海馬CA1樹状突起における入力統合機能の解析

(1)

平成 25 年度

学位論文（博士）

海馬 CA1 樹状突起における入力統合機能の解析

玉川大学大学院脳情報研究科

脳情報専攻

学籍番号

112715003

氏名

近藤将史

(2)

Ph.$D$Thesis$

Doctor$of$Philosophy$in$Engneering$

Specialized*in*Neuroscience*

Modulations of spike-timing dependent

plasticity by spatially distinct and

co-activating synaptic inputs in

hippocampal CA1 neurons

Masashi KONDO

Graduate School of Brain sciences, Tamagawa University

Mentor Prof. Takeshi Aihara Grad. Sch. Brain Sci., Tamagawa Univ.

Reviewers Prof. Yoshikazu Isomura Grad. Sch. Brain Sci., Tamagawa Univ.

Prof. Hiroshi Kojima Grad. Sch. Brain Sci., Tamagawa Univ.

(3)

Acknowledgement

I would like to especially thank:

Prof. Takeshi Aihara, my mentor, for all of your support, all of your encouragement

and for sharing your great scientific knowledge to me.

Prof. Yoshikazu Isomura, Prof. Hiroshi Kojima and Prof. Kazuyuki Samejima, my

reviewers of the thesis, for being my inspiration through your attentive and scientific comments.

To all colleagues in Aihara laboratory, I would like to sincerely thank for interesting discussions, and friendly chats and great kindness to me.

Thanks to the Grant-in-Aid for JSPS Fellowship for the financial support and funding.

Father and Mother for all of your love, kindness and support over the years.

Shimada, our great lovely chipmunk, for all of your marvelous healing helps me a lot.

Finally, to my wife Tomomi for supporting lives everyday and everything, for all of your love and encouragement with the thesis.

This work was supported by the Global COE Program for Tamagawa University and a Grant-in-Aid for JSPS Fellows (23-11250) from the Japan Society for the Promotion of Science and KAKENHI grants (19200014, 20500278) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

(4)

謝辞

指導教官である相原威教授には，本研究を進めるに当たっての研究計画立案，実験方法，学会発表準備や論文執筆など，熱心に御指導いただきました．思い返せば学部１年のときから足掛け 9 年間もお世話になりました．研究以外のことにも様々便宜を計っていただいたことも，一度や二度ではありません．この場を借りて，そのひとつひとつに深く御礼申し上げます．本論文の審査に際して，一線の研究者としての厳しく，また有意義なコメントを数多くくださった主査の磯村宜和教授，副査の小島比呂志教授，鮫島和行教授に感謝致します．私に神経科学を研究するきっかけを与えてくださるだけでなく，早くから研究の面白さとそこへの情熱を教えてくださった，松本深志高校 OB でもある塚田稔名誉教授にも，この場を借りて感謝致します．また共に研究室で研究を行った脳情報研究科の上條中庸さん，早川博章さん，研究員の佐村俊和さん，杉崎えり子さん，吉田美奈子さん，他たくさんの相原研究室メンバーには，研究に行き詰まったときにアドバイスをいただいたり，日々の他愛のない話を聞いてもらったり，研究面・精神面ともに助けられました．ありがとうございます．妻・友美には，学生である私を日々様々な面で支えてくれました．お陰様で，ここまで辿り着くことができました．ありがとう．最後に父・祥人と母・悟子には，この博士課程三年間だけでなく，それ以前にわたって金銭的に大きな負担をかけました．修士課程への進学にあたり，『大学院へ進学して，研究者として生きていきたい』という非常に贅沢で我儘なことを言い放った息子に対し，厳しいことを言いつつも，好きにさせてくれたこと，それらは深い愛情から来ていたものであったと感じています．いくら感謝をしても足りないでしょうけれど，『ありがとう』と言わせてください．ここに全員の氏名を記すことはできませんが，他にも数多くの方々にお世話になりました．皆さまのご助力なくして，この論文が完成することはなかったと思います．この場を借りて感謝致します．本当にありがとうございました．

(5)

Chapter 1. Preface

In our daily lives, we can retrieve various things, for example, the way from home to office, the faces and names of friends and so on. It is a gift by the function of learning and memory of brain.

Neuron is a specific input-output device―When certain input pattern is incoming to neurons, the neurons generate spikes as the output depend on the input pattern. Therefore, if information were expressed on a certain input pattern, neurons received the input act as the information processor.

How does a brain build various memories? In recent studies, investigators have thought that the memories have been embedded in strengths of synapses, which the sites of connection between one neuron and another one. In other words, the neurons build up “memory” on thousands synapses on their dendrites by synaptic plasticity.

This neuronal function is hypothesized by Hebb (1949) and at a later date, a lot of experimental studies have shown as Long-term potentiation (LTP) and Long-term Depression (LTD) (Bliss and Lømo, 1973; Ito, 1989). However, looking at the most of previous studies about LTP and LTD, LTP and LTD have been induced by artificial stimuli that are not observed in vivo. For example, High-frequency stimulation and Low-frequency stimulation have been used for induction of LTP and LTD, respectively.

One problem, how neurons change the synaptic strength with activities in vivo, was proposed by the above study. Whereas, Markram, his

(8)

colleagues and also Bi & Poo proposed experimentally a new learning rule that the synaptic strength is determined by the difference of pre- and post-synaptic activities (Bi and Poo, 1998; Markram et al., 1997). Today, it is called “spike-timing-dependent plasticity (STDP)”. This learning rule has indicated that the synaptic strength is changed by the plausible neural activities in vivo and been thought a predominant candidate as a neural basis of learning and memory.

Nevertheless, the conventional LTP, which are induced by tetanus (high-frequency) stimuli, has properties called as “cooperativity”, “input specificity” and “associativity” (Malenka, 2003; McNaughton, 2003). As a matter of fact, the weak inputs can also induce LTP when strong inputs are simultaneously activated. This phenomenon is “associativity” in LTP. By existing these properties in LTP, neurons could self-organize the “cell assembly” in their networks (Hebb, 1949). However, in STDP rule, the method of this property implemented has been almost unknown.

Accordingly, in this study, we focused on back-propagating action potentials (bAPs), in which is actively propagated from the soma to dendrites when neurons fire, and used optical imaging method with a voltage-sensitive dye. In other words, this approach could be discussed on synaptic plasticity based on dendritic properties and the power of information processing for STDP. As a result, this study might provide new

(9)

aspects of information processing on dendrites and the definition of “associativity” in spike-timing-dependent learning rule.

