in Boys with Autism. Biological Psychiatry, 62(9), 1022-1029. doi:https://doi.org/10.1016/j.biopsych.2006.12.029 Ossipov, M. H., Morimura, K., & Porreca, F. (2014). Descending pain modulation and chronification of pain. Current opinion in supportive and palliative care, 8(2), 143-151. doi:10.1097/SPC.0000000000000055 Overstreet-Wadiche, L., & McBain, C. J. (2015). Neurogliaform cells in cortical circuits. Nature Reviews Neuroscience, 16(8), 458. 229

Pacico, N., & Mingorance-Le Meur, A. (2014). New in vitro phenotypic assay for epilepsy: fluorescent measurement of synchronized neuronal calcium oscillations. PloS one, 9(1), e84755. Packer, A. M., & Yuste, R. (2011). Dense, unspecific connectivity of neocortical parvalbumin- positive interneurons: a canonical microcircuit for inhibition? The Journal of Neuroscience, 31(37), 13260-13271. Pan, Y.-X., Bolan, E., & Pasternak, G. W. (2002). Dimerization of morphine and orphanin FQ/nociceptin receptors: generation of a novel opioid receptor subtype. Biochemical and biophysical research communications, 297(3), 659-663. doi:https://doi.org/10.1016/S0006-291X(02)02258-1 Pan, Y. X., Xu, J., Bolan, E., Abbadie, C., Chang, A., Zuckerman, A., . . . Pasternak, G. W. (1999). Identification and characterization of three new alternatively spliced mu-opioid receptor isoforms. Mol Pharmacol, 56(2), 396-403. doi:10.1124/mol.56.2.396 Pang, K., & Rose, G. M. (1989). Differential effects of methionine5-enkephalin on hippocampal pyramidal cells and interneurons. Neuropharmacology, 28(11), 1175-1181. doi:10.1016/0028-3908(89)90208-6 Papaleo, F., Kieffer, B. L., Tabarin, A., & Contarino, A. (2007). Decreased motivation to eat in µ- opioid receptor-deficient mice. 25(11), 3398-3405. doi:10.1111/j.1460- 9568.2007.05595.x Parra, P., Gulyas, A. I., & Miles, R. (1998). How many subtypes of inhibitory cells in the hippocampus? Neuron, 20(5), 983-993. Pasternak, G. W., & Pan, Y.-X. (2011). Mix and match: heterodimers and opioid tolerance. Neuron, 69(1), 6-8. doi:10.1016/j.neuron.2010.12.030 Pasternak, G. W., & Pan, Y.-X. (2013). Mu opioids and their receptors: evolution of a concept. Pharmacological Reviews, 65(4), 1257-1317. doi:10.1124/pr.112.007138 Pasternak, G. W., Snowman, A. M., & Snyder, S. H. (1975). Selective enhancement of [3H]opiate agonist binding by divalent cations. Mol Pharmacol, 11(6), 735-744. Patel, J. P., Sengupta, R., Bardi, G., Khan, M. Z., Mullen-Przeworski, A., & Meucci, O. (2006). Modulation of neuronal CXCR4 by the μ-opioid agonist DAMGO. Journal of neurovirology, 12(6), 492-500. doi:10.1080/13550280601064798 Pathak, D., Guan, D., & Foehring, R. C. (2016). Roles of specific Kv channel types in repolarization of the action potential in genetically identified subclasses of pyramidal neurons in mouse neocortex. J Neurophysiol, 115(5), 2317-2329. Pathan, H., & Williams, J. (2012). Basic opioid pharmacology: an update. British journal of pain, 6(1), 11-16. doi:10.1177/2049463712438493 Penn, Y., Segal, M., & Moses, E. (2016). Network synchronization in hippocampal neurons. Proceedings of the National Academy of Sciences of the United States of America, 113(12), 3341-3346. doi:10.1073/pnas.1515105113 Persson, A. I., Thorlin, T., Bull, C., Zarnegar, P., Ekman, R., Terenius, L., & Eriksson, P. S. (2003). Mu‐and delta‐opioid receptor antagonists decrease proliferation and increase neurogenesis in cultures of rat adult hippocampal progenitors. European Journal of Neuroscience, 17(6), 1159-1172. Pert, C. B., & Snyder, S. H. (1973). Opiate receptor: demonstration in nervous tissue. Science, 179(4077), 1011-1014. Peters, H. C., Hu, H., Pongs, O., Storm, J. F., & Isbrandt, D. (2005). Conditional transgenic suppression of M channels in mouse brain reveals functions in neuronal excitability, resonance and behavior. Nature Neuroscience, 8(1), 51-60. Petrovic, P., Kalso, E., Petersson, K. M., & Ingvar, M. (2002). Placebo and opioid analgesia-- imaging a shared neuronal network. Science, 295(5560), 1737-1740. 230

