ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2016

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1190

Functional Imaging of SpinalLocomotor Networks

CHETAN NAGARAJA

ISSN 1651-6206ISBN 978-91-554-9500-8urn:nbn:se:uu:diva-280062

Dissertation presented at Uppsala University to be publicly examined in B22, BiomedicalCentre, Uppsala, Wednesday, 27 April 2016 at 09:00 for the degree of Doctor of Philosophy(Faculty of Medicine). The examination will be conducted in English. Faculty examiner:Associate Professor Gilad Silberberg (Karolinska Institute, Department of Neuroscience).

AbstractNagaraja, C. 2016. Functional Imaging of Spinal Locomotor Networks. DigitalComprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1190.47 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9500-8.

Movement is necessary for the survival of most animals. The spinal cord contains neuronalnetworks that are capable of motor coordination and of producing different movements. Inparticular, a very reduced neuronal network in the spinal cord can produce simple rhythmicoutputs even in the absence of descending or sensory inputs. This basic circuit was discoveredby Thomas Graham Brown (reported in 1911) and is termed central pattern generator. For overa century a large number of studies have been carried out in order to identify the neuronalcomponents that are part of these networks.

In project 1 we focused on Renshaw cells, which are a population of spinal interneuronsexpressing the alpha-2 subunit of the nicotinic acetylcholine receptors (Chrna2). Renshawcells are the only identified central targets for motor neuron inputs, and in turn they mediateinhibition of the motor neurons. We analyzed the activity pattern of Renshaw cells on a cell-population level in neonates when the circuit is still developing. At segment 1 of the lumbarspinal cord, Renshaw cells show significantly greater activity response to functional sensoryand motor inputs from rostral compared to the caudal segments. Contrarily, the suppression ofthe monosynaptic stretch reflex was more pronounced when caudal roots were stimulated. Ourdata underline the importance of sensory input during motor circuit development and help tounderstand the functional organization of Renshaw cell connectivity.

Several neurons that play distinct roles in locomotor central pattern generation have beenidentified with the help of genetics. For instance, the V0 population of spinal interneurons areidentified by the expression of transcription factor developing brain homeobox 1 (Dbx1). V0neurons are necessary for producing an alternating rhythm at all locomotor speeds. In project2 we have looked at a population of dorsally derived ventrally projecting interneurons thatexpress the transcription factor doublesex and mab-3 related transcription factor 3 (Dmrt3).On a behavioral level Dmrt3 neurons are involved in regulating coordination across differentlocomotor speeds. On a microcircuit level, we have shown that individual Dmrt3 neurons showdistinct frequencies of oscillations for a constant locomotor rhythm. In addition, removal ofinhibitory neurotransmission from Dmrt3 neurons results in uncoupling of rhythm in motorneurons.

In project 3 the activity patterns in populations of flexor related motor neurons arecharacterized during fictive locomotion in neonatal mice. An interesting and intriguing findingin project 3 is the presence of multiple rhythmicities in motor neurons. Multiple rhythmicitiesare seen even when the locomotor output shows a constant frequency.

Keywords: Central pattern generators, two-photon microscopy, locomotor rhythm, multiplerhythmicities, inhibitory neurotransmission

Chetan Nagaraja, , Department of Neuroscience, Box 593, Uppsala University, SE-75124Uppsala, Sweden.

© Chetan Nagaraja 2016

ISSN 1651-6206ISBN 978-91-554-9500-8urn:nbn:se:uu:diva-280062 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-280062)

 

You never change things by fighting the existing reality. To change some-

thing, build a new model that makes the existing model obsolete

Buckminster Fuller

 

List of papers

This thesis is based on the following papers, which are referred to in the text by their roman numerals

I Nagaraja, C., Helmchen, F., Peuckert, C., Kullander, K. Topographical organization of functional sensory and motor inputs to Renshaw cells. Manuscript

II Larhammar, M*., Perry, S*., Nagaraja, C., Tafreshiha, A.,

Hilscher, M., Edwards, S., Caxieta, F., Kullander, K. Dmrt3 derived spinal cord neurons regulate locomotor coordination across different speeds. Manuscript

III Nagaraja, C., Kullander, K. A microcircuit analysis of motor

neuron behavior during fictive locomotion. Manuscript

IV Langer, D., van ’t Hoff, M., Keller, A.J., Nagaraja, C., Pfäffli, O.A., Göldi, M., Kasper, H., Helmchen, F. HelioScan: A software framework for controlling in vivo microscopy set-ups with high hardware flexibility, functional diversity and ex-tendibility. Journal of Neuroscience Methods. 2013 Feb 6; 215, 38-52.