Dawkins defined the “meme” as the unit of information in brain (Dawkins, 1976). If the meme truly exist in brain, these are represented with neural systems based on billions neurons, glial cells and them networks. My ultimate goal on the scientific research is how the “meme” is stored and transferred in complex neuronal systems. This principle is emerged in the future, at the moment, we will govern the information in the brain and be able to voluntarily control (decode and encode) the information in the brain (The world would be almost as if the “Ghost in the

Shell”!). I strongly wish that this thesis gets ahead start of my scientific

(10)

! Abbreviations

ACSF, artificial cerebrospinal fluid;

AP5, D-2-amino-5-phosphonovaleric acid; bAP, back-propagating action potential; CA1, cornu ammonis field I

CA3, cornu ammonis field III

CMOS, complementary metal oxide semiconductor; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; DD, distal dendrite;

DG, dentate gyrus EC, entorhinal cortex

EPSP, excitatory postsynaptic potential; fEPSP, field potential of EPSP;

IPSP, inhibitory postsynaptic potential; LTD, long-term depression; LTP, long-term potentiation; NMDA, N-methyl-D-aspartate PD, proximal dendrite; PP, parforant path PSP, post-synaptic potential SC, schaffer collaterals

(11)

SEM, standard error of the mean; STDP, spike-timing-dependent plasticity; STDPS, STDP-induction stimuli; TS, test stimulus; TTX, tetrodotoxin; VSD, voltage-sensitive dye

(12)

! List of figures

Figure 2-1: Hippocampal formation and other cortical area ______________________________ 7 Figure 2-2: Hippocampal circuit ______________________________________________________ 8 Figure 2-3 Hebb’s Postulate. ________________________________________________________ 11 Figure 2-4: Three key properties of long-term potentiation ______________________________ 13 Figure 2-5: Relationship between the synaptic modification and pre-post spike timings ____ 16 Figure 2-6: Action potential backpropagation and local dendritic Ca2+ spike ______________ 17 Figure 2-7: Voltage-sensitive dye imaging from Layer 2/3 pyramidal cell __________________ 19 Figure 3-1: Schematic drawing and actual image of a hippocampal slice showing the

stimulation and optical recording sites _______________________________________________ 25 Figure 3-2: Relative timings of the paired stimulation ___________________________________ 27 Figure 3-3: Pharmacological separation of VSD signals _________________________________ 36 Figure 3-4: Latency to peak of excitatory and inhibitory components in the VSD signal _____ 39 Figure 3-5: Examples of bAP modulation by PD inputs _________________________________ 42 Figure 3-6: Summarized results of bAP modulation ____________________________________ 44 Figure 3-7: Simultaneous electrophysiological and VSD recordings of long-term modifications. _________________________________________________________________________________ 46 Figure 3-8: Influence of the timing of PD inputs on the distal STDP induced by E-phase timing _________________________________________________________________________________ 48 Figure 3-9: Influence of the timing of PD inputs on distal STDP induced by I-phase timing __ 49 Figure 3-10: Influence of inhibitory effects induced by PD inputs on distal STDP modulation 51 Figure 3-11: Influence of the timing of DD inputs on the proximal STDP induced by the E- and I-phase ___________________________________________________________________________ 54

(13)

! List of figures

Figure 2-1: Hippocampal formation and other cortical area ______________________________ 7 Figure 2-2: Hippocampal circuit ______________________________________________________ 8 Figure 2-3 Hebb’s Postulate. ________________________________________________________ 11 Figure 2-4: Three key properties of long-term potentiation ______________________________ 13 Figure 2-5: Relationship between the synaptic modification and pre-post spike timings ____ 16 Figure 2-6: Action potential backpropagation and local dendritic Ca2+_{spike ______________ 17} Figure 2-7: Voltage-sensitive dye imaging from Layer 2/3 pyramidal cell __________________ 19 Figure 3-1: Schematic drawing and actual image of a hippocampal slice showing the

stimulation and optical recording sites _______________________________________________ 25 Figure 3-2: Relative timings of the paired stimulation ___________________________________ 27 Figure 3-3: Pharmacological separation of VSD signals _________________________________ 36 Figure 3-4: Latency to peak of excitatory and inhibitory components in the VSD signal _____ 39 Figure 3-5: Examples of bAP modulation by PD inputs _________________________________ 42 Figure 3-6: Summarized results of bAP modulation ____________________________________ 44 Figure 3-7: Simultaneous electrophysiological and VSD recordings of long-term modifications. _________________________________________________________________________________ 46 Figure 3-8: Influence of the timing of PD inputs on the distal STDP induced by E-phase timing _________________________________________________________________________________ 48 Figure 3-9: Influence of the timing of PD inputs on distal STDP induced by I-phase timing __ 49 Figure 3-10: Influence of inhibitory effects induced by PD inputs on distal STDP modulation 51 Figure 3-11: Influence of the timing of DD inputs on the proximal STDP induced by the E- and I-phase ___________________________________________________________________________ 54

(14)

Chapter 2. Background

2.1. Hippocampus

A brain can be separated into some different functional and morphological areas. Investigators have being believed that, particularly, hippocampus might act as important roles in learning and memory, especially the episodic, semantic, and spatial memory.

Figure 2-1: Hippocampal formation and other cortical area

Adapted from Figure 3 of Cheung & Cardinal, 2005. CA1, CA2, CA3: cornu ammonis fields 1–3; DG: dentate gyrus; EC: entorhinal cortex; f: fornix; s: septal pole of the hippocampus; S: subiculum; t: temporal pole of the hippocampus.

(15)

Hippocampus is located in medial temporal lobe of bilateral cerebral cortex (Figure 2-1). Many cortical areas, visual, auditory, somatosensory and association area of cerebral cortex, amygdala and so on, are connected to hippocampus. If a part of hippocampus is affected, memories cannot be formed after the affection.

Figure 2-2: Hippocampal circuit

Neural circuit of hippocampus is mainly involved dentate gyrus, CA3 and CA1 area. The circuit consisted these area have been called as “tri-synaptic circuit” because the number of their synaptic connections are three times. EC: entorhinal cortex, DG: dentate gyrus, CA3 and 1: cornu ammonis field 3 and 1, Sub: subicurum. (Image: Santiago Ramón y Cajal (1911), Wikipedia)

Rat’s hippocampal circuit is illustrated at Figure 2-2. The starting point of hippocampal neural circuits is entorhinal cortex (EC), which are received inputs from other cortical areas. Then, dentate gyrus (DG) is received neural inputs from EC. It is a worthy of special mention in DG that neurons are born at adult age and the adult neurogenesis are deeply

(16)

related to the memory system (Farioli-Vecchioli et al., 2008; Piatti et al., 2011). A next stage in the hippocampal circuits is CA3 in rats. CA3 has a recurrent connection (CA3-CA3 connections) so various studies have mentioned that the relationship of the associative memory, the ability of memory to associate people with their belongings (Shing et al., 2011). CA3 neurons output the spikes to CA1 area, the final part of hippocampal circuits, via the schaffer collaterals (SC) pathway. For decades, CA3-CA1 synapses have been observed to understand the synaptic plasticity (Engert and Bonhoeffer, 1999; Harris et al., 1984; Tsien et al., 1996). The outputs of CA1 neurons project to EC. This information flow of DG-CA3-CA1 is termed “tri-synaptic circuits of hippocampus” and the famous synaptic connections of hippocampus (Yeckel and Berger, 1990).

The layers of CA1 area generally consist of alveus, stratum pyramidale,

stratum radiatum, and stratum lacnosum-moleculare. The alveus layer

contains the axon to EC and the basal dendrites of CA1 pyramidal cells. The stratum pyramidale layer contains the cell body of pyramidal cells, which are the principle excitatory neurons of CA1 area. CA1 pyramidal cells put their apical dendrites to stratum radiatum and stratum lacnosum-moleculare. The SC pathway and CA1 apical dendrites form the synapses in stratum radiatum. In the stratum lacnosum-moleculare layer, the dendrites are so thin and form the synaptic connection from EC layer 3

(17)

via perforant path (PP) fibers (Witter, 1993; Witter et al., 2000). Again, because the architecture of layers formation in CA1 area is very clearer than other hippocampal and cortical areas, many studies, which are about the synaptic transmission, plasticity, cellular morphology and so on, have been performed (Brankack et al., 1993; Dvorak-Carbone and Schuman, 1999; Inoue et al., 2001; Tsukada et al., 2005).