Pettinati, H. M., O'Brien, C. P., Rabinowitz, A. R., Wortman, S. P., Oslin, D. W., Kampman, K. M., & Dackis, C. A. (2006). The Status of Naltrexone in the Treatment of Alcohol Dependence: Specific Effects on Heavy Drinking. Journal of clinical psychopharmacology, 26(6). Phillips, L. H., MacPherson, S. E., & Della Sala, S. (2002). Age, cognition and emotion: the role of anatomical segregation in the frontal lobes: the role of anatomical segregation in the frontal lobes. In Handbook of Neuropsychology: the frontal lobes (pp. 73-98): Elsevier Science. Pi, H.-J., Hangya, B., Kvitsiani, D., Sanders, J. I., Huang, Z. J., & Kepecs, A. (2013). Cortical interneurons that specialize in disinhibitory control. Nature, 503(7477), 521-524. doi:10.1038/nature12676 Pierce, R. C., & Kumaresan, V. (2006). The mesolimbic dopamine system: the final common pathway for the reinforcing effect of drugs of abuse? Neuroscience Biobehavioral Reviews, 30(2), 215-238. Piomelli, D., & Greengard, P. (1990). Lipoxygenase metabolites of arachidonic acid in neuronal transmembrane signalling. Trends in Pharmacological Sciences, 11(9), 367-373. doi:https://doi.org/10.1016/0165-6147(90)90182-8 Piros, E. T., Prather, P. L., Law, P. Y., Evans, C. J., & Hales, T. G. (1996). Voltage-dependent inhibition of Ca2+ channels in GH3 cells by cloned mu- and delta-opioid receptors. Mol Pharmacol, 50(4), 947-956. Piros, E. T., Prather, P. L., Loh, H. H., Law, P. Y., Evans, C. J., & Hales, T. G. (1995). Ca2+ channel and adenylyl cyclase modulation by cloned mu-opioid receptors in GH3 cells. Mol Pharmacol, 47(5), 1041-1049. Pla, R., Stanco, A., Howard, M. A., Rubin, A. N., Vogt, D., Mortimer, N., . . . Rubenstein, J. L. R. (2017). Dlx1 and Dlx2 Promote Interneuron GABA Synthesis, Synaptogenesis, and Dendritogenesis. Cerebral Cortex, 28(11), 3797-3815. doi:10.1093/cercor/bhx241 Pomerleau, O. F. (1998). Endogenous opioids and smoking: A review of progress and problems. Psychoneuroendocrinology, 23(2), 115-130. doi:https://doi.org/10.1016/S0306- 4530(97)00074-7 Ponterio, G., Tassone, A., Sciamanna, G., Riahi, E., Vanni, V., Bonsi, P., & Pisani, A. (2013). Powerful inhibitory action of mu opioid receptors (MOR) on cholinergic interneuron excitability in the dorsal striatum. Neuropharmacology, 75, 78-85. doi:https://doi.org/10.1016/j.neuropharm.2013.07.006 Porter, J. T., Cauli, B., Staiger, J. F., Lambolez, B., Rossier, J., & Audinat, E. (1998). Properties of bipolar VIPergic interneurons and their excitation by pyramidal neurons in the rat neocortex. Eur J Neurosci, 10(12), 3617-3628. Pradhan, A. A., Befort, K., Nozaki, C., Gavériaux-Ruff, C., & Kieffer, B. L. (2011). The delta opioid receptor: an evolving target for the treatment of brain disorders. Trends Pharmacol Sci, 32(10), 581-590. doi:10.1016/j.tips.2011.06.008 Premoli, I., Bergmann, T. O., Fecchio, M., Rosanova, M., Biondi, A., Belardinelli, P., & Ziemann, U. (2017). The impact of GABAergic drugs on TMS-induced brain oscillations in human motor cortex. Neuroimage, 163, 1-12. doi:https://doi.org/10.1016/j.neuroimage.2017.09.023 Price, C. J., Cauli, B., Kovacs, E. R., Kulik, A., Lambolez, B., Shigemoto, R., & Capogna, M. (2005). Neurogliaform Neurons Form a Novel Inhibitory Network in the Hippocampal CA1 Area. The Journal of Neuroscience, 25(29), 6775. doi:10.1523/JNEUROSCI.1135- 05.2005

231

Price, C. J., Scott, R., Rusakov, D. A., & Capogna, M. (2008). GABA(B) receptor modulation of feedforward inhibition through hippocampal neurogliaform cells. J Neurosci, 28(27), 6974-6982. doi:10.1523/jneurosci.4673-07.2008 Prinster, S. C., Hague, C., & Hall, R. A. (2005). Heterodimerization of g protein-coupled receptors: specificity and functional significance. Pharmacol Rev, 57(3), 289-298. doi:10.1124/pr.57.3.1 Przewlocki, R., Parsons, K. L., Sweeney, D. D., Trotter, C., Netzeband, J. G., Siggins, G. R., & Gruol, D. L. (1999). Opioid Enhancement of Calcium Oscillations and Burst Events Involving NMDA Receptors and L-Type Calcium Channels in Cultured Hippocampal Neurons. The Journal of Neuroscience, 19(22), 9705-9715. doi:10.1523/jneurosci.19-22- 09705.1999 Purves, D. G. J. A. D. F. W. C. H. A.-S. L. J. O. M. L. E. W. (2008). Neuroscience, 4th ed. Sunderland, MA, US: Sinauer Associates. Puumala, T., & Sirviö, J. (1998). Changes in activities of dopamine and serotonin systems in the frontal cortex underlie poor choice accuracy and impulsivity of rats in an attention task. Neuroscience, 83(2), 489-499. Qiu, Y.-H., Wu, X.-Y., Xu, H., & Sackett, D. (2009). Neuroimaging study of placebo analgesia in humans. Neuroscience bulletin, 25(5), 277-282. Qu, C.-L., Huo, F.-Q., Huang, F.-S., & Tang, J.-S. (2015). Activation of mu-opioid receptors in the ventrolateral orbital cortex inhibits the GABAergic miniature inhibitory postsynaptic currents in rats. Neuroscience Letters, 592, 64-69. doi:https://doi.org/10.1016/j.neulet.2015.02.045 Raehal, K. M., Schmid, C. L., Groer, C. E., & Bohn, L. M. (2011). Functional selectivity at the μ- opioid receptor: implications for understanding opioid analgesia and tolerance. Pharmacol Rev, 63(4), 1001-1019. doi:10.1124/pr.111.004598 Rakic, P. (2009). Evolution of the neocortex: a perspective from developmental biology. Nature reviews. Neuroscience, 10(10), 724-735. doi:10.1038/nrn2719 Ramon y Cajal, S. (1995). Histology of the Nervous System of Man and Vertebrates (History of Neuroscience, No 6). In: Oxford University Press. Ramsey, N. F., Gerrits, M. A. F. M., & Van Ree, J. M. (1999). Naltrexone affects cocaine self- administration in naı̈ ve rats through the ventral tegmental area rather than dopaminergic target regions. European Neuropsychopharmacology, 9(1), 93-99. doi:https://doi.org/10.1016/S0924-977X(98)00009-1 Raveh, A., Cooper, A., Guy-David, L., & Reuveny, E. (2010). Nonenzymatic rapid control of GIRK channel function by a G protein-coupled receptor kinase. Cell, 143(5), 750-760. doi:10.1016/j.cell.2010.10.018 Reith, M. E., Sershen, H., Vadasz, C., & Lajtha, A. (1981). Strain differences in opiate receptors in mouse brain. Eur J Pharmacol, 74(4), 377-380. doi:10.1016/0014-2999(81)90059-5 Ren, J., Aika, Y., Heizmann, C., & Kosaka, T. (1992). Quantitative analysis of neurons and glial cells in the rat somatosensory cortex, with special reference to GABAergic neurons and parvalbumin-containing neurons. Experimental Brain Research, 92(1), 1-14. Rios, C. D., Jordan, B. A., Gomes, I., & Devi, L. A. (2001). G-protein-coupled receptor dimerization: modulation of receptor function. Pharmacology & Therapeutics, 92(2), 71- 87. doi:https://doi.org/10.1016/S0163-7258(01)00160-7 Rivero, G., Llorente, J., McPherson, J., Cooke, A., Mundell, S. J., McArdle, C. A., . . . Kelly, E. (2012). Endomorphin-2: a biased agonist at the μ-opioid receptor. Mol Pharmacol, 82(2), 178-188. doi:10.1124/mol.112.078659 Robbins, C. A., & Tempel, B. L. (2012). Kv1.1 and Kv1.2: Similar channels, different seizure models. Epilepsia, 53, 134-141. doi:10.1111/j.1528-1167.2012.03484.x 232

