Mark David Kvarta [email protected]

Synaptic Blues: Corticosteroids and AMPA Receptor Function in Depression and Antidepressant Action

Item Type dissertation

Authors Kvarta, Mark

Publication Date 2015

Abstract The pathophysiology underlying depression remains elusive, despite decades of research. Depression is most commonly treated, often unsatisfactorily, by manipulating synaptic serotonergic tone, yet the key substrates that mediate antidepressant effica...

Keywords chronic stress; synaptic strength; Anhedonia; Corticosterone; Depression; Metyrapone; Stress, Psychological

Download date 30/09/2021 10:06:03

Link to Item http://hdl.handle.net/10713/4853 Mark David Kvarta [email protected]

Degree and date to be conferred: Ph.D., 2015 University of Maryland School of Medicine Program in Neuroscience, Graduate Program in Life Sciences Education 2009-present University of Maryland School of Medicine, MSTP (M.D./Ph.D. Program) Graduate Program in Life Sciences, Program in Neuroscience University of Maryland - Baltimore, Baltimore, MD

2008 B.A. Biological Sciences Magna Cum Laude Cornell University, Ithaca, NY Concentration in Neurobiology and Behavior

Research and Professional Experience 2014 Teaching Assistant – “Proseminar in Experimental Design” GPLS737 Program in Neuroscience, University of Maryland Baltimore, Baltimore, MD 2011-2015 Graduate Research Assistant. Program in Neuroscience, UMB. Baltimore, MD 2008-2009 Research Assistant. Department of Neurobiology and Behavior, Harris-Warrick Lab, Cornell University. Ithaca, NY 2007-2008 Honors Thesis & Undergraduate Research. Department of Neurobiology and Behavior, Harris-Warrick Lab, Cornell University, Ithaca, NY 2007-2008 Teaching Assistant - Undergraduate. “Introduction to Neurophysiology” BIONB491 Department of Neurobiology and Behavior, Cornell University, Ithaca, NY 2007 Chemistry Teacher – High School. Hun School of Princeton Summer Program, Princeton, NJ

Honors and Awards 2015 Phi Kappa Phi 2014 Society for Neuroscience Graduate Student Travel Award 2014 Neurobiology of Stress Young Investigator Travel Award 2013 Ph.D. Scholar Award – UMB Graduate Program in Life Sciences 2013,2014 AAAS / Science Program for Excellence in Science 2012 Journal of Neurophysiology - Editor's Choice– Kvarta et al. 2012 2012 Program in Neuroscience - Highest Attendance Award 2010 Medical Student Research Day - 3rd Prize Presentation 2009 Medical Student Research Day - 2nd Prize Poster 2008 Faculty for Undergraduate Neuroscience Travel Award 2008 Magna Cum Laude Bachelors of Arts, Cornell University 2007-2008 Cornell University Biological Sciences Honors Program

Publications * indicates equal contribution Kvarta MD*, Sharma D*, Castellani RJ, Morales R, Reich SG, Shin R. Primary central nervous system lymphoma preceded by sentinel demyelination: a case study and review. [Submitted] Kvarta MD, Bradbrook KE, Dantrassy HM, Bailey AM, Thompson SM. Corticosterone mediates the synaptic effects of chronic stress at rat hippocampal temporoammonic synapses. [Provisional acceptance] Fischell J, LeGates TA, Kvarta MD, Van Dyke AM, Thompson SM. Rapid antidepressant action and restoration of excitatory synaptic strength after chronic stress by negative

modulators of alpha5-containing GABAA receptors. Neuropsychopharmacology. 2015 Apr 22 Thompson SM, Kallarackal AJ, Kvarta MD, Van Dyke AM, Cai X. An excitatory synapse hypothesis of depression (review). Trends in Neurosciences. 2015 May; 38(5):279- 294 Kallarackal AJ*, Kvarta MD*, Camaratta E, Jaberi L, Cai X, Bailey AM, Thompson SM. Selective stress-mediated decrease in AMPA receptor-mediated synaptic excitation at hippocampal temporoammonic-CA1 synapses in a rat model of depression. Journal of Neuroscience. 2013 Oct; 33(40):15669-15674. PMC3787493 Cai X, Kallarackal AJ, Kvarta MD, Goluskin S, Gaylor KE, Bailey AM, Lee HK, Huganir RL, Thompson SM. Local potentiation of excitatory synapses by serotonin and its dysregulation in rodent models of depression. Nature Neuroscience. 2013 Apr;16(4):464-72. PMC3609911 Kvarta MD, Harris-Warrick RM, Johnson BR. Neuromodulator-Evoked Synaptic Metaplasticity Within a Central Pattern Generator Network. Journal of Neurophysiology Nov 2012 108:2846-2856. PMC3545119 Johnson BR, Brown J, Kvarta MD, Lu JY, Schneider LR, Nadim F, Harris-Warrick RM. Differential Modulation of Synaptic Strength and Timing Regulate Synaptic Efficacy in a Motor Network. Journal of Neurophysiology Jan 2011 105:293-304. PMC3023374

Selected Abstracts Kvarta MD, Fischell J, Hesselgrave N, LeGates TA, Van Dyke AM, Thompson SM. From synapses to depression and back: Restoring plasticity of excitatory drive in cortico- mesolimbic synapses. 45th Annual Meeting of the Society for Neuroscience; 2015 Oct. Chicago, IL. [Submitted] Cassady M, Grimes K, Yin S, Turlington KW, Richert ME, Kvarta MD, Thompson SM, Hopp J. Epileptic seizures produce a transient improvement in mood in patients with epilepsy and depression. 69th Annual Meeting of the American Epilepsy Society; 2015 Dec. Philadelphia, PA. [Submitted] Turlington KW, Richert ME, Kvarta MD, Thompson SM, Hopp J. Epileptic seizures produce a transient improvement in mood in epileptic patients with depression. 45th Annual Meeting of the Society for Neuroscience; 2015 Oct. Chicago, IL. [Submitted]

Richert ME, Kvarta MD, Thompson SM, Hopp J. Evaluating Peri-ictal Mood Changes in Patients in an Epilepsy Monitoring Unit. 67th Annual Meeting of the American Academy of Neurology. 2015 April. Washington, D.C. Kvarta MD, Bradbrook KE, Dantrassy HM, Bailey AM and Thompson SM. Elevated corticosterone mediates the behavioral and neurobiological effects of chronic stress at rat hippocampal temporoammonic-CA1 synapses. 44th Annual Meeting of the Society for Neuroscience; 2014 Nov 15-19. Washington, D.C. Fischell J, LeGates TA, Kvarta MD, Van Dyke AM, Thompson SM. Rapid antidepressant action of an alpha5 selective benzodiazepine partial inverse agonist: restoration of excitatory synaptic strength. 44th Annual Meeting of the Society for Neuroscience; 2014 Nov 15-19. Washington, D.C. Fischell J, LeGates TA, Kvarta MD, Van Dyke AM, Thompson SM. Rapid antidepressant action of an alpha5 selective benzodiazepine partial inverse agonist: restoration of excitatory synaptic strength. 24th Neuropharmacology Conference; 2014 Nov 13-14. Arlington, VA Kvarta MD, Sharma D, Castellani RJ, Morales R, Reich SG, Shin R. Primary central nervous system lymphoma preceded by sentinel demyelination: a case study and hypotheses. Annual Meeting for the American Neurological Association, 2014 Oct 12- 14. Baltimore, MD. Abstract nr. S804 Richert ME, Thompson SM, Kvarta MD, Hopp J. Evaluating Peri-ictal Mood Changes in Patients in an Epilepsy Monitoring Unit. Summer Research Training Programs Student Research Forum; 2014 Jul 25. Baltimore, MD Kvarta MD and Thompson SM. Chronic elevation of corticosteroids: innocent bystander or villain in the genesis of depression? 29th Annual National MD/PhD Student Conference; 2014 Jul 18-20. Keystone, CO Kvarta MD and Thompson SM. Neurobiological correlates of depressive behavior at hippocampal temporoammonic – CA1 synapses are mediated by elevated corticosterone. 4th Biennial Neurobiology of Stress Workshop, 2014. University of Cincinatti; 2014 Jun 17-20. Cincinatti, Ohio Kvarta MD, Bradbrook KE, Dantrassy HM, Bailey AM and Thompson SM. Elevated corticosterone mediates the behavioral and neurobiological effects of chronic stress at rat hippocampal temporoammonic-CA1 synapses. 17th Annual Program in Neuroscience Retreat - UMB; 2014 Jun 6. Baltimore, MD Kvarta MD and Thompson SM. Attenuating corticosterone synthesis prevents chronic stress- induced behavioral and neurobiological changes at temporoammonic – CA1 synapses in the rat hippocampus. Greater Baltimore Chapter of the Society for Neuroscience Annual Meeting; 2013 Nov 20; Baltimore, MD. Abstract nr. 53 Kvarta MD and Thompson SM. Attenuating corticosterone synthesis prevents chronic stress- induced behavioral and neurobiological changes at temporoammonic – CA1 synapses in the rat hippocampus. 43rd Annual Meeting of the Society for Neuroscience; 2013 Nov 9- 13; San Diego, CA. Abstract nr. 814.13 Kvarta MD and Thompson SM. Elevated corticosterone is sufficient to trigger the neurobiological correlates of chronic stress at temporoammonic - CA1 cell synapses in the rat hippocampus. 16th Annual Program in Neuroscience Retreat – University of Maryland Baltimore; 2013 June 6; Baltimore, MD. Abstract nr. 001

Kvarta MD and Thompson SM. Elevated corticosterone is sufficient to trigger the neurobiological correlates of chronic stress at temporoammonic - CA1 cell synapses in the rat hippocampus. Greater Baltimore Chapter of the Society for Neuroscience Annual Meeting; 2012 Nov 16; Baltimore, MD; Abstract nr. P72 Kvarta MD and Thompson SM. Elevated corticosterone is sufficient to trigger the neurobiological correlates of chronic stress at temporoammonic - CA1 cell synapses in the rat hippocampus. 42nd Annual Meeting of the Society for Neuroscience; 2012 Oct 13- 17; New Orleans, LA. SfN; Abstract nr 775.20/FF2 Kvarta MD, Johnson BR, Harris-Warrick RM. Synaptic depression and recovery dynamics: neuromodulator-evoked metaplasticity within a central pattern generator network. 39th Annual Meeting of the Society for Neuroscience; 2009 Oct 17-21; Chicago, IL. SfN; 2009 Abstract nr. 270.7/CC20 Kvarta MD, Johnson BR, Harris-Warrick RM. Synaptic depression and neuromodulator- evoked metaplasticity within a central pattern generator network. 38th Annual Meeting of the Society for Neuroscience; 2008 Nov 15-19; Washington, DC. SfN; 2008 Abstract nr. 574.13/OO12

Selected Presentations Kvarta MD and Thompson SM (2014) Chronic elevation of corticosteroids: innocent bystander or villain in the genesis of depression? 29th Annual National MD/PhD Student Conference. Keystone, CO Kvarta MD (2014) Neurobiological correlates of depressive behavior at hippocampal temporoammonic – CA1 synapses are mediated by elevated corticosterone. Young Investigator Travel Awardee Data Blitz, 4th Biennial Neurobiology of Stress Workshop. University of Cincinatti, Cincinatti, OH Kvarta MD (2014) “Is corticosterone sufficient and necessary to the behavioral and synaptic effects of chronic stress in the hippocampus?” Membrane Biology Student Seminar Series, UMB. Baltimore, MD Kvarta MD, Li TP, Wilson K (2013) “Graduate school survival guide: Choosing a mentor, navigating the program & the grant writing process” UMSOM 5th annual MSTP retreat. Baltimore, MD Kvarta MD (2012) “The role of glucocorticoids in the neurobiological sequelae of chronic stress at temporoammonic-CA1 synapses in the rat hippocampus”. UMB Membrane Biology Program Student Seminar Series. Baltimore, MD Kvarta MD (2012) “Do glucocorticoids mediate the neurobiological effects of chronic stress on the hippocampus?” UMSOM 4th annual MSTP retreat. Baltimore, MD Kvarta MD, Wied H. (2012) “The fellowship grant writing process”. UMSOM 4th annual MSTP retreat. Baltimore, MD Kvarta MD, Wittenberg GF. (2011) “Transcranial Magnetic Stimulation of Primary Motor Cortex and Neuroplasticity of Practiced Reaching Movements in a Robotic Rehabilitation Environment”. UMSOM MSTP Summer Research Symposium. Baltimore, MD Kvarta MD, Li TP (2011) “MS1 Survival Guide”. UMSOM 3rd Annual MSTP Retreat. Baltimore, MD Kvarta MD, Thompson SM (2010) "Corticosterone: Necessary and Sufficient Mediator of the "Chronic Unpredictable Stress" Model of Clinical Depression". UMSOM 33rd Annual Medical Student Research Day. Baltimore, MD

Kvarta MD, Thompson SM (2010) "Relating the rat corticosterone feedback loop via the HPA axis to the “Chronic Unpredictable Stress” model of clinical depression" UMSOM MSTP Summer Research Symposium. Baltimore, MD Kvarta MD, Johnson BR and Harris-Warrick RM (2009), Synaptic Depression and Recovery Dynamics: Neuromodulator-Evoked Metaplasticity within a Central Pattern Generator Network. UMSOM 32nd Annual Medical Student Research Day. Baltimore, MD

University Service & Committee memberships 2013 Clinical Case Studies workgroup MSTP 2012 MSTP Transition Committee 2013 Stephen Max Memorial Lecture Selection Committee 2011-2013 Student Interviewer & MSTP Admissions Committee 2013 MSTP Student Representative, University of Maryland School of Medicine Council 2011 MSTP Student Council Representative to the Annual University of Maryland School of Medicine MSTP Advisory Committee 2011 Planning committee member: MSTP “Clinical Case Studies” course for MD/PhD students in their graduate research years, now a required part of the curriculum 2009-2012 MSTP Student Council Member

Conferences & Professional Events Attended 2014 44th Annual Meeting of the Society for Neuroscience, Washington D.C. 2014 24th Neuropharmacology: GABAergic Signaling in Health and Disease. Arlington, VA 2014 Annual Meeting for the American Neurological Association. Baltimore, MD. 2014 29th National MD/PhD Student Conference. Keystone, CO 2014 Combining Clinical and Research Careers in Neuroscience. Association of University Professors of Neurology, National Institute of Neurological Disorders and Stroke, American Neurological Association, Child Neurology Society. Washington D.C. 2014 4th Biennial Neurobiology of Stress Workshop. Cincinatti, OH 2013 Greater Baltimore Chapter of the Society for Neuroscience Annual Meeting. Baltimore, MD 2013 43rd Annual Meeting of the Society for Neuroscience. San Diego, CA 2013 New York Academy of Sciences: “Treatment-Resistant Depression: Glutamate, Stress-Hormones and their Role in the Regeneration of Neurons”. New York, NY 2012 42nd Annual Meeting of the Society for Neuroscience. New Orleans, LA 2012 3rd Biennial Neurobiology of Stress Workshop. University of Pennsylvania; Philadelphia, PA 2011 41st Annual Meeting of the Society for Neuroscience. Washington D.C. 2009 39th Annual Meeting of the Society for Neuroscience. Chicago, IL 2008 38th Annual Meeting of the Society for Neuroscience. Washington D.C.

Professional Society Memberships 2014-2015 American Academy of Neurology (#234225) 2013-present American Association for the Advancement of Science (#20310116) 2012-2014 American Medical Association 2008-present Society for Neuroscience (#210252712)

Student Club Memberships 2009-2011 Student Interest Group in Neurology 2009-2011 Psychiatry Interest Group 2009-2011 Physical Medicine & Rehabilitation Interest Group 2009-2011 Emergency Medicine Interest Group 2010-2011 Anesthesiology Interest Group

Service to Community, Volunteer Work & Mentorship 2012-2013 Small Miracles Cat & Dog Rescue. Volunteer. Ellicott City MD. 2012-2013 MSTP Student Mentorship Program 2012-2013; 2014-2015 UMB Graduate Program in Life Sciences Big Sibling Program. 2012-2014 Mentored three high school students who shadowed or worked in the lab and are interested in scientific research 2014-2015 Mentored medical students in designing and conducting a clinical research project in the UMMC Epilepsy Monitoring Unit, successfully applying for funding, and presenting data at national and international conferences (See abstracts by Turlington, Richert, Cassady et al.) 2010-2011 Meals on Wheels of Central Maryland. Columbia, MD 2009-2010 A Bridge to Academic Excellence. Tutor. University of Maryland School of Pharmacy. Baltimore, MD

Abstract

Title of Dissertation: “Synaptic Blues: Corticosteroids and AMPA Receptor Function in

Depression and Antidepressant Action”

Mark Kvarta, Doctor of Philosophy, 2015

Dissertation Directed by: Scott Thompson, Chair and Professor, Department of Physiology

The pathophysiology underlying depression remains elusive, despite decades of research.

Depression is most commonly treated, often unsatisfactorily, by manipulating synaptic serotonergic tone, yet the key substrates that mediate antidepressant efficacy are unknown.

Chronic stress-based models recapitulate a depressive-like phenotype, but it is unclear how stress leads to specific neurobiological changes to induce a depressive-like neurobehavioral state.

I used electrophysiology, molecular biology, and behavioral assays in rats and mice to address the inextricably linked questions “how do we become depressed?” and “how do antidepressants work?” I first contributed to the discovery that 5HT1BR-mediated increase in

AMPA receptor function, via CaMKII phosphorylation of the S831 residue of GluA1, is a necessary action of serotonergic antidepressants, and contributed to the identification of hippocampal temporoammonic(TA)-CA1 synapses as a key site of dysregulation and serotonergic action in depression. Using the chronic unpredictable stress (CUS) model, I demonstrated that the

TA-CA1 deficits caused by chronic stress are mediated by decreased synaptic GluA1 expression and AMPA receptor strength specific to distal apical dendrites. Decreased AMPA:NMDA stoichiometry at TA-CA1 synapses correlated with decreased sucrose preference scores, a model of anhedonia. In asking how CUS causes these synaptic and behavioral deficits, I found that chronic elevations in corticosterone(CORT) are both sufficient and necessary by administering exogenous CORT in lieu of CUS and by administering metyrapone (a corticosteroid synthesis inhibitor) during CUS, respectively.

My results support a hypothesis that depression results from synaptic deficits at key excitatory synapses in cortico-mesolimbic circuitry, that these deficits are induced by chronic stress via corticosterone, and that antidepressant efficacy is mediated by reversing these synaptic deficits. This hypothesis suggests a novel antidepressant strategy by which other compounds or interventions can rapidly restore affective state by reversing chronic-stress- and corticosterone- induced deficits in excitatory synapses. Consistent with this hypothesis, I also participated in a study showing that negative modulators of GABAA receptor function act as rapid antidepressants and restore AMPAR-mediated excitatory strength at TA-CA1 synapses.

Synaptic Blues: Corticosteroids and AMPA Receptor Function in Depression and Antidepressant Action

By Mark David Kvarta

Dissertation submitted to the Faculty of the Graduate School of the University of Maryland Baltimore in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2015

Dedication

To those who just haven’t yet been saved

iii

Acknowledgments

I would not be here today without my parents, and would not be who I am without my family. My Mom and Dad have repeatedly given everything they have for me and my siblings. Through example, they have taught me perseverance, resilience, and the importance of happiness. They have always been my biggest fans and have always ensured I have the resources I need to succeed and find happiness. There are many times my trajectory was in danger of straying off course, especially as an adolescent, and they have always fought for me and succeeded through love and sheer force of will. All four of my siblings have been major influences on me. Growing up without my two brothers to look up to, Andrew and Daniel, would just not have been the same. They have always been my heroes. My sisters, Jess and Patty, instilled in me the need to think outside myself and to approach the world with an open heart and mind.

I have had many meaningful teachers and mentors along the way in high school and beyond. However, I have two people, specifically, to blame for infecting me with the research bug: Drs. Bruce Johnson and Ronald Harris-Warrick. Their enthusiasm, mentorship, and support changed the way I thought about the world of biology and my own place in it while I already had one foot in the door towards a career goal in medicine. During my time in Bruce’s course, first as a student then a teaching assistant, I fell in love with synaptic physiology—wire-tapping living signals!—and I have never looked back. While working in Ron’s laboratory, my focus shifted from one of exhaustive accumulation of existing knowledge to one of exploration and discovery. Together, they instilled in me the idea that a novel experimental result, no matter how seemingly small, is a pleasure and a privilege to observe.

The University of Maryland MD/PhD program has been and continues to be a critical part of my personal and professional development. If not for Dr. Terry Rogers and Ms. Nancy Malson, there are no guarantees I would have found my way to be here. Terry has always been a tremendous and encouraging mentor, and continues to care about all of his recruits, despite himself “graduating” from the MSTP. I would also like to acknowledge Jane Bacon, without whom I would have quite certainly gone mad attempting to navigate the vagaries of transitioning between medical school, graduate school, and now back again. I will rue the day that Jane realizes she could make a fortune charging for therapy sessions. Finally, I thank Drs. Michael Donnenberg and Achsah Keegan for taking the reins of the program and continuing to foster a training environment that always keeps the best interests of the students first and foremost.

Ms. Jenn Aumiller and Dr. Renee Cockerham have been instrumental in maximizing the amount of time as a student that I could actually be a student. At least each semester, it seems, I have had some kind of unique problem that required their assistance, and I am sure they are happy to see me move beyond their jurisdiction. I also thank Program in Neuroscience directors Drs. Bruce Krueger and Jessica Mong for their input and guidance when needed, and for maintaining PIN as the best program on campus.

No good science emerges a vacuum, and I would like to thank Drs. Thomas Blanpied, Mary Kay Lobo, Sandra Jurado, Todd Gould, Brian Mathur, and their laboratories for all of their insight into my own projects over my tenure here, as well as for broadening my own view of the synapses and circuits we are all interested in. Todd and Tom deserve an extra special thank you for agreeing to be readers for this document. If anything within offends you, please go confront them.

My entire thesis committee has been invaluable for my journey as a graduate student. Thank you very sincerely to Drs. George Wittenberg, Tom Blanpied, Todd Gould, and Brad Alger. George epitomizes the physician scientist I wish to emulate, and he very nearly snatched me from the clutches of basic science to join his lab. Todd has also greatly influenced my career goals and perspective, although he probably does not realize it. If there is one person that could stand in for all of the faculty in a training program, it is Brad Alger; I cannot thank him enough for instilling in me the expectation for only the most rigorous scientific thinking. Brad has a unique way of impressing quite profound questions into a challenge of basic truths; he has made me think deeply with a quick quip while passing by during a coffee break more times than I could count. Tom is perhaps the most humbly brilliant person I have ever met, along with his trainee and my good friend Peter Li; they are both very special people. Getting to interact with both of them on a regular basis has truly been a gift.

They say that the moment your mentor owes you more than you owe him or her is the moment you are ready to move on. Unfortunately, that moment will never come for me. Dr. Scott Thompson is an utterly phenomenal mentor. He has a unique talent for constantly challenging his own and others’ ideas and biases in ever more constructive ways. Scott is capable of generating inspiriation, curiousity, and a sense of belonging in anyone. He truly cares about and supports all of his colleagues and friends. And he has a LOT of friends. In four years working together, I cannot recall a single pessimistic comment. Because of how he pours all of his energy—all of his heart—into his trainees, I would drop anything for him in return. Someday I aspire to be as outstanding a mentor, but I know it’s not possible. He is one of a kind.

Scott also has a knack for attracting exceptional people with whom to work. I owe Drs. Angy Kallarackal, Stephanie Cerceo Page, and Stephanie Parrish Aungst thanks for making a continuing effort to make this lab an excellent training environment for an annoying new graduate student. I owe Angy extra thanks for ensuring I had the skills to continue on in her stead and entrusting me to carry the torch onward. I would also like to thank our former technicians Leelah Jaberi and Kaitlin Gaylor for helping things run smoothly, and always with a smile.

I used to think soon-to-be-Dr. Adam Van Dyke and I had too much in common, and could never be friends. The first part is certainly untrue. As my perennial partner in general shenanigans, I want to extend my thanks to Adam for making our lab a hilarious and exciting place to work. I could not ask for a better person with whom to share a foxhole, especially when that foxhole is a lab or a rental car, and there may be free drinks afterwards. I used to wonder whether he was a nice guy who pretended to be crass, or whether he was an insensitive guy who pretended to be thoughtful. I never came up with an answer, but at the very least he has not murdered me despite ample opportunity.

Finally, my acknowledgements are nowhere near completed until I thank one of the most important people in my life, Dr. Tara LeGates. Tara embodies everything the modern scientist should be: brilliant, creative, humble, dedicated, and supportive. I’m thrilled to have been a tiny part of her career, as it is sure to bring me accolades by association. Science cred aside, she has been my rock and my best friend through thick and thin. Tara inspires me to be the best scientist, the best friend, the best person I can be. Her smile and the laughter of my best little buddy Connor can melt a scowl off my face like nothing else. She is more than I could have ever asked for. I hope she keeps me around to spend many, many years together.

Table of Contents

1) Introduction ...... 1

2) General methods ...... 42

3) 5HT1BR-mediated plasticity of TA-CA1 synapses ...... 48

4) Chronic stress selectively decreases AMPAR-mediated excitation at temporoammonic-CA1 synapses ...... 69

5) Corticosterone mediates the synaptic and behavioral effects of chronic stress at TA-CA1 synapses ...... 87

6) Discussion: towards an excitatory synapse hypothesis of depression ...... 120

7) Appendix A: Protocols & recipes ...... 130

8) Appendix B: TA-CA1 plasticity – role of NMDARs and 5HT1BRs ...... 138

9) Literature cited ...... 142

List of Figures

Figure 1.1 - Animal models of depression ...... 21

Figure 1.2: Corticomesolimbic circuitry relevant for depression ...... 27

Figure 1.3 - Simplified Hippocampal Circuit ...... 33

Figure 2.1 – Transverse hippocampal slice preparation and electrode setup ...... 43

Figure 3.1 - Potentiation of TA-CA1 fEPSPs is irreversible in response to endogenous serotonin elevation, but transient following selective 5HT1BR activation ...... 53

Figure 3.2- 5HT1BR antagonist isamoltane reverses persistent potentiation of TA-CA1 fEPSPs by imipramine...... 54

Figure 3.3 – Effect of anpirtoline, fluoxetine and wash on phosphorylation of S831 site of GluA1 and T286 site of CaMKII ...... 56

Figure 3.4 – Anpirtoline increases CaMKII T286 and GluA1 S831 phosphorylation in NAc core ...... 57

Figure 3.5 – Anpirtoline has no effect on SC-CA1 fEPSPs...... 58

Figure 3.6 – Anpirtoline-induced potentiation of TA-CA1 fEPSPs is occluded by HFS-induced LTP ...... 59

Figure 3.7 - HFS-induced LTP at TA-CA1 synapses is occluded by anpirtoline ...... 60

Figure 3.8 – Effect of chronic stress and chronic fluoxetine anpirtoline-induced modulation of TA-CA1 fEPSPs ...... 63

Figure 4.1 – CUS reduces AMPA:NMDA transmission at TA-CA1 synapses, which is reversed by chronic fluoxetine treatment ...... 74

Figure 4.2 - CUS reduces GluA1 protein expression in SLM of chronically stressed rats exhibiting anhedonia, reversed by chronic fluoxetine treatment ...... 75

Figure 4.3 - Western blots of SLM tissue from CUS rats reveal decrease of PSD-95 expression, but no effect on NR1 or GluA2 ...... 76

Figure 4.4 - Layer-specific downregulation of GluA1 protein following CUS...... 77

Figure 4.5 - Layer-specific loss of AMPAR-mediated excitatory strength following CUS ...... 78

Figure 4.6 - Correlation between SLM, but not SR, A:N ratio and SPT scores ...... 79

Figure 4.7 - CUS does not alter ability of TA-CA1 to undergo LTP ...... 80

vii

Figure 5.1 – Experimental timeline for chronic CORT experiments ...... 95

Figure 5.2 - Loss of sucrose preference by chronic exogenous CORT ...... 95

Figure 5.3 - Effect of chronic CORT on NSF test ...... 96

Figure 5.4 - Chronic CORT decreases AMPAR-mediated excitation at TA-CA1 ...... 97

Figure 5.5 - Chronic CORT decreases GluA1 expression in SLM and GR expression in SP ...... 98

Figure 5.6 - Chronic CORT treatment decreases 5HT1BR expression in SLM ...... 99

Figure 5.7 - Acute CORT is not sufficient to decrease TA-CA1 AMPA:NMDA...... 100

Figure 5.8 - Chronic CORT alters 5HT1BR mediated potentiation of TA-CA1 fEPSPs ...... 102

Figure 5.9 - Effect of acute in vivo CORT exposure on in vitro anpirtoline response at TA-CA1 synapses ...... 103

Figure 5.10 - Experimental timeline for MET+CUS experiments ...... 104

Figure 5.11 - MET significantly attenuates CORT increases in response to CUS ...... 105

Figure 5.12 - MET prevents induction of anhedonia in SPT caused by CUS ...... 107

Figure 5.13 - MET treatment prevents CUS-induced neohypophagia ...... 108

Figure 5.14 - MET prevents CUS-induced decreases in AMPAR-mediated TA-CA1 signaling 109

Figure 5.15 - MET prevents CUS-induced decreases in GluA1 GR e ...... 110

Figure 5.16 - MET treatment prevents stress-induced changes in the potentiation of TA-CA1 fEPSPs by the 5-HT1BR agonist anpirtoline ...... 111

Figure 5.17 - MET treatment prevents stress-induced impairment of spatial memory consolidation ...... 113

Figure 6.1 - Rapid, persistent antidepressant effect of negative allosteric modulation of alpha-5 containing GABAARs ...... 128

Figure 6.2 - Negative allosteric modulation of alpha5-containing GABAARs rapidly reverses AMPAR strength deficit at TA-CA1 synapses after chronic stress ...... 129

Figure 8.1 - LTP of TA-CA1 fEPSPS is NMDAR-dependent ...... 139

Figure 8.2 - TA-CA1 LTP does not depend on 5HT1BR signaling ...... 140

Figure 8.3 - 5HT1BR antagonism does not result in submaximal LTP ...... 141

viii

1) Introduction “The pain is unrelenting, and what makes the condition intolerable is the foreknowledge that no remedy will come, not in a day, an hour, a month, or a minute. It is hopelessness even more than pain that crushes the soul.” –William Styron, Darkness Visible; A Memoir of Madness

Profound sadness. Pervasive lack of pleasure. These core features characterize depression, a devastating neuropsychiatric disease which will afflict over 50 million of the current

US population at some point in their life (Kessler et al., 2005). Depression is a leading risk factor for the estimated one million suicide deaths per year worldwide (World Health Organization,

2012). The burden of depression is rising and is projected to become the second worldwide leading cause of disability by 2020 (Murray and Lopez, 1997); it already has the greatest impact of all biomedical diseases on disability in the United States (Flint and Kendler, 2014).

Worldwide, depression is responsible for almost 12% of total years lived with disability (Ustün et al., 2004).

One undeniable contribution to the enormous individual and socioeconomic cost that depression exacts is the inadequacy of its treatment, medical and otherwise. Current medications for depression are effective in only a subset of patients and act slowly, complicating treatment

(Nestler et al., 2002). In the largest prospective clinical trial of depression to date, in which the effect of first-line treatments in practical settings was examined, only 47% of the nearly 3000 depressed patients examined responded to a single course of treatment with citalopram, a selective-serotonin reuptake inhibitor (SSRI) (Gaynes and Warden, 2009). Only one third of patients achieved remission. In that subset of patients that eventually exhibited remission, average time to remission was 47 days. Furthermore, fewer than half (in some countries, less than 10%) of those affected by depression worldwide receive treatment (World Health Organization, 2012), perhaps in part because of the cost and sustained effort of continuous dosing required by the latency to improvement and the well-known ineffectiveness of SSRI treatment in many patients.

A better understanding of the neurobiological basis of depression is desperately needed to

1 develop more effective and faster acting therapies. In this thesis, I pursue this understanding by primarily focusing on better understanding the nature of the changes that occur in the brain that underlie the symptoms of depression and on how those changes are triggered.

The first clue came from an effort to answer the question “How do antidepressants work?” I and others in the lab discovered that a key action of antidepressant action is to potentiate synapses via serotonin receptor subtype 1B signaling that results in increased AMPA-type glutamate receptor phosphorylation at a specific serine residue, S831 (Chapter 3). This led us to ask the next question, “Why would potentiation of synapses contribute to the efficacy of antidepressants?” I used a chronic stress model of depression to examine first what happens to temporoammonic-CA1 synapses following chronic stress in a depressive-like phenotype (Chapter

4). I demonstrated that this population of excitatory synapses in the hippocampus is specifically weakened in a depressive-like phenotype, driven by a loss of AMPAR strength and GluA1 subunit expression. That is, antidepressants need to potentiate synapses because they are weak in a depressive-like state. I then asked the question “How does chronic stress cause these synapses to become weaker?” I demonstrated that corticosterone, the predominant glucocorticoid in rodents, is necessary and sufficient for chronic stress to translate into a depressive-like synaptic and behavioral phenotype (Chapter 5). Ultimately, my results contribute to a synthesis suggesting that signs and symptoms of depression are caused by weakening of key excitatory synapses and that this weakening is rescued as a key action of antidepressant action. I and others demonstrate this rescue as a shared effect of different antidepressant approaches (Chapter 6). I conclude that protecting these synapses from the effects of chronic stress and corticosteroids, and strengthening these synapses to reverse dysfunction, are critical to maintaining and restoring normal affect.

These thus present appealing targets for therapeutic intervention.

Major Depressive Disorder

Symptoms & Diagnosis

Also known as Major Depressive Disorder (MDD) or Major Depression, depression is a disorder of persistently depressed mood that interferes with daily social or occupational activities, and, when severe, can lead ultimately to a feeling that life is not worth living. The criteria for diagnosis of depression according to the Diagnostic and Statistical Manual of Mental Disorders,

Fifth Edition (DSM-5) include five or more symptoms that represent a change from previous functioning, at least one of which is either depressed mood or loss of interest or pleasure

(anhedonia). Depressed mood and anhedonia are thus considered core symptoms. Symptoms are present every day or nearly every day over at least a 2-week period and may also include: significant weight loss or weight gain or change in appetite; insomnia or hypersomnia; psychomotor agitation or retardation; fatigue or loss of energy; feelings of worthlessness or inappropriate or delusional guilt; diminished ability to think or concentrate or indecisiveness; recurrent thoughts of death, suicidal ideation or suicide attempt.

As with many mood disorders, there are no known biological hallmarks for depression, and therefore physicians must rely on the presentation of symptoms for diagnosis. Several questionnaires and scales have been created, and periodically updated, to aid in screening and diagnosis for depression, including the Hamilton Depression Rating Scale (Hamilton, 1960) the

Beck Depression Inventory (Beck et al., 1961), and the Montgomery-Åsberg Depression Rating

Scale (Montgomery and Åsberg, 1979). By standardizing diagnosis and providing quantitative measures of disease severity, these scales have been useful in screening potential therapies.

It should be noted that patients who are depressed do not comprise a homogenous population of symptomatology. As noted above, either one of the two core symptoms –depressed mood or anhedonia—are required, along with several others for a total of 5 out of 9 listed symptoms (in the DSM). In the International Statistical Classification of Diseases and Related

Health Problems (ICD-10), two of the three core symptoms (depressed mood, anhedonia, and increased fatigability) are required in addition to two additional symptoms for the duration of two weeks to classify as a depressive episode. Clearly, there are hundreds of combinations of symptoms that would lead to diagnosis in either statistical manual. It seems most likely that a single pathology is not responsible for such a broad range of combinations of symptoms. Efforts have been made to further sub-classify clinical subtypes of depression (e.g. (Sharpley and Bitsika,

2014) depressed mood, anhedonic depression, somatic depression and cognitive depression), echoing the criticism of a unitary diagnosis of what is most likely a collection of varying neuropathology, perhaps reflecting differences in the underlying brain circuits affected.

Clinically, a limited distinction is typically made between endogenous depression, characterized by melancholic features including anhedonia and lack of mood reactivity, and atypical depression which is generally characterized by depressed mood with intact mood reactivity (i.e. positive change in affect in response to positive events), however these distinctions often bear little weight in clinical research design and treatment approaches. Poorly recognized depression syndromes may also contribute to the heterogeneity in live-imaging studies and post-mortem tissue studies.

Depression is a heterogeneous condition, but we are able to examine the neurobiological substrates that underlie specific sets of symptoms that transcend traditional clinical statistical- based diagnoses, just as the dysfunctional circuitry underlying specific symptoms can be altered in more than one mood or neurological disorder.

Demographics & Risk Factors

Epidemiologists have identified many factors that put individuals at risk of depression with the hope of providing insight into the pathophysiology of the disease. The overall lifetime prevalence of depression is estimated to be 16.9% in the United States, with studies consistently showing a higher prevalence in women than men (Weissman et al., 1996; Flint and Kendler,

2014). However, an emerging body of literature suggests that men may experience similar rates

4 of depression, but the phenotype is masked by the fact that traditional depressive symptoms are at odds with gender and cultural norms, and can be revealed when taking into account potential alternative symptoms of depression. In a large co-morbidity survey, sex differences in prevalence were eliminated when allowing for the alternative symptoms of irritability, aggression, substance abuse, and risk taking (Martin et al., 2013). Other risk factors include emotional trauma, neurotropic viral infections (e.g. Borna disease), chronic medical conditions, stochastic processes during development, and shift work or altered light environment (Fava and Kendler, 2000;

Nestler et al., 2002; LeGates et al., 2014).

Depression and depressive symptoms are highly heritable. Having a first degree relative with depression imparts a 2.84-fold lifetime risk (Sullivan et al., 2000). Age of onset and likelihood of recurrence after relapse are most heritable (Levinson, 2006). Overall heritability is estimated to be about 38%, with most interest revolving around genetic mechanism. Indeed, many potential candidate genes involving the central nervous system (CNS) have been identified in large genetic studies comparing depressed and healthy individuals, but no single gene has particularly striking penetrance. Non-genetic mechanisms of heritability also play a role (e.g. shared environmental exposures, learned behavior, and parental care).

Several genes responsible for the metabolism and signaling of the monoamine neuromodulators serotonin, dopamine and noradrenaline have been implicated in the genesis of depression, along with BDNF and corticotropin-releasing hormone receptor genes. Specifically, the Val66Met allele in the coding region of BDNF has been associated with reduced hippocampal volume, depression and suicidality (Verhagen et al., 2010; Hajek et al., 2012; Zai et al., 2012).

Other leading candidates are polymorphisms in the serotonin transporter promoter region (5-

HTTLPR), the serotonin receptor gene HTR2A, and the monoamine oxidase A gene, all which interact with stressful life events and confer small positive associations to suicidal risk (Antypa et al., 2013). However, the effect size and significance of these polymorphisms and many others that showed promise have been contested by subsequent meta-analysis (Schild et al., 2013). No single

5 gene candidate seems to rise to the top as explanatory, as MDD seems to be a complex and heterogeneous collection of disorders caused by variations in multiple genes, each responsible for a small effect on risk or symptoms (Mandelli and Serretti, 2013). Convergence onto common molecular pathways is a theme that may suggest there are multiple ways to disrupt the same molecular and cellular targets and neuronal circuits. A wealth of evidence supports the involvement of heritable traits putting an individual at greater risk for depression in response to further environmental factors.

Does the way in which the brain influences behavior and dictates how we interact with the environment contribute to the pathogenesis of depression? Human epidemiologists have attempted to characterize trait contributions to risk to answer this question. Indeed, there are several personality traits that are associated with mood disorders generally and depressive episodes or their onset specifically. The most striking risk factor regarding depression is neuroticism, frequently found to be the most powerful predictor of first or subsequent depressive episodes (Enns and Cox, 1997). Neuroticism is a personality trait characterized by anxiety, frustration, worry, fear, and inappropriately high frustration or emotion in response to minor stressors. High levels of neuroticism are associated with early-onset depression (age<30) (Bukh et al., 2011), a negative information-processing bias (Chan et al., 2007), and resistance to antidepressant treatment (Bock et al., 2010). Interestingly, neuroticism has been found to have significant non-genetic heritability (Eley et al., 2015), likely by either vicarious learning of suboptimal coping strategies or as a result of direct effects of neurotic parenting (e.g. overprotection, intrusiveness, and expressed anxiety).