* Equal contribution Reprint was made with permission from Elsevier for paper IV.

 

Contents

Introduction ..................................................................................................... 9 Two-photon calcium imaging of neuronal network activity ...................... 9 

Principle of two-photon microscopy ................................................... 10 Advantages of two-photon over one-photon excitation fluorescence microscopy ........................................................................................... 10 

Calcium imaging for recording neuronal activity .................................... 11 Spinal motor circuits ................................................................................ 12 Motor neurons .......................................................................................... 14 

Motor pools and their topographical organization ............................... 14 Proprioception ..................................................................................... 15 

Spinal reflexes .......................................................................................... 16 Renshaw cells ........................................................................................... 17 

Renshaw cell discovery ....................................................................... 17 Location, anatomy and electrophysiological properties of Renshaw cells .................................................................................. 18 Immunohistochemical identification of Renshaw cells ....................... 18 Synaptic inputs and outputs of Renshaw cells ..................................... 18 Renshaw cell inputs to other neurons .................................................. 19 Role of Renshaw cells in motor control ............................................... 19 

Central pattern generators (CPG) ............................................................. 20 Genetics to define components of the locomotor CPG ........................ 21 Glutamate as the main neurotransmitter for central pattern generation ............................................................................................ 24 Effect of inhibitory neurotransmitters on central pattern generation .... 24 

Aims of the projects ...................................................................................... 25 

Materials and Methods .................................................................................. 26 

Results and Discussions ................................................................................ 32 

Acknowledgements ....................................................................................... 38 

References ..................................................................................................... 41 

Abbreviations

CGDA Calcium Green-1-conjugated dextran amine Chrna2 Nicotinic acetylcholine receptor subunit α2 CNS Central nervous system CPG Central pattern generator Dmrt3 Doublesex and mab-3 related transcription factor 3 DRG Dorsal root ganglion DTA Diphtheria toxin Allatostatin GABA Gamma-Aminobutyric acid GTO Golgi tendon organ KO Knock out NA Numerical aperture OPA One-photon absorption PD Progenitor domain TPA Two-photon absorption TPI Two-photon imaging TPM Two-photon microscopy VAChT Vesicular acetylcholine transporter VGLUT 1/ 2 Vesicular glutamate transporters 1/2 V0V / V0D Ventral / dorsal subpopulations of V0 interneurons WT Wild type

9

Introduction

Functional locomotor output measured from the ventral roots or muscles is a summation of activity in individual motor neurons. Movement is brought about by a cascade of contractions of different muscles in a coordinated manner. Muscle activity is driven by populations of motor neurons that show a spatiotemporal distribution and are in turn interacting with premotor inter-neurons (Song et al., 2016) to produce motor output. Hence, knowledge of the characteristic features of motor neurons and interneurons during locomotor activity is essential to understand spinal locomotor networks in health, and aid in identifying the cause of locomotor disorders.

Investigations on neuronal circuits have been greatly facilitated by several technological advancements. These include achievements in molecular ge-netics, development of activity indicators, instrumentation to visualize struc-ture and function of the nervous system, and optogenetics enabling light activation or inactivation of activity in populations of neurons in any desired spatiotemporal order. With the tools at hand great strides have been made in understanding the CNS is general and the locomotor system in particular (Goulding, 2009).

The Kullander lab has identified markers that have aided genetic targeting of Renshaw (Perry et al., 2015) and Dmrt3 cells (Larhammar et al., manu-script). The aim of this thesis is to characterize activity patterns in popula-tions of spinal interneurons and motor neurons during locomotor like activity and other electrophysiological assays.

Two-photon calcium imaging of neuronal network activity Two photon imaging has enabled investigations on nervous system structure and function with good spatial and temporal resolution. It is a fluorescence imaging technique that has benefitted from advancements in microscopy, generation of cellular activity indicators, and molecular genetics (Denk et al., 1990; Grienberger and Konnerth, 2012; Jessell, 2000). Structures as small as a few microns to several millimeters (Svoboda and Yasuda, 2006), and activ-ity over milliseconds (Grewe et al., 2010) can be visualized.