Robinson, H. P., Kawahara, M., Jimbo, Y., Torimitsu, K., Kuroda, Y., & Kawana, A. (1993). Periodic synchronized bursting and intracellular calcium transients elicited by low magnesium in cultured cortical neurons. J Neurophysiol, 70(4), 1606-1616. doi:10.1152/jn.1993.70.4.1606 Rodríguez-Muñoz, M., Sánchez-Blázquez, P., Vicente-Sánchez, A., Berrocoso, E., & Garzón, J. (2012). The Mu-Opioid Receptor and the NMDA Receptor Associate in PAG Neurons: Implications in Pain Control. Neuropsychopharmacology, 37(2), 338-349. doi:10.1038/npp.2011.155 Rojas, D. C., & Wilson, L. B. (2014). γ-band abnormalities as markers of autism spectrum disorders. Biomarkers in medicine, 8(3), 353-368. Rola, R., Jarkiewicz, M., & Szulczyk, P. (2008). Modulation of Ca2+ channel current by mu opioid receptors in prefrontal cortex pyramidal neurons in rats. Acta Neurobiol Exp (Wars), 68(1), 10-25. Rolls, E. T., & Grabenhorst, F. (2008). The orbitofrontal cortex and beyond: from affect to decision-making. Progress in neurobiology, 86(3), 216-244. Rose, J. E., & Woolsey, C. N. (1948). The orbitofrontal cortex and its connections with the mediodorsal nucleus in rabbit, sheep and cat. Res Publ Assoc Res Nerv Ment Dis, 27 (1 vol.), 210-232. Rossier, J., Vargo, T. M., Minick, S., Ling, N., Bloom, F. E., & Guillemin, R. (1977). Regional dissociation of beta-endorphin and enkephalin contents in rat brain and pituitary. Proceedings of the National Academy of Sciences, 74(11), 5162-5165. Rotaru, D. C., Barrionuevo, G., & Sesack, S. R. (2005). Mediodorsal thalamic afferents to layer III of the rat prefrontal cortex: Synaptic relationships to subclasses of interneurons. 490(3), 220-238. doi:10.1002/cne.20661 Royce, G. J. (1983). Cortical neurons with collateral projections to both the caudate nucleus and the centromedian-parafascicular thalamic complex: A fluorescent retrograde double labeling study in the cat. Experimental Brain Research, 50(2), 157-165. doi:10.1007/BF00239179 Rubenstein, J. L., & Merzenich, M. M. (2003). Model of autism: increased ratio of excitation/inhibition in key neural systems. Genes Brain Behav, 2(5), 255-267. doi:10.1034/j.1601-183x.2003.00037.x Rubia, K., Cubillo, A., Smith, A. B., Woolley, J., Heyman, I., & Brammer, M. J. (2010). Disorder‐specific dysfunction in right inferior prefrontal cortex during two inhibition tasks in boys with attention‐deficit hyperactivity disorder compared to boys with obsessive–compulsive disorder. Human brain mapping, 31(2), 287-299. Rubovitch, V., Gafni, M., & Sarne, Y. (2003). The mu opioid agonist DAMGO stimulates cAMP production in SK-N-SH cells through a PLC–PKC–Ca++ pathway. Molecular Brain Research, 110(2), 261-266. doi:https://doi.org/10.1016/S0169-328X(02)00656-3 Rudebeck, P. H., & Rich, E. L. (2018). Orbitofrontal cortex. Curr Biol, 28(18), R1083-r1088. doi:10.1016/j.cub.2018.07.018 Rudy, B., Fishell, G., Lee, S., & Hjerling-Leffler, J. (2011). Three groups of interneurons account for nearly 100% of neocortical GABAergic neurons. Developmental Neurobiology, 71(1), 45-61. doi:10.1002/dneu.20853 Rudy, B., Maffie, J., Amarillo, Y., Clark, B., Goldberg, E., Jeong, H., . . . Zagha, E. (2009). Voltage gated potassium channels: Structure and function of Kv1 to Kv9 subfamilies. Encyclopedia of Neuroscience, 10, 397-425. Rusin, K. I., Giovannucci, D. R., Stuenkel, E. L., & Moises, H. C. (1997). Kappa-opioid receptor activation modulates Ca2+ currents and secretion in isolated neuroendocrine nerve terminals. J Neurosci, 17(17), 6565-6574. 233

Sabatini, B. L., & Regehr, W. G. (1997). Control of neurotransmitter release by presynaptic waveform at the granule cell to Purkinje cell synapse. Journal of Neuroscience, 17(10), 3425-3435. Saitoh, A., & Yamada, M. (2012). Antidepressant-like Effects of δ Opioid Receptor Agonists in Animal Models. Curr Neuropharmacol, 10(3), 231-238. doi:10.2174/157015912803217314 Sánchez-Blázquez, P., Gómez-Serranillos, P., & Garzón, J. (2001). Agonists determine the pattern of G-protein activation in μ-opioid receptor-mediated supraspinal analgesia. Brain Research Bulletin, 54(2), 229-235. doi:https://doi.org/10.1016/S0361-9230(00)00448-2 Santella, T. M., & Triggle, D. J. (2007). Opium: Infobase Publishing. Sarmiere, P. D., Weigle, C. M., & Tamkun, M. M. (2008). The Kv2. 1 K+ channel targets to the axon initial segment of hippocampal and cortical neurons in culture and in situ. BMC neuroscience, 9(1), 112. Sauriyal, D. S., Jaggi, A. S., & Singh, N. (2011). Extending pharmacological spectrum of opioids beyond analgesia: Multifunctional aspects in different pathophysiological states. Neuropeptides, 45(3), 175-188. doi:https://doi.org/10.1016/j.npep.2010.12.004 Scemes, E., & Giaume, C. (2006). Astrocyte calcium waves: what they are and what they do. Glia, 54(7), 716-725. doi:10.1002/glia.20374 Scherrer, G., Tryoen-Tóth, P., Filliol, D., Matifas, A., Laustriat, D., Cao, Y. Q., . . . Gavériaux- Ruff, C. (2006). Knockin mice expressing fluorescent δ-opioid receptors uncover G protein-coupled receptor dynamics invivo. Proceedings of the National Academy of Sciences, 103(25), 9691-9696. Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., . . . Schmid, B. (2012). Fiji: an open-source platform for biological-image analysis. Nature methods, 9(7), 676-682. Schmidt, P., Schmolke, C., Musshoff, F., Menzen, M., Prohaska, C., & Madea, B. (2003). Area- specific increased density of μ-opioid receptor immunoreactive neurons in the cerebral cortex of drug-related fatalities. Forensic Science International, 133(3), 204-211. doi:https://doi.org/10.1016/S0379-0738(03)00067-7 Schreff, M., Schulz, S., Wiborny, D., & Höllt, V. (1998). Immunofluorescent identification of endomorphin-2-containing nerve fibers and terminals in the rat brain and spinal cord. Neuroreport, 9(6), 1031-1034. doi:10.1097/00001756-199804200-00014 Schreibmayer, W., Dessauer, C. W., Vorobiov, D., Gilman, A. G., Lester, H. A., Davidson, N., & Dascal, N. (1996). Inhibition of an inwardly rectifying K+ channel by G-protein alpha- subunits. Nature, 380(6575), 624-627. doi:10.1038/380624a0 Schulz, S., Mayer, D., Pfeiffer, M., Stumm, R., Koch, T., & Höllt, V. (2004). Morphine induces terminal μ‐opioid receptor desensitization by sustained phosphorylation of serine‐375. The EMBO journal, 23(16), 3282-3289. Schweimer, J., Saft, S., & Hauber, W. (2005). Involvement of catecholamine neurotransmission in the rat anterior cingulate in effort-related decision making. Behavioral neuroscience, 119(6), 1687. Scott, D. J., Stohler, C. S., Egnatuk, C. M., Wang, H., Koeppe, R. A., & Zubieta, J.-K. (2008). Placebo and Nocebo Effects Are Defined by Opposite Opioid and Dopaminergic Responses. Archives of General Psychiatry, 65(2), 220-231. doi:10.1001/archgenpsychiatry.2007.34 %J Archives of General Psychiatry Selleck, R. A., Lake, C., Estrada, V., Riederer, J., Andrzejewski, M., Sadeghian, K., & Baldo, B. A. (2015). Endogenous Opioid Signaling in the Medial Prefrontal Cortex is Required for the Expression of Hunger-Induced Impulsive Action. Neuropsychopharmacology, 40(10), 2464-2474. doi:10.1038/npp.2015.97 234