2.2. Synaptic plasticity

Hebb hypothesized how the connection from a pre-synaptic neuron to a post-synaptic neuron should be modified: “When an axon of cell A is near

enough to excite cell B or repeatedly or persistently takes part in firing it, some growth process or metabolic change takes place in one or both cells such that A's efficiency, as one of the cells firing B, is increased”(Hebb,

1949). Nowadays, his postulate is commonly described as the synaptic plasticity in the cellular neuroscience (Figure 2-3).

(18)

Figure 2-3 Hebb’s Postulate.

A: Initial state. The neuron A is fired, postsynaptic potential will be

revealed in the postsynaptic cell B. B: Correlated firing state. The neurons A and B simultaneously are firing. C: Potentiated state. After the correlated firing state, the strength of synapse between the cell A and B will be potentiated.

A

B

A

B

Potentiation

A

B

Post synaptic potential

A

B

(19)

2.3. Long-term potentiation and depression

By Hebb’s postulate, in other words, the synaptic strength between a pre- and a post-synaptic neuron is modified depending on the simple rule: if the neurons fire together, the neurons are wired together, otherwise the connection should be disappeared. More than 20 years later from Hebb’s postulate, Bliss and Lomo reported that high-frequency electrical stimulations (tetanus stimuli) to pre-synaptic axons raised long-lasting potentiation of synaptic efficacies in vivo rabbits’ brains (Bliss and Lømo, 1973). This phenomenon is known as long-term potentiation (LTP). In contrast, long-lasting depression of synaptic efficacy was found by Lynch and his colleagues(Lynch et al., 1977). It is called as long-term depression (LTD) and experimentally induced by repetitive low-frequency stimulations to pre-synaptic pathways in various cortical areas. Almost neuroscientists should believe that these long-lasting changes are neural bases of learning and memory.

About LTP, three famous properties have been described: as input specificity, cooperativity and associativity (Barrionuevo & Brown, 1983; Lee, 1983; McNaughton, 2003; Figure 2-4). Input specificity is described when LTP is elicited at one set of synapses and other set of synapses that are not inactivated during the induction of LTP do not show LTP. Second, cooperativity is described that certain and definite numbers (strong) of

(20)

Figure 2-4: Three key properties of long-term potentiation

A: Cooperativity. When using high frequency stimulation to induce LTP, a crucial A: Cooperativity. When using high frequency stimulation to induce LTP, a crucial number of presynaptic fibers must be simultaneously activated — they must 'cooperate' to elicit LTP. B: Input specificity. When LTP is elicited at one set of synapses on a postsynaptic cell, adjacent synapses that were not activated during the induction protocol do not show LTP. C: Associativity. LTP can be elicited at synapses that are activated by low-frequency, sub-threshold stimuli if their activation is temporally concurrent with an LTP-inducing stimulus at another set of synapses on the same cell. This figure adapted from Malenka, 2003.

A

B

(21)

synaptic inputs are required for the induction of LTP. Since LTP is not elicited by the small number (weak) of synaptic inputs. Third, associativity is described that LTP can be elicited at synapses that are activated by weak inputs if their activation is temporally concurrent with an LTP-inducing stimulus (strong inputs) at another set of synapses on the same cell.

To flow Ca2+_{into via N-methyl-D-aspartate (NMDA) receptors, or not}

to flow: that is the question... This is the most dominant hypothesis of inducing LTP. A D-(-)-2-amino-5-phosphonopentanoic acid (D-AP5), an antagonist of NMDA receptors, block the induction of LTP (Harris et al., 1984). The NMDA receptor is an inotropic glutamate receptor and permeates the cations non-selectively. However, the ionophore normally blocked by extracellular Mg2+_{and the blockade of Mg}2+_{is released}

depending on the membrane potentials (Garaschuk and Schneggenburger, 1996; Kampa et al., 2004). In brief, only when post-synaptic depolarization and the glutamate release from pre-synaptic terminals simultaneously occurred, NMDA receptors permeate the Ca2+_{and other cations into cells.}

After that, the increase of Ca2+_{in spines triggers some intracellular}

cascades (the activation of kinases etc.), and thus LTP is induced at the synapses.

(22)

2.4. Spike-timing-dependent plasticity

Spike-timing-dependent plasticity (STDP) is the learning theory including LTP and LTD. Although LTP and LTD are experimentally induced by the high-frequency stimulation and the low-frequency stimulations, respectively, the synaptic modification on STDP follows relative timings of pre- and post-synaptic cell’s firing(Markram et al., 1997; Bi & Poo, 1998; Figure 2-5).

Pre

Pre Pre-before-Post

Post

Post t t Pre Post-before-Pre

Depresion

Potentiation

Post t t

(23)

Figure 2-5: Relationship between the synaptic modification and pre-post spike timings

Across the synapse, when postsynaptic cell fires precede presynaptic cell (spike timings are > 0 ms), the synaptic strength is potentiated. Conversely, when presynaptic cell fires precede postsynaptic cell (spike timings are < 0 ms), the synaptic strength is depressed. This figure is adapted from Bi & Poo, 1998.

2.5. Dendrites

Neurons have some cellular architecture. The axon and the dendrite are the parts of output and input, respectively. Thousands of synaptic inputs are integrated in dendrites so that spikes via axons transmit the information to other neurons.

2.5.1. Passive properties of dendrites

Neurons, of course also other types of cell, have lipid bi-layer cell membranes. They separate between the outer and inner chemical environments so that the membranes act as capacitors because the electrochemical potential yield across the membranes. In addition, in their cellular space, various cellular organs exist. These physiological structures are implemented three passive electric properties in dendritic trees: the specific membrane resistivity, the specific membrane capacitance, and the intracellular resistivity. Based on these passive electrical properties, the shape and propagation of post-synaptic potentials (PSPs) are influenced drastically and these are not uniform over the dendritic trees(Golding et al.,

(24)

2005; Stuart and Spruston, 1998; Williams and Stuart, 2002).

2.5.2. Active properties of dendrites

Dendrites are “active”. This term means that the cellular membrane has voltage-dependent conductance, in other words, variable resistance in the membrane. Generally, axonal cables have active conductance, voltage-dependent Na+_{, K}+_{channels, so that spikes are generated and}

propagated to the other neurons. For about 10 years, many studies have reported that spikes are initiated by the active properties in dendrites (Stuart & Sakmann, 1994; Schiller et al., 1997; Stuart et al., 1997; Golding

et al., 1999; Figure 2-6).

Figure 2-6: Action potential backpropagation and local dendritic Ca2+

spike

In each traces of membrane potentials, the colors (green, cyan, orange, red) correspond with somatic and dendritic locations. Adapted from Häusser et

al., 2000.

axon soma

(25)

Consequently, when the axon and/or cell body fire, the spikes actively propagate to dendrites. These spikes are known as back-propagating action potentials (bAPs). These are signals to inform the synapses post-synaptic firing and an important factor on Hebbian synaptic plasticity and STDP (Kampa et al., 2007; Magee and Johnston, 1997).