Activity in neuronal populations can be measured in intact tissue longitu-dinally (Helmchen and Denk, 2005). Of the various cellular activity indica-

10 

tors, synthetic (Stosiek et al., 2003), or genetically encoded fluorescent cal-cium indicators (Miyawaki et al., 1997) are most preferred. Calcium signal-ing is involved in the regulation of various cellular functions, from neuro-transmitter release occurring over microseconds to gene transcription occur-ring over several minutes to hours (Grienberger and Konnerth, 2012). Hence, imaging calcium transients has emerged as a popular tool to investigate the activity in populations of neurons.

Two-photon microscopy overcomes the limitations of techniques such as electrophysiology, that only allow investigation of single cell or groups of cells (but lacking specificity), and functional magnetic resonance imaging or optical imaging, probing activity in large regions of the CNS and thereby lacking cellular resolution. Thus two-photon imaging has emerged as a very useful technique to investigate nervous system function in health and dis-ease.

Principle of two-photon microscopy Two-photon microscopy (TPM) is based on the principle of two-photon ab-sorption (TPA) as proposed by Maria-Goeppert Mayer in her doctoral thesis (Göppert-Mayer, 1931). TPA occurs when two photons with a specific wavelength (e.g. 800nm) arrive nearly simultaneously (delta t ~ 0.5 fs) at a molecule, resulting in an excitation of the molecule followed by an emission of light with shorter wavelength (e.g. 400nm) thereby returning to the ground energy level (Figure 1A) (Helmchen and Denk, 2005). For TPA to occur, a large photon density is necessary, which is supplied by femtosecond pulsed lasers (Denk et al., 1990). This type of microscopy is called nonlinear microscopy (Denk et al., 1990; Helmchen and Denk, 2005). A large flux of photons translates to a high intensity of exciting light, which would damage the tissue if delivered continuously. This problem is overcome by the use of pulsed lasers that deliver very high instantaneous intensities necessary for achieving TPA, while keeping the average intensities within the tolerable limits for biological tissues (Diaspro et al., 2006). Another requirement to reach a high photon density is the use of objectives with high numerical ap-erture (Helmchen and Denk, 2005).

Advantages of two-photon over one-photon excitation fluorescence microscopy TPM uses wavelengths in the near-infrared region, biological tissue inher-ently scatters less at these wavelengths allowing a penetration depth of up to 1 millimeter (Grienberger and Konnerth, 2012; Helmchen and Denk, 2005), which is not possible with one-photon absorption (OPA). As TPE depends quadratically on the intensity of the excitation light, the excitation is limited

11

to a small focal volume, whereas with OPA excitation occurs in the regions above and below the focal volume (Fig. 1B).

The limitation of the excitation site to the focal volume and the possibility to use long wavelength light for excitation results in a reduction of photo damage and photobleaching compared to one-photon excited fluorescence (Benninger and Piston, 2013). In the confocal microscope, a pinhole is used to reject emitted light from off focus planes, which is dispensable when TPA is used for excitation (Helmchen and Denk, 2005). Hence, with TPM it is possible to image populations of neurons with cellular resolution deep in intact tissue and because of limited photo damage and bleaching, longitudi-nal studies can be performed (Denk et al., 1990; Helmchen and Denk, 2005).

Figure 1. Principle of two-photon absorption

A. Illustration of one- and two-photon absorption giving rise to one-photon emis-sion, using Jablonski diagram. B. Fluorescence excitation in a specimen by one- and two-photon absorption. As two-photon absorption is nonlinear the emission decreas-es as 1/z2 from the focal center, where z is the axial distance.

Calcium imaging for recording neuronal activity Calcium signaling is involved in many cellular functions such as contraction of cardiac, smooth and skeletal muscles, gene transcription, regulation of ion channels, ATP production and exocytosis (Berridge et al., 2000; Grienberger and Konnerth, 2012). Intracellular calcium is either free or bound to calcium buffering proteins such as calretinin, calmodulin, calbindin and parvalbumin. The intracellular calcium concentration is around 100 nM and when this rises to 1000 nM, the cell is activated (Berridge et al., 2000). Increase in concentration of cytosolic calcium is caused by entry from the outside or release from intracellular stores (Berridge et al., 2000). The entry of calcium into the cytosol occurs by the opening of channels: voltage operated calcium channels (VOC), receptor operated calcium channels (ROC), second mes-senger operated calcium channels (SMOC), mechanically operated calcium channels (MOC), or tonically active calcium channels (Berridge et al., 2000; Tsien and Tsien, 1990). The release of calcium from internal stores is medi-ated by Inositol-tri-phosphate receptors (InsP3R) and Ryanodine receptors

12 

(RyRs). The InsP3 receptors form intracellular calcium channels that are regulated by InsP3 concentration as well as cytosolic calcium (Taylor, 1998).