Services, T. U. S. D. o. H. a. H. (2020). Retrieved from https://www.cdc.gov/drugoverdose/data/index.html Sheffield, M. E., Edgerton, G. B., Heuermann, R. J., Deemyad, T., Mensh, B. D., & Spruston, N. (2013). Mechanisms of retroaxonal barrage firing in hippocampal interneurons. J Physiol, 591(19), 4793-4805. doi:10.1113/jphysiol.2013.258418 Sheffield, M. E. J., Best, T. K., Mensh, B. D., Kath, W. L., & Spruston, N. (2011). Slow integration leads to persistent action potential firing in distal axons of coupled interneurons. Nature Neuroscience, 14(2), 200-207. doi:10.1038/nn.2728 Shen, M., Piser, T. M., Seybold, V. S., & Thayer, S. A. (1996). Cannabinoid Receptor Agonists Inhibit Glutamatergic Synaptic Transmission in Rat Hippocampal Cultures. 16(14), 4322- 4334. doi:10.1523/JNEUROSCI.16-14-04322.1996 %J The Journal of Neuroscience Shi, M.-M., Fan, K.-M., Qiao, Y.-N., Xu, J.-H., Qiu, L.-J., Li, X., . . . Liu, Z.-Q. (2020). Hippocampal µ-opioid receptors on GABAergic neurons mediate stress-induced impairment of memory retrieval. Molecular Psychiatry, 25(5), 977-992. doi:10.1038/s41380-019-0435-z Shippenberg, T. S. (2009). The dynorphin/kappa opioid receptor system: a new target for the treatment of addiction and affective disorders? Neuropsychopharmacology, 34(1), 247. doi:10.1038/npp.2008.165 Signorini, S., Liao, Y. J., Duncan, S. A., Jan, L. Y., & Stoffel, M. (1997). Normal cerebellar development but susceptibility to seizures in mice lacking G protein-coupled, inwardly rectifying K channel GIRK2. Proceedings of the National Academy of Sciences, 94(3), 923-927. doi:10.1073/pnas.94.3.923 Simon, A., Olah, S., Molnar, G., Szabadics, J., & Tamas, G. (2005). Gap-junctional coupling between neurogliaform cells and various interneuron types in the neocortex. J Neurosci, 25(27), 6278-6285. doi:10.1523/jneurosci.1431-05.2005 Simon, A., Oláh, S., Molnár, G., Szabadics, J., & Tamás, G. (2005). Gap-Junctional Coupling between Neurogliaform Cells and Various Interneuron Types in the Neocortex. The Journal of Neuroscience, 25(27), 6278-6285. doi:10.1523/jneurosci.1431-05.2005 Simon, E. J., Hiller, J. M., & Edelman, I. (1973). Stereospecific binding of the potent narcotic analgesic [3H] etorphine to rat-brain homogenate. Proceedings of the National Academy of Sciences, 70(7), 1947-1949. Simons, L., Pielech, M., Erpelding, N., Linnman, C., Moulton, E., Sava, S., . . . Berde, C. (2014). The responsive amygdala: treatment-induced alterations in functional connectivity in pediatric complex regional pain syndrome. Pain, 155(9), 1727-1742. Singer, W. (1993). Synchronization of cortical activity and its putative role in information processing and learning. Annual review of physiology, 55(1), 349-374. Sirota, A., Csicsvari, J., Buhl, D., & Buzsáki, G. (2003). Communication between neocortex and hippocampus during sleep in rodents. Proceedings of the National Academy of Sciences, 100(4), 2065-2069. Smart, D., & Lambert, D. G. (1996). delta-Opioids stimulate inositol 1,4,5-trisphosphate formation, and so mobilize Ca2+ from intracellular stores, in undifferentiated NG108-15 cells. J Neurochem, 66(4), 1462-1467. doi:10.1046/j.1471-4159.1996.66041462.x Smart, D., Smith, G., & Lambert, D. G. (1994). mu-Opioid receptor stimulation of inositol (1,4,5)trisphosphate formation via a pertussis toxin-sensitive G protein. J Neurochem, 62(3), 1009-1014. doi:10.1046/j.1471-4159.1994.62031009.x Smith, J. L., Mattick, R. P., Jamadar, S. D., & Iredale, J. M. (2014). Deficits in behavioural inhibition in substance abuse and addiction: A meta-analysis. Drug and alcohol dependence, 145, 1-33. doi:https://doi.org/10.1016/j.drugalcdep.2014.08.009