2.6. Optical imaging

History of neuroscience might be compered to history of technical advances of optics. Neural structures are microscopic in a single neuron and/or the networks including huge neurons are so complicated in order to measure with just a single electrode. Thus, in recent years, a lot of researchers are using the optical imaging methods to probe the neural activities in thin dendrites and net responses of the neuronal network.

2.6.1. Voltage-sensitive dye

Voltage-sensitive dyes (VSD) are fluorescent molecules that change their fluorescence in response to changes in the membrane potentials (Figure 2-7). The fluorescent signals are fast following the membrane potential and the magnitudes linearly reflect to the amplitudes of membrane potentials(Grinvald et al., 1983; Gupta et al., 1981; Jin et al., 2002). By using the VSD, we could get signals everywhere the change of membrane potential exists: in the cell body, along the axon and the dendrite.

(26)

Figure 2-7: Voltage-sensitive dye imaging from Layer 2/3 pyramidal cell

Voltage-sensitive dye imaging allows us to probe the change of membrane potentials (postsynaptic potentials and spikes) in axon, soma and dendrites. Image is adapted from Zecevic et al., 2009.

(27)

Chapter 3. Modulation of spike-timing-dependent

plasticity mediating back-propagating action

potentials ~ cooperativity on STDP ~

3.1. Introduction

Neurons “communicate” each other with their spikes. When neurons fire, the spikes are transferred to other neurons via axon. At the same time, the spikes are propagated to soma to their own dendrites. These spikes are called back-propagating actin potential (bAP). Using bAPs, neurons communicate their thousands synapses on the dendrites. Detecting somatic spiking in synapses might important for neurons, which obtain appropriate input-output relations with the Hebbian synaptic plasticity.

Some properties of bAPs in CA1 pyramidal neurons in the hippocampus have revealed in several reports (Bernard and Johnston, 2003; Hoffman et al., 1997; Spruston et al., 1995; Williams, 2003). bAPs are gradually decreased along dendrites but excitatory postsynaptic potentials (EPSPs) induced by dendritic inputs enhanced the bAPs (Magee and Johnston, 1997; Sjöström and Häusser, 2006; Stuart and Häusser, 2001); however, in contrast, they are depressed by inhibitory postsynaptic potentials (IPSPs) induced by inhibitory neurons (shunting effect) (Tsubokawa and Ross, 1996; Williams, 2003). Thus, bAPs are modulated during their propagation from proximal dendrite (PD) to distal dendrites (DD).

(28)

postsynaptic inputs, and this is known as spike-timing-dependent synaptic plasticity (STDP) (Bi and Poo, 2001, 1998; Magee and Johnston, 1997; Markram et al., 1997; Nishiyama et al., 2010, 2000; Stuart and Häusser, 2001). Recent studies on the dependency of STDP on dendritic location, which were performed in rat cortical layers 2/3 and 5 (Froemke et al., 2005; Letzkus et al., 2006; Sjöström and Häusser, 2006), have identified distinct time windows for STDP at different dendritic locations. Previously, we classified STDP profiles into two types depending on their layer-specific location along the dendrite (proximal and distal) in the CA1 area of hippocampal slices (Aihara et al., 2007; Tsukada et al., 2005). Those results have suggested that this location dependency is due to the feedforward and feedback inhibitory connections of GABAergic neurons in the CA1 neurons. The role of inhibitory inputs in STDP induction have been shown to depend on input frequency (Nishiyama et al., 2010).

bAP modulation by synaptic inputs during propagation from PD to DD is still unclear, and how this influence information processing, which is referred to as coding modulation, remains elusive. Reports exist on the influence of the interactions of dendritic inputs on synaptic plasticity (Wang et al., 2003; Xu et al., 2006); however, the influence of such interactions on the information integration that occurs at dendrites remains unclear. In other words, this is a problem that: Does the plasticity of

(29)

associativity from distinct inputs pathways on STDP rule exist?

To investigate how excitatory postsynaptic inputs at PD influence information processing of the synaptic inputs at DD by means of bAPs, stimulation was simultaneously applied to PD during long-term potentiation (LTP) or long-term depression (LTD) induction at DD. To induce LTP or LTD, the STDP induction protocol was used. The influence of the PD input on LTP or LTD at DD was examined by optical imaging by applying three relative timings of PD inputs for bAPs. One involved the coincidence of bAP and EPSP in which bAP was facilitated. The second involved the coincidence of bAP and IPSP in which bAP was suppressed. The third involved bAP preceding EPSP in which bAP remained unchanged. The results suggested that bAP functions as an information carrier or a causality detector of temporal information of PD inputs to DD. The opposite direction of influence from DD to PD was measured, and the influence of the interactions of dendritic inputs on information integration at the dendrites was demonstrated. In conclusion, the emergence of associativity was depended on the dendritic location: about the impact to the synaptic plasticity at another site, the PD inputs’ were superior to the DD inputs’.

(30)

3.2. Materials & Methods

3.2.1. Slice preparation

All experiments were performed in accordance with the European Communities Council Directive of 1986 (86/609/EEC) and the Tamagawa University guidelines for the care and use of laboratory animals and were approved by the Animal Experiment Ethics Committee at Tamagawa University. All efforts were made to minimize the number of animals used and their suffering.

Three- to four-week old male Wister rats were deeply anesthetized by isofulrane and decapitated. The brain was quickly isolated and chilled in ice-cold artificial cerebro-spinal fluid for cutting (cACSF; 124 mM NaCl, 5.0 mM KCl, 2.6 mM NaH₂PO₄, 4.0 mM MgSO₄, 0.5 mM CaCl₂, 26 mM NaHCO₃, and 10 mM glucose, oxygenated with 95% O2 and 5% CO2 gas mixture). The tissue was sliced with micro slicer at 400 µm thickness. After slicing, the slices were shortly (maximum 10 min) maintained at 35℃ in a submerged chamber containing gassed cACSF. After the short recovery, the slices were transferred to holding chamber containing gassed normal medium of ACSF. The nACSF have same compositions for cACSF but the concentrations of MgSO₄ and CaCl₂ were modified to 2.0 mM and 1.0 mM, respectively.

(31)

3.2.2. Drugs

In some experiments, we used a selective GABA_A channel antagonist, gabazine (SR-95531, 10 µM, Tocris Bioscience, Bristol, UK); an NMDA channel antagonist, AP5 (50 µM, Wako Pure Chemical Industries, Ltd., Osaka, Japan); an AMPA channel antagonist, CNQX (10 µM, Wako Pure Chemical Industries, Ltd.); and a voltage-gated Na+_{channel blocker, TTX}

(1 µM Wako Pure Chemical Industries, Ltd.). All the drugs were added to normal ACSF.

3.2.3. Electrical stimulation

As shown in Figure 3-1, two bipolar tungsten electrodes were placed in fixed positions at dendritic regions that were proximal and distal to the soma in the stratum radiatum (DD-stim and PD-stim electrodes, respectively) to stimulate SC, which induced EPSPs at PD (<150 µm from the soma) and DD (>200 µm from the soma). The slices were cut by a thin razor blade to avoid convergence of the two synaptic inputs (Fig. 3-1, right). In the end of each experiment, we confirmed that the responses induced by the PD- and DD-stim electrodes did not converge on each dendritic site in the recording area with a Ca2+_{-free solution (Fig. 3-3 and Results section}

3.3.1). The other bipolar electrode (bAP-stim) used to induce bAPs was placed in a fixed position in the stratum oriens bordering the alveus.