There are several cellular events that give rise to an increase in intracellu-lar calcium, which can be restricted to sites within a cell, the entire cell or be a propogating wave of synchronized population activity (Cheng et al., 1993; Kirkby et al., 2013; Thomas et al., 1998). Localized calcium release events occur at and are restricted to sites of intracellular stores. These calcium re-lease events are mediated by Ryanodine receptors (RyRs) termed ‘calcium sparks’ (Cheng et al., 1993) or by InsP3 receptors called ‘calcium blips’ for smaller calcium release events and ‘calcium puffs’ for larger calcium events (Thomas et al., 1998). The calcium release events can be stereotypic or vari-able in both amplitude and duration (Cheng et al., 1993; Thomas et al., 1998). These variable calcium release events are believed to reflect interac-tions between multiple calcium channels made of either InsP3 or Ryanodine receptor cluster. When the frequency of the calcium release events increases and sufficient intracellular calcium concentration is reached, a calcium wave is generated (Thomas et al., 1998). Intracellular calcium can also increase by influx of calcium. Of particular interest are propagating waves of synchro-nized activity in populations of neurons early in development. These propa-gating waves of calcium activity are generated throughout the nervous sys-tem across different species (Blankenship and Feller, 2010; Kirkby et al., 2013; Rosenberg and Spitzer, 2011). They are mediated through ATP, nico-tinic acetylcholine, Glutamate, GABAA receptors in different parts of the nervous system and are implicated in refinement of synaptic connections in the nervous system (Ackman et al., 2012; Hanson and Landmesser, 2006; Kirkby et al., 2013).

Suitable dyes have been developed that can bind to cytosolic free calcium and transduce a change in intracellular calcium concentration reported as a change in fluorescence intensity (Grienberger and Konnerth, 2012; Miyawaki et al., 1997). Such dyes, combined with two-photon microscopy have permitted measurements of cellular activity from populations of neu-rons with good temporal and spatial resolution (Garaschuk et al., 2000; Gobel et al., 2007).

Spinal motor circuits The spinal cord contains neuronal circuits that can, in the absence of de-scending or sensory inputs, generate simple rhythmic movements and simple reflex movements such as myotactic, reciprocal, withdrawal and crossed-extensor reflexes (Goulding, 2009; Guertin, 2009; Kjaerulff and Kiehn, 1996). These neuronal networks are called central pattern generators or CPGs (Goulding, 2009; Grillner, 2003; Marder and Bucher, 2001). A CPG consists of premotor interneurons that drive motor neurons (Andersson et al., 2012; Guertin, 2009; Hägglund et al., 2013). The locomotor CPG has been

13

extensively investigated using behavioural, in vivo and in vitro experiments (Andersson et al., 2012; Gosgnach et al., 2006; Talpalar et al., 2013; Wallén-Mackenzie et al., 2006). A widely used locomotor assay known as fictive locomotion has been employed to characterize the locomotor CPG across species (Figure 2A) (Cowley and Schmidt, 1994; Gabriel et al., 2008; Kriellaars et al., 1994; Kudo and Yamada, 1987; Wallén and Williams, 1984; Zhong et al., 2007). Fictive locomotion can be elicited in isolated spi-nal cords, hindlimb-attached spinal cords or decerebrate preparations by bath application of drug cocktails or by electrical stimulation of descending tracts or sensory afferents (Cowley and Schmidt, 1994; Kudo and Yamada, 1987; Noga et al., 2003). The locomotor-like rhythm in fictive locomotion is a sequence of rhythmic bursts that is alternating between the left and the right as well as between flexors and extensors (Figure 2B). The ventral spinal cord has been found to contain circuits necessary and sufficient for locomotor-like activity (Cina and Hochman, 2000; Dai et al., 2005; Goulding, 2009; Kjaerulff and Kiehn, 1996) using activity dependent labeling (Cina and Hochman, 2000; Dai et al., 2005) or by eliciting fictive locomotion in dorsal horn removed preparations (schematics of a dorsal horn removed preparation shown in Figure 2C) (Dougherty and Kiehn, 2010; Kjaerulff and Kiehn, 1996). Further it has been shown that several genetically identified premotor interneurons necessary for normal locomotion are located in the ventral part of the spinal cord (Crone et al., 2008; Gosgnach et al., 2006).