235

Sombati, S., & Delorenzo, R. J. (1995). Recurrent spontaneous seizure activity in hippocampal neuronal networks in culture. 73(4), 1706-1711. doi:10.1152/jn.1995.73.4.1706 Soyka, M., & Rosner, S. (2008). Opioid antagonists for pharmacological treatment of alcohol dependence-a critical review. Current drug abuse reviews, 1(3), 280-291. Spahn, V., Fischer, O., Endres-Becker, J., Schäfer, M., Stein, C., & Zöllner, C. (2013). Opioid withdrawal increases transient receptor potential vanilloid 1 activity in a protein kinase A-dependent manner. Pain, 154(4), 598-608. Spampinato, S., Bedini, A., & Baiula, M. (2013). Chapter 218 - Prodynorphin-Derived Peptides. In A. J. Kastin (Ed.), Handbook of Biologically Active Peptides (Second Edition) (pp. 1596-1601). Boston: Academic Press. Spitzer, N. C., Olson, E., & Gu, X. (1995). Spontaneous calcium transients regulate neuronal plasticity in developing neurons. Journal of neurobiology, 26(3), 316-324. Stein, C. (2016). Opioid Receptors. Annual Review of Medicine, 67(1), 433-451. doi:10.1146/annurev-med-062613-093100 Stein, C., & Machelska, H. (2011). Modulation of peripheral sensory neurons by the immune system: implications for pain therapy. Pharmacological Reviews, 63(4), 860-881. Stein, C., & Zöllner, C. (2009). Opioids and sensory nerves. In Sensory Nerves (pp. 495-518): Springer. Steriade, M. J. J. o. n. (2001). Impact of network activities on neuronal properties in corticothalamic systems. 86(1), 1-39. Storm, J. F. (1987). Action potential repolarization and a fast after-hyperpolarization in rat hippocampal pyramidal cells. J Physiol, 385, 733-759. doi:10.1113/jphysiol.1987.sp016517 Striedter, G. F. (2005). Principles of brain evolution. Sunderland, MA, US: Sinauer Associates. Stubbs, B., Thompson, T., Acaster, S., Vancampfort, D., Gaughran, F., & Correll, C. U. (2015). Decreased pain sensitivity among people with schizophrenia: a meta-analysis of experimental pain induction studies. Pain, 156(11), 2121-2131. Stumm, R. K., Zhou, C., Schulz, S., & Höllt, V. (2004). Neuronal types expressing μ- and δ- opioid receptor mRNA in the rat hippocampal formation. J Comp Neurol, 469(1), 107- 118. doi:10.1002/cne.10997 Suffczynski, P., Crone, N. E., & Franaszczuk, P. J. (2014). Afferent inputs to cortical fast-spiking interneurons organize pyramidal cell network oscillations at high-gamma frequencies (60–200 Hz). J Neurophysiol, 112(11), 3001-3011. Sugita, S., & North, R. A. (1993). Opioid actions on neurons of rat lateral amygdala in vitro. Brain Res, 612(1), 151-155. doi:https://doi.org/10.1016/0006-8993(93)91655-C Sullivan, M. D. (2018). Depression Effects on Long-term Prescription Opioid Use, Abuse, and Addiction. The Clinical Journal of Pain, 34(9). Suzuki, N., & Bekkers, J. M. (2010). Distinctive classes of GABAergic interneurons provide layer-specific phasic inhibition in the anterior piriform cortex. Cereb Cortex, 20(12), 2971-2984. doi:10.1093/cercor/bhq046 Suzuki, N., Tang, C., & Bekkers, J. (2014). Persistent barrage firing in cortical interneurons can be induced in vivo and may be important for the suppression of epileptiform activity. 8(76). doi:10.3389/fncel.2014.00076 Svoboda, K. R., Adams, C. E., & Lupica, C. R. (1999). Opioid receptor subtype expression defines morphologically distinct classes of hippocampal interneurons. J Neurosci, 19(1), 85-95. doi:10.1523/jneurosci.19-01-00085.1999 Svoboda, K. R., & Lupica, C. R. (1998). Opioid inhibition of hippocampal interneurons via modulation of potassium and hyperpolarization-activated cation (Ih) currents. Journal of Neuroscience, 18(18), 7084-7098. 236

Szűcs, E., Büki, A., Kékesi, G., Horváth, G., & Benyhe, S. (2016). Mu-Opioid (MOP) receptor mediated G-protein signaling is impaired in specific brain regions in a rat model of schizophrenia. Neuroscience Letters, 619, 29-33. doi:http://dx.doi.org/10.1016/j.neulet.2016.02.060 Takahashi, A. (2016). Subchapter 7A - Enkephalin. In Y. Takei, H. Ando, & K. Tsutsui (Eds.), Handbook of Hormones (pp. 55-e57A-52). San Diego: Academic Press. Taki, K., Kaneko, T., & Mizuno, N. (2000). A group of cortical interneurons expressing μ-opioid receptor-like immunoreactivity: a double immunofluorescence study in the rat cerebral cortex. Neuroscience, 98(2), 221-231. doi:http://dx.doi.org/10.1016/S0306- 4522(00)00124-X Tamás, G., Buhl, E. H., Lörincz, A., & Somogyi, P. (2000). Proximally targeted GABAergic synapses and gap junctions synchronize cortical interneurons. Nature Neuroscience, 3(4), 366-371. Tamás, G., Simon, A. L., Anna, & Szabadics, J. (2003). Identified Sources and Targets of Slow Inhibition in the Neocortex. Science, 299(5614), 1902-1905. doi:10.1126/science.1082053 Tanaka, E., & North, R. A. (1994). Opioid actions on rat anterior cingulate cortex neurons in vitro. The Journal of Neuroscience, 14(3), 1106. doi:10.1523/JNEUROSCI.14-03- 01106.1994 Tanaka, T., Saito, H., & Matsuki, N. (1996). Intracellular calcium oscillation in cultured rat hippocampal neurons: a model for glutamatergic neurotransmission. The Japanese Journal of Pharmacology, 70(1), 89-93. Tanowitz, M., Hislop, J. N., & von Zastrow, M. (2008). Alternative splicing determines the post- endocytic sorting fate of G-protein-coupled receptors. Journal of Biological Chemistry, 283(51), 35614-35621. Tanowitz, M., & von Zastrow, M. (2003). A novel endocytic recycling signal that distinguishes the membrane trafficking of naturally occurring opioid receptors. Journal of Biological Chemistry, 278(46), 45978-45986. Tejeda, H. A., Counotte, D. S., Oh, E., Ramamoorthy, S., Schultz-Kuszak, K. N., Bäckman, C. M., . . . Shippenberg, T. S. (2013). Prefrontal cortical kappa-opioid receptor modulation of local neurotransmission and conditioned place aversion. Neuropsychopharmacology, 38(9), 1770-1779. Tejeda, H. A., Hanks, A. N., Scott, L., Mejias-Aponte, C., Hughes, Z. A., & O'donnell, P. (2015). Prefrontal cortical kappa opioid receptors attenuate responses to amygdala inputs. Neuropsychopharmacology, 40(13), 2856-2864. Terasawa, E., Schanhofer, W. K., Keen, K. L., & Luchansky, L. (1999). Intracellular Ca Oscillations in Luteinizing Hormone-Releasing Hormone Neurons Derived from the Embryonic Olfactory Placode of the Rhesus Monkey. The Journal of Neuroscience, 19(14), 5898. doi:10.1523/JNEUROSCI.19-14-05898.1999 Terenius, L. (1973). Characteristics of the “receptor” for narcotic analgesics in synaptic plasma membrane fraction from rat brain. Acta pharmacologica et toxicologica, 33(5‐6), 377- 384. Tershner, S. A., & Helmstetter, F. J. (2000). Antinociception produced by mu opioid receptor activation in the amygdala is partly dependent on activation of mu opioid and neurotensin receptors in the ventral periaqueductal gray. Brain Res, 865(1), 17-26. doi:https://doi.org/10.1016/S0006-8993(00)02179-X Terskiy, A., Wannemacher, K. M., Yadav, P. N., Tsai, M., Tian, B., & Howells, R. D. (2007). Search of the human proteome for endomorphin-1 and endomorphin-2 precursor proteins. Life Sci, 81(23-24), 1593-1601. doi:10.1016/j.lfs.2007.09.025 237