(32)

Figure 3-1: Schematic drawing and actual image of a hippocampal slice showing the stimulation and optical recording sites

Two stimulation electrodes, the proximal dendrite (PD)-stimulation and distal dendrite (DD)-stimulation, were placed in the Schaffer-commissural collaterals of the stratum radiatum CA1 area to stimulate PD and DD, respectively. A third stimulation electrode, the back-propagating action potential (bAP)-stimulation, was placed at the striatum oriens bordering the striatum alveus to fire pyramidal cells and induce bAPs. In addition, slices were cut to avoid overlapping stimulations of PDs and DDs.

To determine the pairing stimuli for the relative timings of the presynaptic and postsynaptic activations, we used a programmable stimulator (PG4000A, Cygnus Technology Inc., Delaware Water Gap, PA, USA) and isolators (Iso-Flex, A.M.P.I., Jerusalem, Israel). The intensities of the electric pulses that were transmitted by the PD-stim and DD-stim electrodes and that were used to stimulate the presynaptic fibers for the induction of EPSPs at PD and DD were fixed at a constant value. These values were set at a subthreshold to produce neuronal spiking in the CA1 region and to avoid converging responses on each dendritic site (50–150 µA). However, the subthreshold inputs to PD and DD presumably triggered

Optical imaging area (1 mm ×1 mm) CA3 DG CA1 DD PD PD stim. DD stim. bAP stim. cut cut 200 μm

(33)

CA1 feedforward inhibitory interneurons because Pouille and Scanziani (2001) have reported that the spiking of CA1 feedforward interneurons could be triggered by the unitary EPSPs induced by the activations of CA3 pyramidal cells, while CA1 pyramidal cells were not fired. On the other hand, the intensity that was applied to the bAP-stim electrode for the induction of bAPs was fixed at 300 µA, which was a suprathreshold level for CA1 pyramidal cells.

As shown in Figures 3-2A and B, the following three relative timings of bAP inputs for PD inputs and DD inputs (τP, τD = 5 ms, 20 ms, or −5 ms)

were used: E-phase, in which the peak bAP time followed 5 ms after the rising time of EPSP that was induced by the synaptic input (PD or DD input; τP, τD = 5 ms) to make bAP coincide with EPSP at PD or DD;

I-phase, in which the peak time of bAP followed 20 ms after the rising time

of EPSP (τP, τD = 20 ms) to make bAP coincide with the disynaptic

feedforward IPSP that was induced by stimulating SC; and Ne-phase, in which the peak time of bAP had a negative timing of −5 ms (τP, τD = −5

ms). The reason why these timings were chosen in this study was that we wanted to focus on STDP profiles of the positive timings in the hippocampal CA1 dendrites, as has been detailed in previous studies (Aihara et al., 2005; Nishiyama et al., 2000; Tsukada et al., 2005).

(34)

Figure 3-2: Relative timings of the paired stimulation

A: Three timings, 5 ms, 20 ms, and −5 ms, were used for the relative timings between

bAP and the excitatory postsynaptic potential (EPSP) with the PD-stimulation or DD-stimulation electrodes (τP or τD). The relative timing of 5 ms was defined as the

E-phase timing because it marked the coincidence of bAP and EPSP; the relative timing

of 20 ms was defined as the I-phase because the peak amplitude of bAP was approximately set to the peak amplitude of the feedforward inhibitory postsynaptic potential (IPSP); and the relative timing of −5 ms was defined as the Ne-phase because bAP was not set to EPSP or IPSP. B: Illustrations of PD inputs for bAPs by each relative timing phase. DD inputs were also defined as E-, I-, or Ne-phase DD inputs for bAP (not shown).

-5 ms 5 ms 20 ms PD stim. or DD stim. bAP stim. t t t t E- phase (τP, τD= 5 ms) I-phase (τP, τD= 20 ms) Ne-phase (τP, τD= -5 ms) E-phase (τP˕5 ms) 5 ms t (τP˕20 ms) I-phase 20 ms t (τP˕-5 ms) Ne-phase t -5 ms bAP stim. PD stim. PD DD

B

A

(35)

The time parameters in the E- and I-phase were determined on the basis of the time courses that had been measured for EPSPs and IPSPs by electrophysiological methods(Karnup and Stelzer, 1999; Nishiyama et al., 2010; Sayer et al., 1990). In Result section, the time-courses of VSD signal at each input location by these inputs were shown. The pairing stimuli with one of the three relative timings were used at DD and PD. These relative timings were used for STDP induction, and the pairing stimuli were applied at 1 Hz for 100 s.

3.2.4. Optical imaging with VSD

Membrane activities, bAPs, and EPSPs were measured as optical signals by optical imaging with VSD. The staining method was identical to that described by Tominaga et al. (2000) and Tominaga and Tominaga (2010). Each slice was stained with 50 µL of ACSF and 50 µL of fetal bovine serum containing 200 µM of Di-4-ANEPPS (Life Technologies Corporation, Burlington, ON, CAN) in a plexiglass ring with a fine-mesh membrane filter (Millicell®-CM PICM01250, EMD Millipore Corporation, Billerica, MA, USA) in an interface chamber. After staining, the slices were returned to normal ACSF for at least an additional 1 h. A naive slice without GABAergic transmission blockade was used for each stimulus sequence. A recent study has shown an enhancement of GABA_A

(36)

receptor-mediated currents after the acute perfusion of Di-4-ANEPPS to cultured neurons (Mennerick et al., 2010). However, we did not observe any qualitative differences in this staining procedure, and it was similar to those of previous reports (Tominaga and Tominaga, 2010; Tominaga et al., 2000). In addition, the authors of that study have proposed that this staining procedure can minimize the phototoxicity and chemical toxicity of VSD itself.

Slices were viewed with a 20× objective (XLUMPlanFI, Olympus Corporation, Tokyo, Japan) and a 0.5× camera adapter (U-TV0.5CX-3, Olympus Corporation); the total magnification of the system was 10×. The VSD signals were recorded with a 565-nm dichroic beam splitter and a 610-nm long-pass filter. The MiCAM ULTIMA high-speed complementary metal oxide semiconductor (CMOS)-based imaging system (Brainvision Inc., Tokyo, Japan) was employed in the experiments. This system had a 100 × 100 CMOS sensor array, each with a receptive area of 10 × 10 µm2_{/photopixel, and a time resolution of 0.5 ms/frame was used to}

analyze the spatiotemporal activities of the CA1 network. The optical signals at the dendritic areas contained three components: EPSPs, IPSPs, and bAPs. Therefore, we confirmed these components from the optical signals of each other using the same methods as those used previously (Momose-Sato et al., 1999).

(37)

When the optical recording system was triggered by signals from the external channel of the programmable stimulator, which was the same as that generating the electric pulse stimuli, an electrical controlled shutter (built-in lamp house, MHF-G150LR, SCHOTT MORITEX Corporation, Saitama, Japan) was opened for 100 ms prior to starting the recording to avoid mechanical noise from the shutter and rapid dye bleaching. Double halogen lamps (LM-150, SCHOTT MORITEX Corporation) were used as light sources for the optical imaging. In addition, to reduce optical noise, averaging was performed eight times for each trial, and the obtained values were used as the experimental measurements.