Figure 2. Schematics of an in vitro preparation of a mouse spinal cord for fictive locomotion experiments

A. Ventral roots L2 (flexor related) and L5 (extensor related) are connected to suction electrodes (arrowhead) to measure locomotor output during fictive locomotion. Alternating rhythm (raw trace), generated by bath application of drugs, measured from the right L2 and L5 ventral roots are indicated by arrows. B. The locomotor output (electroneurograms, ENG) measured from

14 

three ventral roots are shown after rectification and smoothing. Note the alternation between left-right (rL2 – lL2) and flexor-extensor (rL2 – rL5). C. Schematics of a dorsal horn removed spinal cord preparation that is suffi-cient to generate fictive locomotion.

Motor neurons Motor neurons can be classified as upper and lower motor neurons. Upper motor neurons are present in the motor regions of the cerebral cortex and brainstem. They innervate lower motor neurons and are not involved in gen-erating movement directly (Stifani, 2014). Lower motor neurons are present in the brainstem and spinal cord and send their axons through the ventral roots and cranial nerves. A subset of lower motor neurons called somatic motor neurons innervate skeletal muscles and hence bring about movement. The somatic motor neurons can be classified into: alpha, beta and gamma motor neurons (Stifani, 2014). The somatic motor neurons innervating hindlimb muscles are located in the lumbar spinal cord at lamina IX. Activi-ty patterns of motor neurons in the lumbar spinal cord during locomotor-like activity have been the focus of recent investigations (Hinckley et al., 2015; Machado et al., 2015).

Alpha, beta and gamma motor neurons Contractions of skeletal muscles are initiated by motor neurons called alpha motor neurons. Alpha motor neurons are the largest of the 3 somatic motor neurons. Gamma motor neurons do not play a role in force generation and innervate specialized organelles interspersed in muscles called the muscle spindles (Friese et al., 2009; Hunt, 1951; Kuffler et al., 1951; Shneider et al., 2009). Muscle spindles are capsular structures contained in muscles called intrafusal fibres, and are sites of a specialized sensory afferent innervation. Muscle fibres outside the muscle spindles are called extrafusal. Further, gamma motor neurons innervate muscle spindles and regulate the sensitivity of the muscle spindle receptors (Shneider et al., 2009). Alpha and gamma motor neurons differ considerably in their soma size (Friese et al., 2009; Shneider et al., 2009) and axonal diameters and conduction velocities (Eccles and Sherrington, 1930; Hunt, 1951; Kuffler et al., 1951). Beta motor neurons form the least understood group of somatic motor neurons and in-nervate both muscle spindles and extrafusal fibres (Bessou et al., 1965; Stifani, 2014).

Motor pools and their topographical organization Muscles are controlled by populations of motor neurons, and motor neurons alone are the output system of the nervous system (Carp and Wolpaw, 2001). Populations of motor neurons innervating a muscle are called a motor pool,

15

and each motor pool innervates a muscle (Figure 3B). Motor pools lie side-by-side along the length of the spinal cord (Figures 3Ai and ii). Motor pools and the muscles they innervate in the hindlimbs show a topographic relation (Romanes, 1951; Sürmeli et al., 2011; Vanderhorst and Holstege, 1997). The proximal muscles for the hindlimbs have their motor pools ventral to those for distal muscles in mice (Sürmeli et al., 2011). That means, when moving from the ventral to the dorsal part of the ventral horn in the spinal cord, mo-tor pools that innervate proximal muscles are followed by those that inner-vate distal muscles.

Proprioception Proprioceptors are sensory units that detect changes in the length and forces generated in muscles. There are at least two types of proprioceptors: golgi tendon organs (GTO) and the muscle spindles (Donoghue and Sanes, 2001). GTOs detect forces generated in the muscles, and convey this information to the CNS through 1b sensory afferents. Muscle spindle afferents detect changes in length of the muscles and convey this information to the CNS. These sensory afferents are called proprioceptive 1a afferents and are the only sensory afferents that make monosynaptic connections onto motor neu-rons (Figures 3B and C).