Terwilliger, R. Z., Beitner-Johnson, D., Sevarino, K. A., Crain, S. M., & Nestler, E. J. (1991). A general role for adaptations in G-proteins and the cyclic AMP system in mediating the chronic actions of morphine and cocaine on neuronal function. Brain Res, 548(1-2), 100- 110. doi:10.1016/0006-8993(91)91111-d The Petilla Interneuron Nomenclature Group. (2008). Petilla terminology: nomenclature of features of GABAergic interneurons of the cerebral cortex. Nature reviews. Neuroscience, 9(7), 557-568. doi:10.1038/nrn2402 Toledo-Rodriguez, M., Goodman, P., Illic, M., Wu, C., & Markram, H. (2005). Neuropeptide and calcium-binding protein gene expression profiles predict neuronal anatomical type in the juvenile rat. The Journal of physiology, 567(Pt 2), 401-413. doi:10.1113/jphysiol.2005.089250 Torres-Reveron, A., Williams, T. J., Chapleau, J. D., Waters, E. M., McEwen, B. S., Drake, C. T., & Milner, T. A. (2009). Ovarian steroids alter mu opioid receptor trafficking in hippocampal parvalbumin GABAergic interneurons. Experimental Neurology, 219(1), 319-327. doi:https://doi.org/10.1016/j.expneurol.2009.06.001 Trafton, J., & Basbaum, A. (2004). [d-Ala2, N-MePhe4, Gly-ol5] enkephalin-induced internalization of the μ opioid receptor in the spinal cord of morphine tolerant rats. Neuroscience, 125(3), 541-543. Traub, R. D., Jefferys, J. G. R., & Whittington, M. A. (1997). Simulation of Gamma Rhythms in Networks of Interneurons and Pyramidal Cells. Journal of Computational Neuroscience, 4(2), 141-150. doi:10.1023/A:1008839312043 Tricoire, L., Pelkey, K. A., Daw, M. I., Sousa, V. H., Miyoshi, G., Jeffries, B., . . . McBain, C. J. (2010). Common origins of hippocampal Ivy and nitric oxide synthase expressing neurogliaform cells. J Neurosci, 30(6), 2165-2176. doi:10.1523/jneurosci.5123-09.2010 Trujillo, K. A., & Akil, H. (1991). Inhibition of morphine tolerance and dependence by the NMDA receptor antagonist MK-801. Science, 251(4989), 85-87. doi:10.1126/science.1824728 Tsu, R. C., Chan, J. S., & Wong, Y. H. (1995). Regulation of multiple effectors by the cloned delta-opioid receptor: stimulation of phospholipase C and type II adenylyl cyclase. J Neurochem, 64(6), 2700-2707. doi:10.1046/j.1471-4159.1995.64062700.x Unterwald, E. M., Rubenfeld, J. M., & Kreek, M. J. (1994). Repeated cocaine administration upregulates kappa and mu, but not delta, opioid receptors. Neuroreport, 5(13), 1613- 1616. doi:10.1097/00001756-199408150-00018 Valentino, R. J., & Volkow, N. D. (2018). Untangling the complexity of opioid receptor function. Neuropsychopharmacology, 43(13), 2514-2520. doi:10.1038/s41386-018-0225-3 Varela, F., Lachaux, J.-P., Rodriguez, E., & Martinerie, J. (2001). The brainweb: phase synchronization and large-scale integration. Nature Reviews Neuroscience, 2(4), 229-239. Vaughan, C. W., Ingram, S. L., Connor, M. A., & Christie, M. J. (1997). How opioids inhibit GABA-mediated neurotransmission. Nature, 390(6660), 611-614. Volk, D. W., Pierri, J. N., Fritschy, J.-M., Auh, S., Sampson, A. R., & Lewis, D. A. (2002). Reciprocal alterations in pre-and postsynaptic inhibitory markers at chandelier cell inputs to pyramidal neurons in schizophrenia. Cerebral Cortex, 12(10), 1063-1070. Volk, D. W., Radchenkova, P. V., Walker, E. M., Sengupta, E. J., & Lewis, D. A. (2012). Cortical opioid markers in schizophrenia and across postnatal development. Cerebral Cortex, 22(5), 1215-1223. Volkow, N. D., & Fowler, J. S. (2000). Addiction, a disease of compulsion and drive: involvement of the orbitofrontal cortex. Cerebral Cortex, 10(3), 318-325.