We analyzed the optical signals offline with BV-Ana software (Ver. 0802, Brainvision Inc.). The optical signals discussed in the following sections were filtered by a 5 × 5 spatial arithmetic mean filter, and the drift components were corrected with the drift–remove function in the software. The 5 × 5 spatial filter substitutes a central value of 5 × 5 pixels with an average value of 5 × 5 pixels. The drift–remove function consisted of the following three steps: first, four points that were considered to have the drift component and that contained none of the evoked responses were chosen manually; second, a fitted curve was calculated by the average waveform of these four data points, which was then considered the drift component; finally, all data points were subtracted from the drift

(38)

component, which removed the drift without attenuating the real signal.

3.2.5. Electrophysiological recording

An extracellular field potential recording was made with a 1–4 Mohm borosilicate glass capillary that was filled with 2 M NaCl using a patch-clamp amplifier (Axopatch 200B, Molecular Devices, LLC, Sunnyvale, CA, USA) on current-clamp mode, which was filtered at 10 kHz and digitized at 250 kHz. Each capillary was pulled with a micropipette puller (P-97, Sutter Instrument Co., Novato, CA, USA). The recording electrode was placed at the proximal (<50 mm from the pyramidal cell layer) stratum radiatum region to monitor the field EPSPs (fEPSPs).

In the long-term modification study, baseline responses were recorded before applying STDP induction pairing stimulations following the delivery of the test stimulus (TS) every 20 s through the PD-stim electrode to establish electrophysiological stability for more than 15 min. After the pairing stimulations, the delivery of TS was resumed once every 20 s for at least 45 min.

3.2.6. Data analysis

(39)

fluorescence to the baseline VSD fluorescence (ΔF/F). For signal analysis, we selected two typical locations on a dendrite for analysis in the same manner as that described in our previous reports (Aihara et al., 2007, 2005). One was PD, which was defined as a point 50 ± 10 µm distant from the stratum pyramidale. The other was DD, which was defined as a point 200 ± 10 µm distant from the stratum pyramidale.

To calculate the inhibitory effect in VSD signals by the presynaptic stimulations (PD and DD), we linearly subtracted the VSD signals blocked GABA_A receptors by SR-95531 from the signal in naïve (intact inhibition) condition (Figure 3-4).

The magnitude of STDP at PD and DD was estimated as the change in the peak values of the optical signal in the same slice. At the beginning of each experiment, TS was applied once every 3 min for at least 15 min to ensure that the peak values of the optical signal had stabilized and that the peak value of the control level was measured. After the control level had been determined, the STDP-induction stimuli (STDPS) were applied, which were followed by TS. A naïve slice was used for each stimulus sequence of TS-STDPS-TS. As a control, we confirmed that the optical signal depending on the response was not changed for at least 15 min, which was the time period set for the experiment. Therefore, the effects of dye bleaching could be ignored. The signals after STDPS were measured

(40)

from 25 to 30 min after the application of STDPS to confirm stability. Following this, the ratio of the peak value for TS before and after STDPS was calculated as STDP. The responses were considered to be LTP and LTD if the optical responses to TS after STDP induction were significantly larger and smaller, respectively, in comparison with the response to TS before STDP induction (baseline).

Each electrophysiological trace was filtered (500-Hz low-pass) with Clampfit 10 software (Molecular Devices, LLC). The initial slopes of fEPSP before and after pairing stimulation were measured to calculate the magnitude of plasticity. The magnitude of plasticity was defined as fEPSP slopes at 30 min after STDP induction stimulus/averaged baseline fEPSP slopes.

All summarized data are presented as mean ± SEM. Statistical significance was evaluated with Excel software (Microsoft, Redmond, WA, USA). When we tested whether the magnitudes of synaptic plasticity were significantly changed in comparison with the baseline and whether the plastic changes of the electrophysiological and VSD responses were significantly different, Student’s t-test was used. The amplitude changes in bAPs and STDP between groups with the application of different parameters were compared using two-way ANOVA and Tukey–Kramer post-hoc test. If a calculated P value was less than 0.05, it was considered

(41)

(42)

3.3. Results

3.3.1. What kind of electrophysiological events do VSD

signals reflect?

Many aspects of neural activities, such as the action potentials of presynaptic fibers, EPSPs and IPSPs, the action potentials of postsynaptic neurons, and bAPs at dendrites, can be detected as optical signals with VSD imaging techniques. Thus, to identify what types of cellular responses resulted in our optical signals, we examined optical signals in the following four different solutions: normal artificial cerebrospinal fluid (ACSF;

control condition), ACSF with 10 µM

6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and 50 µM D-2-amino-5-phosphonovaleric acid (AP5; CNQX + AP5 condition), normal ACSF with all Ca2+ replaced with Mg2+ (Ca2+-free condition), and ACSF with 1 µM tetrodotoxin (TTX; TTX condition).

Figure 3-3 shows the typical optical signals that occurred in response to the application of electrical pulses to PD and DD and to the bAP input in several conditions. Figures 3-3A and B show that the stimuli with subthreshold intensity applied to PD and DD induced EPSPs because a large part of the optical signals in the control condition disappeared in the CNQX + AP5 condition, and this was because the AMPA- and NMDA-related components were blocked. In addition, because the VSD

(43)

Figure 3-3: Pharmacological separation of VSD signals

Traces indicate the voltage-sensitive dye (VSD) responses to the application of three discrete stimuli (A: PD input, B: DD input, C: bAP input). The traces in each row reflect the VSD signals at each location (Str. Pyr., striatum pyramidale; PD, proximal dendrite; DD, distal dendrite) in several solutions. Reagents were added to normal artificial cerebrospinal fluid (ACSF), and the concentrations are indicated below: 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; 10 µM), D-2-amino-5-phosphonovaleric acid (AP5; 50 µM), tetrodotoxin (TTX; 1 µM), and Ca2+-free (MgSO4, 3.6 µM and CaCl2, 0 µM). Calibration: 1.0 × 103% (fractional change in VSD fluorescence to the baseline VSD fluorescence: ΔF/F), 10 ms. Str. Pyr., striatum pyramidale; PD, proximal dendrite; DD, distal dendrite.

3' 6WU3\U 6WU3\U 6WU3\U '' A PD input 3' '' 3' '' B DD input C bAP input

&RQWURO _&14;$3 &D_IUHH _77;

10 ms 1.0 × 10 -3 10 ms 1.0 × 10 -3 10 ms 1.0 × 10 -3

(44)

signals reflected presynaptic fiber volleys directly evoked by electrical stimulation in the Ca2+-free condition, we could identify whether the PD or DD input independently stimulated each location with this procedure.

When we applied bAP stimulation, antidromically induced spike-like activity could be observed at the cell layer and along the dendrites (Figure 3-3C). The optical signals in the three conditions (control, CNQX + AP5, and Ca2+-free) were almost identical, and these responses disappeared in the TTX condition. Therefore, this spike-like response may correlate with the changes in the membrane potential triggered by voltage-gated Na+ channel activities.

Based on these results by the pharmacological experiments, we defined the optical signals induced by EPSP and bAP. After that, in the end of all experiments, we tested whether each optical signal induced by PD, DD and bAP stim. corresponded to EPSPs and bAPs with the same pharmacological application.