16 

Figure 3. Motor pools, proprioceptive 1a afferents and monosynaptic reflex

A. Schematics of a spinal cord with distribution of motor pools and their axons exit-ing through the ventral roots. i) Three different pools each represented in a different colour and occupying different spatial positions along the dorso (d) – ventral (v), medio (m) - lateral (l) and rostro (r) – caudal (c) axis. ii) Transverse section through the spinal cord showing the motor pools. B. Schematics of a spinal cord showing a pool of motor neurons, in red, that innervates a muscle. Motor roots exit the spinal cord through the ventral roots (in red) and sensory roots (in black) enter the spinal cord and have their cell bodies in the dorsal root ganglion (DRG). C. Transverse section of a spinal cord showing direct inputs from proprioceptive 1a afferents (black) onto motor neurons (red). A monosynaptic reflex response (MSRR) meas-ured at the ventral root, is elicited by stimulation of the dorsal root.

Spinal reflexes Stereotyped behaviours generated in response to sensory stimuli are known as reflexes. Reflexes are produced by the central nervous system and involve both the brain and spinal cord. Reflexes are comprised of a sensory and mo-tor component. Sensory inputs are conveyed by the sensory afferents whose cell bodies are located in the dorsal root ganglion. The motor component consists of motor neurons that can induce contraction of muscles or auto-nomic effector organs (Burke, 2001). Motor neurons from one motor pool innervate one muscle (Figure 3B). In turn, the proprioceptive 1a afferents, that detect changes in muscle length project back onto the homonymous

17

(same) motor pool thereby giving rise to a highly specific connectivity pat-tern (Figure 3B) (Mears and Frank, 1997; Sürmeli et al., 2011). Electrical stimulation of the dorsal roots in rodents has been employed to elicit propri-oceptive 1a afferents evoked monosynaptic reflex response from motor neu-rons (Figure 3C). The monosynaptic reflex response can be measured from segmental ventral roots and has a latency of 6-8 ms (Machacek and Hochman, 2006). There could be one or more interneurons interposed be-tween the sensory and motor components, in the case of the presence of a single interneuron the reflex is called a disynaptic reflex.

Renshaw cells Renshaw cells are a group of spinal interneurons that are the only identified targets for motor neuronal axon collaterals into the central nervous system. They receive functional excitatory inputs from motor neuron axon collat-erals. In response to ventral root stimulations (antidromic stimulations) the Renshaw cells produce a high frequency burst of action potentials (almost 1500 Hz) (Alvarez and Fyffe, 2007; Mentis et al., 2006; Renshaw, 1946). The phenomenon of Renshaw cells receiving excitatory inputs from motor neuron axon collaterals and in turn mediating inhibition of the motor neurons is termed recurrent inhibition (Bhumbra et al., 2014; Eccles et al., 1954). Their identification has been aided by their location, anatomy and electro-physiological signature (Alvarez and Fyffe, 2007).

Renshaw cell discovery Antidromic stimulation of motor neuron axons resulted in suppression of the output from non-stimulated motor neurons (Renshaw, 1941, 1946). This phenomenon is known as conditioning and has been investigated in the con-text of the monosynaptic reflex response (Devanandan et al., 1965; Eccles et al., 1961b; Eccles et al., 1963; Machacek and Hochman, 2006; Rudomin and Dutton, 1969). Thus, it was argued that subsequent to the arrival of the antidromic impulses to the motor neuron soma there had to be neural activity that conditioned activity in the other motor neurons (Renshaw, 1946). Fur-ther investigations lead to the discovery of a group of spinal interneurons located in the ventral horn that responded with a high frequency burst of action potentials in response to antidromic stimulation. These neurons were named Renshaw cells by John Carew Eccles, as it was Birdsey Renshaw who had carried out the systematic investigations that lead to the discovery of these interneurons (Eccles et al., 1954; Renshaw, 1946).