238 von Zastrow, M. (2010). Regulation of opioid receptors by endocytic membrane traffic: Mechanisms and translational implications. Drug and alcohol dependence, 108(3), 166- 171. doi:https://doi.org/10.1016/j.drugalcdep.2010.02.014 Wang, D., Sun, X., Bohn, L. M., & Sadée, W. (2005). Opioid Receptor Homo- and Heterodimerization in Living Cells by Quantitative Bioluminescence Resonance Energy Transfer. Mol Pharmacol, 67(6), 2173. doi:10.1124/mol.104.010272 Wang, J., Wang, Y., Guo, F., Feng, Z., Wang, X., & Lu, C. (2016). Nicotinic modulation of Ca2+ oscillations in rat cortical neurons in vitro. American Journal of Physiology-Cell Physiology, 310(9), C748-C754. Wang, W., Kim, H. J., Lv, P., Tempel, B., & Yamoah, E. N. (2013). Association of the Kv1 family of K+ channels and their functional blueprint in the properties of auditory neurons as revealed by genetic and functional analyses. J Neurophysiol, 110(8), 1751-1764. doi:10.1152/jn.00290.2013 Wang, X.-s., & Gruenstein, E. I. (1997). Mechanism of synchronized Ca2+ oscillations in cortical neurons. Brain Res, 767(2), 239-249. doi:https://doi.org/10.1016/S0006-8993(97)00585- 4 Wang, Y.-L., Wang, J.-G., Guo, F.-L., Gao, X.-H., Zhao, D.-D., Zhang, L., . . . Lu, C.-B. (2017). Selective dopamine receptor 4 activation mediates the hippocampal neuronal calcium response via IP3 and ryanodine receptors. Brain Res, 1670, 1-5. doi:https://doi.org/10.1016/j.brainres.2017.05.012 Wang, Y., Dye, C. A., Sohal, V., Long, J. E., Estrada, R. C., Roztocil, T., . . . Rubenstein, J. L. (2010). Dlx5 and Dlx6 regulate the development of parvalbumin-expressing cortical interneurons. J Neurosci, 30(15), 5334-5345. doi:10.1523/jneurosci.5963-09.2010 Wang, Y., Gupta, A., Toledo-Rodriguez, M., Wu, C. Z., & Markram, H. (2002). Anatomical, Physiological, Molecular and Circuit Properties of Nest Basket Cells in the Developing Somatosensory Cortex. Cerebral Cortex, 12(4), 395-410. Wang, Y., Toledo-Rodriguez, M., Gupta, A., Wu, C., Silberberg, G., Luo, J., & Markram, H. (2004). Anatomical, physiological and molecular properties of Martinotti cells in the somatosensory cortex of the juvenile rat. J Physiol, 561(Pt 1), 65-90. doi:10.1113/jphysiol.2004.073353 Wassum, K., Ostlund, S., Maidment, N., & Balleine, B. (2009). Distinct opioid circuits determine the palatability and the desirability of rewarding events. Proceedings of the National Academy of Sciences, 106(30), 12512-12517. Weiser, M., De Miera, E. V.-S., Kentros, C., Moreno, H., Franzen, L., Hillman, D., . . . Rudy, B. (1994). Differential expression of Shaw-related K+ channels in the rat central nervous system. Journal of Neuroscience, 14(3), 949-972. Westlund, K. N., & Willis, W. D. (2015). Chapter 25 - Pain System. In G. Paxinos (Ed.), The Rat Nervous System (Fourth Edition) (pp. 703-731). San Diego: Academic Press. Whistler, J. L., & Von Zastrow, M. (1998). Morphine-activated opioid receptors elude desensitization by β-arrestin. Proceedings of the National Academy of Sciences, 95(17), 9914-9919. Whistler, J. L., & Von Zastrow, M. (1999). Dissociation of functional roles of dynamin in receptor-mediated endocytosis and mitogenic signal transduction. Journal of Biological Chemistry, 274(35), 24575-24578. Whittington, M., Traub, R., Faulkner, H., Jefferys, J., & Chettiar, K. (1998). Morphine disrupts long-range synchrony of gamma oscillations in hippocampal slices. Proceedings of the National Academy of Sciences, 95(10), 5807-5811.

239

Whittington, M. A., Traub, R. D., & Jefferys, J. G. R. (1995). Synchronized oscillations in interneuron networks driven by metabotropic glutamate receptor activation. Nature, 373(6515), 612-615. doi:10.1038/373612a0 Wickman, K. D., Iñiguez-Lluhi, J. A., Davenport, P. A., Taussig, R., Krapivinsky, G. B., Linder, M. E., . . . Clapham, D. E. (1994). Recombinant G-protein βγ-subunits activate the muscarinic-gated atrial potassium channel. Nature, 368(6468), 255-257. Williams, J. T., Ingram, S. L., Henderson, G., Chavkin, C., von Zastrow, M., Schulz, S., . . . Christie, M. J. (2013). Regulation of μ-opioid receptors: desensitization, phosphorylation, internalization, and tolerance. Pharmacological Reviews, 65(1), 223-254. doi:10.1124/pr.112.005942 Wimpey, T. L., & Chavkin, C. (1991). Opioids activate both an inward rectifier and a novel voltage-gated potassium conductance in the hippocampal formation. Neuron, 6(2), 281- 289. Winters, B. L., Gregoriou, G. C., Kissiwaa, S. A., Wells, O. A., Medagoda, D. I., Hermes, S. M., . . . Bagley, E. E. (2017). Endogenous opioids regulate moment-to-moment neuronal communication and excitability. Nature Communications, 8, 14611-14611. doi:10.1038/ncomms14611 Wise, S. P., Murray, E. A., & Gerfen, C. R. (1996). The frontal cortex-basal ganglia system in primates. Critical Reviews in Neurobiology, 10(3-4). Witkowski, G., & Szulczyk, P. (2006). Opioid μ receptor activation inhibits sodium currents in prefrontal cortical neurons via a protein kinase A- and C-dependent mechanism. Brain Res, 1094(1), 92-106. doi:http://dx.doi.org/10.1016/j.brainres.2006.03.119 Wolf, R. C., Plichta, M. M., Sambataro, F., Fallgatter, A. J., Jacob, C., Lesch, K. P., . . . Grön, G. (2009). Regional brain activation changes and abnormal functional connectivity of the ventrolateral prefrontal cortex during working memory processing in adults with attention‐deficit/hyperactivity disorder. Human brain mapping, 30(7), 2252-2266. Wong, F. K., Bercsenyi, K., Sreenivasan, V., Portalés, A., Fernández-Otero, M., & Marín, O. (2018). Pyramidal cell regulation of interneuron survival sculpts cortical networks. Nature, 557(7707), 668-673. doi:10.1038/s41586-018-0139-6 Xie, C. W., & Lewis, D. V. (1997). Involvement of cAMP-dependent protein kinase in mu-opioid modulation of NMDA-mediated synaptic currents. J Neurophysiol, 78(2), 759-766. doi:10.1152/jn.1997.78.2.759 Xie, W., Samoriski, G. M., McLaughlin, J. P., Romoser, V. A., Smrcka, A., Hinkle, P. M., . . . Wu, D. (1999). Genetic alteration of phospholipase C β3 expression modulates behavioral and cellular responses to μ opioids. 96(18), 10385-10390. doi:10.1073/pnas.96.18.10385 %J Proceedings of the National Academy of Sciences Xie, Y.-f., Huo, F.-q., & Tang, J.-s. (2009). Cerebral cortex modulation of pain. Acta Pharmacologica Sinica, 30(1), 31-41. doi:10.1038/aps.2008.14 Xu, J., Xu, M., Rossi, G. C., Pasternak, G. W., & Pan, Y. X. (2011). Identification and characterization of seven new exon 11-associated splice variants of the rat μ opioid receptor gene, OPRM1. Mol Pain, 7, 9. doi:10.1186/1744-8069-7-9 Xu, Q., Cobos, I., De La Cruz, E., Rubenstein, J. L., & Anderson, S. A. (2004). Origins of Cortical Interneuron Subtypes. The Journal of Neuroscience, 24(11), 2612-2622. doi:10.1523/jneurosci.5667-03.2004 Yang, J.-M., Zhang, J., Chen, X.-J., Geng, H.-Y., Ye, M., Spitzer, N. C., . . . Li, X.-M. (2013). Development of GABA Circuitry of Fast-Spiking Basket Interneurons in the Medial Prefrontal Cortex of erbb4-Mutant Mice. The Journal of Neuroscience, 33(50), 19724- 19733. doi:10.1523/jneurosci.1584-13.2013