3.3.2. Time-course of VSD signals of excitatory and

inhibitory inputs

In hippocampal CA1 pyramidal cells, excitatory and inhibitory synaptic inputs temporally overlap in dendrites (Karnup and Stelzer, 1999). In addition, most feedforward post-synaptic inhibitory responses were

(45)

observed late for excitatory responses (Pouille and Scanziani, 2001; Turner, 1990). Therefore, we tried to identify the each latency for peak amplitudes of excitation and inhibition from VSD signals.

To calculate the time-course of excitatory and inhibitory synaptic inputs to dendrites in VSD imaging, we performed the comparative experiment for the VSD signals in naïve (intact inhibition) and GABA_A receptor- blocked conditions. After that, the time-course of inhibition was calculated by subtraction the VSD signal of GABA_A receptor-blocked condition from that of naïve condition (Figure 3-4A and B). The mean latencies of excitatory and inhibitory peak amplitudes were 6.8 ± 0.7 ms [n = 8, mean ± standard error of the mean (SEM)] and 21.6 ± 1.3 ms (n = 5, mean ± SEM), respectively, at PD and 7.1 ± 0.8 ms (n = 7, mean ± SEM) and 19.7 ± 1.7 ms (n = 6, mean ± SEM), respectively, at DD (Figure 3-4C). These values were almost confirmed to electrophysiologial data (Karnup and Stelzer, 1999; Turner, 1990). From these reasons, when the pairing stimulations with the E- and I-phase, in which we applied to paired stimulations of bAP input and PD/DD inputs, it is consistent with that bAPs might be paired with EPSPs and IPSPs at each input location by optical imaging with VSD.

(46)

Figure 3-4: Latency to peak of excitatory and inhibitory components in the VSD signal

A1-3: Schematic diagrams of the calculation of inhibitory component in naïve VSD

signals. (A1; VSD signal in naïve (intact inhibition) condition, A2; the signal in GABAA receptor-blocked condition, A3; the inhibitory component calculated by subtraction the naïve signal (A1) from the GABAA receptor-blocked signal (A2)). B1-3: Sample traces of VSD responses and the calculated inhibitory component. Scale bar: 0.2%, 10 ms. C: Summary of the latency from stimulation to peak responses of excitatory and inhibitory components. Circles and square indicate the latency of peak excitation at PD and DD, respectively. Triangles and diamonds indicate that of peak inhibition at PD and DD, respectively. Filled and open symbols indicate the mean and individual values, respectively. t t t A1 B1 C A3 A2 B2 B3 stim. stim. stim. -0.15 -0.10 -0.05 0.00 0.05 0.10 VSD signal ( F/F; %) 25 20 15 10 5 0 Latency (ms) 25 20 15 10 5 Excitatory, PD Excitatory, DD Inhibition, PD Inhibition, DD

(47)

3.3.3. bAPs were differently modulated by synaptic inputs

at various relative timing of pre- and post-synaptic

activations

To investigate the modulation of bAPs by synaptic inputs at PD (PD input), three sets of pairing stimuli were applied with the relative timings of τP = 5

ms (E-phase), τP = 20 ms (I-phase), or τP = −5 ms (Ne-phase). Figure 3-5

shows examples of the VSD signals that were elicited by the spread of bAP and the modulation by the PD input. Each trace in each data set indicated VSD signals at the following locations (from upper to lower): the pyramidal cell somata, apical PDs, apical middle dendrites (the middle of the proximal and distal points), apical DDs, and apical tufts (lacunosum moleculare layer). The second and fourth data points were labeled as PD and DD, respectively.

First, the superposed amplitudes of bAPs and EPSPs that were induced by the E-phase PD input were measured along the somatodendritic axis (Fig. 3-5A). The left panel shows bAPs along the dendrite when a stimulus (bAP input) was applied at the alveus. bAPs were elicited and delivered to DD. The amplitude decreased with the distance from the soma. The middle panel shows EPSP that was elicited at PD by a synaptic input to the Schaffer-commissural collaterals (SC). Here, EPSP that was propagated to DD was markedly reduced. The right panel shows the results of bAP

(48)

3' '' E E$3 LQSXW PD input E$3 LQSXW3'LQSXW

A

E-phase (τ

P

= 5 ms)

B

I-phase (τ

P

= 20 ms)

C

Ne-phase (τ

P

= -5 ms)

3' ''

E$3 LQSXW PD input E$3 LQSXW3'LQSXW

3'

''

(49)

Figure 3-5: Examples of bAP modulation by PD inputs

A: bAP modulation by E-phase PD inputs (τP = 5 ms). The traces indicate the optical imaging responses along the somatodendritic axis. Each trace indicates voltage-sensitive dye (VSD) signals from the pyramidal cell somata to the stratum lacunosum moleculare. VSD signals in the apical proximal and distal dendrites (indicated by thick black lines) were used for detailed analysis of PD and DD, respectively (Figs. 3–7). The arrows indicate the stimulus sites. Left: bAP propagation (control bAP). A bAP stimulus was applied at the alveus, and bAP that was generated propagated from the soma to the dendritic ends along the dendrite. The amplitudes of bAPs decreased with increasing distance from the soma. Middle: EPSPs elicited by PD inputs. Right: bAP modulation by PD inputs. bAPs were amplified at PD and then propagated to DD. The amplitudes of the modulated bAPs at DD were larger than those of the control bAPs. B: bAP modulation by I-phase PD inputs (τP = 20 ms). bAPs were attenuated at PD and propagated to DD. The amplitudes of the modulated bAPs at DD were smaller than those of the control bAP. C: bAP modulation by Ne-phase PD inputs (τP = −5 ms). bAPs were not significantly changed, and the amplitudes of bAPs at DD remained unchanged. Calibration: 2.0 × 10−3 (ΔF/F), 20 ms.

______________________________________________________________________

modulation by the PD stimulus. The synaptic input was coincidently applied to PD at the E-phase (E-phase PD input). bAPs were amplified (124.9%) and delivered to DD. The modulated bAP at DD was larger (128.3%) than the control bAP at DD without the PD input, as shown in the left panel.

Second, the superposed amplitudes that were obtained by pairing bAP and the I-phase PD input were measured at PD. Figure 3-5B shows an example of the results. The left and middle panels show the same results as those for the E-phase PD input. The right panel shows the results of bAP modulation by the PD input. An I-phase PD input was applied so that bAP was changed because of the timing of IPSP. As a result, bAPs were

(50)

attenuated (86.0%) and delivered to DD. The modulated bAP at DD was smaller (73.1%) than the control bAP at DD, as shown in the left panel.

Third, the superposed amplitudes that were obtained by pairing bAP and the Ne-phase PD input were measured at PD. Figure 3-5C shows an example of the results. The left and middle panels show the same results as those for the E- and I-phase PD inputs. The right panel shows that bAP modulated by the Ne-phase PD input was not significantly different from the control without the PD input and that bAP delivered to DD was not different from that without the PD input.