How is it that neuroticism contributes to depression? Approximately 55% of the genetic liability of major depression is shared with neuroticism (Kendler et al., 1993). Although the etiological factors shared by neuroticism and depression are tightly entangled, including genetic and early environment risk factors, neuroticism is believed to have its own causative effects via the greater distress that high-neuroticism subjects report (Kendler and Gardner, 2011). Proposed

6 mechanisms include greater exposure to stressful events and greater reactivity to those events.

There is evidence for both mechanisms increasing stress load, with the greater effect caused by reaction to stressful situations (Bolger and Schilling, 1991). In addition to exacerbating the physiological and psychological response to discrete episodes of stress, neuroticism as a personality trait contributes substantially to temporally stable symptoms of both anxiety and depression (Kendler and Gardner, 2011).

Depression and depressive symptoms can be secondary to other medical conditions, or share significant comorbidities, particularly with other neuropsychiatric diseases. Generally, chronic illness leads to increased rates of depression, potentially through the chronic physiological and psychological stresses of illness, shared risk factors, or unknown direct or indirect causes. Indeed, increased rates of depression are found in adults in inpatient settings compared to outpatient settings. Hypothyroidism is a common cause of depressive symptoms, and treatment of the underlying medical disorder is often a cure. Patients with diseases of hypercortisolism, either primary or secondary, and those treated with chronic steroids, also suffer from higher rates of depression. Chronic endogenous or exogenous steroid exposure can approximate external stress and subsequent activation of the hypothalamic-pituitary-adrenal

(HPA) axis, and exact internal stress on the body. A common factor for all of these diseases contributing to depression is through chronic physiological and psychological stress of being ill.

Despite all of these known contributors to risk for depressive illness, no one genetic trait, personality trait, or medical condition is specific or sensitive enough to be satisfying from an etiological perspective. However, many environmental factors are also associated with depression, and they primarily involve some form of stress.

Stress, Allostasis & Depression

Stress: good gone bad

The stress response is critical for survival in response to perceived threats to homeostasis; it is a powerful tool to quickly and maximally mobilize an organism’s resources to overcome an environmental challenge (Sapolsky, 1998). In the short term, stress shifts the priority of physiological systems to immediately increase fitness and survivability in the face of a threat.

Acute stress can heighten sensory and cognitive function, increase metabolic capacity via increased oxygen and glucose delivery to muscles, and dampen pain sensitivity, all at the expense of putting expensive long-term “repair” or “construction” projects on hold, like growth, immune function, digestion, and reproductive function. However, repeatedly activating energy reserves and putting off growth and repair processes can eventually lead to chronic stress-related disease.

Chronic or severe stress is intimately entwined with many forms of biological disease, including major depression. In one dramatic example, the odds ratio for the onset of major depression in the month of a stressful life event was greater than 5.6, with about two thirds of the risk by stress estimated to be causal (Kendler et al., 1999). The remainder of the association was attributed to individuals predisposed to depression were also predisposed to selecting higher-risk environments, which may also contribute indirectly to later onset. This prospective study replicated previous findings that depressed patients report more stressful life events than nondepressed subjects (Billings et al., 1983), while eliminating potential confound of a negative valence memory bias in depression that would not be properly controlled in a post-hoc case- control study.

How does stress contribute to depression? Hans Selye championed the idea that chronic physiological over-activation exacted a toll on human health and contributed to disease, while recognizing the general adaptive quality of having a stress response – in fact he named chronic stress-based illness “general adaptation syndrome” (Selye, 1950). In fact, Selye posited that the benefit of antiquated non-specific therapies, like blood-letting and fever- or foreign protein-

8 therapy, is due to kick-starting the stress response. The stress response was viewed as a drive towards restoring the milieu intérieur, in respect to the concept of homeostasis, propounded by legendary 19th century physiologist Claude Bernard and elaborated by his scientific “grandson”

Walter Cannon, respectively (Bernard, 1878; Cannon, 1929). That is, in health, the stress response increases fitness by promoting adaptation in this way to maintain homeostasis in response to varying environmental demands.

Allostasis is the modern concept that several homeostatic set points can be differentially regulated to meet different demands in the internal and external environment over time (e.g. in response to physical or psychological stress) (Sterling and Eyer, 1988; McEwen, 1998). However, an allostatic adaptation to environmental challenge exacts a cost on the brain and body by repeatedly meeting these varying demands. This wear and tear is referred to as the allostatic load

(McEwen, 1998). Especially in genetically susceptible individuals the cumulative allostatic load may become deleterious, becoming allostatic overload and leading to disease states (McEwen,

2008). The CNS in general has a dynamic range of activity that is exercised over the course of daily life. This is most adaptive for facing environmental challenges when circuits required for homeostasis can be rapidly mobilized to meet demands, then turned down again quickly when the threat to homeostasis, whether physical or psychosocial, is no longer present. Therefore facing stress per se is not harmful, rather it is either chronic activation or the inability for the systems underlying these responses to turn down again at the conclusion of the stressor that is detrimental

(McEwen, 1998). This concept is not restricted to the CNS, as the metabolic, cardiovascular, and immune systems can all contribute to disease when chronically challenged or over-activated

(Sapolsky, 1998).

There is considerable physiological evidence to support the allostatic load hypothesis and the consequences for health and neuropsychological function. In one poignant illustration, the allostatic load of adult human subjects in a large prospective cohort was quantified, based on several physiologic parameters of regulatory systems including cardiovascular and metabolic

9 profiles, sympathetic nervous system, and HPA activation and reactivity (Seeman et al., 1997).

Higher allostatic load was associated with poorer baseline cognitive performance overall

(memory, spatial ability, and abstract reasoning), and over the subsequent three years, greater cognitive decline, greater decline in physical performance, and greater incidence of cardiovascular disease.

Selye proposed that chronic stress prematurely depletes a finite “adaptational energy” existing in each organism (Selye and Tuchweber, 1976), thereby leading to the aforementioned

“adaptation syndrome”. Building upon this work, Sapolsky, Krey and McEwen argued that the overall cost of adaptation and response to stress and the environment was the ultimate cause of aging (Sapolsky et al., 1986). Joining the ideas of finite adaptational capability with another popular view, that aging could be conceptually thought of as an ultimately decreased adaptivitiy to stress, they described HPA dysregulation as both a consequence and cause of pathophysiological consequences in a feed-forward cascade, as well as a biomarker of aging

(Sapolsky et al., 1984). This logic has since been applied to the HPA dysregulation hypothesis of chronic stress and depression.

Moving beyond general concepts, I will explore changes to the brain that likely underlie the hypothesized transformation of an adaptive, self-limiting stress response by activation of the

HPA axis to chronic activation and potential contribution to depression.

The hypothalamic-pituitary-adrenal axis in chronic stress and depression

The brain is a powerful threat detector that interprets environmental stimuli and classifies them as either benign or potentially threatening, the latter of which activates the stress response.

Neurons in the paraventricular nucleus of the hypothalamus secrete corticotropin-releasing hormone (CRH, sometimes called corticotropin-releasing factor), targeting so-called corticotropes in the anterior pituitary via hypothalamo-hypophysial portal vessels. Following CRH receptor activation, these initial target neurons cleave pre-pro-opiomelanocortin to produce tens of

10 potential polypeptides (depending on post-translational modification and cleavage site) including

α-melanocyte stimulating hormone, β-lipotropin, pro-opiomelanocortin (POMC), and adrenocorticotropic hormone (ACTH). These peptides are then secreted into the systemic circulation. ACTH, with a systemic half-life of about 10 minutes (Yalow et al., 1964), stimulates adrenocortical cells in the zona fasciculate of the adrenal gland to produce and secrete glucocorticoids, including 11-deoxycorticosterone, corticosterone, and cortisol. These steroid hormones are also called corticosteroids, named for their origin in the adrenal cortex. In humans, cortisol is the predominant glucocorticoid, while corticosterone (CORT) predominates in rodents.

Both of these steroid hormones are produced by 11β-hydroxylase, and target glucocorticoid receptors (GRs), which are present in virtually cell with a nucleus, including all neurons. GRs are nuclear steroid hormone receptors, and act in the nucleus through transactivation and transrepression of target genes. Mineralocorticoid receptors (MRs) also exist throughout the brain, but because of a 10-fold higher affinity for corticosteroids than GRs, MRs are thought to be saturated at nearly all physiological concentrations, while GRs are only saturated in times of stress and increased corticosteroid concentration. In the brain, corticosteroids increase activity in areas that play key roles in attention, vigilance, memory formation, and behavioral strategy selection (Krugers et al., 2010b).

A significant subset of depressed patients display abnormal activity of the HPA axis, resulting in increased levels of glucocorticoids (Quarton et al., 1955; Pariante, 2009).

Glucocorticoids regulate neuronal survival, excitability, proliferation, metabolism, and memory, and high glucocorticoid levels may promote depression by impairing these processes (Holsboer,

2000). Glucocorticoid receptors (GRs) throughout the HPA axis, including the HPA-moderating hippocampus and PFC, exert negative feedback on glucocorticoid production. A dysfunction of

GRs in HPA feedback areas like the hippocampus may cause HPA dysregulation and failure to limit the stress response (Anacker et al., 2011b), as seen in depressed patients (Heuser et al.,

1994). This is thought to be both a consequence of, and further contribute to, allostatic load and

11 overload. How can this system, necessary for effective yet controlled stress response, be perturbed by chronic stress and contribute to depression?

The hippocampus, a brain area associated with both mood regulation and higher cognitive functions like learning and memory, is very reactive to stress (e.g. (Magariños et al.,

1996)) because of these high levels of GR expression (McEwen et al., 1968). Hippocampal enrichment of GRs allows for negative feedback to the hypothalamus, limiting the endocrine stress response (Reul and de Kloet, 1985). Thus, activation of hippocampal GRs by high levels of

CORT after stress mediates homeostatic feedback actions that restrain HPA system activation (De

Kloet et al., 1998).

Could a dysfunction of GR function in the brain contribute to the genesis of depression?

In human depression, GR expression is altered in many brain regions involved in affect and in

HPA regulation, including the hippocampus (e.g. (Wang et al., 2013)). It is unknown whether this is compensatory or causative; however there is evidence for both (Juruena et al., 2004; Anacker et al., 2011b). GR expression in the hippocampus is decreased in depression (Medina et al., 2013) and in chronic stress models of depression (Guidotti et al., 2013), and can be normalized following subsequent antidepressant treatment. Mice with a brain-specific knock-out or knock- down of GR display reduced depression-like behavior and altered HPA activity, while a forebrain-regional knockout produces depression-like behavior (Montkowski et al., 1995;

Karanth et al., 1997; Tronche et al., 1999; Boyle et al., 2005; Ridder et al., 2005). Although these results are mixed, perhaps reflecting developmental effects of GRs and the role of GR activation in other brain areas, the conclusions that can be made are that HPA abnormalities are common in depression, and can be recapitulated by chronic stress with some validity in animal models. The

HPA axis itself is also altered in depressed brains: the pituitary and adrenal glands are increased in size and activity in depressed patients (Nemeroff et al., 1992), further implicating aberrant stress exposure and reactivity.

Not only are HPA abnormalities common in depression, but depression is also common in patients with HPA abnormalities, specifically primary abnormalities of GC production, and patients receiving exogenous GCs (Quarton et al., 1955; Jeffcoate et al., 1979). There is compelling evidence of a causal link between altered HPA function and with a large subset of severely depressed patients, with specific regard to the core symptom of anhedonia. HPA dysregulation is correlated with severity of symptoms and is common (~80% of patients) in treatment-resistant depression (Anacker et al., 2011b). The correlation is even greater in melancholia patients (up to 90%), a subtype of major depression that can be identified using modified criteria relying centrally on anhedonia as the only core symptom, leaving depressed mood in the absence of anhedonia a separate entity with weaker correlation to HPA activity

(Taylor and Fink, 2008). Abnormal HPA function also correlates with a nine-fold increase in suicide risk (Coryell and Schlesser, 2001). This further supports the hypothesis that HPA dysregulation can be not only a consequence of depression, but a causative factor. It is important to note that not all depressed patients exhibit HPA derangement, nor is every depressive episode due to stress, and therefore any model of depression can only model individual facets or small constellations of symptoms, rather than all of human depression.

Animal models

A valid, useful animal model is critical for conducting insightful preclinical mechanistic studies regarding biology and human disease. Depression is a human condition, but a depressive- like behavioral state that is accompanied by many behavioral and biomolecular changes related to human depression can be induced in rodents. As with the development of any animal model, it is important to assess etiological, face, construct and predictive validity of candidate models and the neurobehavioral state they induce. Even the earliest calls for a reliable depression model in animals laid out the minimum criteria of 1) analogous symptoms between model and disease 2) behavioral changes that can be objectively evaluated 3) objective criteria about subjective states

13 induced 4) effective treatment modalities in humans should be effective in the model 5) reproducibility (McKinney et al., 1969). As I will discuss, some models fail to meet even these initial criteria, let alone the inclusion of subsequent criteria aiming to best model human disease.

Before evaluating different models, I will briefly define the various types of validity that a researcher should aspire to meet with the use of their model. Etiological validity implies similarity between the trigger used to precipitate neurobehavioral abnormalities and suspected etiological factors in human depression (Berton and Nestler, 2006). In the context of chronic stress and depression, in which depressive episodes are often precipitated by stressful life events, animal models in which a depressive-like neurobehavioral state is induced by exposure to stress has the greatest etiological validity. Construct validity implies that the neurobehavioral abnormalities observed in the model exhibit homology to symptoms of human depression. A key relevant example of construct validity given by Berton and Nestler (2006) is that evidence of reduced reward in animal models might be related to anhedonia in humans in regards to underlying mechanism. Face validity more superficially describes the similarity between signs or symptoms in humans and their corresponding behavioral state induced in animal models, the phenomenological similarity or “face value” (Willner et al., 1992). Finally, predictive validity embodies the ability of the model to predict therapeutic efficacy. As we shall see, predictive validity in depression research raises unique challenges, since humans respond to SSRIs only after several weeks of treatment.

What behavioral changes are best used to evaluate depression models? Behavioral despair tests are some of the most widely used because of their ease of use, low cost, reliability, and potential for high throughput. Common examples include the forced-swim test in rats or mice and tail suspension test in mice (Cryan et al., 2002). Both tests interpret immobility as

“behavioral despair”, with animals that spend less time struggling as exhibiting greater hopelessness or helplessness. Animals treated with an acute injection of an antidepressant drug spend more time struggling, and this is interpreted as antidepressant efficacy. However, these

14 tests lack etiological and construct validity. Furthermore they suffer from incomplete predictive validity since animals respond to a single, acute injection of a traditional antidepressant, even though human symptoms are only relieved by repeated, chronic antidepressant treatment. The greatest use for these tests is in rapidly screening antidepressant efficacy of therapeutics (Petit-

Demouliere et al., 2005), although they yield little evidence on how quickly treatments would act in other tests or in humans, a condition some use to judge proper predictive validity of the model

(Willner et al., 1992). Some argue that these tests are not models of depression at all, but rather

“black box” screening tests that should not be used for definitive evidence of a depressive phenotype (Nestler and Hyman, 2010).

A number of genetic models for depression have been developed, however, their evaluation is made challenging by the lack of reliable endophenotypes in depression. Selective breeding and forward genetic screens have resulted in H/Rouen mice, Flinders Sensitive rats, congenital Learned Helplessness rats, and high 8-OH-DPAT-sensitive rats, to name a few (El

Yacoubi and Vaugeois, 2007). These models have proved insightful for several potential mechanisms underlying depression pathophysiology, but vary in their etiological and construct validity. Many genetically altered mice have been developed around specific receptors, transporters, enzymes, and signal transduction proteins, greatly increasing the wealth of our knowledge about contributions to depressive or anti-depressant related behaviors (Cryan et al.,

2002). Most of these studies are based on immobility time in behavioral despair models, the forced swim and tail suspension tasks, in which increased immobility time is considered pro- depressive reflecting despair and decreased immobility time is considered anti-depressive reflecting resilience. Examples include antidepressant-like phenotypes of 5HT1A receptor knockouts, pro-depressive phenotype of 5HT1B receptor knockouts (Mayorga et al., 2001), antidepressant-like phenotype of monoamine oxidase A and B knockouts (Cases et al., 1995;

Grimsby et al., 1997) and antidepressant-like phenotype of GR-impaired transgenics

(Montkowski et al., 1995). Major criticisms to the use of genetic models in studying depression

15 are the lack of penetrance of specific genetic variants in the depressed population, the promiscuous nature of risk imparted from one gene toward multiple neuropsychiatric disorders, and the marginal increase in risk in disease from any one variant to be modeled (Nestler and

Hyman, 2010).

Chronic stress based models

Although many humans are exposed to stress and do not become depressed, episodes of depression often occur in the context of stress or are thought to be precipitated by stressful life experiences (Billings and Moos, 1984; Nestler et al., 2002; Hammen, 2005). In the context of chronic stress and depression, in which depressive episodes are often precipitated by stressful life events (Kendler et al., 1999), animal models in which a depressive-like neurobehavioral state is induced by exposure to stress have the greatest etiological validity, and many researchers consider stress-based models to have the greatest validity, generally, for depression research.

Chronic stress-based models in rodents come in several flavors, but the common factor is repeated exposure to psychological and physiological stress beyond an acute period, thought to have solid etiological basis to the human condition. The stressfulness of a given paradigm is usually confirmed with measures of HPA activation (adrenal weight, serum CORT levels, GR expression in HPA feedback areas) or changes in body weight. These models, lasting several weeks, are generally considered to have good face and construct validity by inducing depressive- like states and altering the putative underlying brain circuits in parallel with findings in depressed humans (Strekalova and Steinbusch, 2009; Lammel et al., 2014). Importantly, these changes are reversed by chronic, but not acute, administration of antidepressants (Papp et al., 1996; Grippo et al., 2006; Bondi et al., 2008), comparable to the delay for the therapeutic response in humans.

These models thus demonstrate satisfactory etiological, construct, face, and predictive validity.

Limitations of stress based models include the induction of several anxiety-related behaviors,

16 perhaps mirroring the co-morbidity and overlap of anxiety disorders and depression and some of their respective symptoms.

Repeatedly exposing an animal to a single stress over time (for example chronic restraint stress or foot shock) is the simplest chronic-stress based method to generate a depressive phenotype. One potential disadvantage is the habituation to a single stressor after repeated exposure (Magariños and McEwen, 1995a; Herman, 2013). This is observed as dampened HPA reactivity in response to stress. This possibility of habituation leads to variability in induction of a depressive-like neurobehavioral state, as animals might not fully be in an altered phenotypic state by the time the stressor becomes attenuated. Furthermore, these models rely predominantly on physiological stress, while psychological and psychosocial stress is more relevant to the human condition.

A more robust version of chronic stress exposure of a repeated single stressor, with greater ethological relevance, is the social defeat stress paradigm, in which an experimental animal is chronically defeated by a socially dominant conspecific, daily for several days to weeks

(Berton et al., 2006). Typically, the experimental animal is placed in the cage of an aggressive resident animal (often a retired breeder). The intruder is quickly attacked and defeated by the resident. After a few minutes of physical contact, the animals are separated by a divider, but kept in the same cage for a period of time (typically an hour). This promotes both physiological and psychological/psychosocial stress. A subset of animals do not develop depressive-like behavior

(reduced social interaction, anhedonia) despite chronic defeat, perhaps because they successfully habituate and adapt to the single repeated stress, making this model a hotbed for resilience research [e.g. (Krishnan et al., 2007; Francis et al., 2015)]. In this and other chronic stress models, the “resilient” group sometimes serves as the control for effects of stress itself, with the so-called

“susceptible” animals studied for their hedonic or psychosocial deficits (Strekalova and

Steinbusch, 2009).

The chronic unpredictable stress (CUS) or chronic variable stress model involves stressing rodents for several weeks using a rotation of daily stressors, with the goal of preventing adaptation to any one stressor, and imparting continual psychological stress by unpredictability.

In contrast to repeated single stress, CUS often increases stress responsiveness, marked by greater

HPA reactivity, thought to be driven by the perception of uncontrollability in addition to unpredictability (Herman, 2013). Typically, the rotation of stressors includes restraint, social isolation, food/water deprivation, forced swim, cage or littermate rotation, exposure to white noise or a strobe light (Willner et al., 1987). Three weeks of CUS induces daily increases in serum CORT and causes depressive-like changes in animal behaviors, such as anhedonia, sleep disturbances, and alterations in sexual behavior that resemble the symptoms of human depression

(e.g. (Grønli et al., 2004)), Furthermore, these changes are reversed by chronic treatment with traditional antidepressants (e.g. SSRIs), but not by acute antidepressant treatment or other compounds like anxiolytics or neuroleptics (Katz, 1982). The modern version of CUS focuses on mild stressors, to be more realistic compared to original versions of CUS that included more extreme stressors like foot shock and prolonged heat or cold exposure (Strekalova and

Steinbusch, 2009). After chronic mild stress, the induced state has been shown to be more lasting and stable, measured by tests for anhedonia and decreased reward (Willner, 1997). I have chosen to rely centrally on CUS to induce a depressive-like state, while acknowledging we cannot hope for any animal model to fully embody all facets, symptoms and signs of human depression.

A curious, decades-old clinical observation has helped inspire yet another model of depression, that finds its etiological and construct validity in the endocrine stress response.

Patients with diseases of hypercortisolism, like Cushingoid diseases (characterized by intractable corticosteroid release due to a pituitary or adrenal tumor), as well as patients chronically administered high doses of corticosteroids, have higher rates of depression than patients of other chronic diseases (Wolkowitz et al., 1993). Furthermore, depression scores correlate with severity of HPA derangement in Cushing’s patients with depression, and there are hippocampal structural

18 and functional changes (i.e. memory dysfunction) that correlate with cortisol levels (Starkman et al., 1992a, 1992b). Furthermore, hippocampal volume deficits can be at least partially reversed in patients that are treated to reduce hypercortisolism (Starkman et al., 1999). These findings illustrate how chronic elevation of corticosteroids can lead to several of the same neurobehavioral changes observed in depression. However, it should be acknowledged that the glucocorticoid- stress-depression link cannot underlie generally all depression, as only about half of depressed patients have excess glucocorticoid levels. The link appears to be critical, though, in specific subtypes of depression (Shorter and Fink, 2010) (such as “typical”, “melancholia” or

“endogenous” depression, as discussed above), or perhaps only in initial depressive episodes.

Despite its shortcomings, this link between excess glucocorticoids and chronic stress is one of the strongest biological findings to date regarding depression. In animals, exogenous corticosteroid administration has been used to recapitulate the effects of chronic stress and depression systemically or in the brain. There is preliminary evidence that chronic CORT administration mimics many aspects of CUS, since it shares the etiological and construct framework based on a hormonal stress response, and reproduces face effects of anhedonia and helplessness-like behaviors (Gourley et al., 2009). These glucocorticoid-induced changes in animals are reversed by chronic antidepressants, indicating predictive validity (Gourley and

Taylor, 2009). Left unanswered in this work, however, is the mechanistic explanation for the observed behavioral effects. Because I and others had previously observed that chronic stress results in a weakening of excitatory synapses in reward circuits, potentially accounting for the deficits in hedonic state, it was important to ask whether chronic CORT would produce similar changes in synapses. If so, it would provide further evidence strengthening the hypothesis that the synaptic changes are causative of the behavioral changes. I will use this model in Chapter 5 to test the hypothesis that chronic CORT elevation mediates the neurobiological effects of CUS discussed in Chapter 4 and 5, demonstrating that it is both sufficient and necessary for synaptic effects, as well as neurobehavioral changes.

Not all facets of human depression can be recapitulated in animal models, due to the subjective and internally-directed nature of many symptoms. For example, changes in appetite, circadian disturbances, and psychomotor agitation could all be modeled as readouts of a depressed phenotype in animals, while other symptoms which rely on verbal report likely cannot

(thoughts of death, suicidality, or feelings of worthlessness). Regarding the two core symptoms of depression, it has been argued that depressed mood cannot be measured in animals, and therefore anhedonia is the indispensable symptom to realistically model depression (Willner et al., 1992).

That said, what signs in animal models indicate that an animal is in a depressive-like state?

Although some easily administered tests are thought to provide a measure of ‘behavioral despair’, such as the forced swim and tail suspension tests, they lack some validity because changes can be detected following a single administration of antidepressants (Detke et al., 1997; Petit-

Demouliere et al., 2005) as discussed above. Another behavioral measure that is affected in these models is the sucrose preference test (SPT) (Rygula et al., 2006), in which rodents are presented with a choice between normal water and a dilute sucrose solution. The naturally high preference of rodents for the sucrose is diminished in depression models and recovered with chronic, but not acute, antidepressant treatment (Pothion et al., 2004; Cai et al., 2013). Measures of anhedonia and decreased reward are two of the most widely assayed depression-related behaviors.

The novelty suppressed feeding test (NSFT), in which the latency for animals to eat food in the center of a brightly lit novel arena is measured (Dulawa & Hen, 2005), is also affected in models of depression (neohypophagia) and sensitive to chronic, but not acute, antidepressant treatment. I will use the SPT and NSFT to evaluate whether animals are exhibiting a depressed phenotype.

An overview of triggers and neurobehavioral endpoints for various animal models of depression is summarized in Figure 1.1 [from (Berton and Nestler, 2006)].

Figure 1.1 - Animal models of depression Overview of etiological triggers (left) that are used to generate a depressive-like phenotype, and the neurobehavioral endpoints that are used to assess them (right). Not all endpoints are responsive to antidepressants, while others respond to acute treatment. Sucrose preference and hyponeophagia respond only to chronic treatment with antidepressants, best modeling the human condition. Adapted from Berton and Nestler, 2006.

Brain circuitry affected by chronic stress, depression, and antidepressants

To begin to understand what substrates in the brain are altered by chronic stress and might directly generate depressive signs and symptoms, I will start with a brief survey of brain changes in depression. Scientists and clinicians alike have long asked "what does the depressed brain look like?", and there are likely thousands of studies published to date on this topic.

Depression affects many brain regions, as would be expected to underlie the diverse collection of

21 symptoms, from classical emotional ones to cognitive deficits common in depression (Millan et al., 2012; Trivedi and Greer, 2014).

Decreased size and activity in cortical areas: Hippocampus and PFC

In respect to specific circuits and areas, structural studies in post-mortem suicide completers, as well as imaging studies in severely depressed patients have reported reduced volume and decreased glial density in hippocampus and prefrontal cortex (PFC), regions thought to play a role in the cognitive aspects of depression (Drevets, 2001; Sheline, 2003; Trivedi and

Greer, 2014).

One of the most robust findings in human imaging studies is decreased hippocampal volume in depression (Frodl et al., 2002; Sheline et al., 2003; MacQueen and Frodl, 2011), the magnitude of which correlates with duration of untreated depressive illness (Sheline et al., 2003).

Successful AD treatment reverses these decreases in hippocampal volume (i.e. (Fales et al.,

2009)). Interestingly, suicide attempters that are depressed have smaller hippocampal volumes than non-suicide attempters that are depressed, and suicide completers have up to a 15% reduction in hippocampal volume compared to non-suicidal depressed patients who died from other causes (Colle et al., 2015). It has been suggested that these volumes rebound between depressive episodes, thus obscuring the findings of many imaging studies (Campbell and

MacQueen, 2004).

What processes could underlie this decrease in volume? An impairment of normal, ongoing neurogenesis in the dentate gyrus of the hippocampus, as observed in several models of chronic stress (Yun et al., 2010), and atrophy of dendrites (Magariños and McEwen, 1995b) might underlie this decrease in hippocampal volume. Neurogenesis could explain the differences in structural plasticity, and intact hippocampal neurogenesis may be required for monoaminergic antidepressant action in animal models (Santarelli et al., 2003). Neurogenesis is also thought to be necessary for the antidepressant action of electroconvulsive therapy, an effective therapy reserved

22 for patients who have proved resistant to pharmacological treatment (Rotheneichner et al., 2014).

Using suicide again as an extreme example within the depressed population with readily available post-mortem samples, decreased neurotrophins are found in the brains of suicide completers, consistent with a critical role of neurogenesis or synaptogenesis (Dwivedi et al., 2003). Because

(1) neurogenesis is downregulated in stress models of depressive-like behavior and possibly in humans (Yun et al., 2010), (2) neurogenesis is upregulated by traditional antidepressant treatments (Malberg et al., 2000; Boldrini et al., 2009; Anacker et al., 2011a), and may be necessary for antidepressant efficacy (Santarelli et al., 2003; Li et al., 2008), (3) the BDNF

Val66Met polymorphism confers risk of neuroticism, depression, and hippocampal volume loss

(Sen et al., 2003; Schumacher et al., 2005; Frodl et al., 2007) and (4) the likely neurotrophin- based mechanism of dendritic atrophy and volume loss in the hippocampus in chronic stress and depression (Warner-Schmidt and Duman, 2006; Frodl et al., 2007; Yun et al., 2010), a neurogenesis hypothesis of depression has been posited (Duman and Li, 2012).

However, this hypothesis is not complete, as disruption of neurogenesis does not induce depressive-like phenotypes, and is not required for antidepressant action under all circumstances

(Hanson et al., 2011). In fact, dendritic complexity and synaptic plasticity rather than cell proliferation seems to be more important for maintenance and restoration of mood (Bessa et al.,

2009). It remains unclear how neurogenesis may alter activity in the hippocampus at the neuronal and circuit levels to mediate the antidepressant effects of monoaminergic antidepressants, but it there is mounting evidence that plasticity mechanisms, prominent and well-studied in the hippocampus, are required (Femenía et al., 2012). Furthermore, there is considerable overlap between mechanisms of neurogenesis and synaptogenesis and their reliance on BDNF and its receptor trkB, which opposes the atrophic effects of chronic stress in cortical areas like the hippocampus (Duman and Monteggia, 2006; Warner-Schmidt and Duman, 2006). Because of how robustly this brain area changes in depression, in response to antidepressant treatment, and in animal models of stress and depression, and because of its extensive interconnectedness with

23 cortical, limbic, and reward structures, the hippocampus is one of the most widely studied brain areas regarding depression.

The PFC is an important site at which cognitive evaluations, such as the controllability of a stressor (Amat et al., 2005) or the positive valence of a stimulus (Berridge and Kringelbach,

2013), can influence affect and reward. Measured with functional magnetic resonance imaging

(fMRI), depressed patients display evidence of reduced activity in the PFC, and SSRI treatment restores normal activity levels and morphology (Kennedy et al., 2001; Fales et al., 2009;

Koolschijn et al., 2009). In animal models, chronic stress results in atrophy of distal dendrites of pyramidal cells in the PFC and loss of dendritic spines (Liston et al., 2006; Radley et al., 2008).

Seven days of chronic restraint stress produces a decrease in synaptic excitation in layer V pyramidal cells in the PFC (Yuen et al., 2012), driven by glucocorticoid-dependent increase in proteasome-mediated degradation of synaptic proteins, including AMPA and NMDA type glutamate receptors. Consistent with decreased excitation, expression of several activity- dependent genes is decreased in the PFC of depressed humans (Covington et al., 2010).

Experimentally stimulating the medial PFC in animal models of depression has an antidepressant effect (Covington et al., 2010), while clinically the PFC is one target for deep brain stimulation in treatment resistant depression patients (Mayberg et al., 2005).

Bad apples: Increased size and activity in amygdala and lateral habenula

The amygdala is responsible for several negative affective processes, fear, and perception of emotional and social stimuli. Depressed patients are reported to have increased volume compared to controls (Prather et al., 2001; Frodl et al., 2003), providing a contrasting effect to

PFC and hippocampus. This may be driven by dendritic hypertrophy (Vyas et al., 2002), and result in hyperactivity in depressed patients (Drevets, 2003). Optogenetic activation of amygdala inputs to the hippocampus can impair social behavior (Felix-Ortiz and Tye, 2014), while normal amygdala function is necessary for a dynamic range of fear behavior in response to social

24 interactions or to novelty (Prather et al., 2001). Excessive activation of the amygdala may thus contribute to negative affect, either by improper modulation by higher brain areas, or by negatively modulating reward circuitry (Drevets, 2001).

Another “negative mood” area is the lateral habenula, a region responsible for negative valence, which contributes to disappointment and negative thinking. The habenula is connected both directly and indirectly to serotonergic and dopaminergic pathways (Sego et al., 2014).

Activation of this region contributes to social avoidance behavior (Stamatakis and Stuber, 2012), while lateral habenula neurons are activated by either absence of expected reward or by presence of punishment (Matsumoto and Hikosaka, 2009). Excitability is increased in the learned helplessness model of depressive behavior (Li et al., 2011a), suggesting a role in depressive illness (Proulx et al., 2014).

In summary, depression symptoms may be attributed to either overactive amygdala and/or lateral habenula (areas associated with negative mood and valence), or to underactive hippocampus and PFC (areas associated with higher cognitive functions like cognitive function, attention and learning, and with moderating negative mood). As there is more than one way to skin a cat, it appears there is more than one way to perturb mood-related circuitry. Simplistically, activating “negative mood” areas, and suppressing “good mood” areas might lead to the same result, especially as all of these areas converge on reward areas.

The definition of the core symptom anhedonia would seem to convict the brain's reward circuitry. Indeed, anhedonia has been linked to dopaminergic dysfunction in the basal ganglia

(Miller et al., 1996), while patients with MDD have been observed to have reductions in striatal size and function (Lorenzetti et al., 2009). The monoamine dopamine is classically considered the

“reward hormone”, and is released from dopaminergic neurons in the ventral tegmental area

(VTA) in response to reward stimuli, like food or sex. Stimulation of dopamine release is potently rewarding, as animals trained to self-stimulate will do so at the expense of all else, even to the exclusion of food or life-saving warmth, if given the opportunity (Carlezon and Chartoff, 2007).

The dopaminergic VTA and the GABAergic population of dopamine-receptor expressing medium spiny neurons in the nucleus accumbens (NAc), which reciprocally innervate each other and both of which receive projections from limbic and cortical structures. Thus, layering these connections on top of excitatory glutamatergic interconnections between amygdala, PFC, and hippocampus, an integrated view of so-called corticomesolimbic circuitry and its systems-level role in depression is beginning to emerge (Figure 1.2).

Consistent unidirectional results on the role of the reward circuitry have not been observed in animal models. In mice, optogenetic stimulation of the ventral tegmental area (VTA) relieves a depressive-like phenotype after chronic stress, while inhibition of VTA induces a depressive-like behavioral phenotype (Tye et al., 2013). However, in a social defeat model, VTA stimulation induces a depressive-like phenotype in animals otherwise resilient to social-defeat stress, with specificity of this effect for VTA-NAc, but not VTA-mPFC, projections (Chaudhury et al., 2013). In a chronic restraint stress model, behavioral anhedonia was shown to be mediated by the loss of strength of excitatory synapses onto dopamine receptor (D1R, specifically) - expressing NAc medium spiny neurons via melanocortin 4 receptor activation (Lim et al., 2012).

Pathophysiological changes in reward circuits are thus thought to be a major contributor to the biological underpinning of anhedonia, and play a role in other symptoms of depression (Akiskal and McKinney, 1973; Nestler and Carlezon, 2006). In sum, it is unclear exactly how changes in affect are mediated by the reward circuitry (are more computational parameters necessary to full understanding?), but mounting evidence suggests that the effect of the underlying changes are sourced in or at least pass through the striatum (Nestler and Carlezon, 2006; Russo and Nestler,

2013). Dysfunction in several aspects of reward and motivated behavior that are observed in depression and other mood disorders, most notably resulting in the core symptom of anhedonia, are thought to arise ultimately through a corticomesolimbic mechanism.

Figure 1.2: Corticomesolimbic circuitry relevant for depression Cortical and limbic areas have a complex relationship with the reward circuitry. For example, the hippocampus and PFC have dense glutamatergic projections to the nucleus accumbens, which in turn di- synaptically disinhibits dopaminergic VTA neurons and results in dopamine release at many projection sites, chemically signaling reward. For simplicity, many connections are not shown, such as reciprocal innervation of hippocampus and PFC. From Russo and Nestler, 2013.

Serotonin and the monoamine hypothesis of depression

Serotonin, or 5-hydroxytryptamine (5-HT), is produced by neurons in the raphe nuclei in the brainstem, and is released from serotonergic projections that innervate multiple cortical areas.

Serotonergic neuromodulation influences a wide array of functions including: trophic activity in cell survival or synaptogenesis, motor function, mood regulation, pain sensitivity, and neuroendocrine functions like feeding behavior and circadian rhythmicity (Gaspar et al., 2003;

Martinowich and Lu, 2008). Serotonin is also produced by enterochromaffin cells and is abundant systemically, (in fact “serotonin” was named for its presence in serum as a vasoconstrictor) but I will not discuss its roles outside the CNS (Rapport et al., 1948).

Regarding higher cognition, serotonin regulates cognitive flexibility, attention, and impulsivity (Puig and Gulledge, 2011). The serotonergic system has been closely examined in depression, driven primarily by the usefulness of serotoninergic modulators, most prominently

SSRIs, as antidepressant treatments. Furthermore, neuromodulatory disruption including that of

27 serotonin is thought to contribute to cognitive impairments associated with depression (Millan et al., 2012). Serotonergic signaling can promote BDNF signaling and enhance synaptic plasticity

(Martinowich and Lu, 2008).

An original, but outdated, viewpoint was that depression was primarily a derangement of monoaminergic signaling (Schildkraut, 1965). This was driven primarily by the early success of iproniazid, a monoamine oxidase inhibitor (MAOI) in exerting beneficial effects on mood. Later imipramine, a tricyclic compound, was also found to boost mood, although both drugs were being used for other indications. A placebo-controlled imipramine study in depressed patients found therapeutic effect in 74% of patients (Ball and Kiloh, 1959), lighting the fire for tricyclic antidepressant (TCA) development. TCAs block serotonin and norepinephrine transporters and prevent clearance from the synaptic cleft, which is believed to be the mechanism of therapeutic efficacy of these drugs. However, side effects range from dry mouth and urinary retention to potentially fatal cardiovascular problems and serotonin syndrome (Jefferson, 1975).

This initial golden age of drug development soon included more targeted compounds to reduce the number of serious side effects: SSRIs like fluoxetine and serotonin-norepinephrine reuptake inhibitors like duloxetine (SNRIs) (Slattery et al., 2004). Unfortunately efficacy of these designer drugs is no greater than for TCAs, highlighting the limitation of the monoamine hypothesis of depression. The largest to-date randomized clinical trial demonstrated that only half of depressed patients treated with a single 3-month long course of SSRIs responded fully, with nearly ¾ failing to achieve full remission (Gaynes and Warden, 2009). TCAs are now generally reserved only for use after failure of the less harmful SSRIs and SNRIs. A major limitation for the use of these drugs, and complication for unravelling their mechanisms, is that monoaminergic treatments all require weeks to months to achieve symptomatic improvement, despite elevating serotonin levels immediately.

The monoamine hypothesis, despite over five decades of appeal and hundreds of millions of patients treated (albeit with mixed success), has met with a mixture of successful and failed

28 predictions. For example, a catecholamine depleting drug, reserpine, was shown to induce depressive symptoms in hypertensive patients (Lemieux et al., 1956; Quetsch et al., 1959), but later analysis of the same data revealed that the rate of depression was roughly the same as for the general population ~10% (Baumeister et al., 2003). Depletion of tryptophan, the serotonin precursor, would be predicted by the monoamine hypothesis to cause depressed mood. This was true for depressed patients being treated, but not for healthy controls, and not for depressed patients in full remission that were no longer being treated (Delgado et al., 1999). Another significant flaw in the monoamine hypothesis is the prolonged latency (weeks to months) to therapeutic effect, despite rapid (minutes to hours) increase of monoamines in the synaptic cleft

(Katz et al., 2004). This leads to the conclusion that caused not by an abnormality of 5-HT function itself, but by dysfunction of other systems or brain regions modulated by 5-HT over time.