18 

Location, anatomy and electrophysiological properties of Renshaw cells Renshaw cells are located in the lamina VII of the ventral horn (Alvarez et al., 1999; Alvarez et al., 2005; Benito-Gonzalez and Alvarez, 2012; Perry et al., 2015). They have dendritic arborisations that extend dorsoventral or mediolateral, further they are a morphologically unclassified group of inter-neurons (Perry et al., 2015). Distinctions can be made between motor neu-rons, 1a inhibitory interneurons and Renshaw cells based on the dendritic trees and the path lengths – which is the distance from the soma to the den-drite terminal (Bui et al., 2003). Motor neurons have elaborate dendritic trees with complex branching patterns; the path length of 1a inhibitory interneu-rons is similar to those in motor neurons but the branching not as complex; and Renshaw cells have smaller and simpler dendritic trees. The average path length in motor neurons is 5 times greater than in 1a interneurons and 18 times that of Renshaw cells (Bui et al., 2003). The Renshaw cells are electrophysiologically homogeneous, and show hyperpolarization-activated cation current ( ), and small calcium-activated potassium currents ( ), (Perry et al., 2015).

Immunohistochemical identification of Renshaw cells Renshaw cells are identified by the combined expression of the calcium buffering protein, calbindin and the presence of large Gephyrin clusters on the soma and proximal dendrites (Alvarez et al., 1997; Carr et al., 1998; Geiman et al., 2000). Gephyrin clusters are scaffolds onto which receptors for Glycine and GABAA colocalize (Tyagarajan and Fritschy, 2014). Renshaw cells are the only group of spinal interneurons that show unusually large Gephyrin clusters on soma and proximal dendrites whereas most spinal interneurons and motor neurons show the opposite, i.e., large Gephyrin clus-ters on distal dendrites (Alvarez et al., 1997). Both soma size of motor neu-rons and the cluster sizes of Gephyrin increase during development (Geiman et al., 2000). The size of Gephyrin cluster on Renshaw cells correlates with the inhibitory synaptic currents and is regulated by excitatory inputs from motor neurons (Gonzalez-Forero et al., 2005).

Synaptic inputs and outputs of Renshaw cells Renshaw cells receive motor neuronal, descending and interneuronal (in-cluding other Renshaw cells) inputs and in turn synapse onto motor neurons and interneurons (Alvarez and Fyffe, 2007).

Excitatory synaptic inputs to Renshaw cells Individual adult Renshaw cells receive cholinergic inputs from multiple mo-tor neurons as revealed by vesicular acetylcholine transporter (VAChT)-

19

immunohistochemistry and most of these inputs are localized on the den-drites (Alvarez et al., 1999). Cholinergic inputs from motor neurons to Renshaw cells can be detected as early as E18 and remain functional across development into adulthood (Eccles et al., 1954; Mentis et al., 2006; Perry et al., 2015). Primary afferent inputs to Renshaw cells as revealed by vesicular glutamate transporter 1 (VGLUT1)-immunohistochemistry indicates there are no VGLUT1 synapses on Renshaw cells at E18, by P0 about 60%, and by P15 all Renshaw cells have received these synapses (Mentis et al., 2006). However, adult Renshaw cells do not respond to inputs from primary affer-ents, whereas neonatal Renshaw cells do so (Mentis et al., 2006). While there is a loss of functional primary afferent inputs there is no elimination of synapses on Renshaw cells. In adult Renshaw cells, the density of VAChT inputs is two to four times greater than the density of VGLUT1 (primary afferents) or VGLUT2 (interneuronal) inputs (Alvarez and Fyffe, 2007). In addition to monosynaptic excitatory inputs, Renshaw cells also show poly-synaptic activity in response to cutaneous and high threshold afferent inputs (Piercey and Goldfarb, 1974).

Inhibitory synaptic inputs to Renshaw cells Renshaw cells receive functional inhibitory synaptic inputs, which are GABAergic or Glycinergic as identified by immunohistochemical (Geiman et al., 2002) and electrophysiological studies (González-Forero and Alvarez, 2005; Nishimaru et al., 2010). It is assumed that the inhibitory control of Renshaw cells occurs in order to modulate the motor output (Hultborn and Pierrot-Deseilligny, 1979; Nishimaru et al., 2010). The evidence for this hypothesis is that during spontaneous activity, Renshaw cells receive inhibi-tory input that is ‘symmetrical’ with the excitatory input to synaptically con-nected or unconnected motor neurons belonging to the same segment (Nishimaru et al., 2010).

Renshaw cell inputs to other neurons Apart from their role in the recurrent inhibition circuit, Renshaw cells also synapse onto -motor neurons, other interneurons including 1a inhibitory interneurons and Renshaw cells. Of particular interest is the inputs to 1a inhibitory interneurons, which gives rise to recurrent facilitation by disinhibition of motor neurons by inhibition of tonically firing 1a inhibitory interneurons (Hultborn et al., 1971).