Figure 3-6 summarizes the results of bAP modulation by various PD inputs. Figure 3-6A shows that the average amplitudes of bAPs at PD were amplified to 135.7 ± 3.0% (+35.7 ± 3.0%, n = 15) by the E-phase PD input, attenuated to 66.1 ± 5.8% (−33.9 ± 5.8%, n = 15) by the I-phase PD input, and not significantly changed (105.6 ± 1.5%, n = 12) by the Ne-phase PD input in comparison with those of normal bAPs without PD inputs. Figure 3-6B shows that the average amplitudes of propagated bAPs were modulated by E-phase, I-phase, and Ne-phase inputs at DD; the average amplitudes were 129.2 ± 9.2% (+29.2 ± 9.2%, n = 15), 61.4 ± 3.1% (−33.9 3.1%, n = 15), and 101.5 ± 4.4% (+1.5 ± 4.4%, n = 15), respectively, in comparison with those of normal bAPs without PD inputs. These findings suggested that bAPs were enhanced or suppressed depending on the timing

(51)

of PD inputs and the delivery to DD. Thus, modulation of bAPs may affect STDP at DD.

Figure 3-6: Summarized results of bAP modulation

bAPs were modulated by the timing of PD inputs. The amplitude changes of bAPs at PD (A) and DD (B). Single and double asterisks indicate significant differences of P < 0.05 and P < 0.01, respectively.

3.3.4. Long-term synaptic modifications determined with

conventional electrophysiological recording and VSD

imaging

We verified the use of optical recording to monitor the long-term modifications of synaptic responses (LTP or LTD) in comparison with the results obtained with the conventional electrophysiological recording method of extracellular field recording. Figure 3-7 indicates the results of the long-term synaptic modifications induced by the pre and postsynaptic

-40 -20 0 20 40

E-phase I-phase Ne-phase -40

-20 0 20 40

E-phase I-phase Ne-phase

DD Am pl itu de c ha ng e of b AP (% ) _** ** ** ** ** * PD (τP= 5 ms) (τP= 20 ms) (τP= -5 ms) Am pl itu de c ha ng e of b AP (% ) (τP= 5 ms) (τP= 20 ms) (τP= -5 ms) PD input PD input

A

B

(52)

pairing stimulation paradigm with simultaneous electrophysiological and VSD recordings. The relative timings of 5 and 20 ms (E-phase and I-phase,

160 140 120 100 80 60 40 20 0 Piaring stim (τP = +5 ms) Time (min) Relative timings Time (min) 175 150 125 100 75 45 30 15 0 -15 110 100 90 80 70 45 30 15 0 -15 Piaring stim (τD = +20 ms) VSD rec. before VSD rec. before VSD rec. after VSD rec. after

A

B

C

D

E

fEPSP slope (%) fEPSP slope (%) Plasticity (%) +5 ms +20 ms slope N.S. N.S.

(53)

Figure 3-7: Simultaneous electrophysiological and VSD recordings of long-term modifications

A, B: Long-term monitoring of field EPSPs (fEPSPs) with the application of E-phase

(A) or I-phase (B) pairing stimulation. The white arrows indicate the time points for the

acquisition of the VSD signal before and after the pairing stimulation. C, D: Sample traces of fEPSPs before and after E- (C) or I-phase (D) pairing stimulations. Gray and black traces indicate the responses to the test stimulus before and after the pairing stimulation, respectively. Calibration: 0.2 mV, 10 ms. E: The plastic changes of the fEPSP slope and VSD peak indicated a similar tendency. The gray and white bars indicate the plastic changes of the VSD peak and fEPSP slope, respectively. Comparison of the responses before pairing with those after pairing revealed that the peak of the VSD signal as well as the fEPSP slope was significantly potentiated or depressed (Student’s paired t-test, P < 0.05, n = 6 and 5 pairs for potentiation and depression, respectively). In addition, statistical significance was not found between the changes in the VSD peak and fEPSP slope for the two relative timings (+5 or +20 ms) of the pairing conditions. N.S.: not significant.

respectively) were used for the pairing stimulation. These timing parameters were also used in our previous STDP studies (Aihara et al., 2007, 2005; Tsukada et al., 2005).

These results confirmed that LTP and LTD were observed with the extracellular recordings of the STDP-induction pairing protocols, the relative timings of which were 5 and 20 ms, respectively (LTP: 146.77 ± 10.52%, LTD: 92.59 ± 2.22%, paired t-test, P < 0.05, n = 6 for LTP and 5 for LTD). The modifications persisted for at least 45 min (Figures 3-7A and 3-7B). Moreover, the magnitudes of the plasticity were almost identical between the field EPSP slope and the VSD peak in the case of LTP and LTD (Figure 3-7E, not significant, Student’s t-test, P < 0.05).

(54)

3.3.5. Modulations of STDP by synaptic inputs at other site

Influence of proximal synaptic inputs to distal STDP

Next, to investigate the influence of the different relative timings of PD inputs during STDP induction at DD, STDP at DD were measured with PD inputs at three relative timings (E-, I-, and Ne-phases) (E-phase at DD in Fig. 3-8 and I-phase at DD in Figure 3-9). STDP was measured as the amplitude change of EPSP before and after the pairing stimuli. Figure 3-8 shows that STDP was induced with the E-phase (STDPE) at DD with and

without the PD input during the STDP induction protocol.

Diagrams in the upper part of Figure 3-8A shows the experimental procedure in which DD inputs were applied with E-phase timing (τD = 5

ms) following the PD pairing stimuli with E-phase (τP = 5 ms), I- phase (τP

= 20 ms), or Ne-phase timing (τP = -5 ms) as well as without the PD input.

Lower traces in Figure 3-8A show DD input-induced VSD signals of each induction patterns before and after the pairing stimuli as gray and black lines, respectively. The results are summarized in Figure 3-8B. Without PD inputs, STDPE was induced as LTP (LTP, P < 0.01; 120.4 ± 6.7%, n = 12).

In addition, when the E-phase PD input was applied, STDPE was facilitated

(LTP, P < 0.01; 149.3 ± 9.8%, n = 12). However, STDPE that was

modulated by the I-phase PD input (92.7 ± 6.6%, n = 8) was smaller than the control STDPE. Moreover, STDPE following the Ne-phase PD input

(55)

Figure 3-8: Influence of the timing of PD inputs on the distal STDP induced by E-phase timing

A: Schematic drawing of the pairing stimuli and STDP measuring sites. PD inputs were applied with the three relative timings of the

E-, I-, and Ne-phases. Modulation of the distal STDPE (i.e., STDP induced with the E-phase in DD) was measured. Lower traces indicate examples of optical imaging responses. The gray and black traces correspond with the VSD signals before and after the distal STDPE induction, respectively. Calibration; 0.05% and 20 ms. B: Modulation of distal STDPE by the timing of PD inputs. Distal STDPE was more potentiated with the E-phase PD input than without the PD input (i.e., the control). It was depressed with I-phase PD inputs compared with controls. Finally, it was not significantly modulated by Ne-phase PD inputs. The single and double asterisks indicate significant differences of P < 0.05 and P < 0.01, respectively. N.S. denotes no significant difference.

0 20 40 60 80 100 120 140 160 ** * * Distal STDP E (%) N.S. PD input DD input bAP input 'LVWDO 67'3( bAP EPSP + EPSP-phase Proximal input + EPSP-phase Proximal input Control (w/o Proximal input) + IPSP-phase Proximal input + IPSP-phase Proximal input + Negative-phase Proximal input

+ Negative-phase_{Proximal input}

EPSP EPSP EPSP bAP

Control (w/o Proximal input)

EPSP bAP IPSP bAP τP= 5 ms w/o PD input τP= 20 ms τP= -5 ms

海馬CA1樹状突起における入力統合機能の解析

平成 25 年度

学位論文（博士）