What properties of serotonin might explain both the therapeutic efficacy of serotonergic- based treatment but inadequacy as a complete explanation? Matching the complexity of serotonin effects in depression is the complexity of serotonin action on neurons and physiology, suggested by the diverse array of behaviors affected by serotonin signaling. The serotonin receptor family

(5-HT1-7) is the largest family of related G-protein coupled receptors (GPCRs), with 13 distinct

GPCR genes and an additional ligand-gated ionotropic serotonin receptor (5HT3R) encoded by 5 distinct genes. Within the family of 5HT GPCRs, there are several subgroups with high homology to one another, for example 5HT1R subtypes (A-F) and 5HT5R share high phylogenetic homology

(Nichols and Nichols, 2008). Consequently, only this subfamily of receptors amongst all 5HTRs couples to the heterotrimeric protein Gαi, classically an inhibitor of adenylate cyclase. 5HT2R couples to Gαq, activating phospholipase C (PLC), while the remaining G-protein coupled 5HT receptors couple to Gαs, which activates adenylate cyclase and increases cyclic adenosine monophosphate (cAMP). Receptors are either self-regulatory autoreceptors or heteroreceptors

(Barnes and Sharp, 1999). Serotonin is ultimately transported back into the presynaptic

29 transporter SERT or metabolized by monoamine oxidases, targets of SSRIs and MAOIs respectively. Because of the varied population of 5HTRs, identifying the receptors critical for antidepressant efficacy is challenging.

A particularly important 5HTR is the rodent 5HT1BR, which has high homology to the

5HT1DβR/h5HT1BR in humans (Adham et al., 1992). These receptors are autoreceptors on serotonergic nerve terminals (Sari et al., 1997), where they act in a feedback manner to inhibit serotonin release. In cell lines, these receptors inhibit cAMP via Gαi, although there is in vitro evidence for activation leading to intracellular calcium elevation (Giles et al., 1996). 5HT1BRs are also expressed on postsynaptically on dendrites, and can contribute to increased and decreased excitability (Berger and Huynh, 2002; Stean et al., 2005). In tissue from depressed patients, there is decreased p11 protein expression, which is necessary for 5HT1B/DR surface expression

(Svenningsson et al., 2006). A h5HT1BR polymorphism has been identified that is associated with depression and substance abuse, but not other psychiatric disorders (Huang et al., 2003). In animal models, 5HT1BR knockout mice do not respond to antidepressants in the forced swim test, but exhibit increased sensitivity to SSRIs in the tail suspension test (Mayorga et al., 2001; Chenu et al., 2008). The 5HT1BR agonist anpirtoline produces antidepressant-like effects in wild-type animals, yet 5HT1B autoreceptors limit the effects of serotonin reuptake blockade in the synaptic cleft (Schlicker et al., 1992; Malagié et al., 2001). These suggest a critical, but complicated role, for 5HT1BRs in antidepressant action. Disentangling 5HT1BR’s distinct roles as auto- and heteroreceptors may explain some of the confusion. In animals with serotonergic neurons pharmacologically lesioned, anpirtoline still has an antidepressant-like effect in the FST, suggesting potential antidepressant actions of 5HT1BRs are via heterosynaptic activation, rather than via auto-receptor mediated inhibition.

It now appears that the role of serotonin in depression is more nuanced than the outdated monoamine hypothesis of depression had originally postulated (Krishnan and Nestler, 2008;

Kallarackal, 2011). Despite lack of evidence of an insufficiency of serotonin synthesis or release,

30 serotonin-based medications have been the principal pharmacotherapy to treat depression for decades. Therefore, I posit that circuits exist that become dysfunctional as the result of chronic stress, due to corticosteroid-induced weakening of excitatory synapses, and that these circuits can be repaired by SSRIs. A lack of serotonin is not the primary deficit, or at least not the sole deficit, underlying affective deficits in depression, but it can play a role in the therapeutic response. An analogy might be helpful. Imagine a leaky pipe that is not performing to full potential, for which a repairman from Serotonin Repairmen, Inc., arrives to help restore the pipe to full function. The problem with the pipe (key brain circuitry) was not a deficit in plumbers (serotonin), but some other problem that the presence of the plumber was able to address and restore the pipe to health.

In the same way, the usefulness of enhancing serotonergic signaling at the depressed synapse does not directly implicate an absence of serotonin as the reason the synapses are depressed.

Certainly, there are lessons to be learned from the story of SSRIs, their successes and shortcomings. Logically, if there are circuits that underlie depressive behavior that can become dysfunctional after chronic stress and subsequently restored by SSRI treatment, then these circuits become appealing targets to use as models to attempt to answer the questions “How do antidepressants work?” “How do we become depressed in the first place?” and “How do we begin the process of identifying a new generation of antidepressants?”

Given all of the changes to the brain in depression and in chronic-stress based models of depression, great care should be taken in selecting brain regions for investigation. There is a particularly compelling brain area which fulfills the growing list of criteria- the hippocampus. It is affected by chronic stress and depression and antidepressant treatment, is densely innervated by serotonergic fibers, interacts with the reward circuitry, dialogues with—and is affected by—the stress response, and is accessible to molecular, electrophysiological, and behavioral investigation.

The intersection of stress, cognition, antidepressants and reward

A dialogue between the hippocampus, the neocortex, and striatum is important for maintaining normal affective and cognitive function (Morris et al., 1982; Gray, 1998; Remondes and Schuman, 2004). The hippocampus executes functions of spatial navigation, declarative memory encoding and recall, novelty detection, emotional regulation, and attention control

(Lisman and Otmakhova, 2001; Andersen et al., 2006). The entorhinal cortex (EC) is the principal conduit of inputs and outputs between the hippocampus and neocortex, but the hippocampus is innervated by many regions including basal forebrain, hypothalamus, and brain stem (Amaral and Lavenex, 2006). This central hub of brain activity is innervated by serotonergic, cholinergic, dopaminergic, and noradrenergic projections (Otmakhova and Lisman,

2000). The hippocampus is one of the most strikingly organized brain regions, with several prominent subfields. The hippocampal formation includes the dentate gyrus (DG), where adult neurogenesis occurs, the subiculum, and CA areas 1-3, named in Latin for resembling the elegant spiral shape of a ram’s horn: cornu ammonis. In the simplified view of the classical hippocampal trisynaptic circuit, information arrives in the DG from the entorhinal cortex via the perforant path, and is then propagated sequentially to CA3 and CA1, before being passed back to EC either directly or after another synaptic transmission in the subiculum. However, CA1 and CA3 both receive direct inputs from EC, neuromodulatory regions, and nucleus reuniens. Furthermore, there are many input and output connections between subfields intrinsic to the hippocampus, and others which form connections with other brain regions extrinsic to the hippocampus (Buhl and

Whittington, 2006). A simplified summary diagram of connections within the hippocampus is presented in figure 1.3 [from (Amaral and Lavenex, 2006)].

Figure 1.3 - Simplified Hippocampal Circuit Diagram summarizing intrinsic and extrinsic hippocampal connectivity, containing serial and parallel projections. Pre and para refer to the pre- and para-subiculum. Many connections are not shown for simplicity, particularly extrinsic connections including neuromodulatory projections. From Amaral and Lavenex 2006.

In addition to the highly organized collection of hippocampal subfields, there is an elegant laminar organization within subfields that gives rise to layers that are identifiable by structure and function. Pyramidal cells in CA1 and CA3 align neatly in the cell body layer, or stratum pyramidale. Apical and basal dendrites branch off from the pyramidal cell somata. The apical dendrites of CA1 pyramidal neurons receive two main inputs. The Schaffer collateral (SC) pathway, comprising excitatory input from area CA3, terminates in stratum radiatum (SR). The direct EC input, or temporoammonic (TA) pathway, terminates on distal apical dendrites in stratum lacunosum moleculare (SLM), several hundred microns from the cell body (Steward and

Scoville, 1976; Remondes and Schuman, 2002). The TA pathway provides sensory information

33 to CA1 and also assists CA1 in its role as a comparator (Vinogradova, 2001) and is important for consolidation of long-term spatial memory (Brun et al., 2002; Remondes and Schuman, 2004).

Within CA1, SLM is a key region for neuromodulation due to dense innervation by serotonergic and dopaminergic projections, and enrichment of serotonin and dopamine receptors (Ihara et al.,

1988; Otmakhova and Lisman, 1999). Indeed, many physiological effects of neuromodulation are unique to this layer even compared to other regions within CA1 (Otmakhova and Lisman, 2000).

This makes SLM of paramount interest in the context of depression and substrates on which

SSRIs may act to underlie antidepressant action.

To study how chronic stress can lead to depression, and because it is an important locus of antidepressant action, it is important to characterize the changes in the hippocampus that are caused by chronic stress and may contribute to depressive-like symptoms and behavior. In animal models based on chronic stress, there are dramatic alterations in neuronal structure and function

(Krugers et al., 2010b), including altered synaptic currents (Karst and Joëls, 2003), deficits in synaptic plasticity in CA1 and DG (Alfarez et al., 2003), decreased neurogenesis and cell turnover in DG (Joëls et al., 2004; Mirescu and Gould, 2006), decreased dendritic complexity in

CA3 (Magariños and McEwen, 1995a), and dysfunctional circuit function (Airan et al., 2007).

These deficits match high resolution imaging observations in unmedicated depressed patients, in which detailed volume loss is observed in each hippocampal subfield (Huang et al., 2013). These alterations in hippocampal function are among some of the clearest signs yet found in chronic stress-based models of depression, but are they merely epiphenomenological correlates or do they contribute to the genesis of depression? How do effects of chronic stress interact with serotonergic modulation of these pathways and set the stage for antidepressant action?

Excitatory transmission, synaptic plasticity, and depression

Glutamate is an essential amino acid for mammals, with a key role in cellular metabolism throughout the body, but it is also the most abundant excitatory neurotransmitter in the CNS

(Meldrum, 2000). Glutamate is packaged and stored in vesicles, released by axon terminal depolarization presynaptically, and activates metabotropic GPCRs and ionotropic receptors, before being cleared by diffusion, reuptake or metabolism. Ionotropic receptors include the α- amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR), the N-Methyl-D- aspartate receptor (NMDAR), and the kainite receptor. These are the major direct players in excitatory synaptic transmission in the central nervous system, although there are hundreds of compounds that can modulate excitatory transmission.

Glutamate receptors are critical for synaptic plasticity. NMDA and AMPA receptors are known for their roles in the induction and expression of the form of synaptic plasticity known as long-term potentiation (LTP), respectively (Bliss and Collingridge, 1993). LTP is considered the neurobiological substrate of learning and memory, and is mediated via trafficking and regulation of AMPA receptors at the synaptic membrane (Malenka, 2003). Although there is ongoing debate about the details, the overall accepted model of LTP is that NMDAR activation results in activation of calcium calmodulin-dependent kinase II (CaMKII), which then recruits AMPARs to the synapse (Nicoll and Roche, 2013). There are two proposed mechanisms, not mutually exclusive, which describe exactly how the synapse captures AMPARs following LTP, lateral diffusion and exocytosis from an intracellular pool. A mechanism is still required that would explain how AMPARs become preferentially captured, either by increase in a number of available slots or by some change in the AMPAR that makes it more amenable to capture (Opazo et al., 2012). Critical roles for LTP are served by scaffolding proteins and their organization

(MacGillavry et al., 2013), modification of receptors via phosphorylation or protein-protein interactions (Lee et al., 2000), and the role of specific auxiliary subunits of AMPARs. The key

AMPAR phosphorylation sites for increasing trafficking and conductance are the CaMKII/Protein

Kinase C (PKC) target serine 831 and the Protein Kinase A (PKA) target serine 845. Most of the studies that have led to this body of work have been performed at CA1 pyramidal neuron synapses. However, LTP has been described at other synapses, with some variation in absolute

35 mechanism. For example, LTP at mossy fiber synapses is NMDAR-independent and is presynaptically expressed (Nicoll and Malenka, 1995).

What does this have to do with depression? A growing body of biological psychiatry supports the hypothesis that alterations in synaptic structure and function underlie changes in the function of neuronal circuits and ultimately affect and behavior (Krishnan and Nestler, 2008;

Kavalali and Monteggia, 2012). Furthermore, synaptic plasticity is a critical mediator of these changes (Pittenger and Duman, 2008). Chronic stress impairs LTP in the hippocampus and PFC, with disruptions in associated behavior of these regions. Evidence exists for mechanisms including dendritic reorganization (atrophy or spine loss) and weakening of synapses through loss or decreased conductance of glutamate receptors (Popoli et al., 2012). NMDARs and AMPARs are lost in the PFC due to increased turnover following chronic stress (Yuen et al., 2012). In depressed humans, mRNA of GLUR1 and GLUR3 genes for GluA1 and GluA3 proteins are downregulated in CA1 and DG (Duric et al., 2013). These changes can be recapitulated in chronically stressed animals and reversed by antidepressant treatment. Thus, it is hypothesized that the mechanism by which chronic stress can lead to depressive-like behavior is by impairment of synaptic strength and plasticity.

Conversely, it can be predicted that successful antidepressant treatment would act on these substrates and reverse deficits in synaptic function.

Monoamine-based antidepressants result in reconfiguration of synaptic structure and function, as neuromodulation of glutamatergic transmission is one of the key actions of serotonin.

In the hippocampus, 5-HT application in acute slices and chronic fluoxetine treatment in vivo result in increased S831 and S845 phosphorylation (Svenningsson et al., 2002). AMPAR subunits

GluA1-4 exhibit time dependent increases in mRNA and protein expression in response to chronic fluoxetine treatment, matching the time period required for monoaminergic antidepressants to show clinical efficacy in patients (Barbon et al., 2011).

Blocking NMDARs also results in a rapid, long-lasting antidepressant effect. In humans, this has been observed in treatment-resistant patients (that is, patients resistant to treatment with

SSRIs) with an acute injection of ketamine (Berman et al., 2000). This clinical effect, observed within a few hours, persists for up to seven days following a single treatment (Zarate et al., 2006).

These results have been modeled in rodents with the forced swim, novelty suppressed feedings, learned helplessness, and sucrose preference tests, using naïve unstressed rodents and also rodents subjected to chronic stress (Garcia et al., 2008; Autry et al., 2011; Li et al., 2011b). Ketamine, despite being an NMDAR antagonist, enhances glutamate release (Moghaddam et al., 1997) by producing a mild disinhibition and increase in coherent oscillatory activity that is thought to lead to enhanced AMPAR activation. The expression of ketamine’s effect is blocked by AMPAR antagonists, suggesting both glutamate receptors are critically involved (Maeng et al., 2008).

These findings of the rapid antidepressant action of NDMAR antagonists have sparked great interest in siring a new generation of antidepressants.

Different mechanisms of antidepressant action ultimately converge on the same final synaptic mechanisms that are also affected by chronic stress and in depression. All effective mechanisms of antidepressant treatment (SSRIs, TCAs, ECT, ketamine) critically involve BDNF

(Duman and Monteggia, 2006; Martinowich et al., 2013), implicating neuronal and synaptic actions of antidepressant treatment and dysfunction in depression.

Summary of logic

Chronic stress models provide the means to begin the search for brain defects underlying depression, by presenting us with neurobiological phenotypes that change in parallel with depressive-like signs and symptoms in brain regions that are known to be involved in regulating those behaviors. Many changes also occur in the same brain regions in depressed humans where we observe synaptic and circuit deficits in animal models, increasing confidence in the face and construct validity of the findings in these models. SSRIs further narrow our search for brain

37 defects underlying depression, by allowing us to concentrate effort on brain regions that are repaired by antidepressant treatment, in parallel to apparent reversal of mood deficits. Although unlikely to be fully inclusive of all depression and all depressive symptoms, this gives us a foothold that reflects the validity of the models, techniques, and observations that point to these areas. Identifying deficits in a depressive-like state that are also downstream targets of antidepressant effects should enable a search for alternative approaches to reverse these deficits.

This should invite workarounds to the ineffectiveness of current standard-of-care approaches by suggesting alternative mechanisms to reach these same downstream endpoints.

Hypothesis and predictions

I hypothesize that some of the symptoms of depression, specifically the core symptom of anhedonia, are caused by corticosteroid-dependent reduction of excitatory synaptic strength at key synapses in cortico-mesolimbic circuitry, which can be reversed by monoaminergic antidepressant treatment because of 5HT1BR-mediated serotonergic potentiation, as well as by other compounds or treatments that also promote potentiation of these synapses.

Hypothesis 1: Monoaminergic 5HT1BR-mediated antidepressant efficacy is caused by increased excitatory synaptic transmission at cortico-mesolimbic sites via AMPAR phosphorylation by

CaMKII.

Prediction 1.1. AMPAR and CaMKII phosphorylation in the hippocampus and NAc will be increased following SSRI, TCA, or anpirtoline application.

Prediction 1.2. AMPAR and CaMKII phosphorylation following SSRI or anpirtoline application will coincide with synaptic potentiation at TA-CA1.

Hypothesis 2: Depressive behavior is caused by chronic-stress induced dysfunction of key excitatory synapses in the corticomesolimbic circuitry, specifically a loss of AMPAR-mediated synaptic strength.

Prediction 2.1: CUS will cause a decrease in AMPA:NMDA ratios at TA-CA1 synapses, which is reversed by SSRI treatment.

Prediction 2:2: CUS will cause a decrease in expression of AMPAR subunit expression in SLM.

This down-regulation will be reversed by SSRI treatment.

Prediction 2.3: AMPAR-mediated synaptic strength at TA-CA1 will correlate with sucrose preference test scores.

Hypothesis 3: Chronic stress alters synapses and induces depressive behavior via chronic stress- induced elevations of corticosterone; corticosterone is both sufficient and necessary for the

39 neurobiological correlates of chronic stress: impaired TA-CA1 synaptic function, altered responsiveness to 5HT1BR agonists, anhedonia, and hyponeophagic behavior.

Prediction 3.1: If chronic elevations of corticosterone are sufficient for the synaptic and behavioral effects of chronic stress, then chronic exogenous CORT administration will result in a depressive-like behavioral phenotype measured with SPT and NSFT, and recapitulate the stress- induced weakening and alteration of TA-CA1 synapses.

Prediction 3.2: If stress-induced elevations of corticosterone are necessary for the synaptic and behavioral effects of chronic stress, then blocking elevated corticosterone secretion throughout

CUS will prevent the neurobiological behavioral and synaptic correlates of chronic stress.

In the following chapters of this thesis, I will hopefully contribute to the knowledge surrounding the inextricably linked questions “how do we become depressed?”, “how do antidepressants work?”, and “how do we begin the process of identifying a new generation of antidepressants?” In chapter 3, I have begun our investigation by testing hypothesis 1 and examining a necessary action of serotonin for antidepressant action, focusing primarily on my contributions to a study led by Xiang Cai and Angy Kallarackal (Cai et al., 2013). These results led to identification of the TA-CA1 synapses as a key site of dysregulation in depression and as an archetype of critical synaptic function for antidepressant action. In chapter 4, I test hypothesis

2, and further characterize the deficits caused by chronic stress at TA-CA1 synapses, and demonstrate that loss of excitatory synaptic strength, and their reversal with successful antidepressant treatment, are critical biomarkers of a depressed phenotype before and after chronic stress (Kallarackal*, Kvarta* et al., 2013). In chapter 5, I demonstrate that corticosterone mediates the synaptic and behavioral effects of chronic stress in the CUS model, by demonstrating that it is both sufficient and necessary (Kvarta et al., 2015). In chapter 6, I explore the emerging excitatory synapse hypothesis of depression that is consistent with my and other’s

40 results, which resulted in our conceptualization and testing of a novel strategy for rapid antidepressant treatment (Fischell et al., 2015; Thompson et al., 2015).

Science is a collaborative effort, and none of these studies could have been completed and published without contributions by others. Here I have focused nearly exclusively on data I personally collected. I have taken careful effort to acknowledge the efforts of my colleagues and collaborators when presenting data I did not collect in full.

The disclaimer of Max Hamilton, who sounded a little anhedonic himself in the abstract for his widely cited A Rating Scale for Depression, applies herein as it does for many scientific endeavors: “Unfortunately, it cannot be said that perfection has been achieved, and indeed, there is considerable room for improvement”.

2) General methods

All protocols were submitted to and approved by the University of Maryland School of

Medicine Institutional Animal Care and Use Committee.

Animals

Male Sprague-Dawley rats (Harlan Laboratories, 3-4 weeks old at start) kept on a regular

12 hour light/dark cycle were group housed, with food and water available ad libitum. All rats sacrificed for in vitro electrophysiology and molecular biology experiments were 7-8 weeks old.

Acute hippocampal slice preparation

All tissue for electrophysiology and western blot experiments were prepared acutely, on the same day they were used. At 7-8 weeks old, rats were deeply anesthetized using 0.025mL/kg

Somnasol (390 mg pentobarbital sodium, 50 mg phenytoin sodium per mL). After rapid decapitation, both hippocampi were removed and placed into ice cold artificial cerebrospinal fluid

(ACSF). ACSF contained 120mM NaCl, 3 mM KCl, 1.0 mM NaH2PO4, 1.5 mM MgSO4*7H20,

2.5 mM CaCl2, 25 mM NaHCO3 and 20mM glucose, and was bubbled with carbogen

(95%O2/5%CO2). Hippocampi were mounted in 2% Agar on a vibratome slicing chamber, with the long axis perpendicular to the direction of cut. Transverse hippocampal slices were then cut on a vibratome in 400µm sections. Slices were allowed to recover for at least 1 hour at the interface of physiological medium on ashless filter paper, humidified with carbogen.

Acute slice electrophysiology

After recovery, the subiculum and area CA1 were microdissected away from the dentate gyrus and area CA3 with two straight cuts, to isolate the temporoammonic-CA1 pathway for synaptic physiology and prevent spontaneous epileptiform discharge. Slices were transferred to a submersion-type recording chamber that was perfused at room temperature with ACSF

(1mL/min). ACSF was continuously bubbled with carbogen. Picrotoxin (100µM) and CGP52432 or CGP54626 (2µM) were included to block GABAA and GABAB receptors, respectively.

Concentric bipolar tungsten electrodes were placed in either SLM, to stimulate TA afferents, or

SR, to stimulate SC afferents (Figure 2.1). Capillary glass recording pipettes were filled with

ACSF (3-5MΩ), and placed >500µm from the stimulating electrodes. 100µs stimuli were delivered at a frequency of 0.05Hz. Stimulus intensity was generally set so that the elicited fEPSP was 0.1-0.2mV in amplitude. In AMPA:NMDA experiments, stimulus intensity was varied to elicit fEPSPs with fiber volleys from 0.05 to 0.5mV. All compounds used in field recording experiments were dissolved in ACSF and applied via the perfusion system. Concentrations of these compounds are noted at their final experimental concentration in ACSF applied to the slice.

Field EPSPs were recorded using an n.p.i. amplifier, filtered at 3kHz, and amplified 1000x prior to digitization at 10kHz.

Figure 2.1 – Transverse hippocampal slice preparation and electrode setup

For AMPA:NMDA experiments, MgCl2 was omitted from the ACSF. Electrode pairs were placed to record either SC-CA1 fEPSPs or TA-CA1 fEPSPs. Stimulus intensity was varied to elicit fEPSPs with fiber volleys from 0.05 to 0.5mV. 6-10 fEPSPs were recorded at each intensity and averaged post-hoc prior to analysis. Next, the AMPAR antagonist DNQX was washed in (50µM) for 10 minutes. After the response stabilized, the same range of stimulus intensities was again delivered to elicit an NMDAR-mediated response. After completion of the

43 experiment, the competitive NMDAR antagonist 2-Amino-5-phosphonopentanoic acid (APV,

80µM) was washed in to verify NMDAR responses. For analysis, the first 2ms of the initial slope of the fEPSP during the initial period was used an indicator of the strength of the AMPAR- mediated component of the fEPSP, while the subsequent 2-5 ms of the slope in the presence of

DNQX was used an indicator of the strength of the kinetically slower NMDAR-mediated component.

For LTP experiments TA-CA1 fEPSPs were elicited for 30 minutes as a baseline period, at a stimulus intensity of 150% of the lowest threshold that elicited a post-synaptic response. To induce LTP, a high frequency stimulus train was delivered (four 1 second-long trains at 100 Hz, delivered one minute apart). Stimulation then resumed at 0.05 Hz.

In some experiments, a serotonin receptor antagonist or a serotonin reuptake inhibitor was used while continuously stimulating at 0.05Hz. In these experiments, after a 30 minute baseline period, drugs were perfused over the slice via the ACSF. For anpirtoline (50 µM,

Tocris), a 5HT1BR agonist, and Fluoxetine (10 µM, National Institutes of Health), a selective serotonin reuptake inhibitor, drug was applied for 60 minutes. Slices were then washed with drug- free ACSF for 60 minutes. For imipramine (2 µM, Sigma-Aldrich), a tri-cyclic antidepressant, the drug was applied for 120 minutes, followed by 60 minutes of washout. Isamoltane (10 µM,

Tocris) has a 27-fold higher affinity for 5HT1BRs (IC50=39nmol/L) compared to 5HT1ARs

(IC50=1070nmol/L)(Waldmeier et al., 1988), and was continuously perfused throughout the entire recording when it was used to test the necessity of 5HT1BR signaling for drug- or HFS-induced potentiation.

Western blotting

Western blotting was used to quantify the expression of specific synaptic proteins and measure their phosphorylation states. After the recovery period following sectioning, slices were incubated in oxygenated control drug-free ACSF for at least 5 minutes at room temperature. In some experiments, slices were incubated in either fluoxetine- or anpirtoline-containing ACSF.

Slices were then placed on glass slides resting on dry ice to quickly freeze, and micropunches

(1mm in diameter) were targeted to either SLM, SR, or SP in area CA1. Punches were pooled (~3 punches per sample) and homogenized in 80µL of ice-cold lysis buffer (protease and phosphatase inhibitor cocktail, Sigma-Aldrich). Homogenates were kept at -80°C.

Protein quantification for each sample was performed using a standard Bradford Assay

(Coomassie reagent, Thermo Scientific, Rockford, IL). 10µg of protein was mixed with sample buffer (Laemmli, Sigma-Aldrich), boiled, and loaded into 4-12% Bis-Tris gel (Novex, Life

Technologies). Control lanes were included in all experiments. After running in 1X NuPage

MOPS SDS running buffer, the gel was transferred onto PVDF membranes in 1X NuPage transfer buffer in 10% methanol. The PVDF membrane was blocked with 0.5-2.5% nonfat dry milk in buffer containing 1M Tris-buffered saline and 0.05% Tween 20. Membranes were then probed with primary antibodies for 1 hour at room temperature, rinsed with TBS-Tween, then incubated in secondary antibodies, before being developed with enhanced chemiluminescence

(SuperSignal™ West Pico Chemiluminescent Substrate, Life Technologies). If they were first probed for phosphoproteins, membranes were stripped and reprobed with pan-protein antibodies.

Phospho-specific antibodies used were Ser831-phosphorylated GluA1 (1:1000, Thermo

Scientific, Pierce) and Thr286-phosphorylated CaMKII (1:3000; Cell Signaling). Pan-antibodies used were rabbit anti-GluA1 (0.5 µg/ml; Chemicon), anti-glucocorticoid receptor (GR, 1:1000;

Millipore), anti-CaMKII (1:2000, Cell Signaling), and anti-β–actin (1:5000; Cell Signaling), and an HRP-linked anti-rabbit IgG secondary antibody (1:1000; Cell Signaling). Total expression is quantified as signal intensity normalized to β-actin, while phospho-expression is quantified as phospho-protein normalized to total protein (e.g. pS831-GluA1:total GluA1). Densitometric data is transformed, so I treated it as ordinal for statistical analysis.

Chronic unpredictable stress (CUS) procedure

Rats were exposed to 2 stressors per day for 3-4 weeks, as described previously (Willner,

Muscat, & Papp, 1992). Light cycle stressors included restraint (30 minutes), strobe light (30

45 minutes), forced swim in cold water (5 minutes), cage rotation (3 hours), cage tilt (3 hours), white noise (3 hours). Overnight stressors included food deprivation (16 hr), overnight social isolation

(16 hr), and overnight water deprivation (16hr). In metyrapone experiments, overnight stressors were not used, except for food deprivation preceding novelty-suppressed feeding, and single housing during sucrose preference tests.

Sucrose preference

Sucrose preference tests were generally performed weekly on all rats. While singly housed for a single dark cycle, rats were presented with two bottles, containing either tap water or a dilute sucrose solution for 16 hours. Rats were trained once with 2% sucrose, then tested with

1% sucrose for subsequent tests. Tests lasted from 2 hours before the dark cycle until two hours after. Single housing during this period is potentially a stressor. Data are quantified as sucrose solution consumed as a percent of total fluid consumption. In each subsequent test, the location of the sucrose-containing bottle was alternated relative to the water bottle, to eliminate a learned side-preference. The bottles (Critter Canteen, Leith Petwerks, Salem OR) used were equipped with spring-loaded ball-bearing spouts to minimize leakage.

Novelty-suppressed feeding (NSF)

Rat chow pellets were placed in the center of a brightly lit arena in a dark room, as described previously (Dulawa, 2009; Santarelli et al., 2003). Control and treated rats were placed in the corner of the arena and latency to feed was measured. After one bite, the rat was returned to the familiar home cage to feed ad libitum for 5 minutes, to ensure sufficient hunger (all consumed

>0.05g). Rats that did not feed in the arena were assigned the maximum time allowed. This data was therefore treated as ordinal. The chamber was cleaned between rats. In the chronic CORT experiment, rats were first food deprived for 24 hours to instigate feeding behavior, and the maximum time was 400 seconds. In the CUS±MET experiment, rats were food deprived for 16 hours (as part of the CUS paradigm) and the maximum time was 600 seconds.

Statistics

Data are presented as mean ± SEM. Statistics were calculated using SPSS (IBM, Armonk,

NY) and Graphpad (Graphpad Software, Inc., La Jolla, CA). All data tested with parametric tests

(t-tests for 2 groups, ANOVA for >2 with Bonferroni post-hocs) were normally distributed and homoscedastic. Non-parametric statistics were used for ordinal data (NSF, Western blots).

3) 5HT1BR-mediated plasticity of TA-CA1 synapses

A portion of the work in this chapter has been published at Nature Neuroscience. My own contributions are discussed with the pronoun “I”. For the work done by my lab mates, I will use “we.”

Xiang Cai*, Angy J. Kallarackal*, Mark D. Kvarta, Sasha Goluskin, Kaitlin Gaylor, Aileen M. Bailey, Hey-Kyoung Lee, Richard L. Huganir & Scott M. Thompson. Local potentiation of excitatory synapses by serotonin and its alteration in rodent models of depression. Nature Neuroscience 2013 Apr;16(4):464-72. *- denotes equal contribution

Introduction

Current treatments for depression remain unsatisfactory, partly because the underlying biological changes in brain function in depression remain poorly understood. The discovery that changes in monoamine levels alter the affective state of patients led to the hypothesis that a dysfunction of monoamine signaling, particularly serotonin (5-HT), causes depression (Heninger et al., 1996). Elevation of serotonin levels with conventional antidepressants, such as SSRIs and

TCAs, may modulate neuronal excitability and plasticity by altering the transcription, translation, and phosphorylation state of target proteins (Krishnan and Nestler, 2008). Neither the principal neural circuits nor the key proteins that are modulated by serotonin are known, however.

Serotonin is capable of regulating the glutamatergic system, and evidence of changes in excitatory synaptic transmission in models of depression is accumulating (Sanacora et al., 2008;

Duman and Aghajanian, 2012), but it remains unclear how these two neurotransmitter systems interact. Strikingly, there is considerable overlap between molecular and neuronal changes due to antidepressant treatment and molecular mechanisms of synaptic plasticity.

Depression is likely caused by dysfunction in a variety of brain regions and cell types, including the hippocampus (Campbell and MacQueen, 2004). Altered hippocampal function may also influence adversely the activity of the cortex, amygdala, and other structures associated with reward, motivation, and emotionality and can contribute to the diverse and complex changes in neurobehavioral state that underlie depression in humans. For example, the hippocampal formation provides a major source of excitatory input to the NAc (Phillipson and Griffiths, 1985),

48 where chronic stress weakens AMPAR-mediated excitatory synaptic transmission (Lim et al.,

2012), which is a necessary deficit to induce anhedonia following chronic stress exposure.

To synthesize the emerging evidence of a critical role of synaptic dysfunction in depressive-like behaviors with the existing body of work on monoamine signaling, the hippocampus is one of the brain areas most inviting to study because of (1) thoroughly studied mechanisms of synaptic plasticity, (2) manifestation of antidepressant treatment by serotonergic modulation and ascribed cellular and molecular changes, (3) critical ties to other brain areas (e.g. the NAc) which are affected in depression and are likely to underlie symptoms, and (4) substantial vulnerability to stress corresponding with changes in depression and recapitulated in animal models.

As discussed in the introduction, the 1B subtype of serotonin receptor (5HT1BR) is an intriguing actor in regards to depression and antidepressant treatment. A h5HT1BR polymorphism has been identified that is associated with depression and substance abuse, but not other psychiatric disorders (Huang et al., 2003). In tissue from depressed patients, there is decreased p11 protein expression, which necessary for 5HT1B/DR surface expression (Svenningsson et al.,

2006). Above, I discussed a critical role for 5HT1BRs in animal models of antidepressant action.

Briefly, 5HT1BR knockout mice respond poorly to SSRIs in the FST (Mayorga et al., 2001;

Chenu et al., 2008), while wild-type mice exhibit antidepressant-like effects in response to administration of a 5HT1BR agonist (anpirtoline), even if their serotonergic neurons (and any effect of presynaptic autoreceptors) are pharmacologically eliminated (Schlicker et al., 1992;

Malagié et al., 2001). This suggests an attractive discrete role for 5HT1BRs activated heterosynaptically (Mathur et al., 2011) that may completely differ from the effect of inhibitory autoreceptors, which tend to inhibit excitability even in key cortico-striatal areas. 5HT1BRs are expressed postsynaptically on some dendrites, and can contribute to increased and decreased excitability (Berger and Huynh, 2002; Stean et al., 2005). Where might we look for 5HT1BR- mediated actions in the hippocampus?

5-HT1B receptor mRNA is present in CA1 cells at high levels and 5-HT1BR binding sites are concentrated in stratum lacunosum-moleculare (SLM) (Aït Amara et al., 1995; Doucet et al.,

1995; Sari et al., 1999), implying that SLM 5-HT1BRs are postsynaptically expressed at TA-CA1 synapses. In fact, there is evidence that serotonin interacts with mesolimbic dopaminergic system via postjunctional 5HT1B heteroreceptors in the ventral hippocampus, modulates activity of glutamatergic hippocampo-accumbens pathways, and secondarily alters DA in the NAc

(Boulenguez et al., 1996), demonstrating direct interaction with the greater cortico-mesolimbic circuitry. Meanwhile, there is a paucity of 5-HT1BR expression in other layers of CA1 and CA3 subfields, with virtually no expression detected in stratum radiatum, limning TA-CA1 as an interesting candidate at which to study 5HT1BR-mediated modulation and more broadly the role of serotonin and SSRIs.

Indeed, the highest concentration of serotoninergic fibers in the forebrain is SLM of hippocampal areas CA1 and CA3 (Ihara et al., 1988), where the axons of layer III neurons in the entorhinal cortex form excitatory synapses with the distal apical dendrites of pyramidal cells. The

TA pathway is required for some spatial recognition tasks and for long-term consolidation of spatial memory (Remondes and Schuman, 2004).

5HT1BRs are appealingly situated at the confluence of stress, serotonin, depressive-like behavior and antidepressant action at a major locus of synaptic plasticity, cognitive function, and affective regulation in a cortico-mesolimbic structure. In order to understand the function of serotonin and its role in antidepressant efficacy better, I studied its actions at TA-CA1 synapses, in conjunction with a study led by Xiang Cai and Angy Kallarackal (Cai et al., 2013).

Hypothesis 1: Monoaminergic 5HT1BR-mediated antidepressant efficacy is caused by increased excitatory synaptic transmission at cortico-mesolimbic sites via AMPAR phosphorylation by

CaMKII.

Prediction 1.1. AMPAR and CaMKII phosphorylation in the hippocampus and NAc will be increased following SSRI, TCA, or anpirtoline application.

Prediction 1.2. AMPAR and CaMKII phosphorylation following SSRI or anpirtoline application will coincide with synaptic potentiation at TA-CA1.

Methods

Western blots and electrophysiology experiments were performed as described in general methods. Antibodies used were anti-Ser831-phosphorylated GluR1 (1:1000; Chemicon), anti-

Thr286-phosphorylated CaMKII (1:5000; Promega), anti-GluR1 (0.5 µg/ml; Chemicon), and anti-

CaMKII (1:2000; Cell Signaling Technology).

Sucrose preference test. SPTs were performed somewhat differently than elsewhere in this document. After 3 hours of water deprivation, rats were given a choice between two bottles for 3-4 hrs, one with 1% sucrose solution and another with normal drinking water. To prevent the possible effects of side preference in drinking, the position of the bottles was reversed after 2 hrs.

Mice were tested over 12 hours (6 hrs light/6hrs dark) with bottle positions switched after 6 hrs.

The consumption of water and sucrose were measured by weighing the bottles. Preference for sucrose was calculated as a percentage of consumed sucrose-containing solution relative to the total amount of liquid intake. 50% means that they drank equally from both bottles, i.e. they had no preference for sucrose. Naïve animals were tested repeatedly in group housing until the cage had consistent preference for 1% sucrose, but were always tested individually thereafter. For experiments with only wild type animals, any individual that did not demonstrate a sucrose preference >65% was discarded. In some experiments, animals were first trained to the task using

2% sucrose. Tests were repeated once per week during the course of CUS or antidepressant administration.

Antidepressant treatment. Animals were given antidepressants via their drinking water in order to minimize stress. The concentrations of antidepressants were: imipramine, 100 mg/liter; fluoxetine, 80 mg/liter. Animals were housed singly and drinking water was changed every 3 days. Animals were given antidepressants continually for 3 - 4 weeks. Control animals received

51 water only. Experiments were then performed and analyzed with the experimenter blinded to the experimental condition of the animals.

Data analysis. Experiments were performed with the investigators blinded to the genotype of the animals or the identity of the substance being applied whenever possible. The blind was not broken until data analysis was complete. For quantification of anpirtoline and antidepressant actions in electrophysiological experiments, fEPSP or EPSC slope values were averaged and quantified over a three minute period preceding application of the substance

(control) and a three minute period at the end of the substance application (effect).

Results

In our study, field excitatory post synaptic potentials (fEPSPs) were recorded in SLM of area CA1 of acutely prepared rat hippocampal brain slices in response to stimulation of the TA pathway. Bath application of the tricyclic antidepressant imipramine (2 µM) or the SSRIs fluoxetine (20 µM) or citalopram (10 µM), in order to acutely elevate extracellular serotonin, all increased the slope of TA-CA1 fEPSPs over the course of ~60 minutes of drug application. We asked whether 5HT1BRs mediated this potentiation, by pretreating slices with 5HT1BR antagonist isamoltane (10 µM), and observed no effect of imipramine application. My colleagues observed that in mice lacking 5HT1BRs, imipramine and fluoxetine failed to have any effect. The 5HT1BR- selective agonist anpirtoline (50 µM) on the other hand, did increase the slope of TA-CA1 fEPSPs. Unlike the acute potentiating effects of the monoaminergic antidepressants, the effects of anpirtoline on fEPSP slope were fully reversed after 60 min of drug wash-out Figure 3.1.

Anpirtoline-induced potentiation was unaffected by depletion of serotonin with p- chlorophenylalanine, indicating that anpirtoline’s effect was mediated by postsynaptic heteroreceptors and not from an action on presynaptic inhibitory autoreceptors on serotonergic nerve terminals.