240

Yang, Y.-M., Arsenault, J., Bah, A., Krzeminski, M., Fekete, A., Chao, O. Y., . . . Wang, L.-Y. (2020). Identification of a molecular locus for normalizing dysregulated GABA release from interneurons in the Fragile X brain. Molecular Psychiatry, 25(9), 2017-2035. doi:10.1038/s41380-018-0240-0 Yang, Y.-M., & Wang, L.-Y. (2006). Amplitude and Kinetics of Action Potential-Evoked Ca2+ Current and Its Efficacy in Triggering Transmitter Release at the Developing Calyx of Held Synapse. Journal of Neuroscience, 26(21), 5698-5708. Yizhar, O., Fenno, L. E., Prigge, M., Schneider, F., Davidson, T. J., O’shea, D. J., . . . Paz, J. T. (2011). Neocortical excitation/inhibition balance in information processing and social dysfunction. Nature, 477(7363), 171-178. Yokota, E., Koyanagi, Y., Yamamoto, K., Oi, Y., Koshikawa, N., & Kobayashi, M. (2016). Opioid subtype- and cell-type-dependent regulation of inhibitory synaptic transmission in the rat insular cortex. Neuroscience, 339, 478-490. doi:https://doi.org/10.1016/j.neuroscience.2016.10.004 Yu, Y., Zhang, L., Yin, X., Sun, H., Uhl, G. R., & Wang, J. B. (1997). Mu opioid receptor phosphorylation, desensitization, and ligand efficacy. J Biol Chem, 272(46), 28869- 28874. doi:10.1074/jbc.272.46.28869 Yu, Y. J., Arttamangkul, S., Evans, C. J., Williams, J. T., & Von Zastrow, M. (2009). Neurokinin 1 receptors regulate morphine-induced endocytosis and desensitization of μ-opioid receptors in CNS neurons. Journal of Neuroscience, 29(1), 222-233. Zadina, J., Kastin, A., Ge, L., & Hackler, L. (1994). Mu, delta, and kappa opiate receptor binding of Tyr-MIF-1 and of Tyr-W-MIF-1, its active fragments, and two potent analogs. Life Sciences, 55(24), PL461-PL466. Zadina, J. E., Hackler, L., Ge, L. J., & Kastin, A. J. (1997). A potent and selective endogenous agonist for the mu-opiate receptor. Nature, 386(6624), 499-502. doi:10.1038/386499a0 Zamponi, G. W., & Snutch, T. P. (2002). Modulating modulation: crosstalk between regulatory pathways of presynaptic calcium channels. Mol Interv, 2(8), 476-478. doi:10.1124/mi.2.8.476 Zhang, J., Ferguson, S. S., Barak, L. S., Bodduluri, S. R., Laporte, S. A., Law, P.-Y., & Caron, M. G. (1998). Role for G protein-coupled receptor kinase in agonist-specific regulation of μ- opioid receptor responsiveness. Proceedings of the National Academy of Sciences, 95(12), 7157-7162. Zhang, J., Qian, H., Zhao, P., Hong, S.-S., & Xia, Y. (2006). Rapid hypoxia preconditioning protects cortical neurons from glutamate toxicity through δ-opioid receptor. Stroke, 37(4), 1094-1099. Zhang, Y., Ahmed, S., Neagu, G., Wang, Y., Li, Z., Wen, J., . . . Vreugdenhil, M. (2019). μ- Opioid receptor activation modulates CA3-to-CA1 gamma oscillation phase-coupling. IBRO Reports, 6, 122-131. doi:https://doi.org/10.1016/j.ibror.2019.01.004 Zheng, H., Patterson, L. M., & Berthoud, H.-R. (2007). Orexin Signaling in the Ventral Tegmental Area Is Required for High-Fat Appetite Induced by Opioid Stimulation of the Nucleus Accumbens. The Journal of Neuroscience, 27(41), 11075. doi:10.1523/JNEUROSCI.3542-07.2007 Zieglgansberger, W., French, E. D., Siggins, G. R., & Bloom, F. E. (1979). Opioid peptides may excite hippocampal pyramidal neurons by inhibiting adjacent inhibitory interneurons. Science, 205(4404), 415. doi:10.1126/science.451610 Zilbovicius, M., Garreau, B., Samson, Y., Remy, P., Barthelemy, C., Syrota, A., & Lelord, G. (1995). Delayed maturation of the frontal cortex in childhood autism. The American journal of psychiatry.

241

Zimprich, A., Simon, T., & Höllt, V. (1995). Transfected rat mu opioid receptors (rMOR1 and rMOR1B) stimulate phospholipase C and Ca2+ mobilization. Neuroreport, 7(1), 54-56. Zor, U., & Reiss, N. (1991). Regulation of Plasma Membrane Phospholipase A2 Activity by Phosphorylation/Dephosphorylation: Is Glucocorticoid Action Mediated by Induction of Protein Phosphatase? In J. M. Bailey (Ed.), Prostaglandins, Leukotrienes, Lipoxins, and PAF: Mechanism of Action, Molecular Biology, and Clinical Applications (pp. 135-139). Boston, MA: Springer US. Zsiros, V., & Maccaferri, G. (2005). Electrical coupling between interneurons with different excitable properties in the stratum lacunosum-moleculare of the juvenile CA1 rat hippocampus. J Neurosci, 25(38), 8686-8695. doi:10.1523/jneurosci.2810-05.2005 Zubieta, J.-K., Gorelick, D. A., Stauffer, R., Ravert, H. T., Dannals, R. F., & Frost, J. J. (1996). Increased mu opioid receptor binding detected by PET in cocaine–dependent men is associated with cocaine craving. Nature medicine, 2(11), 1225. Zubieta, J.-K., Smith, Y. R., Bueller, J. A., Xu, Y., Kilbourn, M. R., Jewett, D. M., . . . Stohler, C. S. (2001). Regional mu opioid receptor regulation of sensory and affective dimensions of pain. Science, 293(5528), 311-315. Zur Nieden, R., & Deitmer, J. W. (2005). The Role of Metabotropic Glutamate Receptors for the Generation of Calcium Oscillations in Rat Hippocampal Astrocytes In Situ. Cerebral Cortex, 16(5), 676-687. doi:10.1093/cercor/bhj013 %J Cerebral Cortex

242