Figure 3.1 - Potentiation of TA-CA1 fEPSPs is irreversible in response to endogenous serotonin elevation, but transient following selective 5HT1BR activation Single represantative experiments recording fEPSPs at TA-CA1 in acute hippocampal slices in response to bath applied drugs. Data from these experiments demonstrate reversibility of potentiation induced by the 5HT1BR-selective agonist anpirtoline (50µM, filled circles) in response to drug washout, and irreversibility of potentiation induced by endogenous serotonin elevation with the traditional antidepressant imipramine (10µM, open circles).

I asked whether the persistent potentiating effect of increasing endogenous serotonin acutely with antidepressant application in vitro was different from the reversible effect of anpirtoline application because of persistent increase in serotonin levels. An alternate explanation is that endogenous serotonin and anpirtoline differ in the manner in which they activate 5HT1BR.

Another alternative is that endogenous serotonin activates additional serotonin receptors that result in the sustained potentiation observed with SSRI or TCA application. To test this, I applied isamoltane only after imipramine-induced potentiation occurred and imipramine had been washed out. If serotonin remained elevated and this was the cause of the seeming irreversibility of the potentiation, then isamoltane would antagonize 5HT1BRs that were being continuously activated at this population of synapses. Following imipramine-induced potentiation and an hour of washout, isamoltane application (10 µM) decreased the potentiation (Figure 3.2), suggesting persistent elevation of serotonin and resulting 5HT1BR-mediated signaling contributes at least in

53 part to the irreversible effect of elevating endogenous serotonin, compared to the reversible action of 5HT1BR agonists.

Figure 3.2- 5HT1BR antagonist isamoltane reverses persistent potentiation of TA-CA1 fEPSPs by imipramine. Bath application of the tricyclic antidepressant imipramine potentiates fEPSPs at TA-CA1 measured by slope. The 5-HT1BR antagonist isamoltane reverses the normally irreversible potentiation caused by elevation of endogenous serotonin, suggesting a role for persistent presence of serotonin and its action at 5-HT1BR that can be competed against by antagonism.

We then asked whether 5-HT1BR-mediated potentiation of TA-CA1 fEPSPs is mediated pre- or postsynaptically. Whole-cell recordings of CA1 cells following pairs of TA stimulation separated by 50 ms demonstrated that 5-HT1BR-mediated potentiation occurred with no change in paired-pulse ratio, an inversely proportional proxy for release probability, nor any change in fiber volley, a measure of the number of pre-synaptic fibers stimulated. We concluded that 5-HT1BR- induced potentiation is thus inconsistent with classical presynaptic expression mechanisms.

Following anpirtoline application in other whole-cell recordings, photolysis of caged glutamate demonstrated a post-synaptic increase in AMPAR function in SLM, but not SR, indicating that activation of 5-HT1BRs thus specifically and selectively increases the number or efficacy of postsynaptic AMPA receptors in the distal apical dendrites of CA1 pyramidal neurons in a highly localized manner.

We hypothesized that 5HT1BR-mediated potentiation of AMPAR function at TA-CA1 synapses was caused by activation of calcium/calmodulin-dependent protein kinase (CaMKII), one of the kinases that can be activated by 5HT1BRs that has been growingly implicated in neuropsychiatric disease (Robison, 2014). Based on this hypothesis I predicted that AMPAR and

CaMKII phosphorylation would both be elevated along the time-course of potentiation. That is, following anpirtoline application, AMPAR phosphorylation at the CaMKII/PKC site S831 of the

GluA1 subunit would exhibit a reversible increase in phosphorylation, while following antidepressant application, the increase in phosphorylation would not reverse after wash. CaMKII is activated by increases in calcium, and autophosphorylates at the threonine 286 residue, which allows activation to persist beyond calcium elevation (Miller and Kennedy, 1986; Giese et al.,

1998). If CaMKII is responsible for phosphorylation of the GluA1 subunit, then its autophosphorylation at the T286 site should also increase along the same time course as GluA1. I tested this prediction with Western blotting of tissue micro-dissected from SLM following incubation of slices in ACSF for each of the following conditions: ACSF alone, 1 hour in anpirtoline, 1 hour in fluoxetine, 1 hour wash (regular ACSF) following 1 hour in anpirtoline, or

1 hour wash following 1 hour in fluoxetine. Acute administration of fluoxetine (50 μM, 60 min) resulted in an activation of CaMKII (Figure 3.3, left graph, n=6 samples each, Kruskal‐Wallis H,

χ2=10.986, df=4, p=0.027) and phosphorylation of S831 of GluA1 (Figure 3.3 right graph, n=4 samples each, χ2=10.602, df=4, p=0.031) in tissue from SLM that persists after 60 min of washout, whereas both recover after washout of anpirtoline (50 μM, 60 min; *=significantly different from control (post‐hoc Mann‐Whitney U tests, p<0.05). These phosphorylation levels are normalized to total pan-protein levels of the respective proteins (See General methods). In similarly treated tissue micro-dissected from SR, there was no increase in phosphorylation of

GluA1 at S831.

Figure 3.3 – Effect of anpirtoline, fluoxetine and wash on phosphorylation of S831 site of GluA1 and T286 site of CaMKII Western blot of tissue microdissected from SLM after hippocampal slices were incubated in ACSF±drug at different timepoints. Drug conditions (Anp-anpirtoline or Fluox-fluoxetine) are after 1 hour of incubation. Wash conditions are 1 hour of drug-free wash following 1 hour of incubation in drug-containing ACSF. Con-control in drug free ACSF. Anpirtoline reversibly increases CaMKII T286 and GluA1 S831 phosphorylation, while fluoxetine irreversibly increases phosphorylation, following the time course of their respective duration of potentiation of TA-CA1 synapses.

To answer the question of whether the same 5HT1BR-mediated mechanism can increase excitability in other cortico-mesolimbic areas, I examined the NAc in response to a similar paradigm. I took parasagittal acute brain slices, and incubated them in anpirtoline for 1 hour. In tissue punches taken from the NAc core, anpirtoline (50 µM for 60 min) increased phosphorylation of CaMKII T286 (Figure 3.4, left, Mann Whitney U z=2.09 p=0.04) and GluA1

S831 (Figure 3.4, right, Mann-Whitney U z=2.09 p=0.04; n=6 slices each). These results suggest that local regulation of synaptic excitation by serotonin occurs in multiple regions important for cognitive function and affect in the context of serotonin and antidepressant treatment.

Figure 3.4 – Anpirtoline increases CaMKII T286 and GluA1 S831 phosphorylation in NAc core Tissue punches were taken from parasaggital rat brain sections targeting NAc core, following an hour of incubation in ACSF with or without anpirtoline. Western blot with phospho-specific antibodies reveals increased CaMKII T286 and GluA1 S831 phosphorylation, relative to total amount of each respective protein.

CaMKII is a critical kinase enzyme involved in regulation of synaptic strength and induction of synaptic plasticity. Classically, following calcium entry and CaMKII activation, the kinase translocates to the synapse and effects potentiation by phosphorylating principal and auxiliary units of AMPARs in a synapse-specific manner (Lisman et al., 2012). More recently,

CaMKII has been shown to have important roles in structural plasticity and synaptic organization

(Hell, 2014). Regardless, CaMKII has an extensively studied role in conventional LTP, typically studied at SC-CA1 synapses.

Given this evidence of a converging pathway between serotonin-induced potentiation and mechanisms of synaptic plasticity, I tested whether anpirtoline-mediated potentiation of fEPSPs would occlude high-frequency stimulation-induced LTP. Anpirtoline has no effect on SC-CA1 fEPSPs (Figure 3.5, n=7), consistent with the lack of 5HT1BRs in SR. Furthermore, anpirtoline application does not alter LTP at SC-CA1 synapses (p>0.05 compared to drug-free control slices). The effect of anpirtoline on TA-CA1 fEPSP slope remained occluded after induction of

LTP, even when the stimulation intensity was decreased to return the fEPSP to the baseline level, to control for a potential ceiling effect (Figure 3.6, n=5 slices).

Figure 3.5 – Anpirtoline has no effect on SC-CA1 fEPSPs. Anpirtoline neither enhanced SC-CA1 fEPSPs nor occluded LTP of SC-CA1 synapses (grey; n=7 slices). LTP of SC-CA1 fEPSPs in control slices shown in black (n=5 slices). Only every other data point is plotted to allow the two data sets to be distinguished.

Figure 3.6 – Anpirtoline-induced potentiation of TA-CA1 fEPSPs is occluded by HFS-induced LTP The effect of anpirtoline on TA-CA1 fEPSP slope remained occluded after induction of LTP, even when the stimulation intensity was decreased to return the fEPSP to the baseline level (n=5 slices). Representative traces are shown below. Scale bar = 0.1 mV, 20 ms.

When I bath applied anpirtoline first, fEPSPs increased in slope as usual. After potentiation following 1 hour of anpirtoline application, I readjusted the stimulus intensity down to achieve a fEPSP slope similar to baseline at the previous intensity, and then applied a high frequency stimulus to induce LTP. As predicted by a shared mechanism between serotonergic potentiation and synaptic plasticity, 5HT1BR-mediated potentiation occluded subsequent induction of LTP (Figure 3.7, n=4 slices).

Figure 3.7 - HFS-induced LTP at TA-CA1 synapses is occluded by anpirtoline The effect of anpirtoline on TA-CA1 fEPSP slope still occluded induction of LTP, even when the stimulation intensity was decreased to return the fEPSP to the baseline level following anpirtoline-induced potentiation (n=4 slices).

Consistent with the hypothesis that monoaminergic 5HT1BR-mediated antidepressant efficacy is caused by increased excitatory synaptic transmission at cortico-mesolimbic sites via

AMPAR phosphorylation by CaMKII, I observed that GluA1 S831 and CaMKII T286 phosphorylation in the hippocampus and NAc was increased following SSRI, TCA, and anpirtoline application. Furthermore, the increased phosphoprotein detected by Western blot coincided with the time course of synaptic potentiation at TA-CA1.

TA-CA1 synapses are strongly activated during cognitive tasks, involving working and associative memory in primates, and including spatial and non-spatial tasks (Sybirska et al.,

2000). In the course of this investigation, my colleagues in our collaborator Aileen Bailey’s laboratory demonstrated an interaction between 5HT1BR-mediated serotonin signaling and long- term consolidation of spatial learning, which is a known function of TA-CA1 synapses

(Remondes and Schuman, 2004). Inhibiting 5HT1BR signaling improved spatial memory consolidation in a spatially-cued version of the Morris Water Maze (Cai et al., 2013). Briefly, rats were trained to find a hidden platform in a water maze over 10 training blocks. Latency to find the target platform decreased significantly between the first and last training trial, demonstrating that the rats learned the location of the platform. This latency remained short during a probe trial administered 24 hrs after the last training trial, demonstrating that the rats remembered the

60 platform location. Finally, this latency remained significantly shorter than naïve when tested in a consolidation probe trial 28 days after the final training trial, demonstrating memory consolidation. Rats administered a 5-HT1BR antagonist (SB216641, 4 mg/kg, i.p.) daily for 14 days starting after the 24 hr probe trial displayed a significantly shorter escape latency in the consolidation trial than the control group (injected daily with 0.9% NaCl, i.p.). Consistent with

–/– previous reports that 5-HT1BR knockout mice have improved short-term learning in the Morris water maze (Buhot et al., 2003), we conclude that activation of 5-HT1BRs by endogenous serotonin interacts with the memory consolidation process, perhaps through competition for synaptic plasticity-related processes between the non-specific serotonin-induced potentiation and the synapse specific potentiation of critical synapses for the memory.

Other critical findings from this investigation, towards which I had only a minor contribution in collecting, revealed a peculiar alteration in 5-HT1BR-mediated signaling in a chronic stress model of depressive-like behavior which was reversed by successful antidepressant treatment (Cai et al., 2013). A possible prediction of the disputed monoamine hypothesis of depression is that altered actions of serotonin could contribute to the neurobiological pathology underlying depressive symptoms. We therefore investigated serotonin-induced potentiation in rats subjected to CUS (Willner et al., 1987), which caused rats to exhibit both a significant decrease in sucrose preference and increased latency in novelty-suppressed feeding, as reported previously for rodent models of depressive-like behavior (Dulawa, 2009). We observed that the anpirtoline- induced potentiation of TA-CA1 EPSPs was much greater in slices from CUS animals than in unstressed littermates, and that EPSP slope remained enhanced even after 100 minutes of washout in slices from rats subjected to CUS, but not in slices from stress-free control rats, in whole-cell and extracellular recordings (Figure 3.8 bottom, extracellular recordings). The persistence of the potentiation was independent of its magnitude, and was resistant to the 5-HT1BR antagonist isamoltane. Despite the increased magnitude of the response in CUS animals, the potentiation remained specific to the TA-CA1 synapse, with no effect at SC-CA1 synapses. We conclude that

61 endogenous serotonin can still be released from dorsal raphe afferents within the hippocampus in animals subjected to CUS, and that the resting level of extracellular serotonin is still regulated by the serotonin transporter, but that serotonin’s downstream signaling at excitatory synapses is increased in animals subjected to CUS.

Because many antidepressants elevate extracellular serotonin levels and this chronic elevation of serotonin increases AMPA receptor phosphorylation in the hippocampus

(Svenningsson et al., 2002), we asked what effect chronic antidepressant treatment of stress-naïve rats might have on 5-HT1BR-mediated potentiation. In hippocampal slices from rats given imipramine or fluoxetine via their drinking water for at least 3 weeks, anpirtoline-induced potentiation of TA-CA1 fEPSP slope was absent (Figure 3.8 top, p>0.05, paired t-test vs before anpirtoline application for imipramine n=9 slices and fluoxetine n=5 slices; p<0.05 paired t-test vs before anpirtoline for controls n=7 slices). It is possible that this result is due to a ceiling effect of synaptic strengthening during treatment. Anpirtoline produced normal potentiation in slices from animals treated for only 36 hrs, insufficient time for most behavioral effects in rodents and depressed humans. Chronic antidepressant treatment thus prevents 5-HT1BR activation from potentiating TA-CA1 synapses; an effect directly opposite to the enhancement of 5-HT1BR- induced potentiation seen after CUS.

How does successful antidepressant treatment interact with the depressive-like state induced by CUS? We took slices from rats that had been chronically stressed to induce an anhedonia-like state in the SPT, and then we chronically treated them with the SSRI fluoxetine, resulting in rescue of normal hedonic behavior in SPT and NSF tasks after 3 weeks of treatment.

In hippocampal slices, this successful AD treatment was accompanied by restoration of normal magnitude, reversible potentiation of TA-CA1 synapses in response to anpirtoline (Figure 3.8 bottom). A brief period of fluoxetine administration (36 hrs), in contrast, failed to restore either sucrose preference or the reversible anpirtoline response in CUS rats. The response of TA-CA1

62 synapses to 5-HT1BR activation thus correlates with the affective state of the animals under several conditions: stress-naïve controls, chronically stressed by CUS, and AD-rescue after CUS.

Figure 3.8 – Effect of chronic stress and chronic fluoxetine anpirtoline-induced modulation of TA- CA1 fEPSPs (Top) In slices from control animals (black n=7 slices), anpirtoline induces a reversible potentiation of fEPSPs at TA-CA1, which was abolished by chronic TCA or SSRI antidepressant treatment (dark grey- fluoxetine n=5 slices, light grey- imipramine n=9 slices) p>0.05, paired t-test vs before anpirtoline for imipramine and fluoxetine; p<0.05 paired t-test vs before anpirtoline for controls). (Bottom) In slices from rats that have undergone chronic unpredictable stress, the response to anpirtoline becomes irreversible (washout period p<0.05 vs control, unpaired t-test). Successful rescue of behavior with chronic SSRI treatment beforehand rescues the anpirtoline response (washout period p>0.05 vs control, unpaired t-test). (Most of this experiment was performed by Angy Kallarackal; From Cai et al., 2013).

In this study, the final critical finding, to which I had no contribution, is that 5-HT1BR- dependent potentiation of GluA1 receptors is necessary for the therapeutic action of

–/– antidepressants. Using transgenic 5‐HT1BR knockout mice and littermate controls, my colleagues demonstrated that reversal of stress-induced deficits in sucrose preference could not be reversed by fluoxetine. Furthermore, in another mutant line of mice in which the S831 residue had been mutated to a phospho-null alanine residue, they demonstrated that these GluA1 S831A had depressive-like behavior in the SPT and NSFT, neither of which could be reversed by chronic fluoxetine treatment. We conclude that 5-HT1BR activation is necessary for the efficacy of SSRIs

63 in this behavioral test, consistent with previous evidence that acute activation of 5-HT1BRs is sufficient for an antidepressant-like effect of SSRIs in the forced swim test (O’Neill and Conway,

2001).

Taken together, 5-HT1BR activation and subsequent phosphorylation of GluA1 subunits by CaMKII is required for the ability of chronically administered antidepressants to reverse anhedonia in these behavioral assays. Because the sucrose preference test does not depend on the hippocampus, these experiments support the suggestion that serotonin-induced potentiation of excitatory synaptic transmission occurs in multiple corticomesolimbic brain regions involved in hedonic behaviors (Russo and Nestler, 2013).

Discussion

SSRIs, tricyclic antidepressants, agonists of 5-HT1BRs, and conventional LTP all potentiate excitatory synaptic transmission at TA-CA1 synapses and mutually occlude each other, suggesting that the molecular events triggered by serotonin and conventional antidepressants overlap with those underlying activity-dependent synaptic plasticity. Activation of 5-HT1BRs enhanced AMPA receptor-mediated synaptic excitation by activating CaMKII and causing phosphorylation of GluA1 at serine 831. Unlike LTP, 5-HT1BR-mediated potentiation was reversed rapidly after withdrawal of agonist. Given the similarity of the biochemical pathways, this difference is surprising and cannot be readily explained. Induction of LTP requires CaMKII activity, whereas maintenance of LTP does not (Buard et al., 2010). Therefore maintenance requires some process activated by LTP induction that 5-HT1BR activation does not recruit.

Further study of the second-messenger pathways downstream of this novel form of potentiation may reveal key biochemical steps distinguishing the induction of potentiation, which seems to occur via a common mechanism between these processes, from the maintenance of potentiation, which only occurs after induction of LTP but is not sustained by 5HT1BR activation.

Interestingly, sparse evidence exists about the role of 5HT1BR as a post-synaptic heteroreceptor. The autoreceptor mechanism is a well-studied mechanism (Sharp et al., 1989), which has been suggesting by some to explain the latency to clinical improvement with sustained monoaminergic antidepressant treatment, requiring downregulation of 5HT1 receptors and release of an autoregulatory serotonin feedback mechanism or “brake” on endogenous serotonergic signaling (Parsey et al., 2006; Albert and François, 2010). However, the 5HT1BR-mediated potentiation observed at TA-CA1 is mediated post-synaptically.

TA-CA1 synapses were potentiated by serotonin, but nearby SC-CA1 synapses were not.

The signaling processes underlying this phenomenon must therefore be highly localized in postsynaptic CA1 cell dendrites. Serotonin release can be predicted to change the flow of information through the hippocampal circuitry, favoring the so-called direct pathway

(temporoammonic) at the expense of the indirect pathway, by potentiating excitatory synaptic strength. A dysregulation of information flow due to altered serotonin receptor signaling or altered regulation of excitatory synapses may contribute to the cognitive dysfunction of depression, while intervention with SSRI or TCA may reverse these deficits. It remains to be determined whether depression is an effect of altered synaptic strength. That is, are synapses weakened by chronic stress, and restored by antidepressant treatment? Regardless, the findings in this chapter still implicate 5HT1BR-mediated phosphorylation of GluA1 by CaMKII as a necessary action of antidepressant effect. The most logical hypothesis is that synaptic strength in key areas is weakened by chronic stress and contributes to depressive-like behavior.

Antidepressant treatment then restores synaptic strength over time at these key synapses that underlie behavior.

Animals subjected to chronic stress exhibit elevated glucocorticoid levels, and elevated glucocorticoids damage neuronal structure and function in the hippocampus (Krugers et al.,

2010b), including impairing Schaffer collateral LTP. We observed that serotonin-induced potentiation of TA-CA1 synapses was increased in magnitude and became irreversible. A

65 decrease in initial synaptic strength, which has been described in the PFC (Yuen et al., 2012), could potentially account for the increase in the magnitude of the potentiation. Chronic antidepressant administration in naïve animals decreased the potentiation and restoration of sucrose preference in CUS animals treated chronically with antidepressants was accompanied by a restoration of the normal, transient potentiation. Short-term administration of antidepressants had no effect on serotonin-induced potentiation in either naïve or CUS animals. The response of

TA-CA1 synapses to 5-HT1BR activation correlates well with the affective state of the animals under five experimental conditions, and thus represents a unique readout of this model of depression.

The normal response of TA-CA1 to anpirtoline-mediated modulation was restored by chronic fluoxetine treatment after induction of a depressive-like neurobehavioral state. Control animals, assumed to represent a normal affective state, exhibit a modest, reversible potentiation of

TA-CA1 synaptic strength in response to 5-HT1RR activation. Animals exhibiting anhedonia following CUS exhibit a large, irreversible potentiation following anpirtoline application.

Unstressed animals chronically treated with fluoxetine or imipramine with a typical length of time necessary for clinical latency (several weeks) exhibit no response to anpirtoline. These results might tempt the conclusion that antidepressant efficacy is caused by elimination of the ability of synapses to respond to 5-HT1RR modulation. However, successful AD treatment of the stressed depressive-like condition with fluoxetine restored a normal response to anpirtoline, a qualitative and quantitative difference from the result of fluoxetine treatment in stress-naïve animals. This demonstrates the importance of recognizing that effects of antidepressants on normal animals cannot always be generalized to the effects of antidepressants on complex neurobiological systems in a depressive-like neurobehavioral state.

How does serotonergic potentiation of hippocampal synapses contribute to reversal of a depressive-like state? Evidence is increasing that excitatory synapses are altered in models of depression (Duman and Aghajanian, 2012; Yuen et al., 2012). Depression of AMPAR-mediated

66 excitation of neurons in the reward circuitry of the nucleus accumbens (NAc) is required for the genesis of some depressive-like behaviors after chronic restraint stress, including anhedonia in the sucrose preference test (Lim et al., 2012). The source of the inputs that become depressed was not identified, but the NAc receives prominent projections from the hippocampus and frontal cortex. The serotonin-mediated plasticity we describe here may act to counteract this depression of NAc excitatory synapses in two ways, either via neuromodulator-evoked changes extrinsic to the NAc (elsewhere in the corticomesolimbic circuitry) or via changes to intrinsic properties of

NAc neurons like synaptic strength (Johnson et al., 2011). First, serotonin-induced potentiation of

TA-CA1 synapses increases hippocampal action potential discharge (Remondes and Schuman,

2002; Cai et al., 2013) and should increase the excitatory drive onto NAc cells, promote de- depression of NAc inputs, and thereby act to reverse anhedonia and restore the rewarding values of peripheral stimuli, such as sucrose. Consistent with this suggestion, activation of ventral hippocampal 5-HT1BRs has been shown to increase dopamine levels in the NAc (Boulenguez et al., 1996). Increased dopamine alleviates depressive-like behaviors (Chaudhury et al., 2013). The enhanced and persistent action of serotonin in the CUS animals (Figure 3.8) may be interpreted not only as a unique stress-induced form of plasticity of excitatory synapses, but also as a potential compensatory response, acting to promote excitatory throughput via hippocampal-NAc inputs. Second, medium spiny cells in the NAc express high levels of mRNA for 5-HT1BRs. If the plasticity we describe here also occurs in NAc D1-expressing neurons, then direct potentiation of excitatory synapses should counteract stress-induced depression of excitatory synapses (Lim et al., 2012). Consistent with this possibility, I observed a 5-HT1BR-induced activation of CaMKII and phosphorylation of S831 of GluA1 receptors in the NAc (Figure 3.4).

Acute application of SSRIs produced a rapid potentiation of excitatory synaptic transmission in brain slices, whereas the therapeutic action of SSRIs in depressed humans and in animal models of depression requires several weeks. It is possible that induction of potentiation is a critical first step that improves affective state by promoting action potential firing which

67 subsequently triggers other slower, synergistic activity-dependent processes, such as growth factor signaling, neurogenesis, and changes in gene expression (Martinowich and Lu, 2008;

Duman and Aghajanian, 2012). Repeated bouts of potentiation may be required to produce lasting changes, as is observed with deep brain stimulation and electroconvulsive therapy.

The plasticity we describe here for TA-CA1 synapses is attractive as a potential explanation for some of serotonin’s therapeutic actions because it is generalizable, which may, in principle, take place at any excitatory synapse at which 5-HT1BRs are expressed. Potentially, this mechanism could thus re-normalize activity in almost any region in the depressed brain that responds similarly to serotonergic modulation via 5-HT1BR activation. This is in contrast to the well documented ability of serotonin to promote dentate gyrus neurogenesis, which does not occur in any other brain region, but has been posited as a key antidepressant action (Malberg et al., 2000;

Warner-Schmidt and Duman, 2006). The TA-CA1 synapse serves as a useful archetype to study the actions of serotonin on excitatory synaptic transmission and the correlates of a depression-like neurobehavioral state.

Potentiation of excitatory synaptic transmission by serotonin in multiple brain regions involved in cognitive function, and its bidirectional alteration by stress and chronic antidepressants (Pittenger and Duman, 2008; Popoli et al., 2012), provides a novel link between the pathology of depression and its treatment, and thus represents a new perspective on depression and antidepressant action. This emerging view begins to give direction in filling the void left by the weaknesses of the monoamine hypothesis of depression (Kallarackal, 2011).

4) Chronic stress selectively decreases AMPAR-mediated excitation at temporoammonic-CA1 synapses

A portion of the work in this chapter has been published at Journal of Neuroscience. Figures shown here are data that I personally collected in whole or at least in part (i.e. some but not all data points in a group average).

Angy J. Kallarackal*, Mark D. Kvarta*, Erin Cammarata, Leelah Jaberi, Xiang Cai, Aileen M. Bailey, Scott M. Thompson. Chronic stress induces a selective decrease in AMPA receptor-mediated synaptic excitation at hippocampal temporoammonic-CA1 synapses. Journal of Neuroscience 2013 Oct 2;33(40):15669-74. *- denotes equal contribution

Introduction

In the previous chapter, I discussed the finding that following chronic stress, in animals exhibiting an anhedonia-like phenotype, serotonergic modulation is qualitatively and quantitatively altered at serotonergically innervated excitatory synapses in SLM of CA1. Exposed to a 5HT1BR-selective agonist, anpirtoline, the TA-CA1 synapses that reside in SLM are reversibly potentiated, as measured by intracellular or extracellular recording techniques (Cai et al., 2013). Following induction of a depressive-like neurobehavioral state, however, anpirtoline induces an irreversible potentiation of excitatory transmission at TA-CA1 which is greater in magnitude than in slices from animals in a stress-naïve state (Figure 3.8). This potentiation correlates with increased GluA1 S831 phosphorylation and activation of CaMKII. Critically, the ability of fluoxetine to have an antidepressant effect depended on intact 5HT1BRs and the ability of the S831 residue of GluA1 to be phosphorylated.

These findings are consistent with previous studies demonstrating that augmenting synaptic serotonin can potentiate glutamatergic transmission. The synaptic fraction of GluA1 expression is increased by the TCA imipramine (Du et al., 2004), while membrane bound

AMPARs are increased by chronic fluoxetine treatment (Martinez-Turrillas et al., 2002).

However, it remains to be determined whether depression is an effect of altered synaptic strength at key synapses that may underlie depressive-like behavior and symptoms of depression, and that these increases in excitatory function from AD treatment are restorative and oppose an existing

69 deficit. A dysregulation of excitatory synapses may contribute to the cognitive and affective dysfunction of depression, while intervention with SSRI or TCA may reverse these deficits.

If excitatory synaptic function is strengthened by antidepressant treatment, is it weakened by chronic stress to contribute to depressive-like symptoms? In humans, there is data that glutamate/glutamine (glx) metabolism is altered in depressed patients following ECT, and that glx levels are increased in the cortex of depression patients and decreased in amygdala, but imbalance of the glx:GABA ratio is reversed by successful antidepressant treatment (Michael et al., 2003; Pfleiderer et al., 2003; Sanacora et al., 2004). In animal models, mice susceptible to chronic social defeat stress, measured by stress reactivity, express lower levels of GluA1 mRNA in CA1 and DG compared to resilient animals in response to chronic stress (Schmidt et al., 2010).

A SNP in GluA1 can also confer greater sensitivity to stress, while AMPAR potentiators (so- called “AMPAkines”) can have stress-protective or antidepressant effects (Li et al., 2001;

Schmidt et al., 2010). Finally, one of the most exciting findings in the translational neuroscience of depression over the last two decades has been the finding that ketamine, an NMDAR antagonist, exerts a rapid antidepressant effect in humans (Berman et al., 2000). This effect can be modeled in animals (Autry et al., 2011), induces synapse remodeling and upregulates GluA1 and PSD95 (Li et al., 2010), and its expression is blocked by AMPAR antagonists (Maeng et al.,

2008).

Our data together with these studies highlight an interaction between antidepressant efficacy and excitatory glutamatergic transmission and posit the question of whether antidepressants are reversing a deficit of excitatory synaptic strength that is induced by chronic stress and in depression. There is evidence in humans and animal models that chronic stress leads to an atrophy of pyramidal cell dendrites in the hippocampus and cortex and that this promotes a depressive-like neurobehavioral state. In the hippocampus, chronic stress causes a loss of distal apical dendritic branches in CA3 and CA1 pyramidal cells and a loss in the size and number of dendritic spines (Watanabe et al., 1992; Conrad et al., 1996; Magariños et al., 1996; Sousa et al.,

2000; Vyas et al., 2002). Although stress-induced impairment of excitatory synapses has been described in the medial prefrontal cortex (Yuen et al., 2012) and nucleus accumbens (Lim et al.,

2012), the evidence of changes in excitatory synaptic transmission in hippocampal pyramidal cells is ambiguous. In area CA3, chronic stress enhances synaptic currents mediated by NMDARs but has no effect on AMPARs (Kole et al., 2002). In contrast, chronic stress has no effect on basal synaptic strength at Schaffer collateral (SC) synapses in area CA1, but does impair LTP (Alfarez et al., 2003). The functional and physiological consequences of the previously characterized stress-induced dendritic atrophy are thus unclear. Interestingly, SC-CA1 synapses do not express

5HT1BRs, which is necessary for the antidepressant action of monoamines (Cai et al., 2013), and therefore SC-CA1 synapses do not have the capacity to exhibit the qualitative difference in

5HT1BR-mediated potentiation observed at TA-CA1 that corresponds with affective neurobehavioral state. This suggests SC synapses are not particularly important for monoaminergic antidepressant action and may not represent depression-related abnormalities underlying depressive-like symptoms, while maintaining TA-CA1 as a prime archetype for study.

I now test the hypothesis that excitatory synaptic transmission at temporoammonic (TA) synapses in the distal dendrites of CA1 pyramidal cells is altered by CUS and restored by chronic SSRI treatment, as a correlate of the behavioral manifestations of CUS. The CUS model causes an alteration of 5HT1BR modulation at TA-CA1 synapses, which are generally potentiated by monoaminergic antidepressants. If antidepressants potentiate synapses that are weakened by chronic stress, as hypothesized, then the TA-CA1 synapse should be weakened by the induction of a depressive-like neurobehavioral state.

Prediction 2.1: CUS will cause a decrease in AMPA:NMDA ratios at TA-CA1 synapses, which is reversed by chronic SSRI treatment.

Prediction 2:2: CUS will cause a decrease in expression of AMPAR subunit expression in SLM.

This down-regulation will be reversed by SSRI treatment.

Prediction 2.3: AMPAR-mediated synaptic strength at TA-CA1 will correlate with sucrose preference test scores.

Methods CUS was performed as described in Chapter 2.

AD treatment. Animals were given fluoxetine (80 mg/ml) via their drinking water in conjuction with the last 3-4 weeks of CUS. Animals were housed singly and drinking water was changed every 3 days. Control animals received water only. Experiments were then performed and analyzed with the experimenter blinded to the prior treatment.

Sucrose preference test. Rats were presented with two bottles containing either normal tap water or water containing 1% sucrose. Animals were first trained with both bottles while group housed. For each subsequent test, including during the baseline period, animals were individually housed. Bottles were inserted 3 hr before the beginning of their dark cycle and removed 3 hr after the beginning of their light cycle.

Novelty suppressed feeding. Novelty suppressed feeding tests were performed as previously described (Santarelli et al., 2003) in a brightly lit arena in a dimly lit room. Food pellets were placed on white paper in the center of the box. Rats were food deprived for 24 hours.

A maximum time allowance was set at 600s. Immediately after the test, animals were returned to their home cage and allowed to feed for 5 min. Food pellets were weighed before and after the 5 min, and the amount of food consumed was calculated.

Acute slice electrophysiology. Generally performed as described in Chapter 2. For measuring AMPA:NMDA ratios, saline was prepared without MgCl2 in order to produce a robust

NMDAR-mediated component of the fEPSP. Three consecutive responses, corresponding to a minute of recording, were averaged and the fEPSP slope was calculated over a 1-3ms window.

For quantification of AMPAR-mediated responses, the window was fixed in the initial rising phase of the response, 2-5 ms after its initiation. For quantification of NMDAR-mediated responses, the AMPAR selective antagonist DNQX (50 µM) was added to the saline for 15 min and the slope was calculated over a 3-5 ms window in the rising phase of the response, to account for the slower kinetics of the NMDA receptor. The AMPA:NMDA ratio was calculated as a simple ratio between these two slopes at the same fiber volley (~0.15mV) before and after DNQX application.

Western blotting. Protein levels were quantified using Western blotting as discussed in

Chapters 2 and 3. See Appendix A for antibody details. Levels of proteins are expressed as the ratio of intensity normalized to β-actin intensity for statistical analyses.

Results

All CUS rats analyzed had sucrose preferences <65% and all control and fluoxetine- treated animals had >65% sucrose preference. That is, we verified that all animals that had undergone CUS were indeed exhibiting a depressive-like neurobehavioral phenotype, and animals that were treated with antidepressants indeed had a measure of restored affective state.

Selective attenuation of AMPAR-mediated responses at TA-CA1 synapses in rats subjected to CUS

We first compared extracellularly recorded fEPSPs elicited in response to stimulation of

TA afferents in SLM of CA1 in acute hippocampal slices from control rats, rats subjected to

CUS, and rats given 3 weeks of fluoxetine treatment during the last three weeks of 6 weeks of

CUS. Sucrose preference tests were performed in all animals.

We observed no significant difference in the paired-pulse ratio (PPR) of TA-CA1 fEPSPs in slices from control and CUS animals (control = 1.51 ± 0.1, CUS = 1.44 ± 0.1; n=19,14; p>0.5,

T-test), suggesting that presynaptic release probability at TA nerve terminals was not affected by chronic stress. This was consistent with other studies examining the effect of chronic stress on

73 excitatory synapses, which reported similar lack of effect on release probability in CA3- commisural connections and in mPFC (Kole et al., 2002; Yuen et al., 2012). We normalized

AMPAR-mediated synaptic responses by comparing the slopes of the AMPAR- and NMDAR- mediated components at a stimulation intensity producing identical fiber volleys of ~0.15 mV before and after application of the AMPAR antagonist DNQX. The AMPA:NMDA ratio was significantly lower in slices from CUS animals compared to controls (Figure 4.1, n=8 control, 6

CUS; p<0.01, Mann-Whitney U-test). The difference in AMPA:NMDA ratio did not result from a difference in NMDAR-mediated responses because the relationship between NMDAR- mediated responses and fiber volley amplitude was not significantly different in any group.

Figure 4.1 – CUS reduces AMPA:NMDA transmission at TA-CA1 synapses, which is reversed by chronic fluoxetine treatment

CUS can induce anhedonia in the SPT, consistent with a depressive-like neurobehavioral state, and this can be reversed by chronic, but not acute, antidepressant administration. This antidepressant effect is specific for drugs that have clinical antidepressant efficacy (Papp et al.,

1996). In slices from animals with a strong sucrose preference that were administered the SSRI fluoxetine following CUS, the AMPA:fiber volley and AMPA:NMDA ratios were not significantly different in slices from control rats and slices from rats subjected to CUS and then administered fluoxetine (Figure 4.1, p>0.05, Mann-Whitney U test. n=8,8 slices).

To determine whether the decrease in AMPAR-mediated excitation was due to stress- induced changes in receptor number or function, we harvested tissue samples from the SLM region of slices from controls, rats subjected to CUS, and rats subjected to CUS and then treated with fluoxetine. A loss of TA-CA1 AMPAR-mediated transmission, with no change in NMDAR- mediated transmission, might imply a loss of AMPAR expression in SLM following CUS.

Indeed, I observed a significant decrease in GluA1 protein in SLM in CUS tissue compared to control tissue and fluoxetine-treated CUS tissue (Figure 4.2, n=6,6,4 slices), whereas GluN1 expression was unaffected (Figure 4.3). Consistent with previously described atrophy of dendritic spines, there was a corresponding significant decrease in expression of the postsynaptic protein,

PSD-95 in SLM in CUS tissue compared to controls (Figure 4.3).

Figure 4.2 - CUS reduces GluA1 protein expression in SLM of chronically stressed rats exhibiting anhedonia, reversed by chronic fluoxetine treatment Western blots quantifying GluA1 expression in SLM, normalized to β–actin levels. Samples taken from SLM in slices from control rats (n=6), CUS rats (n=6), and CUS rats administered fluoxetine (CUS+AD, n=4). Representative blots (top) and group data (bottom) show that CUS causes a decrease in GluA1 protein expression, which is reversed by chronic fluoxetine treatment. *p<0.05 Kruskal-Wallis H test and Mann-Whitney U post-hoc.

Figure 4.3 - Western blots of SLM tissue from CUS rats reveal decrease of PSD-95 expression, but no effect on NR1 or GluA2 SLM tissue from CUS and control rats revealed that PSD-95 (n=5 control, 6 CUS), but not GluN1 (n=8, 9) or GluA2 (n=8, 9) expression, was decreased in SLM in CUS tissue compared to controls.

The rationale for examining the potential effect on synaptic strength of chronic stress was suggested by the results in Chapter 3 and from other studies in the literature that antidepressants cause an increase in synaptic strength. Specifically our findings pointed to TA-CA1, because of its unique expression and modulation by 5HT1BR, which is necessary for antidepressant action.

However, my hypothesis made no predictions about SC-CA1 synapses. In fact, previous studies have demonstrated a lack of effect of chronic stress on SC-CA1 synaptic strength(Alfarez et al.,

2003). Since 5HT1BR is not significantly expressed in SR leading to a lack of antidepressant effect at these synapses, I asked whether these SC-CA1 synapses would be unaffected by chronic stress, even though the synapses arise from the same pyramidal cell population.

In slices from CUS and control rats, I obtained tissue with micro-punches targeted to either SLM or SR in CA1, and quantified GluA1 expression. Remarkably, GluA1 was downregulated only in SLM in tissue from CUS rats, with no difference in expression observed in

SR between CUS and control rats. There was no significant difference in GluA1 between layers within control rats.

Figure 4.4 - Layer-specific downregulation of GluA1 protein following CUS GluA1 expression was decreased selectively in SLM (n=7,6) in CUS tissue compared to controls, but not in SR (n=5,5). *, p<0.05, Mann-Whitney U-test.

I compared AMPA:NMDA ratios at TA- and SC-CA1 synapses using a two pathway experimental design (see electrode positioning in Figure 2.1) to compare ratios within the same slices. Unlike at TA-CA1 synapses, which had a significant decrease in AMPA:NMDA ratio following CUS, there were no significant differences between SC-CA1 AMPA:NMDA ratios in control and CUS animals (Figure 4.5). This correlated with the lack of differences in GluA1 expression in SR (Figure 4.4).

Figure 4.5 - Layer-specific loss of AMPAR-mediated excitatory strength following CUS Mean AMPA:NMDA ratios for SC-CA1 and TA-CA1 fEPSPs in slices from control and CUS rats (Interaction synapse*group, F(1,8)=7.525, p=0.025). There was no difference between control and CUS at SC-CA1 synapses, whereas the ratio for TA-CA1 responses was significantly different than both CUS SC- CA1 responses and control TA-CA1 responses.

I further characterized the correlation between synaptic strength and chronic stress- induced changes in behavior by comparing AMPA:NMDA ratios at TA-CA1 synapses in slices with individual performance in the sucrose preference test. Rats subjected to CUS exhibited a significantly lower sucrose preference than naïve rats, as expected. This behavioral measure of anhedonia had a significant positive correlation with the TA-CA1 AMPA:NMDA ratios (Figure

4.6, R2=0.5741 p=0.025). Conversely, there was no significant correlation between SPT performance and the AMPA:NMDA ratios of SC-CA1 synapses (p=0.43).

Figure 4.6 - Correlation between SLM, but not SR, A:N ratio and SPT scores (left) Sucrose preference data for the cohort of rats used for the two-pathway experiment in the previous figure, before and after 3 weeks in control and CUS. The difference was significant for CUS (2x2 RM mixed ANOVA interaction group*time F(1,21)=23.33 p=0.0001, CUS at 3 weeks *p<0.05 vs. all other groups Bonferroni test). (right) Mean AMPA:NMDARs for TA-CA1 and SC-CA1 responses in each animal studied are plotted against the sucrose preference of the rat from which the slices were taken. TA- CA1 responses were correlated with sucrose preference (R2=0.5741, p=0.048), whereas SC-CA1 responses were not (R2=0.005, p=0.88)

In summary, chronic stress-induced changes in behavior are strongly correlated with decreases in AMPAR-mediated synaptic excitation at TA-CA1 synapses, but not SC-CA1 synapses, as the result of changes in AMPAR expression and that restoring normal behavior with chronic SSRI administration is accompanied by a recovery of AMPAR function and expression.

CUS does not alter TA-CA1 long-term potentiation

Dudek and Bear demonstrated that AMPAR-mediated excitation at SC-CA1 synapses can reach the same absolute magnitude of the LTP, regardless of whether that those synapses underwent LTD first or were relatively undisturbed. This means that normalized to their lower baseline state, weaker synapses can be potentiated to a greater relative magnitude (Dudek and

Bear, 1993). If this were true at CUS-weakened synapses, then it would be expected that TA-CA1 synapses from CUS rats would exhibit enhanced LTP relative to their own baseline. To determine whether the decrease in basal synaptic strength at TA-CA1 synapses of rats subjected to CUS would also be accompanied by a relative increase in LTP, a 100Hz stimulus (1s train x4) was applied to slices from control and CUS rats after a steady baseline period. There was no

79 difference in either the magnitude or the time course of LTP at TA-CA1 synapses in slices from

CUS rats compared to control rats in response to strong high frequency stimulation trains (Figure

4.7).

Figure 4.7 - CUS does not alter ability of TA-CA1 to undergo LTP TA-CA1 fEPSP slope is plotted before and after induction of LTP in slices from control (black, n=8 slices) and CUS rats (grey, n=11 slices).

CUS leads to a TA-CA1 synapse-specific cognitive deficit

TA-CA1 synapses are required for long-term consolidation of spatial memory (Remondes and Schuman, 2004). In the course of this investigation, my colleagues in our collaborator Aileen

Bailey’s laboratory tested whether subjecting animals to CUS after training them in a spatially cued version of the Morris water maze task would have any effect on memory consolidation, using an experimental protocol that I designed. They demonstrated that CUS, which we now know weakens TA-CA1 synapses by decreased GluA1 expression, approximates a lesion of TA-

CA1. That is, rats trained to learn the location of the hidden water maze platform successfully failed to consolidate the spatial memory of the platform location after 28 days if they underwent

CUS, but not if they were left undisturbed. The performance of individual animals in the consolidation trial in the water maze was correlated positively with their individual results in the

80 sucrose preference test (r(14)=0.50, p=0.07) and negatively with the novelty suppressed feeding test at the end of the experiment (r(14)=-0.74, p=0.002).

Discussion

This data demonstrates that dysfunctions of excitatory synaptic transmission at TA-CA1 result from chronic stress and correlate with impairment in cognitive function and affective state, consistent with the hypothesis that depressive behavior is caused by chronic-stress induced dysfunction of key excitatory synapses in the corticomesolimbic circuitry, specifically a loss of

AMPAR-mediated synaptic strength. As predicted by this hypothesis, I observed that CUS caused a decrease in AMPA:NMDA ratio at TA-CA1 synapses, which was reversed by chronic SSRI treatment. This decrease in AMPAR-mediated synaptic transmission was driven by a loss of

GluA1 subunit expression in SLM, which was also reversed by SSRI treatment. Finally,

AMPAR-mediated synaptic strength at TA-CA1 correlated with sucrose preference test scores, as predicted by my hypothesis.

Chronic stress has been well documented to alter the dendritic architecture of neurons, including hippocampal pyramidal cells (Magariños and McEwen, 1995a), but the functional consequences of these anatomical changes remain poorly understood, with minimal effects reported at the widely studied CA3-CA1 Schaffer Collateral synapses (Alfarez et al., 2003).

Three weeks of CUS led to a decrease in the strength of AMPAR-mediated synaptic transmission at the more distal CA1 apical synapses formed by the temporoammonic input from the entorhinal cortex in SLM. This synaptic weakening was specific to TA-CA1 synapses and was not detected at the more proximal SC-CA1 dendrites. Decreases in the expression of PSD-95 were observed in

SLM in CUS tissue, consistent with previous anatomical descriptions of atrophy of distal dendrites and dendritic spine shrinkage in CA3 and CA1 following chronic stress.

The effects of chronic stress were specifically limited to a downregulation of AMPARs, without affecting NMDAR-mediated transmission at the same population of SLM synapses. The

AMPAR-mediated component of the synaptic responses was decreased by roughly 50%, correlating with ~50% decrease in the expression of GluA1 subunits in SLM. Interestingly, this effect seemed to be subunit-specific, as there was no change in GluA2 expression. There was no difference in AMPA:NMDA after chronic stress at SC-CA1 synapses, and no change in GluA1 expression at SR. Consistent with our electrophysiological results discounting NMDARs from alteration, we observed no changes in GluN1 expression after CUS. In CA3 pyramidal cells, chronic restraint stress has been reported to increase NMDAR-mediated transmission (Kole et al.,

2002). It is possible that there is an inherent difference between CA3 and CA1 cells in their response to stress, that methods between studies biased towards subregions of the hippocampus

(i.e. dorsal vs. ventral), or that chronic restraint stress has different effects than chronic unpredictable stress. In support of the last point, the quality and variety of chronic stress has been directly evaluated in regards to HPA activation and structural plasticity in the hippocampus, with

HPA habituation of a single repeated chronic stressor, but continued HPA activation with varied stressors even after 3 weeks (Magariños and McEwen, 1995b). Those authors found atrophy in

CA3 with either stressor, but suggested that any repeated stressor (they restrained rodents for 6 hours daily) will ultimately be milder in terms of stress load. Regardless, my functional results support existing structural studies that distal synapses of pyramidal neurons are most stress- sensitive, while more proximal SC-CA1 synapses seem to be resistant to the effects of chronic stress. It is possible that few have reported effects of chronic stress in CA1 because most studies focus on the popular SC-CA1 synapses that have been used to so elegantly characterize mechanisms of synaptic plasticity, missing the more robust stress-induced effects observed more distally.

AMPAR-mediated excitation produced by chronic stress was not accompanied by an increase in the relative magnitude of LTP at TA-CA1 synapses, as would be predicted from previous studies on LTD-weakened SC-CA1 synapses (Dudek and Bear, 1993). As TA-CA1 synapses are weakened by chronic stress, but only increase by the same relative magnitude as

82 synapses in unstressed rats following induction of LTP, the absolute synaptic strength achieved by potentiation must be lower in CUS animals. That is, the maximal limit of synaptic strength appears to be reduced, but the machinery involved in the initiation and maintenance of LTP and the relative size of the pool of AMPARs that can be rapidly recruited for membrane insertion seems intact following chronic stress. The loss of maximal synaptic strength should impair the ability of TA inputs to trigger action potential output from area CA1, and also hamper the ability of TA-CA1 input to moderate SC-CA1 plasticity (Remondes and Schuman, 2002).

There is a compelling convergence of mechanisms of chronic stress leading to synaptic dysfunction in corticomesolimbic areas that should be key to depressive-like behavior. Chronic restraint stress causes a comparable decrease in synaptic excitation in layer V pyramidal cells in the medial prefrontal cortex (mPFC) that is mediated by proteosomal degradation of GluA1 subunits (Yuen et al., 2012). However, in mPFC both AMPAR- and NMDAR-mediated transmission equally affected, with GluA1 and GluN1 subunits downregulated in parallel. They observed no change in AMPAR function or GluA1 expression in the hippocampus with seven days of chronic restraint stress, a time at which GluA1 and GluN1 are decreased in mPFC, suggesting that TA-CA1 synapses may be more resistant to the effects of chronic stress than synapses in the mPFC. In the striatum, chronic restraint stress alters excitatory input (likely from

PFC hippocampus) to the NAc, and a loss of AMPAR-mediated strength is necessary and sufficient for anhedonia in mice (Lim et al., 2012).

Our collaborators demonstrated that CUS can approximate a lesion of TA-CA1 synapses, which impairs long-term spatial memory consolidation (Remondes and Schuman, 2004). After first demonstrating significant memory of the target location, rats subsequently subjected to CUS behaved as if they had never learned the task after 4 weeks, performing no better than they did prior to training. In conjunction with the effect of CUS on synaptic strength but intact relative-

LTP, these results suggest that intact transmission and a normal range of synaptic strength is required for TA-CA1 synapses to be able to carry out their role in consolidation of long-term

83 spatial memory. Impairment in spatial memory has also been described in depressed humans, using a videogame virtual reality task in a well-controlled laboratory setting (Gould et al., 2007).

Results of individual animals in the consolidation trial were well correlated with the behavior of the animals in the sucrose preference and novelty suppressed feeding tests, further highlighting the complex functional consequences of stress-induced synaptic weakening. There was also a robust correlation between the TA-CA1 AMPA:NMDA ratio in individual hippocampal slices with the SPT scores of rats from which those slices were harvested. Although animals maintain a high sucrose preference in this test even after bilateral hippocampal lesions

(Gilbert and Kesner, 2002), the correlation we observed in multiple distinct behaviors suggest that multiple types of synapses respond to chronic stress, consistent with results I have discussed on effects in the mPFC (Yuen et al., 2012) and nucleus accumbens (Lim et al., 2012).

In the previous chapter, I discussed how traditional antidepressants like SSRIs and TCAs exert an immediate potentiating action on excitatory synaptic transmission that is selective for

TA-CA1 synapses. In this chapter, I described directly opposite effects of chronic stress.

Consistent with the hypotheses presented thus far, that (1) antidepressant efficacy derives from the ability of serotonin to potentiate AMPAR-mediated excitatory strength, and (2) depressive- like behaviors are caused by loss of AMPAR-mediated transmission at key corticomesolimbic synapses, the data in this chapter demonstrates that chronic AD administration in stressed animals exerts a normalizing action on GluA1 expression and AMPAR-mediated excitation.

In chapter 3, I discussed how the potentiating action of ADs at excitatory synapses is necessary for the ability of traditional antidepressants to restore normal behaviors after chronic stress. Chronic monoaminergic antidepressant treatment increases AMPAR expression with a time course similar to therapeutic efficacy (Martinez-Turrillas et al., 2002), which is several weeks. Ketamine, however, exerts a rapid and lasting therapeutic response in depressed patients

(Berman et al., 2000) and animal models have revealed that NMDAR antagonists induce a rapid strengthening of excitatory synaptic transmission through changes in AMPARs and expression of

84 a variety of synaptic proteins (Autry et al., 2011; Li et al., 2011b). It remains to be explained why

SSRIs and TCAs must be administered to depressed patients for weeks to months before their symptoms are alleviated. Increased expression of GluA1 subunits may be required in order for the acute potentiating actions of serotonin to be able to exert lasting improvements in affect and mood. Additionally, the actions of serotonin may be more regionally restricted than NMDAR antagonism, providing a rationale for the more robust effect of ketamine action, as well as the more severe side effects associated with ketamine use.

Chronic stress causes a decrease in AMPAR function in specific subsets of synapses throughout the brain, of which the TA-CA1 synapse represents a useful stress-sensitive archetype.

Chronic stress certainly affects synapses in multiple brain regions, and depending on the anatomical location of the affected synapses on an individual basis, with individual genetic predisposition and environmental exposures, this may account for the diverse symptoms associated with depression. Although serotonergic agents, such as SSRIs, are capable of producing therapeutic actions in depressed patients, this fact does not necessarily indicate that a primary deficiency in serotonin synthesis or release is the cause of depression (remember

Serotonin Plumbers, Inc. from Chapter 1). Rather, AD-induced elevations in serotonin may exert antidepressant actions indirectly as a result of restoration of the strength of weakened excitatory synaptic transmission. Indeed, SSRIs can increase AMPAR phosphorylation and surface expression (Martinez-Turrillas et al., 2002; Svenningsson et al., 2002) and, here, chronic SSRI treatment was capable of rescuing AMPAR dysfunction in chronically stressed animals that exhibited a depressive-like phenotype. This data provides further rationale for the exploration of new non-monoaminergic therapeutic agents: compounds that directly and indirectly target the strength of excitatory synapses in key brain areas, such as AMPAkines and compounds that promote network excitability (See Chapter 6).

Merely identifying the deficit at excitatory synapses following chronic stress does not fully answer the question “How do we become depressed?”, nor the question “how does CUS cause a

85 depressive-like phenotype?” Many studies have indeed looked at the effects of chronic stress paradigms in the corticomesolimbic circuitry, including the hippocampus, and dozens of mediators of stress have been identified. Changes associated with stress have been attributed to corticosteroids, CRH, noradrenaline, serotonin, dopamine, urocortin, vasopressin, orexin, dynorphin, inflammatory cytokines, endocannabinoids, and neurosteroids (Joëls and Baram,

2009). Having previously identified some of the most robust stress-induced synaptic defects yet reported in CA1, I turned to identify the causative mediator of chronic stress.

5) Corticosterone mediates the synaptic and behavioral effects of chronic stress at TA-CA1 synapses

A portion of the work in this chapter has been provisionally accepted for publication at the Journal of Neurophysiology.

Mark D. Kvarta, Keighly E. Bradbrook, Hannah M. Dantrassy, Aileen M. Bailey, and Scott M. Thompson. Corticosterone mediates the synaptic and behavioral effects of chronic stress at rat hippocampal temporoammonic synapses, 2015. Provisional acceptance. Introduction

Having characterized key changes at excitatory synapses that occur in a chronic stress model of anhedonia and which correlate with depressive-like behavior at an archetypal stress- sensitive, serotonin-responsive synapse that models a necessary action of antidepressant treatment, I now examine how chronic stress translates to a depressive-like phenotype. What is it about chronic stress that alters synapses and contributes to behavior?

In chapter 1, I introduced the concepts of allostasis and allostatic load in relation to chronic stress. In health, the stress response increases fitness by promoting adaptation to maintain homeostasis in response to varying environmental demands in the vein of Walter Cannon and

Claude Bernard (the father and, perhaps, great-grandfather to the concepts of homeostasis).

Allostasis is the modern concept that homeostasis of body systems, including the CNS, operates in a dynamic range rather than a single ubiquitously ideal point, and that the stress response is a major controller that ensures the organism is operating at the appropriate set point as appropriate for environmental conditions (Sterling and Eyer, 1988; McEwen, 1998). Chronic activation of the systems underlying stress responses is detrimental and leads to allostatic overload and leads to disease states, especially in genetically susceptible individuals (McEwen, 2008).

A striking example that may illustrate this concept of the detrimental loss of dynamic range was illustrated in chapter 4, regarding the effect of CUS on TA-CA1 physiology. CUS induces a layer-specific decrease in AMPAR-mediated excitatory synaptic strength. However,

LTP induction at TA-CA1 with high frequency stimulus in slices from CUS rats resulted in a similar relative increase. Together, this likely indicates a lower absolute potentiation, in contrast

87 with previous results with classically depressed SC-CA1 synapses (Dudek and Bear, 1993).

Therefore chronic stress appears to have resulted in a loss of a normal dynamic range that correlated with a depressive-like neurobehavioral state. Further studies would be required to demonstrate that these synapses are similar enough to SC-CA1 synapses, as an alternative explanation for my results could be that a population of TA-CA1 synapses become silenced, rather than all synapses losing some AMPAR-mediated strength.

As discussed at length, chronic stress is a major risk factor for depression via altering brain structure and function, but the role of specific mediators in producing these changes remains unclear. Stress is one environmental trigger that increases the likelihood of, or even precipitates, depressive episodes (Billings et al., 1983; Holsboer, 2000; Anacker et al., 2011b).

Marked morphological, functional, and volumetric brain changes correlate with stress load, depressive episode duration, and response to antidepressants (Sheline, 2000; Sheline et al., 2003;

Koolschijn et al., 2009; Lorenzetti et al., 2009). Two of the more robust changes are dendritic atrophy and spine loss in the PFC and hippocampus (Christoffel et al., 2011). These two cortical areas send glutamatergic projections to reward areas, including the nucleus accumbens (Russo and Nestler, 2013), where activity is positively correlated with cortico-mesolimbic connectivity

(Downar et al., 2014). Conversely, aberrant function in these same regions is a pathophysiological feature of a core symptom of depression, anhedonia (Lim et al., 2012).

Understanding how chronic stress causes these changes is of utmost importance for understanding the etiology of depression and designing preventative and treatment strategies. In animal models, chronic stress triggers depressive-like changes in motivated reward behaviors, such as anhedonia, as well as synaptic and neuronal dysfunction. Chronic, but not acute, antidepressants restore both normal behavior and synaptic function in stressed animals [e.g. (Fales et al., 2009)]. In the rodent hippocampus, chronic stress decreases dendritic spine size and number, and decreases GluA1 mRNA in pyramidal cells, particularly in the most distal apical dendrites (Magariños and

McEwen, 1995b; Sousa et al., 2000; Schmidt et al., 2010).

Taken together with stress-induced decreases in AMPAR-mediated excitation in the prefrontal cortex (Yuen et al., 2012) and nucleus accumbens (Lim et al., 2012), stress may promote depressive symptoms by weakening excitatory synaptic transmission at multiple specific sites in cortico-mesolimbic reward circuitry (Thompson et al., 2015).

So how does chronic stress translate into these neurobiological correlates of a depression- like neurobehavioral phenotype: that is altered behavior and synaptic structure and function?

Chronic stress activates the sympathetic nervous system and the hypothalamic-pituitary-adrenal

(HPA) axis, together resulting in elevation of corticosteroids, epinephrine and norepinephrine, and other stress hormones (Pitman et al., 1988). HPA overactivation and dysregulation is particularly prevalent in depressed patients with melancholia (up to 90%), as identified using modified criteria relying centrally on anhedonia (Taylor and Fink, 2008). Chronic administration of exogenous corticosterone (CORT), the principal corticosteroid in rodents, produces behavioral changes resembling those seen after CUS (Gourley and Taylor, 2009). Furthermore, CORT has been shown to mimic the effects of chronic stress paradigms on dendritic structure in CA3, particularly distal dendrites (Magariños and McEwen, 1995a), and in PFC. However, the chapter

4 discussion highlighted the greater stress resilience of CA1 compared to PFC neurons and other pyramidal neurons in CA3. As there is a dissociation of stress (chapter 4) and antidepressant effects (chapter 3) between proximal and distal synapses even in the same population of neurons

(i.e. no effects observed on SC-CA1 synaptic strength, little change in morphology in other studies), it is possible that the chronic stress mediators are different. Finally, although studies have demonstrated that putative stress mediators can mimic stress effects in different regions of the brain, few demonstrate a critical role that is both sufficient and necessary. One notable, and relevant, exception is a study performed by Chen, Baram and colleagues, demonstrating a layer- specific mediation of spine derangement in SR by corticotropin-releasing hormone (CRH) following acute stress, with no effect at SLM (Chen et al., 2008). Studies that only demonstrate

89 sufficiency do not answer whether the mediator is only one of several or many that could subserve the same function.

Others have demonstrated a corticosterone induced atrophy of distal apical dendrites in

CA3, but reported no effect of chronic CORT injections on CA1 morphology (Woolley et al.,

1990), the same effects observed after chronic restraint stress (Watanabe et al., 1992). The effects of chronic stress in CA3 are blocked by cyanoketone, a steroid synthesis inhibitor. Thus, CORT has an important role in other pyramidal cells in mediating atrophic effects of chronic stress. CA1 neurons may be more vulnerable during early developmental periods as suggested by organotypic culture experiments (Alfarez et al., 2009). There is surprisingly little investigation beyond the structural effects of chronic stress in CA1, probably because structural results in SR have not been particularly compelling or inviting to further investigation. There is evidence of changes noted due to acute stress via glucocorticoids, mineralocorticoids, and adrenal catecholamines on

CA1 physiology [see reviews in (Alfarez et al., 2006; Krugers et al., 2012)], but the changes identified at TA-CA1 are specific to chronic stress. CRH is another stress mediator, but has only been shown to have acute effects, with no effects in SLM (Chen et al., 2008). In the PFC, dendritic remodeling has also been attributed to chronic effects of CORT (Cook and Wellman,

2004; Yuen et al., 2012). It is possible that reports of lack of effect of chronic stress, and therefore lack of identification of mediators of chronic stress in CA1, are based on focusing on SC-CA1 synapses in SR, which do not exhibit 5HT1BR-mediated potentiation necessary to monoaminergic antidepressant efficacy and which are not intrinsically altered following induction of a depressive-like state (Alfarez et al., 2003; Cai et al., 2013). The unique role of TA-CA1 in serotonergic innervation, antidepressant efficacy, and functional physiological alteration corresponding with a depressive-like neurobehavioral state invites more thorough investigation.

Hypothesis 3: I hypothesize that chronic stress alters TA-CA1 synapses and promotes depressive-like behavior via chronic stress-induced elevations of corticosterone. That is, corticosterone is both sufficient and necessary for the neurobiological correlates of chronic stress:

90 impaired TA-CA1 synaptic function, altered responsiveness to 5HT1BR agonists, anhedonia, and hyponeophagic behavior.

Prediction 3.2: If stress-induced elevations of corticosterone are necessary for the synaptic and behavioral effects of chronic stress, then blocking elevated corticosterone secretion throughout

CUS will prevent the neurobiological behavioral and synaptic correlates of chronic stress.

Materials and Methods

Subjects: Male Sprague-Dawley rats (Harlan Laboratories, 3-4 weeks old at start) kept on a regular 12 hour light/dark cycle were group housed. All rats sacrificed for in vitro electrophysiology and molecular biology experiments were 7-8 weeks old and were previously tested for sucrose preference and novelty-suppressed feeding. In the water maze experiment performed by collaborators in Dr. Aileen Bailey’s lab, rats obtained from Charles River were trained in the water maze at 3-4 weeks old. They were tested for memory consolidation at 7-8 weeks, and were tested weekly for sucrose preference.

Corticosterone administration: CORT was dissolved in tap water (50 μg/ml, see

Appendix for details) and administered in water bottles so as to elevate plasma CORT levels in a manner tied to diurnal activity cycles (Gourley and Taylor, 2009). This dose induces several depressive-like behaviors (reduced sucrose intake and impaired forced swim performance) over several weeks, paralleling chronic stress models. After a one-night baseline sucrose preference test, rats were randomized to receive either the CORT solution or regular tap water ad libitum for

3-4 weeks.

CUS and metyrapone (MET) administration: Rats were exposed to 2 stressors per light cycle for 3-4 weeks, as above. Stressors included restraint (30 minutes), strobe light (30 minutes), forced swim in cold water (5 minutes), cage rotation (3 hours), cage tilt (3 hours), white noise (3 hours), and overnight food deprivation. Rats were randomized to receive either MET (50mg/kg i.p.) or vehicle (VEH)(60% sterile saline and 40% PEG), prior to each stressor.

Sucrose preference test: While singly housed for a single dark cycle, control and treated rats were presented with two bottles, containing either tap water or a dilute sucrose solution for

16 hours. Rats were trained once with 2% sucrose (this data is not shown) then tested with 1% for baseline and all subsequent tests. Tests lasted from 2 hours before the dark cycle until two hours after. Single housing during this period is potentially a stressor. Data are quantified as sucrose solution consumed as a percent of total fluid consumption.

Novelty-suppressed feeding (NSF) test: Rat chow pellets were placed in the center of a brightly lit arena in a dark room, as described previously (Santarelli et al., 2003; Dulawa, 2009;

Cai et al., 2013). Control and treated rats were placed in the corner of the arena and latency to feed was measured. After one bite, the rat was returned to the familiar home cage to feed ad libitum for 5 minutes, to ensure sufficient hunger (all consumed >0.05g). Rats that did not feed in the arena were assigned the maximum time allowed. This data was therefore treated as ordinal.

The chamber was cleaned between rats. In the chronic CORT experiment (Figure 1), rats were first food deprived for 24 hours to instigate feeding behavior, and the maximum time was 400 seconds. In the CUS±MET experiment (Figure 4), rats were food deprived for 16 hours (as part of the CUS paradigm) and the maximum time was 600 seconds. Food deprivation preceding the task is potentially a stressor.

Acute slice electrophysiology: AMPA:NMDA ratios were recorded and quantified as in chapter 4, in Mg2+-free ACSF. Six to ten consecutive responses were averaged with fiber volley

(FV) amplitudes nearest to 0.2mV, which is in the linear range of responses to varying stimulus intensity (Kallarackal et al., 2013). When multiple recordings were made from slices from a

92 single animal, average AMPA:NMDA values in that animal were calculated to avoid the potential confound of nested data.

Western blotting: SLM or stratum pyramidale (SP) tissue punches (1 mm diameter) were taken from CA1 in hippocampal slices on a glass slide resting on dry ice, and deposited in standard lysis buffer (protease and phosphatase inhibitor cocktail, Sigma-Aldrich). Protein quantification for each sample was performed using a standard Bradford Assay (Coomassie reagent, Thermo Scientific, Rockford, IL). Antibodies used in this chapter (see Appendix for details) were rabbit anti-GluA1, anti-glucocorticoid receptor, and anti-β–actin, anti-5HT1BR, and an HRP-linked anti-rabbit IgG secondary antibody. Expression is quantified as signal intensity normalized to β-actin. Densitometric data is transformed, so I treated it as ordinal and performed only non-parametric tests.

Corticosterone quantification: Blood samples were collected from tail veins of MET or vehicle-treated rats during restraint stress. Samples were collected in EDTA microtubes (Greiner,

Bio-One North America Inc., Monroe, NC). After centrifugation, plasma CORT was quantified by radioimmunoassay (University of Virginia Ligand Core, Charlottesville, VA). For fecal pellets, enzyme-linked immunosorbent assays (ELISAs) were performed. Briefly, each rat was singly housed in a clean cage during the dark cycle. Two hours after light cycle onset, 2-4 fecal pellets were collected and pooled for each rat. CORT was extracted with ethanol. After ethanol evaporation, extracts were reconstituted with methanol and diluted 1:20 for ELISA analysis alongside CORT standards (Assaypro, Inc., St. Charles, MO). Fecal CORT reflects an integrated measure of several hours of prior serum corticosteroids. Stress-induced CORT in feces is observed within 6-12 hours after stress (Bamberg et al., 2001; Harper and Austad, 2012). This method is minimally invasive and is less susceptible to ultradian and circadian interference than single time points.

Morris Water Maze: Rats in Dr. Bailey’s lab were trained, as described previously

(Remondes and Schuman, 2004; Kallarackal et al., 2013). Rats received 10 blocks of training (4

93 trials/block) over 6 days, with the platform in a fixed location. One day after training, rats completed a probe trial, with path length and latency to target measured. They were then subjected to CUS for 3 weeks, and administered MET or VEH, as above, then left unstressed for

1 week. On day 28 after training, long-term consolidation was tested with a probe trial.

Statistics: Data are presented as mean ± SEM. Statistics were calculated using SPSS

(IBM, Armonk, NY) and Graphpad (Graphpad Software, Inc., La Jolla, CA). All data tested with parametric tests (t-tests for 2 groups, ANOVA for >2 with Bonferroni post-hocs) were normally distributed and homoscedastic. Non-parametric statistics were used for ordinal data.

Results

Is elevated CORT sufficient to generate the neurobiological correlates of chronic stress?

I first tested whether chronic exogenous administration of CORT for 21 days would mimic the effects of CUS in causing reduced sucrose preference and neohypophagia in unstressed animals.

The experimental timeline is shown in Figure 5.1. Chronic CORT reduced sucrose preference

(Figure 5.2; group*time 2x2 mixed ANOVA F(1,54)=8.461, n=57, p=0.005 2x2 mixed ANOVA, p<0.005 vs. all other groups), similar to previous work which reported reduced sucrose intake

(Gourley et al., 2008). The total amount of liquid consumed did not differ by group, suggesting no effect of CORT treatment on water balance. Similarly, after 2 weeks of treatment, CORT- treated animals exhibited a significantly increased latency to feed in the NSF test (Figure 5.3, p<0.05 n=11 Mann-Whitney U). All control rats fed within the allotted time limit, while 2 of 6

CORT-treated rats never fed. There was no significant difference in body weight between the two groups, implying this effect was not likely due to overall changes in metabolism or appetite

(Figure 5.3, C). These results demonstrate that chronic CORT treatment is sufficient to induce anhedonia and neohypophagia.

CUS causes a decrease in the AMPAR-mediated component of excitatory postsynaptic fEPSPs at TA-CA1 synapses (Kallarackal et al., 2013). As with CUS, chronic CORT reduced

AMPA:NMDA ratios (Figure 5.4A,B; p=0.010, n=18 slices, Student’s t-test). AMPAR strength was also reduced significantly following CORT administration when normalized to the fiber volley (FV), a measure of the number of axons activated (Figure 5.4C, n=18, p=0.042). There was no difference in NMDAR:FV responses (Figure 5.4, n=17 p=0.4).

Figure 5.1 – Experimental timeline for chronic CORT experiments After a baseline sucrose preference test (SPT), rats were administered CORT in their drinking water for 4 weeks. Novelty-suppressed feeding (NSF) and SPTs were performed before rats were sacrificed for in vitro electrophysiology and Western blotting.

Figure 5.2 - Loss of sucrose preference by chronic exogenous CORT Sucrose preference scores before and after chronic CORT administration. group*time 2x2 mixed ANOVA F(1,54)=8.461, n=57, p=0.005 2x2 mixed ANOVA, p<0.005 vs. all other groups Bonferroni post-hocs.

Western blots revealed a ~31% reduction in GluA1 protein levels in SLM from CORT- treated animals, compared to SLM in untreated animals (Figure 5.5, p=0.009, n=19 samples,

Mann-Whitney U test). Thus, chronic administration of CORT in unstressed animals was sufficient to cause deficits in TA-CA1 excitatory synapses that are comparable to those produced by CUS (Kallarackal et al., 2013). Chronic CORT also downregulated GR protein levels in CA1 stratum pyramidale (Figure 5.5, n=8, p=0.01, Mann-Whitney U test), as observed following CUS

(Sapolsky et al., 1984; Herman et al., 1995). This is thought to be an indirect measure of stress load over time, and may pathologically contribute to loss of feedback from the hippocampus to the HPA axis, leading to HPA dysregulation and subsequent uncontrollability of the stress response (Anacker et al., 2011b).

Figure 5.3 - Effect of chronic CORT on NSF test (A) Group data summarizing increased latency to feed compared to controls by rats treated with chronic CORT. Chronic CORT elevation caused a significantly increased latency to feed in the novelty-suppressed feeding task, relative to controls (n=11, *p=0.007 Mann-Whitney U test). (B) Survival plot illustrating the proportion of rats that remained unfed vs. time in the novelty–suppressed feeding task. (C) No difference in body weight was detected between the two groups.

Figure 5.4 - Chronic CORT decreases AMPAR-mediated excitation at TA-CA1 Example TA-CA1 fEPSPs recorded in SLM in Mg2+-free ACSF before (black) and after (gray) DNQX wash-in, exemplifying representative AMPAR- and NMDAR-mediated responses, respectively. TA-CA1 fEPSPs from CORT-treated rats exhibited a marked reduction in AMPAR-mediated component relative to control. b, Quantification of group data for experiments from a, in which comparisons of the slope of each response was measured for responses generating a fiber volley (FV) of ~0.2mV. Chronic CORT treatment significantly reduced AMPA:NMDA ratios compared to controls (p=0.010, n=18, Student’s t-test). c, Chronic CORT also reduced AMPAR-mediated signaling when normalized to FV compared to control (p=0.042), but not NMDAR-mediated signaling.

Figure 5.5 - Chronic CORT decreases GluA1 expression in SLM and GR expression in SP Chronic CORT elevation decreases AMPAR-mediated signaling at TA-CA1 synapses by decreasing GluA1 expression. Left: Chronic CORT decreased GluA1 protein expression in SLM, where TA-CA1 synapses are formed (p=0.009, n=19, Mann-Whitney U), mimicking CUS. Right: In stratum pyramidale, where pyramidal cell bodies reside, GR protein expression is also decreased, as in CUS (n=8, p=0.01, Mann- Whitney U).

Interestingly, I also observed a modest, but significant decrease in 5HT1BR expression in tissue from SLM following chronic CORT treatment (Figure 5.6). Since 5HT1BR is most highly enriched in SLM within CA1, this is consistent with dendritic atrophy suggested by decreased

PSD-95 and GluA1 (Figure 4.2, Figure 4.3) following chronic stress.

Figure 5.6 - Chronic CORT treatment decreases 5HT1BR expression in SLM Chronic CORT significantly decreased 5HT1BR protein expression in SLM tissue (p<0.05 Mann-Whitney U, n=7)

Because acute stress and acute elevation of CORT both often exert different effects than chronic stress and chronically elevated CORT (Joëls, 2006), I compared the chronic CORT results to the effects of acute CORT administration. In contrast to CUS and chronic CORT, acute

(~16 hours) CORT increased AMPA:NMDA ratios at TA-CA1 synapses (Figure 5.7, 19.5±3.4 vs. 8.6±1.3 acute CORT, n=11, p=0.023 Mann-Whitney U test). Short exposure to CORT is not sufficient to generate the change in excitatory strength caused by chronic stress. Acute and chronic CORT elevations thus exert opposing actions on excitatory strength at TA-CA1 synapses.

Figure 5.7 - Acute CORT is not sufficient to decrease TA-CA1 AMPA:NMDA In contrast to CUS and chronic CORT, acute (~16 hours) CORT increased AMPA:NMDA ratios at TA- CA1 synapses. (Left) Example fEPSPs before (black) and after (gray) DNQX application in Mg2+-free ACSF. Note the difference in timescales for AMPAR and NMDAR traces, accentuating the slower kinetics of the NMDAR-mediated current flow. (Right) Summary group data (n=11, *p=0.023 Mann-Whitney U test).

Endogenous serotonin and the 5-HT1BR-selective agonist anpirtoline potentiate TA-CA1 synapses as a result of CaMKII-mediated phosphorylation of GluA1 (Cai et al., 2013). In brain slices from animals subjected to chronic stress, 5-HT1BR-mediated potentiation becomes greater in magnitude and irreversible (Cai et al., 2013). In slices from control rats, anpirtoline (50μM) induced an average potentiation of fEPSP slope to 125±8% of the baseline, which reversed to

95±11% of baseline after one hour of washout. In slices from chronic CORT rats, TA-CA1 fEPSPs increased to a maximum of 210±32%, and remained at 196±35% after washout (Figure

5.8, n=18, F(2,32)=6.87, p<0.05 2x3 mixed ANOVA, p<0.05 CORT peak and washout vs. CORT baseline and all control). Therefore, chronic CORT was sufficient to alter TA-CA1 response to serotonergic modulation from reversible to persistent, as occurs after CUS.

Again, as acute and chronic stress often have opposing effects, (Joëls, 2006), I compared the chronic CORT results to the effects of acute CORT administration on anpirtoline-mediated potentiation of TA-CA1 fEPSPs. In contrast to CUS and chronic CORT, acute (~16 hours) CORT

100 administration to rats subsequently resulted in only a modest anpirtoline-mediated potentiation of

TA-CA1 fEPSPs (Figure 5.9). Short exposure to CORT is not sufficient to generate the dramatic change in the ability of 5HT1BR activation to potentiate TA-CA1 synapses caused by chronic stress. Interestingly, this potentiation began to reverse upon washout, but did not fully return to baseline within an hour. This is in contrast to slices from chronic CORT-treated rats, in which potentiation continued despite the beginning of washout.

Taken together, these results show that chronic CORT elevation is sufficient to mimic the both the behavioral effects of CUS and the corresponding neurophysiological and molecular changes in excitatory synaptic transmission at TA-CA1 synapses. Elevated CORT alone is thus sufficient to mimic the behavioral and synaptic consequences of chronic stress in this model.

However, this CORT elevation must be chronic, and not acute.

101

Figure 5.8 - Chronic CORT alters 5HT1BR mediated potentiation of TA-CA1 fEPSPs (A) Example TA-CA1 fEPSPs demonstrating the reversible effect of anpirtoline (50µM) on fEPSPs in slices from control rats (left) and the persistent potentiation in slices from chronic CORT-treated rats (right). Traces taken from the time-points indicated in the graph below. (B) Group data illustrating the difference in time-course of 5-HT1BR-mediated potentiation in slices from control (gray symbols) and CORT-treated (black symbols) rats. Data are normalized to the 10 minutes prior to the beginning of anpirtoline application. (C) Comparison of TA-CA1 fEPSPs at baseline, peak of potentiation, and after washout. Anpirtoline generated a significantly greater peak response in hippocampal slices from chronic CORT-treated rats. This potentiation persisted through washout. Repeated measures ANOVA interaction F(2,32)=6.87, p=0.03. *p<0.05 Bonferroni post-hoc vs. baseline CORT and all control.

102

Figure 5.9 - Effect of acute in vivo CORT exposure on in vitro anpirtoline response at TA-CA1 synapses Activation of 5HT1BRs with anpirtoline caused only a modest potentiation (148.9±0.3% of baseline) consistent with controls. This potentiation began to decrease after an hour of washout (gray trace, 134.2±0.2), but did not fully reverse with washout. Traces above are representative of the group data, illustrating samples at time denoted by (1-3).

Is elevated CORT necessary to generate the neurobiological correlates of chronic stress?

CUS causes a chronic elevation of average CORT levels and repeated spikes in CORT levels in response to each individual stressor (Sapolsky et al., 1984; Magariños and McEwen,

1995b), as part of a complex constellation of concerted neuromodulator interactions [reviewed in

(Joëls and Baram, 2009)]. HPA axis responses to unpredictable or varied stress do not habituate even after 3 weeks [e.g. fig. 1 in (Magariños and McEwen, 1995b), reviewed in (Herman, 2013)].

I predicted that if chronic stress-induced elevations in CORT are necessary for the changes triggered by CUS, then partially inhibiting CORT synthesis prior to each stressor with metyrapone (MET, 50mg/kg i.p.), an inhibitor of the adrenocortical enzyme 11β-hydroxylase

(Holsboer, 2000), would prevent the CUS-induced synaptic and behavioral deficits. The experimental timeline is illustrated in Figure 5.10. During CUS, I tested whether MET attenuated the increase in basal CORT caused by stress-induced dysregulation of the HPA axis (Johnson and

103

Yamamoto, 2009). CORT was extracted from fecal pellets collected overnight, reflecting CORT levels integrated over a day of stress. The CUS-induced increase in CORT, observed in vehicle

(VEH) treated rats, was prevented by MET as measured by ELISA of fecal pellet extractions

(Figure 5.11, 2x2 ANOVA interaction F(1,37)=4.45, p=0.04 *p<0.05 Tukey’s HSD post-hoc vs. all other groups). In unprotected t-tests, MET treatment alone reduced fecal CORT compared to all other groups (p<0.02). Furthermore, MET blunted peak plasma CORT during an individual restraint stressor (Figure 5.11, n=11, p=0.0008 Student’s t-test), as measured by radioimmunoasssay.

Figure 5.10 - Experimental timeline for MET+CUS experiments After a baseline sucrose preference test (SPT), rats were randomized to receive either vehicle (VEH) or MET while undergoing CUS for 3 weeks. Novelty-suppressed feeding (NSF) and SPTs were performed before rats were sacrificed for in vitro electrophysiology and Western blotting. MET or VEH injections were given prior to each stressor throughout the paradigm.

104

Figure 5.11 - MET significantly attenuates CORT increases in response to CUS Left: To determine if MET was effective in reducing time averaged stress-induced CORT levels, fecal pellets were collected overnight from singly-housed rats after a typical day of stressors and compared to values in age-matched stress-naïve rats. CUS significantly elevated CORT in vehicle- but not MET-treated rats (2x2 ANOVA interaction F(1,37)=4.45, p=0.04 *p<0.05 Tukey’s HSD post-hoc vs. all other groups). In unprotected t-tests, MET treatment alone reduced fecal CORT compared to all other groups (‡p<0.02). Right: Rats undergoing CUS received either vehicle or MET treatment prior to restraint stress, and blood samples were taken from their tail veins 10-15 minutes into the stress. Plasma CORT was quantified via RIA. MET significantly reduced plasma CORT levels induced by the stressor (*p=0.0008, n=11, Student’s t-test).

Unlike VEH-treated rats subjected to CUS, MET-treated rats subjected to CUS

(CUS+MET) maintained a high sucrose preference that was not different from the sucrose preference of unstressed MET- or VEH-treated (CUS+VEH) rats (Figure 5.12, n=36,

F(1,20)=18.6, 4x2 mixed ANOVA p<0.001, p<0.05 CUS+VEH vs. baseline and all other groups). In the NSF test, CUS+MET rats exhibited a shorter latency to feed than CUS+VEH rats and were not different from unstressed VEH or MET rats (Figure 5.13A, Kruskal-Wallis H=9.52, p=.0231 n=33, CUS+VEH U-test post-hoc p<0.05 vs. all others). Fewer CUS+VEH rats fed within the allotted time limit than any other group (Figure 5.13B, p<0.05, Fisher’s exact test).

Body weights of stressed and unstressed rats administered VEH or MET reveal a main effect of

CUS on weight (Figure 5.13C, F(1,33)=43.21, **p<0.0001; 2x2 ANOVA), but no main effect of

MET (F(1,33)=0.02, p=0.89), indicating no likely effect of appetite overall, as well as no role of

CORT in mediating the body weight changes in the context of chronic stress, consistent with

105 previous reports (Magariños and McEwen, 1995a). In summary, limiting stress-induced increases in CORT prevented the stress-induction of anhedonia and neohypophagia.

I predicted that MET would also prevent the changes observed at TA-CA1 synapses after

CUS. AMPA:NMDA ratios were significantly lower in slices from CUS+VEH rats compared to

CUS+MET and unstressed MET and VEH rats (Figure 5.14A,B, main effect of CUS

F(1,54)=14.71, p=0.0003; main effect of MET F(1,54)=11.93, p=0.0011; MET*CUS interaction effect (F(1,54)=11.84, p=0.0011). The AMPAR-mediated excitatory component at TA-CA1 was also significantly higher in slices from MET-treated rats when normalized instead to FV (Figure

5.14C, main effect of MET F(1,56)=7.43, p=0.0085), while NMDAR:FV responses were not significantly different (p=0.12). TA-CA1 AMPA:NMDA strength correlated significantly with hedonic state, measured by sucrose preference (R=0.65, p=0.002).

GluA1 expression in SLM decreased by 25% in tissue from CUS+VEH rats as compared to CUS+MET rats (Figure 5.15, n=14, p=0.049, Mann-Whitney U test). GR expression was also lower in SP of CUS+VEH rats, compared to CUS+MET (n=8, p=0.02, Mann-Whitney U test).

Finally, anpirtoline potentiated TA-CA1 fEPSP slope to 155±5% of baseline after 60 minutes in slices from vehicle-injected CUS rats, peaked at 199.5±14% and it remained elevated (166±11%) after washout (Figure 5.16, n=32, mixed ANOVA interaction effect F(2,54)=7.95, p=0.0009, p<0.05 CUS+VEH peak and washout vs. CUS+VEH baseline, and vs. all CUS+MET). In slices from CUS+MET rats, in contrast, TA-CA1 fEPSP potentiation peaked at 149±12% and reversed to 106±27% after washout. MET thus prevented the exaggerated, persistent response to anpirtoline at TA-CA1 synapses observed normally after CUS.

Taken together, these results show that blunting the elevation of CORT produced by the

CUS stressors is sufficient to prevent both the behavioral effects of CUS and the corresponding neurophysiological and molecular changes in excitatory synaptic transmission at TA-CA1 synapses.

106

Figure 5.12 - MET prevents induction of anhedonia in SPT caused by CUS CUS decreased sucrose preference when rats were pretreated with vehicle, but not MET, prior to each stressor. Daily vehicle or MET injections without stress had no effect. Mixed ANOVA 3-way interaction F(1,30)=4.841 (p=0.036), *-p<0.05 vs all other groups Bonferroni post-hoc

107

Figure 5.13 - MET treatment prevents CUS-induced neohypophagia a, MET pretreatment significantly decreased the latency to feed in the NSF task in CUS rats vs. vehicle- treatment, to levels not significantly different than unstressed vehicle or MET groups (Kruskal Wallis H = 9.52, p=.0231 n=33; *p<0.05 vs all others, U-test) b, Survival plot of NSF task in vehicle- and MET- treated rats unstressed or subjected to CUS. 8/10 CUS+MET and 2/9 CUS+vehicle fed within 10 minutes (*p=0.023, Fisher’s exact test). 5/6 and 8/10 of unstressed vehicle- and MET- treated rats fed, respectively. c, Body weights of stressed and unstressed rats administered VEH or MET reveal a main effect of CUS on weight (F(1,33)=43.21, **p<0.0001; 2x2 ANOVA), but no main effect of MET (F(1,33)=0.02, p=0.89).

108

Figure 5.14 - MET prevents CUS-induced decreases in AMPAR-mediated TA-CA1 signaling a, Example TA-CA1 fEPSPs recorded in SLM in Mg2+-free ACSF before (black) and after (gray) DNQX wash-in, exemplifying an AMPAR-mediated and NMDAR-mediated trace, respectively. b, TA-CA1 fEPSPs from MET-treated rats exhibited a higher AMPAR-mediated signaling component relative to vehicle-treated CUS rats, measured by AMPA:NMDA ratio (Main effect of CUS F(1,54)=14.71, p=0.0003; main effect of MET F(1,54)=11.93, p=0.0011; MET*CUS interaction effect (F(1,54)=11.84, p=0.0011). The CUS+VEH group was different from all others (**p<0.01, Tukey’s HSD). c, The AMPAR-mediated excitatory component at TA-CA1 was also significantly higher in slices from MET-treated rats when normalized instead to FV (*-Main effect of MET F(1,56)=7.43, p=0.0085). NMDAR:FV was not different (p>0.05 for both CUS and MET main effects).

109

Figure 5.15 - MET prevents CUS-induced decreases in GluA1 and GR expression SLM punches revealed MET treatment prevented the CUS-induced decrease in GluA1 protein expression observed in the vehicle group (n=14, p=0.049 Mann-Whitney U). SP punches revealed that MET also preventing GR downregulation (p=0.02, Mann-Whitney U).

110

Figure 5.16 - MET treatment prevents stress-induced changes in the potentiation of TA-CA1 fEPSPs by the 5-HT1BR agonist anpirtoline a, Example TA-CA1 fEPSPs demonstrating the effect of anpirtoline (50µM) on slices from vehicle- (left) and MET-treated CUS rats (right). Left: Anpirtoline (red, 2) elicited a robust potentiation from baseline (black, 1), that persisted after washout (gray, 3) in slices from vehicle-treated CUS rats. Right: After MET treatment, anpirtoline elicited a more modest potentiation of TA-CA1 fEPSPs that was reversed upon washout. b, Group data and time course of 5-HT1BR-mediated potentiation. Data are normalized to the 10 minutes prior to the beginning of anpirtoline application. c, Comparison of TA-CA1 fEPSPs at baseline, peak of potentiation, and after washout. Anpirtoline generated a significantly greater peak TA-CA1 response in slices from vehicle-treated rats, and it persisted during washout, whereas a reversible, modest potentiation was elicited in slices from CUS rats treated with MET, as in unstressed controls. Repeated measures mixed ANOVA interaction effect F(2,54)=7.95, p=0.0009. *p<0.05 Bonferroni Tukey’s HSD post-hoc vs. baseline vehicle and from all MET timepoints.

Does MET prevent a CUS-induced TA-CA1 synapse-specific cognitive deficit?

(I was actively involved in the design of this portion of work, but it was performed entirely under the supervision of Dr. Aileen Bailey by her students Keighly Bradbrook and Hannah Dantrassy)

TA-CA1 synapses are required for long-term consolidation of spatial memory (Remondes and Schuman, 2004), and CUS impairs this function (Kallarackal et al., 2013). Impaired spatial

111 memory has been described in depressed humans (Gould et al., 2007). I asked whether blunting

CUS-induced elevation of CORT with MET treatment during CUS would prevent the normal

CUS-induced loss of consolidation in a spatially cued version of the Morris water maze. After naïve rats were trained to learn the location of a hidden platform, confirmed with an initial 24 hour probe trial, they were randomly divided into unstressed control groups and CUS+MET,

CUS+VEH, using the same procedures described above. Twenty-eight days after training, unstressed control rats performed as well as they did in the initial probe trial (Fig 7, n=6, p>0.05 paired t-test), indicating successful consolidation. CUS+VEH rats, in contrast, required a significantly longer latency to reach the platform location (p=0.014, n=6) and traveled a significantly longer path to reach the target (p=0.02), compared to their previous probe trial, indicating that CUS in this group disrupted memory consolidation. However, latency (p>0.05, n=6) and path to target were similar in CUS+ MET rats to their baseline responses (p>0.05), indicating successful memory consolidation despite CUS exposure.

Because MET treatment prevented the effects of CUS on three separate behaviors

(anhedonia, neohypophagia and impaired spatial memory consolidation) and the neurophysiological and molecular deficits of excitatory synaptic transmission at a key stress- sensitive synapse, I conclude that chronically elevated CORT is a necessary mediator of the adverse effects of chronic stress at this pathway.

112

Figure 5.17 - MET treatment prevents stress-induced impairment of spatial memory consolidation Rats were trained to find the location of a hidden platform in a water maze in 10 blocks over 6 days. 24 hrs after the 10th block, they were tested in a probe trial (black) to ensure they had learned the platform location. They were then randomized to control and CUS groups. CUS-subjected rats received either a MET or VEH injection prior to each stressor, while control rats were left undisturbed. After 28 days, rats were tested for spatial memory consolidation in another probe trial (“consolidation”-white) by comparing their performance to their initial probe trial (“24 Hr probe”-black). a, Example diagrams of consolidation trial, with start (open circle), finish (closed circle) and path (connecting line) for rats from each group. b, Path length to platform location (left) and latency to platform location (right) for each group. Control, unstressed rats performed as well after 28 days as after the initial probe trial (p>0.05 for both path length and latency). CUS+VEH rats performed significantly worse after the 28 day consolidation period, with significantly greater latency (t(6)=3.42, p=0.014) and path length (t(6)=3.10, p=0.02) to platform location, indicating impaired memory consolidation. CUS+MET rats successfully consolidated, with no significant difference in latency or path length between trials (p>0.05). n=6 per group. These experiments were performed by my collaborators under the guidance of Dr. Aileen Bailey.

113

Discussion

Physical and psychosocial stressors are significant risk factors for depression, in addition to anxiety and other mood-disorders. Stress affects the hippocampus, a critical region for cognitive processing, mood, and other complex behaviors that are altered in depression, as well as synapses in the prefrontal cortex and nucleus accumbens, two other regions involved in processing reward. Potential mediators of stress include endocannabinoids (Gray et al., 2013), corticotropin-releasing hormone (Chen et al., 2012) and other neuropeptides, opioids such as dynorphin (Mague et al., 2003; McLaughlin et al., 2003), monoamines such as serotonin and norepinephrine, glutamate (Palucha and Pilc, 2005), and glucocorticoids (Joëls and Baram, 2009).

The stress-sensitive TA-CA1 synapse provides a tractable locus at which to examine the neurobiological effects of the stress response and better understand how chronic stress causes altered behavioral phenotypes by changing synapses and circuits. Generalizing the current results from TA-CA1 synapses, I conclude that chronic CORT elevation is both sufficient and necessary to induce the chronic stress phenotype at molecular, synaptic, and behavioral levels.

CORT is necessary and sufficient to mediate the synaptic effects of chronic stress:

Chronic CORT administration induces depressive-like behaviors, including anhedonia and increased helplessness, that are reversed by chronic antidepressants (Gourley and Taylor,

2009). I replicated these previous behavioral results and further observed that chronic CORT induced hyponeophagia in the NSF task, an ethologically relevant choice between feeding motivation and neophobic behavior. These tests are sensitive to chronic, but not acute, SSRI administration (Dulawa et al., 2004). MET treatment significantly reduced the peak CORT response to a stressor and the average CORT levels during CUS. Because CORT administration mimicked, and MET treatment prevented, anhedonia-like and neophobic behaviors in my experiments, I conclude that stress-induced CORT elevation is sufficient and necessary to generate these behavioral changes in the CUS model.

114

Chronic CORT was also sufficient to decrease AMPAR-mediated excitation and GluA1 subunit expression at TA-CA1 synapses in unstressed animals. The effect was specific for

AMPARs, as there was no accompanying change in NMDAR-mediated fEPSPs. Conversely,

MET administration during CUS prevented this dysfunction, illustrating that chronic stress- induced elevations of CORT are necessary for these effects. A decreased AMPA:NMDA ratio at these synapses is consistent with results from synapses in the PFC and NAc (Lim et al., 2012;

Yuen et al., 2012), and I suggest that CORT elevation during chronic stress is a potential mediator at those synapses, as well. I suggest further that weakening of synaptic excitation at multiple synapses in cortico-mesolimbic reward circuits contributes to depressive-like signs after chronic stress. Corresponding restoration of synaptic strength at these stress-sensitive synapses is a critical mechanism of antidepressant action in these models (Popoli et al., 2012; Tizabi et al.,

2012; Kallarackal et al., 2013; Musazzi et al., 2013).

Chronic stress also alters serotonin-induced potentiation at TA-CA1 synapses, both qualitatively and quantitatively (Cai et al., 2013). Mediated by 5-HT1BRs and phosphorylation of

GluA1, serotoninergic modulation co-varies with depressive-like behavior and changes in

AMPAR function. Conversely, chronic antidepressant treatment restores normal serotonin responses at TA-CA1 synapses in conjunction with restoration of normal behavior and synaptic strength. In the present study, the aberrant stress-induced anpirtoline response was mimicked by chronic CORT treatment alone, and prevented by MET treatment in rats subjected to CUS.

TA-CA1 synapses are required for spatial memory consolidation, and impaired spatial memory has been described in humans (Gould et al., 2007). The ability of MET treatment to preserve TA-CA1 synaptic strength and memory consolidation in rats subjected to CUS demonstrates that CORT-induced changes in synaptic strength likely underlie this stress-induced cognitive deficit.

Structurally, chronic restraint stress reduces the number of spines in CA1 cell distal apical dendrites (Pawlak et al., 2005; Maras et al., 2014). Functionally, chronic stress and CORT

115 administration impair LTP at proximal CA1 synapses in vitro in some studies (Alfarez et al.,

2003), but not all [for review see (Joëls et al., 2012)]. Chronic stress or chronic CORT cause atrophy of terminal segments of CA1 dendrites (Sousa et al., 2000), which are more vulnerable to stress than more proximal synapses, but are often overlooked in stress-related studies. In fact, loss of distal dendrites and TA-CA1 synapses may contribute significantly to the atrophy observed after chronic stress in rodents and contribute to the hippocampal volume loss in severe human depression.

Acute vs. chronic effects of CORT

In contrast to the impairment of AMPAR-mediated transmission produced by chronic

CORT (~weeks), acute CORT (~hrs) increased transmission at TA-CA1 synapses, consistent with previous observations at Schaffer collateral synapses (Karst et al., 2005). This short-term effect of

CORT to increase CA1 excitability is consistent with the well-established dichotomy between acute and chronic stress and with the transient mood-boosting effects of steroid treatment (Joëls,

2006). Acute stress or acutely administered CORT in vivo and in vitro can enhance neurotransmission at CA3-CA1 synapses by increasing glutamate release probability, increasing cellular excitability in CA1 pyramidal neurons by modulating IA, and facilitating long-term potentiation, with many of these fast changes attributed to the mineralocorticoid receptor (MR)

(Karst and Joëls, 2005; Wiegert et al., 2006; Olijslagers et al., 2008). Thus, acute stress increases plasticity in short time windows via non-genomic MR, promoting stress-salient memories. In contrast, chronic stress dampens plasticity via GR-mediated genetic regulation, leaving newly formed, stress-salient memories less malleable (Joëls, 2008). I suggest that the synaptic consequences of chronic CORT elevation described here are likely to be mediated by chronic, excess GR activation.

In the introduction and in previous chapters, I discussed how the stress response promotes mobilizing the body’s resources to achieve different homeostatic set points within a dynamic

116 range, as most appropriate for internal and external environmental conditions. However, chronic overactivation of these mechanisms conceptually leads to wear and tear on the stress response system, and ultimately impairs the ability to achieve different points in the dynamic range. My results with CORT illustrate this fact. Short term exposure to CORT greatly increased AMPAR- mediated strength at TA-CA1 synapses, corresponding to the expected response from an acute stressor. Conversely, chronic CORT dampened AMPAR-mediated excitability. In conjunction with the finding that the dynamic range of LTP is impaired following CUS, this presents a hard example of the sometimes overly conceptual process of repeated allostatic demand leading to allostatic-overload.

Glucocorticoids in human disease

There is compelling evidence of a causal link between altered HPA function and a large subset of severely depressed patients, particularly with regard to anhedonia. Depression is common in patients with primary abnormalities of GC production and patients receiving exogenous GCs (Quarton et al., 1955; Jeffcoate et al., 1979). Interestingly, in patients with hypercortisolemia, treating the cause of hypercortisolism (typically via resection of a pituitary tumor hypersecreting ACTH) been effective in reversing hippocampal atrophy (Starkman et al.,

1999). HPA dysregulation is correlated with severity of symptoms and is common (~80% of patients) in treatment-resistant depression (Anacker et al., 2011b). The correlation is even greater in melancholia patients (up to 90%), identified using modified criteria relying primarily on anhedonia (Taylor and Fink, 2008). Abnormal HPA function is also correlated with a nine-fold increase in suicide risk (Coryell and Schlesser, 2001). Downregulation of GRs following chronic stress, as I observed in chronic CORT-treated rats as well as CUS rats, can contribute to loss of

HPA feedback, resulting in higher circulating GCs (Anacker et al., 2011b). This hyposuppressive state results in failure to limit the stress response, as seen in many depressed patients (Quarton et al., 1955; Heuser et al., 1994; Mizoguchi et al., 2001). GR downregulation may represent an

117 adaptive, homeostatic attempt to preserve CA1 neuronal viability from further stress-induced insult such as loss of spines or excitatory synaptic function, but at the same time may contribute maladaptively to HPA dysregulation. HPA dysregulation may be both a causative factor and a consequence of depression.

The dose of MET used here produced effects that are similar to those produced by clinically relevant doses in humans, effectively halving circulating corticosteroids (O’Dwyer et al., 1995). In stressed rats, MET treatment preserved the normal response of TA-CA1 synapses to serotonergic modulation, suggesting that MET treatment protected the serotonin-innervated excitatory synapses upon which SSRIs act (Sigalas et al., 2012; Cai et al., 2013). MET also prevented normal stress-induced downregulation of GRs, suggesting protection from HPA dysregulation. There is clinical evidence that MET enhances antidepressant response (Sigalas et al., 2012; McAllister-Williams et al., 2013) and is effective as a stand-alone therapy in hypercortisolemic patients with depressive symptoms (Jeffcoate et al., 1979). My results suggest several mechanisms by which treatment with MET should oppose pro-depressive changes caused by stress, suggesting potential efficacy as a clinical add-on, particularly for depressed patients with HPA or GC abnormalities.

In conclusion, my data supports my hypothesis that chronic stress alters synapses and induces depressive-like behavior via chronic stress-induced elevations of corticosterone. Chronically elevated corticosterone is both sufficient and necessary for the neurobiological correlates of chronic stress: impaired TA-CA1 synaptic function, altered responsiveness to 5HT1BR agonists, anhedonia, and hyponeophagic behavior. I have demonstrated that glucocorticoids play a critical role as mediators of the synaptic effects of chronic stress, as well as behavioral changes in rats, which are similar to the mood and cognitive changes characterizing human depression, strengthening the hypothesis that excitatory synaptic dysfunction underlies the symptoms of depression (Krugers et al., 2010a; Joëls et al., 2012; Popoli et al., 2012; Duman, 2014). Protecting

118 and restoring excitatory strength at stress-sensitive synapses at key loci throughout the reward systems appears critical for maintaining normal cognitive and emotional behavior.

119

6) Discussion: towards an excitatory synapse hypothesis of depression

Portions of the work in this chapter have been published in Neuropsychopharmacology. I contributed primarily in the conceptualization and experimental design, but I participated in data collection for any figures shown.

Jonathan Fischell, Adam M. Van Dyke, Mark D. Kvarta, Tara A. LeGates, and Scott M. Thompson. Rapid antidepressant action and restoration of excitatory synaptic strength after chronic stress by negative allosteric modulators of alpha5-containing GABAA receptors. Neuropsychopharmacology. 2015 Apr 22.

I am proud to have contributed to the emerging excitatory synapse hypothesis of depression. Our lab has taken a view that has resulted from my own studies and those of the following co-authors. Our resulting review and perspective has been published in Trends in Neurosciences. The earliest forms of this hypothesis was crafted by Drs. Kallarackal, Cai, and Thompson. In many, many ways, this thesis would not exist without them.

Scott M. Thompson, Angy J. Kallarackal, Mark D. Kvarta, Adam M. Van Dyke, Tara A. LeGates, and Xiang Cai. An excitatory synapse hypothesis of depression. Trends in Neurosciences. 2015 Apr 14. In press.

This thesis was structured a bit like a Venn diagram, with experiments targeted to the overlap, at TA-CA1 synapses. Existing antidepressants have effects in many areas of the brain, as would be expected by near-global manipulation of serotonergic transmission. Depression also affects many brain areas (Nestler et al., 2002), likely underlying the diverse array of cognitive and affective symptoms experienced by depressed patients. This approach is more likely to yield novel insight into the pathophysiology of disease than by examining either a treatment or a disease model on its own. Certainly there are brain areas affected by boosting synaptic serotonin that were not defective in some or all patients; effects here likely underlie side effects.

Additionally, there are likely brain areas affected in depression that do not respond to current antidepressant treatment; these areas might embody symptoms that are not adequately treated in some or all patients. I examined a locus at which opposing effects of chronic stress and antidepressant treatment occurred, the TA-CA1 synapse, characterizing aspects of both along the way. Having confirmed both stress-sensitivity and antidepressant-sensitivity, I keyed in on this synapse to ask how chronic stress models translate into depressive-like neurobehavioral state as a proxy for the question “how do we become depressed?”

120

I hypothesized that depression, specifically regarding anhedonia as a core symptom, is caused by corticosteroid-dependent reduction of excitatory synaptic strength at key synapses in cortico-mesolimbic circuitry, which is rescued by monoaminergic antidepressant treatment by

5HT1BR-mediated serotonergic signaling. First, I predicted that excitatory glutamatergic transmission, with TA-CA1 synapses as an archetype, would be potentiated by antidepressant treatment. Second, I predicted that animals that were subjected to a chronic stress paradigm and exhibiting a depressive-like neurobehavioral state would suffer from impaired glutamatergic signaling at some stress-sensitive synapses. Third, I predicted that corticosterone would be necessary and sufficient to mediate the induction of synaptic dysfunction. The experiments I used to test these predictions support my overall hypothesis and provide compelling insight about how to think about depressive-like states and how they can be reversed by antidepressant treatment.

In chapter 1, I discussed how identifying deficits in a depressive-like state that are also downstream targets of antidepressant effects should enable a search for alternative approaches to reverse these deficits. My hypothesis suggests that monoaminergic antidepressants exert antidepressant efficacy via 5HT1BR, by reversing deficits in excitatory synaptic function that are caused by chronic elevations of corticosterone due to chronic stress.

A key point is that any treatment which can adequately address these deficits in excitatory synaptic function in important areas can reverse depressive-like neurobehavioral states without having to act through serotonergic signaling. It still broadly remains a mystery why

SSRIs take so long to act and precisely what seemingly indirect, if any, role that serotonin dysfunction may have in depression (Heninger et al., 1996; Baumeister et al., 2003; Berton and

Nestler, 2006). Therefore it would be a major advancement to be able to target the ultimate downstream targets that underlie reversal of a depressive state.

A new class of antidepressant compounds may (1) act much faster, by immediately targeting the ultimate endpoint of weeks of serotonergic manipulation, (2) exhibit fewer side effects if targeted only to regions that need rescue (3) treat symptoms that do not respond to

121 monoaminergic-based antidepressants, perhaps by ameliorating deficits in areas that are not endogenously equipped with the proper machinery (e.g. areas that do not express 5HT1BRs). In support of this latter point, ketamine, deep-brain stimulation and ECT are effective treatments for patients that are considered treatment-resistant to traditional monoaminergic based antidepressants (Michael et al., 2003; Zarate et al., 2012; Schlaepfer et al., 2014), supporting the idea that there are circuits or synapses that need to be repaired in at least some patients that cannot be mended by augmenting serotonin signaling.

A final complication in augmenting serotonin signaling at synapses that are not affected by depression is the heterogeneity of symptoms that currently merit treatment with monoaminergic-based antidepressants. There are over 900 combinations of symptoms that merit a diagnosis of depression according to the DSM guidelines. It is unlikely that this is a single disease

(Taylor and Fink, 2008), and indeed efforts have been put forth to classify several subtypes of depression that stem from families of symptoms and underlying constructs (Carragher et al.,

2009). It has been argued that grouping so many potentially distinct entities into a single classification has actually greatly hindered research efforts in depression, by drowning meaningful signals in the noise of several distinct entities, and dissociating psychiatry from biology (Shorter and Fink, 2010). Presumably distinct biological effects on the brain underlie distinct symptoms, and monoamine manipulation may either have no effect on circuits that are not defective or even deleterious effects. That is, unintended effects could be caused by traditional antidepressant treatment because the specific subtype of depression does not affect a specific circuit, or an antidepressant is prescribed to someone erroneously diagnosed or for an off-label use. To illustrate the potential detrimental effect of SSRIs on healthy synapses and circuits, recall that LTP at TA-CA1 synapses has been implicated in the consolidation of long- term memory (Remondes and Schuman, 2004). Our studies have demonstrated that antidepressant actions of serotonin can occlude synaptic plasticity processes at TA-CA1 in unstressed animals (e.g. anpirtoline and LTP occlude each other, chapter 3). Furthermore,

122 anpirtoline had no effect on TA-CA1 fEPSPs in rats chronically treated with fluoxetine, presumably because these synapses were potentiated from normal strength in the healthy state to saturation. Intriguingly, intrahippocampal injection of a 5HT1BR agonist causes spatial learning deficits in rats (Buhot et al., 1995), and memory impairments have been reported in healthy volunteers taking antidepressants (Thompson, 1991).

Anti-glucocorticoids: protecting synapses from ongoing damage?

My results demonstrated that there is a glucocorticoid-dependent loss of excitatory synaptic strength at the distal apical dendrites of CA1. Indeed, I demonstrated that chronically elevated corticosterone is both sufficient and necessary to induce weakening of AMPAR- mediated transmission associated with the CUS model of depression. I propose that MET protects the serotonin-innervated excitatory synapses upon which SSRIs act. This is consistent with existing clinical data that MET is effective as an add-on treatment to SSRIs in patients who have not achieved remission (Sigalas et al., 2012; McAllister-Williams et al., 2013). Furthermore,

MET is effective as a stand-alone therapy in hypercortisolemic patients with depressive symptoms (Jeffcoate et al., 1979), which fits in well with reports that treating hypercortisolemia reverses hippocampal atrophy (Starkman et al., 1999). It would be dangerous for all depressed patients to attempt to improve their symptoms with steroid synthesis inhibitors like MET or cyanoketone, as only a subset of depressed patients exhibit HPA hyperactivity. I suggest that chronic stress based models have greater construct validity for this patient subpopulation, and are less relevant for depressed patients who do not exhibit HPA abnormalities or anhedonia. Cortisol reduction in eucortisolemic patients may lead to adrenal insufficiency and an Addisonian crisis.

However, in an emerging era of personalized medicine, it should be possible to identify those likely to respond to MET treatment. The melancholia subset of depression has a high association with anhedonia and HPA dysregulation (>80% for each) and it has been proposed that the dexamethasone suppression test for HPA feedback is particularly sensitive in this population

123

(Taylor and Fink, 2008). Combined with basal CORT levels, a target patient demographic could be identified with reliable biomarkers (Shorter and Fink, 2010). If effective, months of unremitted symptoms while undergoing trials with various monoaminergic antidepressants could be avoided in this population, as nearly half of all depressed patients do not respond to initial treatment with

SSRIs, and ¾ do not achieve full remission (Gaynes and Warden, 2009).

Alternatives to monoamines: targeting glutamatergic transmission

Pathological chronic elevation of glucocorticoids causes depressive-like behavioral changes because it weakens specific subsets of synapses in cortico-mesolimbic circuits, such as the hippocampus (chapter 5, Kvarta et al., 2015), PFC (Duman and Aghajanian, 2012; Yuen et al., 2012), and ventral striatum (Lim et al., 2012), thereby resulting in the anhedonia and the diverse collection of symptoms that characterize human depression. Conventional antidepressants like SSRIs (chapters 3 and 4), fast-acting antidepressants like ketamine, and non-pharmacological treatment like electroconvulsive therapy all act to oppose the deleterious actions of chronically elevated GCs on excitatory synapses (Berman et al., 2000; Tizabi et al., 2012; Kallarackal et al.,

2013; Zarate et al., 2013). Together, these results indicate that chronic excessive glucocorticoids not only cause stress-induced behavioral dysfunction, but also specific synaptic deficits in the hippocampus, a major center for cognitive and emotional processing. Maintaining excitatory strength at stress-sensitive synapses at key loci throughout cortico-mesolimbic reward circuitry appears critical for maintaining normal cognitive and emotional behavior, while restoring excitatory transmission in the depressed state appears critical for antidepressant action.

How else might excitatory transmission be appropriately promoted to reverse a depressed state? There is some limited evidence that directly potentiating AMPARs can have a antidepressant effect. The AMPAkine LY392098 decreases immobility time in the forced swim and tail suspension tests (Li et al., 2001), but has no effect on sucrose preference following a chronic unpredictable stress paradigm (Farley et al., 2010). Another AMPAkine LY451646

124 administered during chronic stress exposure prevented several lasting effects of chronic social stress exposure on physiological, neuroendocrine, and behavioral parameters (Schmidt et al.,

2010). However, AMPAkines are unlikely to be an effective clinical solution unless some sort of regional selectivity can be developed, because the broad expression of AMPARs throughout the brain is at least as problematic as widespread serotonin signaling. A drug that encourages phosphorylation of GluA1 S831 could be beneficial, since S831 phosphorylation is necessary for antidepressant effects of SSRIs and TCAs (chapter 3), but it may not be sufficient for antidepressant efficacy. Furthermore, such a drug might not have any effect at synapses that are not already potentiated by 5HT1BR-mediated serotonin signaling. Paradoxically, the NMDAR antagonist ketamine has been effective in treating depression (Berman et al., 2000), potentially by harnessing mechanisms of homeostatic plasticity or disinhibition, either way resulting in BDNF- dependent and activity-dependent plasticity (Li et al., 2011b; Nosyreva et al., 2013), leading to increased excitatory synaptic strength. However, practical clinical use is limited by abuse potential, psychotomimesis, and dissociative side effects (Machado-Vieira et al., 2009).

A net effect of increased excitation in forebrain circuits can be caused by either promoting glutamatergic transmission, or by impairing GABAergic signaling. A mild disinhibition in cortical areas could increase drive from these areas to mesolimbic circuitry and promote interpretation of reward. Partial inverse agonists at the benzodiazepine site of γ- aminobutyric acid type-A receptors (GABAARs) are negative allosteric modulators that have been shown to promote coherent network activity (Hajós et al., 2004). GABAARs containing the α5 subunit are exclusively expressed by prefrontal cortical neurons and hippocampal pyramidal cells, offering the potential means for selectively targeting cortical inputs to mesolimbic circuits, thereby potentially minimizing psychotomimetic and sedative side effects. My colleagues and I designed an approach to test the antidepressant properties of partial inverse agonists of GABAAR benzodiazepine sites containing the α5 subunit (Atack et al., 2005). α5-selective inverse agonists of the benzodiazepine site are not epileptogenic, hallucinogenic, or anxiogenic in humans (Atack

125 et al., 2009). We hypothesized that by negatively modulating GABAARs, mild disinhibition would promote activity-dependent plasticity, resulting in increased excitatory strength at the synapses disinhibited, and thereby reversing corticosterone-mediated excitatory synapse dysfunction. Initial behavioral experiments were performed by Jonathan Fischell, testing a partial inverse agonist of GABAAR benzodiazepine sites with a 10-100 fold selectivity for GABAARs containing the α5 subunit (L-655,708). A depressive-like neurobehavioral state was induced with chronic restraint stress and verified with SPTs and social interaction tests, which exhibit predictive validity for latency to treatment response. Amazingly, L-655,708 elicited a rapid antidepressant effect (within 24 hours) following a single injection in both the SPT and social interaction test (Fischell et al., 2015). Subsequently, we tested a different α5-GABAAR-selective compound, MRK-016, and observed similar antidepressant effects that lasted up to 7 days despite continued daily exposure to stress (Figure 6.1).

Returning to the TA-CA1 synapses, which are weakened by chronic stress and restored by successful antidepressant treatment, we asked whether L-655,708 would reverse AMPAR weakening as rapidly as it reversed stress-induced behavioral changes. We observed a decrease in the slope of the AMPAR-mediated component of the fEPSP across all stimulation intensities in slices from rats subjected to CRS and given a vehicle injection, compared to slices from unstressed rats and slices from rats subjected to CRS and given an L-655,708 injection (Figure

6.2, One-way ANOVA: F(2,17)=3.675, p=0.047; n=5 unstressed, n=8 CRS+vehicle, n=7

CRS+L-655,708). There was no corresponding difference in the slope of the NMDA component of the fEPSP under these conditions (F(2,17)=0.549, p=0.588). The AMPA:NMDA ratio was significantly higher in slices from CRS rats injected with L-655,708, compared to vehicle-treated animals subjected to CRS (p<0.05 LSD post-hoc), and was not different than responses in in slices from unstressed rats (One-way ANOVA: F(2,17)=4.345, p=0.03). My colleague Adam Van

Dyke demonstrated that this loss of AMPAR-mediated transmission was associated with a

126 decreased GluA1 expression in SLM, consistent with all previous conditions that demonstrated decreased TA-CA1 AMPA:NMDA.

These results replicate our earlier observations that AMPAR-mediated signaling is impaired by chronic stress at TA-CA1 synapses and we now demonstrate that this impairment can be reversed rapidly by treatment with this novel class of compounds. We conclude that the behavioral antidepressant efficacy of L-655,708 is associated with a restoration of excitatory neurotransmission at the archetypal stress-sensitive TA-CA1 synapse. Further studies are needed before it can be established that these compounds will exert clinical efficacy.

Together with the rapid antidepressant efficacy of NMDAR antagonists, these results may herald a new era in approaches for treating depression based on rapidly restoring excitatory synaptic strength (Thompson et al., 2015).

In conclusion, the experiments presented herein support the hypothesis that chronic stress promotes depression via corticosterone-mediated weakening of key excitatory synapses in the cortico-mesolimbic circuitry via impairment of AMPAR expression and function and that antidepressant efficacy is mediated by reversing these synaptic deficits. Many parts of the story remain to be told, but in a clinical field dominated by phenomenology and subjective descriptions of symptoms, elucidating neurobiological mechanisms is critical to answer the questions “How do we become depressed?” and “How do antidepressants work?” Ultimately, it is hoped that scientific inquiry into these questions will, someday, lead to the definitive answer for the depressed patient asking “How do I get well?”

127

Figure 6.1 - Rapid, persistent antidepressant effect of negative allosteric modulation of alpha-5 containing GABAARs a, Quantification of results from one group of rats in the sucrose preference test at five time points noted in the timeline above. Rats were treated with MRK-016 (3 mg/kg i.p.). Mean sucrose preference differed significantly following vehicle injection compared with all other groups (F(3,33)=20.63, p<0.0001, n=12 rats). * p< 0.05 compared to pre-CUS baseline, Tukey’s post-hoc). b, Results in individual animals for the full experiment. C, Quantification of results from one group of rats in the social interaction test at four time points noted in the timeline above. Mean social interaction ratios differed significantly following vehicle injection compared with all other groups (F=(2,14)= 11.84, p=0.0009, n=8 rats). *, p < 0.05 compared to pre-CUS baseline, Tukey’s post-hoc). From Fischell et al., 2015.

128

Figure 6.2 - Negative allosteric modulation of alpha5-containing GABAARs rapidly reverses AMPAR strength deficit at TA-CA1 synapses after chronic stress a, representative traces showing the AMPAR- and NMDAR-mediated components of the fEPSP, recorded extracellularly in SLM of area CA1 in response to simulation of TA afferents in saline lacking added Mg2+. Traces are shown before (black) and after (gray) addition of 50µM DNQX to block AMPARs. Traces were recorded from slices taken from unstressed controls (left) and rats subjected to 10 days of CRS then given an injection 24hrs earlier of either vehicle solution (middle) or L-655,708 (right). b, the mean slope of the relationship between AMPAR- (left) and NMDAR-mediated response (right) and FV amplitude over a range of stimulation intensities from the three groups of animals. The slope of the AMPAR-mediated responses was decreased significantly in slices from vehicle-treated CRS rats (red) compared to slices from unstressed (blue) or L-655,708-treated CRS rats (green) across the range of stimulation intensities (1-Way ANOVA, F(2,17)= 3.675, p= 0.047; n= 5 unstressed, 8 CRS+vehicle, 7 CRS+L-655,708; *, p < 0.05, LSD post hoc test). There were no significant differences in NMDAR-mediated responses (1-Way ANOVA, F(2,17)= 0.549, p = 0.588). c, AMPA:NMDA ratios were computed from the initial slopes of the responses in each slice before and after application of DNQX (shown by dotted lines in A). AMPA:NMDA ratios were decreased significantly in slices from vehicle-treated CRS rats compared to slices from either unstressed or L-655,708-treated CRS rats (1-Way ANOVA F(2,17)= 4.345 p = 0.03; n= 5 unstressed, 8 CRS+vehicle, 7 CRS+L-655,708; *, p < 0.05 compared to unstressed and CRS+L-655,708, LSD post-hoc test). From Fischell et al., 2015.

129

7) Appendix A: Protocols & recipes Sucrose Preference Test

Before testing with 1% sucrose solution in tap water, we find it very useful to ‘train’ rats or mice with a 2% sucrose solution, by providing a choice between 2% sucrose in tap water and plain water. It may be necessary to order special bottles from pet supply stores that do not leak as much as standard animal facility bottles.

3 hour test (used in Cai et al., 2013): After 3 hours of water deprivation, rats were individually housed and given a choice between two bottles for 3-4 hrs, one with 1% sucrose solution and another with normal drinking water. To prevent the possible effects of side preference in drinking, the position of the bottles was reversed halfway through the test. Mice were tested over

12 hours (6 hrs light/6hrs dark) with bottle positions switched after 6 hrs. The consumption of water and sucrose were measured by weighing the bottles, both before and after the test.

Preference for sucrose was calculated as a percentage of consumed sucrose-containing solution relative to the total amount of liquid intake. 50% means that the animal drank equally from both bottles, i.e. they had no preference for sucrose, while 100% means that the rat or mouse only drank from the sucrose-containing solution. Naïve animals were tested repeatedly in a group housing until the cage had a consistent preferences for 1% sucrose, but were always tested individually thereafter. For experiments with only wild type animals, any individual that did not demonstrate a sucrose preference >65% was discarded. In some experiments, animals were first trained to the task using 2% sucrose. Tests were repeated once per week during the course of

CUS or antidepressant administration.

130

Overnight sucrose preference test (used in Kallarackal, Kvarta et al., 2013 and Kvarta et al.,

2015): In order to increase reliability of the measurements, it is desirable to have a longer sucrose preference test, or for it to extend during the entire active (dark) period. This method can also be used to increase the sensitivity of the test because animals will drink more over this time period.

Here, we performed the test for the entire dark cycle (>12 h). Water deprivation beforehand is not necessary in rats.

131

Artificial Cerebrospinal Fluid (ACSF)

Solute Concentration g/L

NaCl 120 mM 7.013

KCl 3 mM 0.223

NaH2PO4 1.0 mM 0.120

MgSO4*7H20 1.5 mM 0.369

CaCl2*2H20 2.5 mM 0.3725

NaHCO3 25 mM 2.10

Glucose 20 mM 3.60

Other drugs used in ACSF: Concentration

APV (NMDAR antagonist) 80µM

Anpirtoline (5HT1BR agonist) 50µM

CGP52432 (GABABR antagonist) 2µM

DNQX (AMPAR antagonist) 50µM

Fluoxetine (SSRI) 20µM

Imipramine (TCA) 2µM

Isamoltane (5HT1BR antagonist) 10µM

Picrotoxin (GABAAR antagonist) 100µM

132

Antibodies

Epitope Dilution Source Species kDa

β–actin 1:2-5000 Cell Signaling #4967 Rabbit 45

CaMKII pThr286 1:5000 Cell Signaling #3361 Rabbit 50

CaMKII (pan) 1:2000 Cell Signaling #3362 Rabbit 60,50

GluA1 pS831 1:1000 Sigma #A4352 Rabbit 101

GluA1 (pan) 1:1000 Thermo Scientific Rabbit 101

#PA1-37776

GluA2 1:1000 Millipore #MAB397 Mouse 102

GR 1:1000 Millipore #PC170 Rabbit 94-97

Mouse IgG HRP 1:1000 Millipore #12-349 Goat

NMDAR1 1:1000 Millipore #AB9864 Rabbit 120

PSD95 1:2000 Millipore #AB9634 Rabbit 95

Rabbit IgG HRP 1:1000 Cell Signaling #7074 Goat

5HT1BR 1:1000 Thermo Scientific Rabbit 40 #PA1-29463

133

Molecular Biology Buffers

RIPA buffer (final concentration)

2.5 ml 1M Tris-Cl pH7.5 (50mM)

1.5 ml 5M NaCl (150mM)

0.1 ml 0.5M EDTA (1mM)

2.5 ml 20% Triton-X100 (1%)

0.5 ml 10% SDS (0.1%)

42.9 ml dH2O

Lysis buffer (RIPA/PPI/0.5%DOC)

1ml RIPA

50ul 10% DOC Deoxycholic acid (DOC – aka sodium deoxycholate)

2ul β-mercaptoEtOH

20ul Protease (Sigma-P8340 stored at -20oC)

20ul PPT (Sigma-P5726 and P2850 located in 4oC)

134

Metyrapone preparation and dosing

40% PEG

60% Sterile saline

25 mg/mL of prepared drug solution

50mg/kg dosage of MET

2 mL vehicle or drug solution / kg rat

Example dosing chart

Rat size (g) Injection amount (mL)

50 0.10

75 0.15

100 0.20

125 0.25

150 0.30

175 0.35

200 0.40

135

CORT for drinking water

For rats, the goal is 50ug/mL of CORT. For mice, use 25 ug/mL.

CORT hemisuccinate (4-pregnen-11β 21-DIOL-3 20-DIONE 21-hemisuccinate) dissolves in water (see Gourley and Taylor 2009). Obtained from Steraloids Inc

(steraloids.com); their catalog number for CORT hemisuccinate is Q1562.

64.45 mg of the compound per liter of tap water will give the desired 50 mg/L of CORT.

Use NaOH to titrate pH up to about 12 and add steroid. Leave it stirring in cold room.

After it’s fully dissolved (~8 hours or overnight), titrate back to 7.2. The solution should be swapped out every 72 hours at room temp. Keep any excess in the refrigerator.

136

Fecal CORT extraction

Collect fecal pellet samples and store in -20°C or -80°C until ready for extraction. I used

Corticosterone ELISA kits from Assaypro, Inc. (St Charles, MO)

1. Frozen samples thawed, dried overnight in centrifugal evaporator

2. Crush remains into a dust-like material

3. Weigh out 0.2 g of material into a 15mL centrifuge tube

4. Add 10mL ethanol to each sample

5. Boil each sample+ethanol in a water bath for 20min

6. Centrifuge for 15 min, then pour supernatant into another tube

7. Add another 5mL ethanol into original sample tube, vortex for 1 min, centrifuge

for 15 min, add supernatant to the first ~10mL of supernatant

8. Allow the ~15mL of supernatant to boil off in a water bath, or allow to evaporate

under air; reconstitute with 1mL of methanol and store at -80°C until assay

9. Proceed with ELISA kit instructions – bring all reagents to room temperature

before use

a. Add 25 uL of standard/sample and 25uL of biotinylated protein per well

(incubate 2hrs)

b. Wash, then add 50uL of SP per well (incubate 30 min)

c. Wash, then add 50uL of Chromogen Substrate per well (incubate 12 min)

d. Add 50uL Stop Solution per well – Read at 450 nm immediately

137

8) Appendix B: TA-CA1 plasticity – role of NMDARs and 5HT1BRs In this work, a major focus has centered on the synaptic strength of the excitatory temporoammonic connection from entorhinal cortex on distal dendrites of CA1 pyramidal neurons in stratum lacunosum moleculare (SLM). A critical role for mechanisms of synaptic plasticity is implicated in monoaminergic-based antidepressant action at TA-CA1 synapses

(Chapter 3). In this Appendix, I present several preliminary experiments that were performed to give confidence in the mechanisms of plasticity at TA-CA1, in light of some confusing reports in the literature.

The TA-CA1 pathway is activated during learning tasks, as metabolic labeling demonstrates activity in SLM, suggesting involvement in the cellular substrate of learning and memory, LTP.

The pathway is sufficient for CA1 place cell firing, and can preserve spatial recognition in the face of a Schaffer collateral lesion (Brun et al., 2002). However, physiological investigation of

LTP at TA-CA1 has yielded a bit of a nuanced story.

A common assumption is that mechanisms concerning synaptic plasticity are considered to be general concepts: that is, the major mediators of synaptic plasticity are constant regardless of the synapse being examined. However, it is obvious that these assumptions only go so far. One would not generalize plasticity at an excitatory synapse in the CNS to the neuromuscular junction without sufficient rationale. The majority of important work characterizing synaptic plasticity in

CA1 has been performed at CA3-CA1 Schaffer collateral synapses.

One study reported that GABAA blockade was required to elicit TA-CA1 LTP due to a heavy inhibitory component present in SLM (Colbert and Levy, 1993), while another study reported no such requirement (Remondes and Schuman, 2003). However, Remondes and Schuman also demonstrated an incomplete contribution of NMDAR receptors, with only partial LTP observed following high frequency stimulation. Blockage of both voltage-gated calcium channels and

138

NMDAR receptors was required to fully inhibit LTP. Furthermore, they reported a requirement of intact GABABRs to observe LTP, believed to be due to auto-inhibition of GABA terminals.

I thus tested the NMDAR-dependence of LTP induction in our hands, by pre-incubating slices in

APV before application of high-frequency stimulation (100Hz, 4x1s trains). Although results were only preliminary with two slices tested, they were clear enough to satisfy us that NMDARs are required for observation of any LTP at TA-CA1 synapses (Figure 8.1).

Figure 8.1 - LTP of TA-CA1 fEPSPS is NMDAR-dependent

5HT1BR-mediated signaling is required for the synaptic potentiation caused by antidepressants or anpirtoline, which results in CaMKII-mediated phosphorylation of AMPARs, in the vein of classically described LTP. Inspired by the possibility that mechanisms of LTP may be different in a layer specific manner, and the role of 5HT1BR signaling in commandeering mechanisms of synaptic plasticity to elicit potentiation, I tested the hypothesis that 5HT1BR activation was required for LTP at TA-CA1 synapses, as electrical stimulation could result in release of endogenous serotonin from raphe terminals in SLM.

To do this, I washed in the 5HT1BR antagonist isamoltane while recording baseline fEPSPs at TA-CA1, and then applied a high frequency stimulus. In two slices, LTP still occurred in the presence of isamoltane (10µM) after 60 minutes, at a concentration that prevented SSRI,

TCA, and anpirtoline-induced potentiation (Chapter 3, Cai et al., 2013).

139

Figure 8.2 - TA-CA1 LTP does not depend on 5HT1BR signaling

In one slice, I subsequently washed out isamoltane, to test the possibility that 5HT1BR antagonism prevented these synapses from fully potentiating. A subsequent HFS resulted in no further potentiating. Thus, I conclude that 5HT1BR antagonism does not result in a sub-maximal potentiation following electrical induction of LTP.

140

Figure 8.3 - 5HT1BR antagonism does not result in submaximal LTP After LTP induction in the presence of isamoltane in Figure 8.2, isamoltane was washed out for 30 minutes and a new baseline was established. A second HFS was applied, demonstrating that the LTP induction in the previous figure in the presence of isamoltane was not sub-maximal.

141

9) Literature cited

Adham N, Romanienko P, Hartig P, Weinshank RL, Branchek T. The rat 5- hydroxytryptamine1B receptor is the species homologue of the human 5- hydroxytryptamine1D beta receptor. Mol Pharmacol 41: 1–7, 1992.

Airan RD, Meltzer L a, Roy M, Gong Y, Chen H, Deisseroth K. High-speed imaging reveals neurophysiological links to behavior in an animal model of depression.

Science (80- ) 317: 819–23, 2007.

Aït Amara D, Segu L, Naïli S, Buhot MC. Serotonin 1B receptor regulation after dorsal subiculum deafferentation. Brain Res Bull 38: 17–23, 1995.

Akiskal HS, McKinney WT. Depressive disorders: toward a unified hypothesis.

Science 182: 20–29, 1973.

Albert PR, François B Le. Modifying 5-HT1A Receptor Gene Expression as a

New Target for Antidepressant Therapy. Front Neurosci 4: 35, 2010.

Alfarez DN, Joels M, Krugers HJ. Chronic unpredictable stress impairs long- term potentiation in rat hippocampal CA1 area and dentate gyrus in vitro. Eur J Neurosci

17: 1928–1934, 2003.

Alfarez DN, De Simoni A, Velzing EH, Bracey E, Joëls M, Edwards F a,

Krugers HJ. Corticosterone reduces dendritic complexity in developing hippocampal

CA1 neurons. Hippocampus 19: 828–36, 2009.

Alfarez DN, Wiegert O, Krugers HJ. Stress, corticosteroid hormones and hippocampal synaptic function. CNS Neurol Disord Drug Targets 5: 521–529, 2006.

142

Amaral D, Lavenex P. Hippocampal Neuroanatomy. In: The Hippocampus

Book, edited by Andersen P, Morris R, Amaral D, Bliss T, O’Keefe J. New York, NY:

Oxford University Press, 2006, p. 37–114.

Amat J, Baratta M V, Paul E, Bland ST, Watkins LR, Maier SF. Medial prefrontal cortex determines how stressor controllability affects behavior and dorsal raphe nucleus. Nat Neurosci 8: 365–371, 2005.

Anacker C, Zunszain P a, Cattaneo A, Carvalho L a, Garabedian MJ,

Thuret S, Price J, Pariante CM. Antidepressants increase human hippocampal neurogenesis by activating the glucocorticoid receptor. Mol Psychiatry 16: 738–50,

2011a.

Anacker C, Zunszain PA, Carvalho LA, Pariante CM. The glucocorticoid receptor: Pivot of depression and of antidepressant treatment? Psychoneuroendocrinology

36: 415–425, 2011b.

Andersen P, Morris R, Amaral D, Bliss T, O’Keefe J. The Hippocampus Book.

Oxford University Press.

Antypa N, Serretti A, Rujescu D. Serotonergic genes and suicide: A systematic review. Eur. Neuropsychopharmacol. 23: 1125–1142, 2013.

Atack JR, Alder L, Cook SM, Smith AJ, McKernan RM. In vivo labelling of alpha5 subunit-containing GABA(A) receptors using the selective radioligand [(3)H]L-

655,708. Neuropharmacology 49: 220–9, 2005.

Atack JR, Maubach KA, Wafford KA, Connor DO, Rodrigues AD, Evans

DC, Tattersall FD, Chambers MS, Macleod AM, Eng W, Ryan C, Hostetler E,

Sanabria SM, Gibson RE, Krause S, Burns HD, Hargreaves RJ, Agrawal NGB,

143

Mckernan RM, Murphy MG, Gingrich K, Dawson GR, Musson DG, Petty KJ. In

Vitro and in Vivo Properties of 3- tert -Butyl-7- ( 5-methylisoxazol- Selective Inverse

Agonist. J Pharmacol Exp Ther 331: 470–484, 2009.

Autry AE, Adachi M, Nosyreva E, Na ES, Los MF, Cheng P, Kavalali ET,

Monteggia LM. NMDA receptor blockade at rest triggers rapid behavioural antidepressant responses. Nature 475: 91–5, 2011.

Ball JR, Kiloh LG. A controlled trial of imipramine in treatment of depressive states. Br Med J 2: 1052–1055, 1959.

Bamberg E, Palme R, Meingassner JG. Excretion of corticosteroid metabolites in urine and faeces of rats. Lab Anim 35: 307–314, 2001.

Barbon A, Caracciolo L, Orlandi C, Musazzi L, Mallei A, La Via L, Bonini

D, Mora C, Tardito D, Gennarelli M, Racagni G, Popoli M, Barlati S. Chronic antidepressant treatments induce a time-dependent up-regulation of AMPA receptor subunit protein levels. Neurochem Int 59: 896–905, 2011.

Barnes NM, Sharp T. A review of central 5-HT receptors and their function.

Neuropharmacology 38: 1083–1152, 1999.

Baumeister AA, Hawkins MF, Uzelac SM. The myth of reserpine-induced depression: role in the historical development of the monoamine hypothesis. J Hist

Neurosci 12: 207–220, 2003.

Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 4: 561–571, 1961.

144

Berger AJ, Huynh P. Activation of 5HT1B receptors inhibits glycinergic synaptic inputs to mammalian motoneurons during postnatal development. Brain Res

956: 380–384, 2002.

Berman RM, Cappiello A, Anand A, Oren D a, Heninger GR, Charney DS,

Krystal JH. Antidepressant effects of ketamine in depressed patients. Biol Psychiatry 47:

351–354, 2000.

Bernard C. Les phenomenes de la vie. Paris, France, France: Librarie J-B Baillier et Fils, 1878.

Berridge KC, Kringelbach ML. Neuroscience of affect: brain mechanisms of pleasure and displeasure. Curr Opin Neurobiol 23: 294–303, 2013.

Berton O, McClung CA, Dileone RJ, Krishnan V, Renthal W, Russo SJ,

Graham D, Tsankova NM, Bolanos CA, Rios M, Monteggia LM, Self DW, Nestler

EJ. Essential role of BDNF in the mesolimbic dopamine pathway in social defeat stress.

Science (80- ) 311: 864–8, 2006.

Berton O, Nestler EJ. New approaches to antidepressant drug discovery: beyond monoamines. Nat Rev Neurosci 7: 137–51, 2006.

Bessa JM, Ferreira D, Melo I, Marques F, Cerqueira JJ, Palha J a, Almeida

OFX, Sousa N. The mood-improving actions of antidepressants do not depend on neurogenesis but are associated with neuronal remodeling. Mol Psychiatry 14: 764–773,

739, 2009.

Billings AG, Cronkite RC, Moos RH. Social-environmental factors in unipolar depression: comparisons of depressed patients and nondepressed controls. J Abnorm

Psychol 92: 119–33, 1983.

145

Billings AG, Moos RH. Chronic and nonchronic unipolar depression. The differential role of environmental stressors and resources. J Nerv Ment Dis 172: 65–75,

1984.

Bliss T V, Collingridge GL. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361: 31–39, 1993.

Bock C, Bukh JD, Vinberg M, Gether U, Kessing LV. The influence of comorbid personality disorder and neuroticism on treatment outcome in first episode depression. Psychopathology 43: 197–204, 2010.

Boldrini M, Underwood MD, Hen R, Rosoklija GB, Dwork AJ, John Mann

J, Arango V. Antidepressants increase neural progenitor cells in the human hippocampus. Neuropsychopharmacology 34: 2376–89, 2009.

Bolger N, Schilling EA. Personality and the problems of everyday life: the role of neuroticism in exposure and reactivity to daily stressors. J Pers 59: 355–386, 1991.

Bondi CO, Rodriguez G, Gould GG, Frazer A, Morilak D a. Chronic unpredictable stress induces a cognitive deficit and anxiety-like behavior in rats that is prevented by chronic antidepressant drug treatment. Neuropsychopharmacology 33: 320–

31, 2008.

Boulenguez P, Rawlins JNPJ, Chauveau J, Joseph MH, Mitchell SN, Gray

JA. Modulation of dopamine release in the nucleus accumbens by 5-HT(1B) agonists:

Involvement of the hippocampo-accumbens pathway. Neuropharmacology 35: 1521–

1529, 1996.

Boyle MP, Brewer JA, Funatsu M, Wozniak DF, Tsien JZ, Izumi Y, Muglia

LJ. Acquired deficit of forebrain glucocorticoid receptor produces depression-like

146 changes in adrenal axis regulation and behavior. Proc Natl Acad Sci U S A 102: 473–8,

2005.

Brun VH, Otnass MK, Molden S, Steffenach H-A, Witter MP, Moser M-B,

Moser EI. Place cells and place recognition maintained by direct entorhinal-hippocampal circuitry. Science (80- ) 296: 2243–6, 2002.

Buard I, Coultrap SJ, Freund RK, Lee Y-S, Dell’Acqua ML, Silva AJ, Bayer

KU. CaMKII “autonomy” is required for initiating but not for maintaining neuronal long- term information storage. J Neurosci 30: 8214–8220, 2010.

Buhl EH, Whittington M. Local Circuits. In: The Hippocampus Book, edited by

Andersen P, Morris R, Amaral D, Bliss T, O’Keefe J. New York, NY: Oxford University

Press, 2006, p. 297–319.

Buhot MC, Patra SK, Naïli S. Spatial memory deficits following stimulation of hippocampal 5-HT1B receptors in the rat. Eur J Pharmacol 285: 221–228, 1995.

Buhot M-C, Wolff M, Benhassine N, Costet P, Hen R, Segu L. Spatial learning in the 5-HT1B receptor knockout mouse: selective facilitation/impairment depending on the cognitive demand. Learn Mem 10: 466–477, 2003.

Bukh JD, Bock C, Vinberg M, Gether U, Kessing LV. Differences between early and late onset adult depression. Clin Pract Epidemiol Ment Health 7: 140–147,

2011.

Cai X, Kallarackal AJ, Kvarta MD, Goluskin S, Gaylor K, Bailey AM, Lee

H-K, Huganir RL, Thompson SM. Local potentiation of excitatory synapses by serotonin and its alteration in rodent models of depression. Nat Neurosci 16: 464–72,

2013.

147

Campbell S, MacQueen GM. The role of the hippocampus in the pathophysiology of major depression. J Psychiatry Neurosci 29: 417–26, 2004.

Cannon W. The wisdom of the body. Physiol Rev 9: 399–431, 1929.

Carlezon W a, Chartoff EH. Intracranial self-stimulation (ICSS) in rodents to study the neurobiology of motivation. Nat Protoc 2: 2987–2995, 2007.

Carragher N, Adamson G, Bunting B, McCann S. Subtypes of depression in a nationally representative sample. J Affect Disord 113: 88–99, 2009.

Cases O, Seif I, Grimsby J, Gaspar P, Chen K, Pournin S, Müller U, Aguet

M, Babinet C, Shih JC. Aggressive behavior and altered amounts of brain serotonin and norepinephrine in mice lacking MAOA. Science 268: 1763–1766, 1995.

Chan SWY, Goodwin GM, Harmer CJ. Highly neurotic never-depressed students have negative biases in information processing. Psychol Med 37: 1281–1291,

2007.

Chaudhury D, Walsh JJ, Friedman AK, Juarez B, Ku SM, Koo JW,

Ferguson D, Tsai H-C, Pomeranz L, Christoffel DJ, Nectow AR, Ekstrand M,

Domingos A, Mazei-Robison MS, Mouzon E, Lobo MK, Neve RL, Friedman JM,

Russo SJ, Deisseroth K, Nestler EJ, Han M-H. Rapid regulation of depression-related behaviours by control of midbrain dopamine neurons. Nature 493: 532–6, 2013.

Chen Y, Andres AL, Frotscher M, Baram TZ. Tuning synaptic transmission in the hippocampus by stress: the CRH system. Front Cell Neurosci 6: 1–7, 2012.

Chen Y, Dubé CM, Rice CJ, Baram TZ. Rapid loss of dendritic spines after stress involves derangement of spine dynamics by corticotropin-releasing hormone. J

Neurosci 28: 2903–11, 2008.

148

Chenu F, David DJP, Leroux-Nicollet I, Le Maêtre E, Gardier AM, Bourin

M. Serotonin1B heteroreceptor activation induces an antidepressant-like effect in mice with an alteration of the serotonergic system. J Psychiatry Neurosci 33: 541–550, 2008.

Christoffel DJ, Golden SA, Russo SJ. Structural and synaptic plasticity in stress-related disorders. Rev Neurosci 22: 535–549, 2011.

Colbert CM, Levy WB. Long-term potentiation of perforant path synapses in hippocampal CA1 in vitro. Brain Res 606: 87–91, 1993.

Colle R, Chupin M, Cury C, Vandendrie C, Gressier F, Hardy P, Falissard

B, Colliot O, Ducreux D, Corruble E. Depressed suicide attempters have smaller hippocampus than depressed patients without suicide attempts. J Psychiatr Res 61: 13–

18, 2015.

Conrad CD, Galea LA, Kuroda Y, McEwen BS. Chronic stress impairs rat spatial memory on the Y maze, and this effect is blocked by tianeptine pretreatment.

Behav Neurosci 110: 1321–1334, 1996.

Cook SC, Wellman CL. Chronic stress alters dendritic morphology in rat medial prefrontal cortex. J Neurobiol 60: 236–248, 2004.

Coryell W, Schlesser M. The dexamethasone suppression test and suicide prediction. Am J Psychiatry 158: 748–753, 2001.

Covington HE, Lobo MK, Maze I, Vialou V, Hyman JM, Zaman S, LaPlant

Q, Mouzon E, Ghose S, Tamminga CA, Neve RL, Deisseroth K, Nestler EJ.

Antidepressant effect of optogenetic stimulation of the medial prefrontal cortex. J

Neurosci 30: 16082–16090, 2010.

149

Cryan JF, Markou A, Lucki I. Assessing antidepressant activity in rodents:

Recent developments and future needs. Trends Pharmacol Sci 23: 238–245, 2002.

Delgado PL, Miller HL, Salomon RM, Licinio J, Krystal JH, Moreno FA,

Heninger GR, Charney DS. Tryptophan-depletion challenge in depressed patients treated with desipramine or fluoxetine: Implications for the role of serotonin in the mechanism of antidepressant action. Biol Psychiatry 46: 212–220, 1999.

Detke MJ, Johnson J, Lucki I. Acute and chronic antidepressant drug treatment in the rat forced swimming test model of depression. Exp Clin Psychopharmacol 5: 107–

12, 1997.

Doucet E, Pohl M, Fattaccini CM, Adrien J, El Mestikawy S, Hamon M. In situ hybridization evidence for the synthesis of 5-HT(1B) receptor in serotoninergic neurons of anterior raphe nuclei in the rat brain. Synapse 19: 18–28, 1995.

Downar J, Geraci J, Salomons T V, Dunlop K, Wheeler S, McAndrews MP,

Bakker N, Blumberger DM, Daskalakis ZJ, Kennedy SH, Flint AJ, Giacobbe P.

Anhedonia and reward-circuit connectivity distinguish nonresponders from responders to dorsomedial prefrontal repetitive transcranial magnetic stimulation in major depression.

Biol Psychiatry 76: 176–185, 2014.

Drevets WC. Neuroimaging and neuropathological studies of depression:

Implications for the cognitive-emotional features of mood disorders. Curr Opin

Neurobiol 11: 240–249, 2001.

Drevets WC. Neuroimaging abnormalities in the amygdala in mood disorders.

Ann N Y Acad Sci 985: 420–444, 2003.

150

Du J, Gray NA, Falke CA, Chen W, Yuan P, Szabo ST, Einat H, Manji HK.

Modulation of synaptic plasticity by antimanic agents: the role of AMPA glutamate receptor subunit 1 synaptic expression. J Neurosci 24: 6578–6589, 2004.

Dudek SM, Bear MF. Bidirectional long-term modification of synaptic effectiveness in the adult and immature hippocampus. J Neurosci 13: 2910–2918, 1993.

Dulawa SC, Holick K a, Gundersen B, Hen R. Effects of chronic fluoxetine in animal models of anxiety and depression. Neuropsychopharmacology 29: 1321–30, 2004.

Dulawa SC. Novelty-Induced Hypophagia. In: Mood and Anxiety Related

Phenotypes in Mice, edited by Gould TD. New York, NY: Humana Press, 2009, p. 247–

259.

Duman RS, Aghajanian GK. Synaptic Dysfunction in Depression: Potential

Therapeutic Targets. Science (80- ) 338: 68–72, 2012.

Duman RS, Li N. A neurotrophic hypothesis of depression: role of synaptogenesis in the actions of NMDA receptor antagonists. Philos Trans R Soc Lond B

Biol Sci 367: 2475–84, 2012.

Duman RS, Monteggia LM. A neurotrophic model for stress-related mood disorders. Biol Psychiatry 59: 1116–27, 2006.

Duman RS. Neurobiology of stress, depression, and rapid acting antidepressants: remodeling synaptic connections. Depress Anxiety 6: 1–6, 2014.

Duric V, Banasr M, Stockmeier CA, Simen AA, Newton SS, Overholser JC,

Jurjus GJ, Dieter L, Duman RS. Altered expression of synapse and glutamate related genes in post-mortem hippocampus of depressed subjects. Int J Neuropsychopharmacol

16: 69–82, 2013.

151

Dwivedi Y, Rizavi HS, Conley RR, Roberts RC, Tamminga CA, Pandey GN.

Altered gene expression of brain-derived neurotrophic factor and receptor tyrosine kinase

B in postmortem brain of suicide subjects. Arch Gen Psychiatry 60: 804–15, 2003.

Eley T, McAdams T a., Rijsdijk F V., Lichtenstein P, Narusyte J, Reiss D,

Spotts EL, Ganiban JM, Neiderhiser JM. The Intergenerational Transmission of

Anxiety: A Children-of-Twins Study. Am J Psychiatry : appi.ajp.2015.1, 2015.

Enns MW, Cox BJ. Personality dimensions and depression: review and commentary. Can J Psychiatry 42: 274–284, 1997.

Fales CL, Barch DM, Rundle MM, Mintun MA, Mathews J, Snyder AZ,

Sheline YI. Antidepressant treatment normalizes hypoactivity in dorsolateral prefrontal cortex during emotional interference processing in major depression. J Affect Disord 112:

206–11, 2009.

Farley S, Apazoglou K, Witkin JM, Giros B, Tzavara ET. Antidepressant-like effects of an AMPA receptor potentiator under a chronic mild stress paradigm. Int J

Neuropsychopharmacol 13: 1207–1218, 2010.

Fava M, Kendler KS. Major Depressive Disorder Review. Neuron 28: 335–341,

2000.

Felix-Ortiz AC, Tye KM. Amygdala inputs to the ventral hippocampus bidirectionally modulate social behavior. J Neurosci 34: 586–95, 2014.

Femenía T, Gómez-Galán M, Lindskog M, Magara S. Dysfunctional hippocampal activity affects emotion and cognition in mood disorders. Brain Res 1476:

58–70, 2012.

152

Fischell J, Van Dyke AM, Kvarta MD, LeGates TA, Thompson SM. Rapid antidepressant Action and Restoration of Excitatory Synaptic Strength after Chronic

Stress by Negative Modulators of alpha5-containing GABAA Receptors.

Neuropsychopharmacology In Press: 1–29, 2015.

Flint J, Kendler KS. The genetics of major depression. Neuron 81: 484–503,

2014.

Francis TC, Chandra R, Friend DM, Finkel E, Dayrit G, Miranda J, Brooks

JM, Iñiguez SD, O’Donnell P, Kravitz A, Lobo MK. Nucleus accumbens medium spiny neuron subtypes mediate depression-related outcomes to social defeat stress. Biol

Psychiatry 77: 212–22, 2015.

Frodl T, Meisenzahl EM, Zetzsche T, Born C, Groll C, Jäger M, Leinsinger

G, Bottlender R, Hahn K, Möller HJ. Hippocampal changes in patients with a first episode of major depression. Am J Psychiatry 159: 1112–1118, 2002.

Frodl T, Meisenzahl EM, Zetzsche T, Born C, Jäger M, Groll C, Bottlender

R, Leinsinger G, Möller H-J. Larger amygdala volumes in first depressive episode as compared to recurrent major depression and healthy control subjects. Biol Psychiatry 53:

338–44, 2003.

Frodl T, Schüle C, Schmitt G, Born C, Baghai T, Zill P, Bottlender R,

Rupprecht R, Bondy B, Reiser M, Möller H-J, Meisenzahl EM. Association of the brain-derived neurotrophic factor Val66Met polymorphism with reduced hippocampal volumes in major depression. Arch Gen Psychiatry 64: 410–416, 2007.

Garcia LSB, Comim CM, Valvassori SS, Réus GZ, Barbosa LM, Andreazza

AC, Stertz L, Fries GR, Gavioli EC, Kapczinski F, Quevedo J. Acute administration

153 of ketamine induces antidepressant-like effects in the forced swimming test and increases

BDNF levels in the rat hippocampus. Prog Neuropsychopharmacol Biol Psychiatry 32:

140–4, 2008.

Gaspar P, Cases O, Maroteaux L. The developmental role of serotonin: news from mouse molecular genetics. Nat Rev Neurosci 4: 1002–1012, 2003.

Gaynes B, Warden D. What did STAR* D teach us? Results from a large-scale, practical, clinical trial for patients with depression. Psychiatry Serv 60: 1439–45, 2009.

Giese KP, Fedorov NB, Filipkowski RK, Silva AJ. Autophosphorylation at

Thr286 of the alpha calcium-calmodulin kinase II in LTP and learning. Science 279: 870–

873, 1998.

Gilbert PE, Kesner RP. The amygdala but not the hippocampus is involved in pattern separation based on reward value. Neurobiol Learn Mem 77: 338–353, 2002.

Giles H, Lansdell SJ, Bolofo ML, Wilson HL, Martin GR. Characterization of a 5-HT1B receptor on CHO cells: functional responses in the absence of radioligand binding. Br J Pharmacol 117: 1119–1126, 1996.

Gould NF, Holmes MK, Fantie BD, Luckenbaugh DA, Pine DS, Gould TD,

Burgess N, Manji HK, Zarate CA. Performance on a virtual reality spatial memory navigation task in depressed patients. Am J Psychiatry 164: 516–9, 2007.

Gourley SL, Kiraly DD, Howell JL, Olausson P, Jane R, Taylor JR. Acute hippocampal BDNF restores motivational and forced swim performance after corticosterone. J Neurosci 64: 884–890, 2009.

Gourley SL, Taylor JR. Recapitulation and reversal of a persistent depression- like syndrome in rodents. Curr Protoc Neurosci Chapter 9: Unit 9.32, 2009.

154

Gourley SL, Wu F, Kiraly DD, Ploski J, Kedves AT, Duman RS, Taylor JR.

Regionally specific regulation of ERK MAP kinase in a model of antidepressant-sensitive chronic depression. Biol Psychiatry 63: 353–359, 2008.

Gray J, Vecchiarelli HA, Hill MN. Endocannabinoid Signaling and Synaptic

Plasticity During Stress. In: Synaptic Stress and Pathogenesis of Neuropsychiatric

Disorders, edited by Popoli M, Diamond D, Sanacora G. New York: 2013, p. 99–124.

Gray JA. Integrating schizophrenia. Schizophr Bull 24: 249–266, 1998.

Grimsby J, Toth M, Chen K, Kumazawa T, Klaidman L, Adams JD,

Karoum F, Gal J, Shih JC. Increased stress response and beta-phenylethylamine in

MAOB-deficient mice. Nat Genet 17: 206–210, 1997.

Grippo AJ, Beltz TG, Weiss RM, Johnson AK. The effects of chronic fluoxetine treatment on chronic mild stress-induced cardiovascular changes and anhedonia. Biol Psychiatry 59: 309–16, 2006.

Grønli J, Murison R, Bjorvatn B, Sørensen E, Portas CM, Ursin R. Chronic mild stress affects sucrose intake and sleep in rats. Behav Brain Res 150: 139–47, 2004.

Guidotti G, Calabrese F, Anacker C, Racagni G, Pariante CM, Riva M a.

Glucocorticoid receptor and FKBP5 expression is altered following exposure to chronic stress: modulation by antidepressant treatment. Neuropsychopharmacology 38: 616–27,

2013.

Hajek T, Kopecek M, Höschl C. Reduced hippocampal volumes in healthy carriers of brain-derived neurotrophic factor Val66Met polymorphism: meta-analysis. world J Biol psychiatry 13: 178–87, 2012.

155

Hajós M, Hoffmann WE, Orbán G, Kiss T, Erdi P. Modulation of septo- hippocampal Theta activity by GABAA receptors: an experimental and computational approach. Neuroscience 126: 599–610, 2004.

Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 23:

56–62, 1960.

Hammen C. Stress and depression. Annu Rev Clin Psychol 1: 293–319, 2005.

Hanson ND, Owens MJ, Nemeroff CB. Depression, Antidepressants, and

Neurogenesis: A Critical Reappraisal. Neuropsychopharmacology 36: 2589–2602, 2011.

Harper JM, Austad SN. Fecal glucocorticoids: a noninvasive method of measuring adrenal activity in wild and captive rodents. Physiol Biochem Zool 73: 12–22,

2012.

Hell JW. CaMKII: Claiming center stage in postsynaptic function and organization. Neuron 81: 249–265, 2014.

Heninger GR, Delgado PL, Charney DS. The revised monoamine theory of depression: a modulatory role for monoamines, based on new findings from monoamine depletion experiments in humans. Pharmacopsychiatry 29: 2–11, 1996.

Herman JP, Adams D, Prewitt C. Regulatory changes in neuroendocrine stress- integrative circuitry produced by a variable stress paradigm. Neuroendocrinology 61:

180–190, 1995.

Herman JP. Neural control of chronic stress adaptation. Front Behav Neurosci 7:

61, 2013.

Heuser I, Yassouridis A, Holsboer F. The combined dexamethasone/CRH test: a refined laboratory test for psychiatric disorders. J Psychiatr Res 28: 341–56, 1994.

156

Holsboer F. The corticosteroid receptor hypothesis of depression.

Neuropsychopharmacology 23: 477–501, 2000.

Huang Y, Coupland NJ, Lebel RM, Carter R, Seres P, Wilman AH,

Malykhin N V. Structural Changes in Hippocampal Subfields in Major Depressive

Disorder: A High-Field Magnetic Resonance Imaging Study. Biol Psychiatry 74: 62–8,

2013.

Huang Y, Oquendo M a, Friedman JMH, Greenhill LL, Brodsky B, Malone

KM, Khait V, Mann JJ. Substance abuse disorder and major depression are associated with the human 5-HT1B receptor gene (HTR1B) G861C polymorphism.

Neuropsychopharmacology 28: 163–9, 2003.

Ihara N, Ueda S, Kawata M, Sano Y. Immunohistochemical demonstration of serotonin-containing nerve fibers in the mammalian hippocampal formation. Acta Anat

(Basel) 132: 335–346, 1988.

Jeffcoate WJ, Silverstone JT, Edwards CR, Besser GM. Psychiatric manifestations of Cushing’s syndrome: response to lowering of plasma cortisol. Q J Med

48: 465–472, 1979.

Jefferson JW. A review of the cardiovascular effects and toxicity of tricyclic antidepressants. Psychosom Med 37: 160–179, 1975.

Joëls M, Baram TZ. The neuro-symphony of stress. Nat Rev Neurosci 10: 459–

66, 2009.

Joëls M, Karst H, Alfarez DN, Heine VM, Qin Y, van Riel E, Verkuyl M,

Lucassen PJ, Krugers HJ. Effects of chronic stress on structure and cell function in rat hippocampus and hypothalamus. Stress 7: 221–31, 2004.

157

Joëls M, Sarabdjitsingh RA, Karst H. Unraveling the time domains of corticosteroid hormone influences on brain activity: rapid, slow, and chronic modes.

Pharmacol Rev 64: 901–38, 2012.

Joëls M. Corticosteroid effects in the brain: U-shape it. Trends Pharmacol Sci 27:

244–50, 2006.

Joëls M. Functional actions of corticosteroids in the hippocampus. Eur J

Pharmacol 583: 312–21, 2008.

Johnson BN, Yamamoto BK. Chronic unpredictable stress augments +3,4- methylenedioxymethamphetamine-induced monoamine depletions: the role of corticosterone. Neuroscience 159: 1233–43, 2009.

Johnson BR, Brown JM, Kvarta MD, Lu JYJ, Schneider LR, Nadim F,

Harris-Warrick RM. Differential modulation of synaptic strength and timing regulate synaptic efficacy in a motor network. J Neurophysiol 105: 293–304, 2011.

Juruena MF, Cleare AJ, Pariante CM. The hypothalamic pituitary adrenal axis, glucocorticoid receptor function and relevance to depression. Rev Bras Psiquiatr 26:

189–201, 2004.

Kallarackal AJ, Kvarta MD, Cammarata E, Jaberi L, Cai X, Bailey AM,

Thompson SM. Chronic stress induces a selective decrease in AMPA receptor-mediated synaptic excitation at hippocampal temporoammonic-CA1 synapses. J Neurosci 33:

15669–74, 2013.

Kallarackal AJ. Serotonergic Modulation of Glutamate Transmission: Bridging

Two Theories of Depression. University of Maryland Baltimore2011.

158

Karanth S, Linthorst ACE, Stalla GK, Barden N, Holsboer F, Reul JMHM.

Hypothalamic-pituitary-adrenocortical axis changes in a transgenic mouse with impaired glucocorticoid receptor function. Endocrinology 138: 3476–3485, 1997.

Karst H, Berger S, Turiault M, Tronche F, Schütz G, Joëls M.

Mineralocorticoid receptors are indispensable for nongenomic modulation of hippocampal glutamate transmission by corticosterone. Proc Natl Acad Sci 102: 19204–

7, 2005.

Karst H, Joëls M. Effect of chronic stress on synaptic currents in rat hippocampal dentate gyrus neurons. J Neurophysiol 89: 625–33, 2003.

Karst H, Joëls M. Corticosterone slowly enhances miniature excitatory postsynaptic current amplitude in mice CA1 hippocampal cells. J Neurophysiol 94:

3479–86, 2005.

Katz MM, Tekell JL, Bowden CL, Brannan S, Houston JP, Berman N,

Frazer A. Onset and early behavioral effects of pharmacologically different antidepressants and placebo in depression. Neuropsychopharmacology 29: 566–579,

2004.

Katz RJ. Animal model of depression: Pharmacological sensitivity of a hedonic deficit. Pharmacol Biochem Behav 16: 965–968, 1982.

Kavalali ET, Monteggia LM. Synaptic mechanisms underlying rapid antidepressant action of ketamine. Am. J. Psychiatry 169: 1150–1156, 2012.

Kendler KS, Gardner CO. A longitudinal etiologic model for symptoms of anxiety and depression in women. Psychol Med 41: 2035–2045, 2011.

159

Kendler KS, Karkowski LM, Prescott CA. Causal relationship between stressful life events and the onset of major depression. Am J Psychiatry 156: 837–841,

1999.

Kendler KS, Neale MC, Kessler RC, Heath AC, Eaves LJ. A longitudinal twin study of personality and major depression in women. Arch Gen Psychiatry 50: 853–862,

1993.

Kennedy SH, Evans KR, Krüger S, Mayberg HS, Meyer JH, McCann S,

Arifuzzman AI, Houle S, Vaccarino FJ. Changes in regional brain glucose metabolism measured with positron emission tomography after paroxetine treatment of major depression. Am J Psychiatry 158: 899–905, 2001.

Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE.

Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National

Comorbidity Survey Replication. Arch Gen Psychiatry 62: 593–602, 2005.

De Kloet ER, Vreugdenhil E, Oitzl MS, Joëls M. Brain corticosteroid receptor balance in health and disease. Endocr Rev 19: 269–301, 1998.

Kole MHP, Swan L, Fuchs E. The antidepressant tianeptine persistently modulates glutamate receptor currents of the hippocampal CA3 commissural associational synapse in chronically stressed rats. Eur J Neurosci 16: 807–816, 2002.

Koolschijn PCMP, van Haren NEM, Lensvelt-Mulders GJLM, Hulshoff Pol

HE, Kahn RS. Brain volume abnormalities in major depressive disorder: a meta-analysis of magnetic resonance imaging studies. Hum Brain Mapp 30: 3719–35, 2009.

Krishnan V, Han M-H, Graham DL, Berton O, Renthal W, Russo SJ,

Laplant Q, Graham A, Lutter M, Lagace DC, Ghose S, Reister R, Tannous P, Green

160

TA, Neve RL, Chakravarty S, Kumar A, Eisch AJ, Self DW, Lee FS, Tamminga

CA, Cooper DC, Gershenfeld HK, Nestler EJ. Molecular adaptations underlying susceptibility and resistance to social defeat in brain reward regions. Cell 131: 391–404,

2007.

Krishnan V, Nestler EJ. The molecular neurobiology of depression. Nature 455:

894–902, 2008.

Krugers HJ, Hoogenraad CC, Groc L. Stress hormones and AMPA plasticity and memory. Nat Rev Neurosci 11: 675–81, 2010a.

Krugers HJ, Karst H, Joels M. Interactions between noradrenaline and corticosteroids in the brain: from electrical activity to cognitive performance. Front. Cell.

Neurosci. 6: 1–10, 2012.

Krugers HJ, Lucassen PJ, Karst H, Joëls M. Chronic stress effects on hippocampal structure and synaptic function: relevance for depression and normalization by anti-glucocorticoid treatment. Front Synaptic Neurosci 2: 24, 2010b.

Kvarta MD, Bradbrook K, Dantrassy H, Bailey AM, Thompson SM.

Corticosterone mediates the synaptic and behavioral effects of chronic stress at rat hippocampal temporoammonic synapses. [Submitted] 0, 2015.

Lammel S, Lim BK, Malenka RC. Reward and aversion in a heterogeneous midbrain dopamine system. Neuropharmacology 76 Pt B: 351–9, 2014.

Lee H-K, Barbarosie M, Kameyama K, Bear MF, Huganir RL. Regulation of distinct AMPA receptor phosphorylation sites during bidirectional synaptic plasticity.

Nature 405: 955–9, 2000.

161

LeGates TA, Fernandez DC, Hattar S. Light as a central modulator of circadian rhythms, sleep and affect. Nat Rev Neurosci , 2014.

Lemieux G, Davignon A, Genest J. Depressive states during Rauwolfia therapy for arterial hypertension; a report of 30 cases. Can Med Assoc J 74: 522–6, 1956.

Levinson DF. The Genetics of Depression: A Review. Biol Psychiatry 60: 84–92,

2006.

Li B, Piriz J, Mirrione M, Chung C, Proulx CD, Schulz D, Henn F, Malinow

R. Synaptic potentiation onto habenula neurons in the learned helplessness model of depression. Nature 470: 535–9, 2011a.

Li N, Lee B, Liu R-J, Banasr M, Dwyer JM, Iwata M, Li X-Y, Aghajanian G,

Duman RS. mTOR-dependent synapse formation underlies the rapid antidepressant effects of NMDA antagonists. Science (80- ) 329: 959–64, 2010.

Li N, Liu R-J, Dwyer JM, Banasr M, Lee B, Son H, Li X-Y, Aghajanian G,

Duman RS. Glutamate N-methyl-D-aspartate receptor antagonists rapidly reverse behavioral and synaptic deficits caused by chronic stress exposure. Biol Psychiatry 69:

754–61, 2011b.

Li X, Tizzano JP, Griffey K, Clay M, Lindstrom T, Skolnick P.

Antidepressant-like actions of an AMPA receptor potentiator (LY392098).

Neuropharmacology 40: 1028–1033, 2001.

Li Y, Luikart BW, Birnbaum S, Chen J, Kwon C-H, Kernie SG, Bassel-Duby

R, Parada LF. TrkB regulates hippocampal neurogenesis and governs sensitivity to antidepressive treatment. Neuron 59: 399–412, 2008.

162

Lim BK, Huang KW, Grueter B a, Rothwell PE, Malenka RC. Anhedonia requires MC4R-mediated synaptic adaptations in nucleus accumbens. Nature 487: 183–9,

2012.

Lisman J, Yasuda R, Raghavachari S. Mechanisms of CaMKII action in long- term potentiation. Nat Rev Neurosci 13: 169–82, 2012.

Lisman JE, Otmakhova N a. Storage, recall, and novelty detection of sequences by the hippocampus: Elaborating on the SOCRATIC model to account for normal and aberrant effects of dopamine. Hippocampus 11: 551–568, 2001.

Liston C, Miller MM, Goldwater DS, Radley JJ, Rocher AB, Hof PR,

Morrison JH, McEwen BS. Stress-induced alterations in prefrontal cortical dendritic morphology predict selective impairments in perceptual attentional set-shifting. J

Neurosci 26: 7870–4, 2006.

Lorenzetti V, Allen NB, Fornito A, Yücel M. Structural brain abnormalities in major depressive disorder: a selective review of recent MRI studies. J Affect Disord 117:

1–17, 2009.

MacGillavry H, Song Y, Raghavachari S, Blanpied T. Nanoscale scaffolding domains within the postsynaptic density concentrate synaptic ampa receptors. Neuron 78:

615–622, 2013.

Machado-Vieira R, Salvadore G, DiazGranados N, Zarate CA. Ketamine and the next generation of antidepressants with a rapid onset of action. Pharmacol Ther 123:

143–150, 2009.

163

MacQueen GM, Frodl T. The hippocampus in major depression: evidence for the convergence of the bench and bedside in psychiatric research? Mol Psychiatry 16:

252–64, 2011.

Maeng S, Zarate C a, Du J, Schloesser RJ, McCammon J, Chen G, Manji

HK. Cellular mechanisms underlying the antidepressant effects of ketamine: role of alpha-amino-3-hydroxy-5-methylisoxazole-4-propionic acid receptors. Biol Psychiatry

63: 349–52, 2008.

Magariños AM, McEwen BS, Flügge G, Fuchs E. Chronic Psychosocial Stress

Hippocampal CA3 Pyramidal Causes Apical Dendritic Neurons in Subordinate Atrophy of Tree Shrews. J Neurosci 76: 3534–3540, 1996.

Magariños AM, McEwen BS. Stress-induced atrophy of apical dendrites of hippocampal CA3c neurons: involvement of glucocorticoid secretion and excitatory amino acid receptors. Neuroscience 69: 89–98, 1995a.

Magariños AM, McEwen BS. Stress-induced atrophy of apical dendrites of hippocampal CA3c neurons: comparison of stressors. Neuroscience 69: 83–8, 1995b.

Mague SD, Pliakas AM, Todtenkopf MS, Tomasiewicz HC, Zhang Y, Stevens

WC, Jones RM, Portoghese PS, Carlezon WA. Antidepressant-like effects of kappa- opioid receptor antagonists in the forced swim test in rats. J Pharmacol Exp Ther 305:

323–30, 2003.

Malagié I, Trillat AC, Bourin M, Jacquot C, Hen R, Gardier AM. 5-HT1B

Autoreceptors limit the effects of selective serotonin re-uptake inhibitors in mouse hippocampus and frontal cortex. J Neurochem 76: 865–871, 2001.

164

Malberg JE, Eisch AJ, Nestler EJ, Duman RS. Chronic antidepressant treatment increases neurogenesis in adult rat hippocampus. J Neurosci 20: 9104–9110,

2000.

Malenka RC. Synaptic Plasticity and AMPA Receptor Trafficking. In: Annals of the New York Academy of Sciences. 2003, p. 1–11.

Mandelli L, Serretti A. Gene environment interaction studies in depression and suicidal behavior: An update. Neurosci Biobehav Rev 37: 2375–97, 2013.

Maras PM, Molet J, Chen Y, Rice C, Ji SG, Solodkin A, Baram TZ.

Preferential loss of dorsal-hippocampus synapses underlies memory impairments provoked by short, multimodal stress. Mol Psychiatry 19: 811–22, 2014.

Martin L a, Neighbors HW, Griffith DM. The experience of symptoms of depression in men vs women: analysis of the National Comorbidity Survey Replication.

JAMA psychiatry 70: 1100–6, 2013.

Martinez-Turrillas R, Frechilla D, Del Río J. Chronic antidepressant treatment increases the membrane expression of AMPA receptors in rat hippocampus.

Neuropharmacology 43: 1230–1237, 2002.

Martinowich K, Jimenez D V, Zarate C a, Manji HK. Rapid antidepressant effects: moving right along. Mol Psychiatry 18: 856–63, 2013.

Martinowich K, Lu B. Interaction between BDNF and serotonin: role in mood disorders. Neuropsychopharmacology 33: 73–83, 2008.

Mathur BN, Capik N a, Alvarez V a, Lovinger DM. Serotonin induces long- term depression at corticostriatal synapses. J Neurosci 31: 7402–11, 2011.

165

Matsumoto M, Hikosaka O. Representation of negative motivational value in the primate lateral habenula. Nat Neurosci 12: 77–84, 2009.

Mayberg HS, Lozano AM, Voon V, McNeely HE, Seminowicz D, Hamani C,

Schwalb JM, Kennedy SH. Deep brain stimulation for treatment-resistant depression.

Neuron 45: 651–60, 2005.

Mayorga AJ, Dalvi A, Page ME, Zimov-Levinson S, Hen R, Lucki I.

Antidepressant-like behavioral effects in 5-hydroxytryptamine(1A) and 5- hydroxytryptamine(1B) receptor mutant mice. J Pharmacol Exp Ther 298: 1101–1107,

2001.

McAllister-Williams RH, Smith E, Anderson IM, Barnes J, Gallagher P,

Grunze HC, Haddad PM, House AO, Hughes T, Lloyd AJ, McColl EM, Pearce SH,

Siddiqi N, Sinha B, Speed C, Steen IN, Wainright J, Watson S, Winter FH, Ferrier

IN. Study protocol for the randomised controlled trial: Antiglucocorticoid augmentation of anti-Depressants in Depression (The ADD Study). BMC Psychiatry 13: 205, 2013.

McEwen BS, Weiss JM, Schwartz LS. Selective retention of corticosterone by limbic structures in rat brain. Nature 220: 911–912, 1968.

McEwen BS. Stress, adaptation, and disease. Allostasis and allostatic load. Ann N

Y Acad Sci 840: 33–44, 1998.

McEwen BS. Central effects of stress hormones in health and disease:

Understanding the protective and damaging effects of stress and stress mediators. Eur J

Pharmacol 583: 174–85, 2008.

166

McKinney WT, Bunney WE, Samuels J, Robinson G, Price R. Animal model of depression. I. Review of evidence: implications for research. Arch Gen Psychiatry 21:

240–248, 1969.

McLaughlin JP, Marton-Popovici M, Chavkin C. Kappa opioid receptor antagonism and prodynorphin gene disruption block stress-induced behavioral responses.

J Neurosci 23: 5674–5683, 2003.

Medina A, Seasholtz AF, Sharma V, Burke S, Bunney W, Myers RM,

Schatzberg A, Akil H, Watson SJ. Glucocorticoid and mineralocorticoid receptor expression in the human hippocampus in major depressive disorder. J Psychiatr Res 47:

307–14, 2013.

Meldrum BS. Glutamate as a neurotransmitter in the brain: review of physiology and pathology. J Nutr 130: 1007S–15S, 2000.

Michael N, Erfurth A, Ohrmann P, Arolt V, Heindel W, Pfleiderer B.

Neurotrophic effects of electroconvulsive therapy: a proton magnetic resonance study of the left amygdalar region in patients with treatment-resistant depression.

Neuropsychopharmacology 28: 720–5, 2003.

Millan MJ, Agid Y, Brüne M, Bullmore ET, Carter CS, Clayton NS, Connor

R, Davis S, Deakin B, DeRubeis RJ, Dubois B, Geyer MA, Goodwin GM, Gorwood

P, Jay TM, Joëls M, Mansuy IM, Meyer-Lindenberg A, Murphy D, Rolls ET, Saletu

B, Spedding M, Sweeney J, Whittington M, Young LJ. Cognitive dysfunction in psychiatric disorders : characteristics, causes and the quest for improved therapy. Nat Rev

Drug Discov 11: 141–68, 2012.

167

Miller HL, Delgado PL, Salomon RM, Berman R, Krystal JH, Heninger GR,

Charney DS. Clinical and biochemical effects of catecholamine depletion on antidepressant-induced remission of depression. Arch Gen Psychiatry 53: 117–128, 1996.

Miller SG, Kennedy MB. Regulation of brain type II Ca2+/calmodulin- dependent protein kinase by autophosphorylation: a Ca2+-triggered molecular switch.

Cell 44: 861–870, 1986.

Mirescu C, Gould E. Stress and adult neurogenesis. Hippocampus 16: 233–8,

2006.

Mizoguchi K, Yuzurihara M, Ishige A, Sasaki H, Chui DH, Tabira T.

Chronic stress differentially regulates glucocorticoid negative feedback response in rats.

Psychoneuroendocrinology 26: 443–59, 2001.

Moghaddam B, Adams B, Verma A, Daly D. Activation of glutamatergic neurotransmission by ketamine: a novel step in the pathway from NMDA receptor blockade to dopaminergic and cognitive disruptions associated with the prefrontal cortex.

J Neurosci 17: 2921–2927, 1997.

Montgomery SA, Åsberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry 134: 382–389, 1979.

Montkowski A, Barden N, Wotjak C, Stec I, Ganster J, Meaney M,

Engelmann M, Reul JMHM, Landgraf R, Holsboer F. Long-term antidepressant treatment reduces behavioural deficits in transgenic mice with impaired glucocorticoid receptor function. J Neuroendocrinol 7: 841–845, 1995.

Morris RG, Garrud P, Rawlins JN, O’Keefe J. Place navigation impaired in rats with hippocampal lesions. Nature 297: 681–683, 1982.

168

Murray C, Lopez A. Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet 349: 1436–42, 1997.

Musazzi L, Treccani G, Mallei A, Popoli M. The action of antidepressants on the glutamate system: regulation of glutamate release and glutamate receptors. Biol

Psychiatry 73: 1180–8, 2013.

Nemeroff CB, Krishnan KR, Reed D, Leder R, Beam C, Dunnick NR.

Adrenal gland enlargement in major depression. A computed tomographic study. Arch

Gen Psychiatry 49: 384–387, 1992.

Nestler EJ, Barrot M, DiLeone RJ, Eisch AJ, Gold SJ, Monteggia LM.

Neurobiology of depression. Neuron 34: 13–25, 2002.

Nestler EJ, Carlezon WA. The mesolimbic dopamine reward circuit in depression. Biol Psychiatry 59: 1151–9, 2006.

Nestler EJ, Hyman SE. Animal models of neuropsychiatric disorders. Nat

Neurosci 13: 1161–1169, 2010.

Nichols DE, Nichols CD. Serotonin receptors. Chem Rev 108: 1614–1641, 2008.

Nicoll RA, Malenka RC. Contrasting properties of two forms of long-term potentiation in the hippocampus. Nature 377: 115–118, 1995.

Nicoll RA, Roche KW. Long-term potentiation: Peeling the onion.

Neuropharmacology 74Elsevier Ltd: 18–22, 2013.

Nosyreva E, Szabla K, Autry AE, Ryazanov AG, Monteggia LM, Kavalali

ET. Acute suppression of spontaneous neurotransmission drives synaptic potentiation. J

Neurosci 33: 6990–7002, 2013.

169

O’Dwyer A-M, Lightman SL, Marks M, Checkley S. Treatment of major depression with metyrapone and hydrocortisone. J Affect Disord 33: 123–128, 1995.

O’Neill MF, Conway MW. Role of 5-HT1A and 5-HT1B receptors in the mediation of behavior in the forced swim test in mice. Neuropsychopharmacology 24:

391–398, 2001.

Olijslagers JE, de Kloet ER, Elgersma Y, van Woerden GM, Joëls M, Karst

H. Rapid changes in hippocampal CA1 pyramidal cell function via pre- as well as postsynaptic membrane mineralocorticoid receptors. Eur J Neurosci 27: 2542–50, 2008.

Opazo P, Sainlos M, Choquet D. Regulation of AMPA receptor surface diffusion by PSD-95 slots. Curr. Opin. Neurobiol. 22: 453–460, 2012.

Otmakhova N a, Lisman JE. Dopamine selectively inhibits the direct cortical pathway to the CA1 hippocampal region. J Neurosci 19: 1437–1445, 1999.

Otmakhova N a, Lisman JE. Dopamine, serotonin, and noradrenaline strongly inhibit the direct perforant path-CA1 synaptic input, but have little effect on the Schaffer collateral input. Ann N Y Acad Sci 911: 462–464, 2000.

Palucha A, Pilc A. The involvement of glutamate in the pathophysiology of depression. Drug News Perspect 18: 262–268, 2005.

Papp M, Moryl E, Willner P. Pharmacological validation of the chronic mild stress model of depression. Eur J Pharmacol 296: 129–36, 1996.

Pariante CM. Risk factors for development of depression and psychosis.

Glucocorticoid receptors and pituitary implications for treatment with antidepressant and glucocorticoids. Ann N Y Acad Sci 1179: 144–52, 2009.

170

Parsey R V, Oquendo MA, Ogden RT, Olvet DM, Simpson N, Huang Y-Y,

Van Heertum RL, Arango V, Mann JJ. Altered serotonin 1A binding in major depression: a [carbonyl-C-11]WAY100635 positron emission tomography study. Biol

Psychiatry 59: 106–13, 2006.

Pawlak R, Rao BSS, Melchor JP, Chattarji S, McEwen BS, Strickland S.

Tissue plasminogen activator and plasminogen mediate stress-induced decline of neuronal and cognitive functions in the mouse hippocampus. Proc Natl Acad Sci U S A

102: 18201–6, 2005.

Petit-Demouliere B, Chenu F, Bourin M. Forced swimming test in mice: a review of antidepressant activity. Psychopharmacology (Berl) 177: 245–55, 2005.

Pfleiderer B, Michael N, Erfurth A, Ohrmann P, Hohmann U, Wolgast M,

Fiebich M, Arolt V, Heindel W. Effective electroconvulsive therapy reverses glutamate/glutamine deficit in the left anterior cingulum of unipolar depressed patients.

Psychiatry Res 122: 185–92, 2003.

Phillipson OT, Griffiths AC. The topographic order of inputs to nucleus accumbens in the rat. Neuroscience 16: 275–296, 1985.

Pitman DL, Ottenweller JE, Natelson BH. Plasma corticosterone levels during repeated presentation of two intensities of restraint stress: chronic stress and habituation.

Physiol Behav 43: 47–55, 1988.

Pittenger C, Duman RS. Stress, depression, and neuroplasticity: a convergence of mechanisms. Neuropsychopharmacology 33: 88–109, 2008.

171

Popoli M, Yan Z, McEwen BS, Sanacora G. The stressed synapse: the impact of stress and glucocorticoids on glutamate transmission. Nat Rev Neurosci 13: 22–37,

2012.

Pothion S, Bizot J-C, Trovero F, Belzung C. Strain differences in sucrose preference and in the consequences of unpredictable chronic mild stress. Behav Brain Res

155: 135–46, 2004.

Prather MD, Lavenex P, Mauldin-Jourdain ML, Mason WA, Capitanio JP,

Mendoza SP, Amaral DG. Increased social fear and decreased fear of objects in monkeys with neonatal amygdala lesions. Neuroscience 106: 653–658, 2001.

Proulx CD, Hikosaka O, Malinow R. Reward processing by the lateral habenula in normal and depressive behaviors. Nat Neurosci 17: 1146–52, 2014.

Puig MV, Gulledge AT. Serotonin and Prefrontal Cortex Function: Neurons,

Networks, and Circuits. Mol. Neurobiol. : 1–16, 2011.

Quarton G, Clark L, Cobb S, Bauer W. Mental disturbances associated with

ACTH and cortisone: a review of explanatory hypotheses. Medicine (Baltimore) 34: 13–

50, 1955.

Quetsch RM, Achor RW, Litin EM, Faucett RL. Depressive reactions in hypertensive patients; a comparison of those treated with Rauwolfia and those receiving no specific antihypertensive treatment. Circulation 19: 366–75, 1959.

Radley JJ, Rocher AB, Rodriguez A, Ehlenberger DB, Dammann M,

McEwen BS, Morrison JH, Wearne SL, Hof PR. Repeated stress alters dendritic spine morphology in the rat medial prefrontal cortex. J Comp Neurol 507: 1141–1150, 2008.

172

Rapport MM, Green AA, Page IH. Serum vasoconstrictor, serotonin; isolation and characterization. J Biol Chem 176: 1243–1251, 1948.

Remondes M, Schuman EM. Direct cortical input modulates plasticity and spiking in CA1 pyramidal neurons. Nature 416: 736–40, 2002.

Remondes M, Schuman EM. Molecular mechanisms contributing to long-lasting synaptic plasticity at the temporoammonic–CA1 synapse. Learn Mem 10: 247–52, 2003.

Remondes M, Schuman EM. Role for a cortical input to hippocampal area CA1 in the consolidation of a long-term memory. Nature 431: 699–703, 2004.

Reul JM, de Kloet ER. Two receptor systems for corticosterone in rat brain: microdistribution and differential occupation. Endocrinology 117: 2505–2511, 1985.

Ridder S, Chourbaji S, Hellweg R, Urani A, Zacher C, Schmid W, Zink M,

Hörtnagl H, Flor H, Henn FA, Schütz G, Gass P. Mice with genetically altered glucocorticoid receptor expression show altered sensitivity for stress-induced depressive reactions. J Neurosci 25: 6243–6250, 2005.

Robison AJ. Emerging role of CaMKII in neuropsychiatric disease. Trends

Neurosci. 2014.

Rotheneichner P, Lange S, O’Sullivan A, Marschallinger J, Zaunmair P,

Geretsegger C, Aigner L, Couillard-Despres S. Hippocampal neurogenesis and antidepressive therapy: Shocking relations. Neural Plast 2014, 2014.

Russo SJ, Nestler EJ. The brain reward circuitry in mood disorders. Nat Rev

Neurosci 14: 609–25, 2013.

173

Rygula R, Abumaria N, Flügge G, Hiemke C, Fuchs E, Rüther E,

Havemann-Reinecke U. Citalopram counteracts depressive-like symptoms evoked by chronic social stress in rats. Behav Pharmacol 17: 19–29, 2006.

Sanacora G, Gueorguieva R, Epperson CN, Wu Y-T, Appel M, Rothman DL,

Krystal JH, Mason GF. Subtype-specific alterations of gamma-aminobutyric acid and glutamate in patients with major depression. Arch Gen Psychiatry 61: 705–13, 2004.

Sanacora G, Zarate CA, Krystal JH, Manji HK. Targeting the glutamatergic system to develop novel, improved therapeutics for mood disorders. Nat Rev Drug

Discov 7: 426–437, 2008.

Santarelli L, Saxe M, Gross C, Surget A, Battaglia F, Dulawa SC, Weisstaub

N, Lee J, Duman RS, Arancio O, Belzung C, Hen R. Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science (80- ) 301: 805–9,

2003.

Sapolsky R, Krey LC, McEwen BS. Glucocorticoid-sensitive hippocampal neurons are involved in terminating the adrenocortical stress response. Proc Natl Acad

Sci U S A 81: 6174–6177, 1984.

Sapolsky R, Krey LC, McEwen BS. The neuroendocrinology of stress and aging: the glucocorticoid cascade hypothesis. Endocr. Rev. 7: 284–301, 1986.

Sapolsky R. Why zebras don’t get ulcers: An updated guide to shess-related diseases and coping. New York: W. H. Freeman, 1998.

Sari Y, Lefèvre K, Bancila M, Quignon M, Miquel MC, Langlois X, Hamon

M, Vergé D. Light and electron microscopic immunocytochemical visualization of 5-

HT1B receptors in the rat brain. Brain Res 760: 281–286, 1997.

174

Sari Y, Miquel MC, Brisorgueil MJ, Ruiz G, Doucet E, Hamon M, Vergé D.

Cellular and subcellular localization of 5-hydroxytryptamine(1B) receptors in the rat central nervous system: Immunocytochemical, autoradiographic and lesion studies.

Neuroscience 88: 899–915, 1999.

Schild AHE, Pietschnig J, Tran US, Voracek M. Genetic association studies between SNPs and suicidal behavior: A meta-analytical field synopsis. Prog Neuro-

Psychopharmacology Biol Psychiatry 46: 36–42, 2013.

Schildkraut JJ. The catecholamine hypothesis of affective disorders: a review of supporting evidence. Am J Psychiatry 122: 509–22, 1965.

Schlaepfer TE, Bewernick BH, Kayser S, Hurlemann R, Coenen VA. Deep brain stimulation of the human reward system for major depression--rationale, outcomes and outlook. Neuropsychopharmacology 39: 1303–14, 2014.

Schlicker E, Werner U, Hamon M, Gozlan H, Nickel B, Szelenyi I, Göthert

M. Anpirtoline, a novel, highly potent 5-HT1B receptor agonist with antinociceptive/antidepressant-like actions in rodents. Br J Pharmacol 105: 732–738,

1992.

Schmidt M V, Trümbach D, Weber P, Wagner K, Scharf SH, Liebl C,

Datson N, Namendorf C, Gerlach T, Kühne C, Uhr M, Deussing JM, Wurst W,

Binder EB, Holsboer F, Müller MB. Individual stress vulnerability is predicted by short-term memory and AMPA receptor subunit ratio in the hippocampus. J Neurosci 30:

16949–58, 2010.

Schumacher J, Jamra RA, Becker T, Ohlraun S, Klopp N, Binder EB,

Schulze TG, Deschner M, Schmäl C, Höfels S, Zobel A, Illig T, Propping P,

175

Holsboer F, Rietschel M, Nöthen MM, Cichon S. Evidence for a relationship between genetic variants at the brain-derived neurotrophic factor (BDNF) locus and major depression. Biol Psychiatry 58: 307–314, 2005.

Seeman TE, Singer BH, Rowe JW, Horwitz RI, McEwen BS. Price of adaptation--allostatic load and its health consequences. MacArthur studies of successful aging. Arch Intern Med 157: 2259–2268, 1997.

Sego C, Gonçalves L, Lima L, Furigo IC, Donato J, Metzger M. Lateral habenula and the rostromedial tegmental nucleus innervate neurochemically distinct subdivisions of the dorsal raphe nucleus in the rat. J Comp Neurol 522: 1454–1484, 2014.

Selye H, Tuchweber B. Stress in relation to aging and disease. In: Hypothalamus,

Pituitary and Aging, edited by Everitt A, Burgess J. Springfield, IL: Charles C Thomas,

1976, p. 557–573.

Selye H. Stress and the general adaptation syndrome. Br Med J 1: 1383–92, 1950.

Sen S, Nesse RM, Stoltenberg SF, Li S, Gleiberman L, Chakravarti A, Weder

AB, Burmeister M. A BDNF coding variant is associated with the NEO personality inventory domain neuroticism, a risk factor for depression. Neuropsychopharmacology

28: 397–401, 2003.

Sharp T, Bramwell SR, Grahame-Smith DG. 5-HT1 agonists reduce 5- hydroxytryptamine release in rat hippocampus in vivo as determined by brain microdialysis. Br J Pharmacol 96: 283–290, 1989.

Sharpley CF, Bitsika V. Validity, reliability and prevalence of four “clinical content” subtypes of depression. Behav Brain Res 259: 9–15, 2014.

176

Sheline YI, Gado MH, Kraemer HC. Untreated depression and hippocampal volume loss. Am J Psychiatry 160: 1516–1518, 2003.

Sheline YI. 3D MRI studies of neuroanatomic changes in unipolar major depression: the role of stress and medical comorbidity. Biol Psychiatry 48: 791–800,

2000.

Sheline YI. Neuroimaging studies of mood disorder effects on the brain. Biol

Psychiatry 54: 338–52, 2003.

Shorter E, Fink M. Endocrine Psychiatry: Solving the Riddle of Melancholia. 1st ed. Oxford University Press, 2010.

Sigalas PD, Garg H, Watson S, McAllister-Williams RH, Ferrier IN.

Metyrapone in treatment-resistant depression. Ther Adv Psychopharmacol 0: 1–11, 2012.

Slattery DA, Hudson AL, Nutt DJ. Invited review: The evolution of antidepressant mechanisms. Fundam Clin Pharmacol 18: 1–21, 2004.

Sousa N, Lukoyanov N V., Madeira MD, Almeida OFX, Paula-Barbosa MM.

Reorganization of the morphology of hippocampal neurites and synapses after stress- induced damage correlates with behavioral improvement. Neuroscience 97: 253–266,

2000.

Stamatakis AM, Stuber GD. Activation of lateral habenula inputs to the ventral midbrain promotes behavioral avoidance. Nat. Neurosci. 15: 1105–1107, 2012.

Starkman MN, Gebarski SS, Berent S, Schteingart DE. Hippocampal formation volume, memory dysfunction, and cortisol levels in patients with Cushing’s syndrome. Biol Psychiatry 32: 756–765, 1992a.

177

Starkman MN, Giordani B, Gebarski SS, Berent S, Schork MA, Schteingart

DE. Decrease in cortisol reverses human hippocampal atrophy following treatment of

Cushing’s disease. Biol Psychiatry 46: 1595–1602, 1999.

Starkman MN, Schteingart DE, Schork MA. Discordant changes in plasma

ACTH and beta-lipotropin/beta-endorphin levels in Cushing’s disease patients with depression. Psychoneuroendocrinology 17: 619–626, 1992b.

Stean TO, Atkins AR, Heidbreder CA, Quinn LP, Trail BK, Upton N.

Postsynaptic 5-HT1B receptors modulate electroshock-induced generalised seizures in rats. Br J Pharmacol 144: 628–635, 2005.

Sterling P, Eyer J. Allostasis: A new paradigm to explain arousal pathology. In:

Handbook of Life Stress, Cognition and Health, edited by Fisher S, Reason J. New York:

John Wiley & Sons, 1988, p. 629–649.

Steward O, Scoville SA. Retrograde labeling of central nervous pathways with tritiated or Evans blue-labeled bovine serum albumin. Neurosci Lett 3: 191–196, 1976.

Strekalova T, Steinbusch H. Factors of Reproducibility of Anhednoia Inductoin in a Chronic Stress Depression Model in Mice. In: Mood and Anxiety Related Phenotypes in Mice, edited by Gould TD. New York, NY: Humana Press, 2009, p. 153–176.

Sullivan PF, Neale MC, Kendler KS. Genetic epidemiology of major depression: Review and meta-analysis. Am J Psychiatry 157: 1552–1562, 2000.

Svenningsson P, Chergui K, Rachleff I, Flajolet M, Zhang X, El Yacoubi M,

Vaugeois J-M, Nomikos GG, Greengard P. Alterations in 5-HT1B receptor function by p11 in depression-like states. Science (80- ) 311: 77–80, 2006.

178

Svenningsson P, Tzavara ET, Witkin JM, Fienberg AA, Nomikos GG,

Greengard P. Involvement of striatal and extrastriatal DARPP-32 in biochemical and behavioral effects of fluoxetine (Prozac). Proc Natl Acad Sci U S A 99: 3182–3187, 2002.

Sybirska E, Davachi L, Goldman-Rakic PS. Prominence of direct entorhinal-

CA1 pathway activation in sensorimotor and cognitive tasks revealed by 2-DG functional mapping in nonhuman primate. J Neurosci 20: 5827–5834, 2000.

Taylor MA, Fink M. Restoring melancholia in the classification of mood disorders. J Affect Disord 105: 1–14, 2008.

Thompson P. Antidepressants and memory: A review. Hum Psychopharmacol

Clin Exp 6: 79–90, 1991.

Thompson SM, Kallarackal AJ, Kvarta MD, Van Dyke AM, LeGates TA,

Cai X. An excitatory synapse hypothesis of depression. Trends Neurosci 38: 279–94,

2015.

Tizabi Y, Bhatti BH, Manaye KF, Das JR, Akinfiresoye L. Antidepressant-like effects of low ketamine dose is associated with increased hippocampal AMPA/NMDA receptor density ratio in female Wistar-Kyoto rats. Neuroscience 213: 72–80, 2012.

Trivedi MH, Greer TL. Cognitive dysfunction in unipolar depression: implications for treatment. J Affect Disord 152-154: 19–27, 2014.

Tronche F, Kellendonk C, Kretz O, Gass P, Anlag K, Orban PC, Bock R,

Klein R, Schütz G. Disruption of the glucocorticoid receptor gene in the nervous system results in reduced anxiety. Nat Genet 23: 99–103, 1999.

Tye KM, Mirzabekov JJ, Warden MR, Ferenczi E a, Tsai H-C, Finkelstein J,

Kim S-Y, Adhikari A, Thompson KR, Andalman AS, Gunaydin L a, Witten IB,

179

Deisseroth K. Dopamine neurons modulate neural encoding and expression of depression-related behaviour. Nature 493: 537–41, 2013.

Ustün TB, Ayuso-Mateos JL, Chatterji S, Mathers C, Murray CJL. Global burden of depressive disorders in the year 2000. Br J Psychiatry 184: 386–392, 2004.

Verhagen M, van der Meij A, van Deurzen P, Janzing J, Arias-Vásquez A,

Buitelaar JK, Franke B. Meta-analysis of the BDNF Val66Met polymorphism in major depressive disorder: effects of gender and ethnicity. Mol Psychiatry 15: 260–71, 2010.

Vinogradova OS. Hippocampus as comparator: Role of the two input and two output systems of the hippocampus in selection and registration of information.

Hippocampus 11: 578–598, 2001.

Vyas A, Mitra R, Shankaranarayana Rao BS, Chattarji S. Chronic stress induces contrasting patterns of dendritic remodeling in hippocampal and amygdaloid neurons. J Neurosci 22: 6810–8, 2002.

Waldmeier PC, Williams M, Baumann PA, Bischoff S, Sills MA, Neale RF.

Interactions of isamoltane (CGP 361A), an anxiolytic phenoxypropanolamine derivative, with 5-HT1, receptor subtypes in the rat brain. Naunyn Schmiedebergs Arch Pharmacol

337: 609–620, 1988.

Wang Q, Van Heerikhuize J, Aronica E, Kawata M, Seress L, Joels M,

Swaab DF, Lucassen PJ. Glucocorticoid receptor protein expression in human hippocampus; stability with age. Neurobiol Aging 34: 1662–73, 2013.

Warner-Schmidt JL, Duman RS. Hippocampal neurogenesis: opposing effects of stress and antidepressant treatment. Hippocampus 16: 239–49, 2006.

180

Watanabe Y, Gould E, McEwen BS. Stress induces atrophy of apical dendrites of hippocampal CA3 pyramidal neurons. Brain Res 588: 341–5, 1992.

Weissman MM, Bland RC, Canino GJ, Faravelli C, Greenwald S, Hwu HG,

Joyce PR, Karam EG, Lee CK, Lellouch J, Lépine JP, Newman SC, Rubio-Stipec

M, Wells JE, Wickramaratne PJ, Wittchen H, Yeh EK. Cross-national epidemiology of major depression and bipolar disorder. JAMA 276: 293–299, 1996.

Wiegert O, Joëls M, Krugers HJ. Timing is essential for rapid effects of corticosterone on synaptic potentiation in the mouse hippocampus. Learn Mem 13: 110–

3, 2006.

Willner P, Muscat R, Papp M. Chronic mild stress-induced anhedonia: A realistic animal model of depression. Neurosci Biobehav Rev 16: 525–534, 1992.

Willner P, Towell A, Sampson D, Sophokleous S, Muscat R. Reduction of sucrose preference by chronic unpredictable mild stress, and its restoration by a tricyclic antidepressant. Psychopharmacology (Berl) 93: 358–364, 1987.

Willner P. Validity, reliability and utility of the chronic mild stress model of depression: A 10-year review and evaluation. Psychopharmacology (Berl). 134: 319–329,

1997.

Wolkowitz OM, Reus VI, Manfredi F, Ingbar J, Brizendine L, Weingartner

H. Ketoconazole administration in hypercortisolemic depression. Am J Psychiatry 150:

810–812, 1993.

Woolley CS, Gould E, McEwen BS. Exposure to excess glucocorticoids alters dendritic morphology of adult hippocampal pyramidal neurons. Brain Res 531: 225–231,

1990.

181

World Health Organization. Depression, Fact Sheet No. 369 [Online]. 2012. http://www.who.int/mediacentre/factsheets/fs369/en/ [1 Jan. 2015].

El Yacoubi M, Vaugeois J-M. Genetic rodent models of depression. Curr Opin

Pharmacol 7: 3–7, 2007.

Yalow RS, Glick SM, Roth J, Berson SA. Radioimmunoassay of human plasma

ACTH. J Clin Endocrinol Metab 24: 1219–1225, 1964.

Yuen EY, Wei J, Liu W, Zhong P, Li X, Yan Z. Repeated stress causes cognitive impairment by suppressing glutamate receptor expression and function in prefrontal cortex. Neuron 73: 962–77, 2012.

Yun J, Koike H, Ibi D, Toth E, Mizoguchi H, Nitta A, Yoneyama M, Ogita K,

Yoneda Y, Nabeshima T, Nagai T, Yamada K. Chronic restraint stress impairs neurogenesis and hippocampus-dependent fear memory in mice: Possible involvement of a brain-specific transcription factor Npas4. J Neurochem 114: 1840–1851, 2010.

Zai CC, Manchia M, De Luca V, Tiwari AK, Chowdhury NI, Zai GC, Tong

RP, Yilmaz Z, Shaikh SA, Strauss J, Kennedy JL. The brain-derived neurotrophic factor gene in suicidal behaviour: a meta-analysis. Int J Neuropsychopharmacol 15:

1037–42, 2012.

Zarate CA, Brutsche NE, Ibrahim L, Franco-Chaves J, Diazgranados N,

Cravchik A, Selter J, Marquardt CA, Liberty V, Luckenbaugh DA. Replication of ketamine’s antidepressant efficacy in bipolar depression: a randomized controlled add-on trial. Biol Psychiatry 71: 939–46, 2012.

Zarate CA, Duman RS, Liu G, Sartori S, Quiroz J, Murck H. New paradigms for treatment-resistant depression. Ann N Y Acad Sci 1292: 21–31, 2013.

182

Zarate CA, Singh JB, Carlson PJ, Brutsche NE, Ameli R, Luckenbaugh DA,

Charney DS, Manji HK. A randomized trial of an N-methyl-D-aspartate antagonist in treatment-resistant major depression. Arch Gen Psychiatry 63: 856–64, 2006.

183