Role of Renshaw cells in motor control Activity in motor neurons has been shown to drive bursts of action potentials in Renshaw cells (Eccles et al., 1954; Renshaw, 1946; Ryall et al., 1971). There are divergent views on the role of Renshaw cells in motor control. Some are of the opinion that recurrent inhibition of motor neurons mediated

20 

by Renshaw cells is not strong enough to exert a strong influence on the motor neuron output (Fyffe, 1991; Lindsay and Binder, 1991). There have been other studies that have implicated a role for Renshaw cells in motor control (Eccles et al., 1954). Renshaw cells have been posited to preferen-tially inhibit smaller motor units compared to larger ones, in the order fast fatiguable < fast fatigue resistant < slow motor units (Friedman et al., 1981). Tonic motor neurons have been shown to have preferential distribution of recurrent inhibition (Granit et al., 1957). There is a linear relation between RIPSP amplitude and the duration of motor neuron activity after hyperpolar-ization (Eccles et al., 1961a). Further, recent data from paired recordings of connected Renshaw cell – motor neurons, have shown that single or trains of action potentials can transiently suppress motor neuron firing (Bhumbra et al., 2014). It still remains to be determined if Renshaw cells can modulate locomotor output.

Central pattern generators (CPG) Central pattern generators are neuronal networks that are capable of produc-ing rhythmic motor output in the absence of descending or sensory input (Marder and Bucher, 2001). The spinal cord and brainstem can produce the rhythms for swimming, walking, scratching, hopping, chewing and breathing (Bellingham, 1998; Morquette et al., 2012; Nakamura and Katakura, 1995; Nishimaru and Kudo, 2000; Smith et al., 2013; Whelan et al., 2000; Whelan, 2010). These networks can function upon suitable stimulation without senso-ry or descending inputs (Brown, 1911). Locomotor CPGs have been the subject of numerous investigations in the field of neuroscience and there is sustained effort to understand the components of the CPG. While the search for the CPG components continues, there have been great advances in terms of novel technologies. Advances in fluorescence microscopy techniques such as two-photon fluorescence microscopy have enabled fast imaging of activi-ty patterns in neurons (Hinckley et al., 2015; Hinckley and Pfaff, 2013). Advances in genetics have permitted targeting of neuronal populations (Gosgnach et al., 2006) and optogenetics have enabled acute light dependent activation or silencing of populations of neurons (Bouvier et al., 2015; Llewellyn et al., 2010). Cre-dependent monosynaptic tracing can label pre-motor interneuronal networks and is an additional example of tools used to delineate the components of the locomotor CPG (Callaway and Luo, 2015; Tripodi et al., 2011). While motor neurons integrate a wide range of inputs, there is no consensus on their role in locomotor CPGs. Rhythmic coordinat-ed output is believed to be generated at the premotor interneuronal level (Goulding, 2009), but an alternative view is that motor neurons together with the interneurons are necessary for rhythm generation (Song et al., 2016). Another feature that remains to be determined is whether CPGs are com-

21

prised of different modular units or whether there is reconfiguration of neu-rons in a task dependent manner (Goulding, 2009).

Walking mammals possess distributed and interconnected networks at the cervical and lumbar regions related to the forelimb and hindlimb movement (Grillner, 1975; Goulding, 2009). The hindlimb CPG located in the lumbar levels of the spinal cord can function without the cervical and rostral thorac-ic segments (Kjaerulff and Kiehn, 1996). In a series of elegant experiments using lesion of spinal cord in-vitro preparations, the anatomical distributions of CPG networks have been uncovered (Kjaerulff and Kiehn, 1996). The connections that are essential for left-right alternation are located in the ven-tral commissure, the CPG is primarily distributed in the ventral third of the spinal cord and the segments extending from Thoracic 12 - Lumbar 6 are sufficient to produce left-right and flexor-extensor alternation.

Does activation of CPGs always produce an alternating rhythm between flexor-extensors as well as left and right? Glutamatergic cells are found to be necessary and sufficient for the generation of fictive locomotion (Hagglund et al., 2013). Several rhythmogenic modules have been identified in the lum-bar cord and these can be activated by stimulating glutamatergic cells in mice (Hagglund et al., 2013). Depending on site of stimulation, rhythmic activity can be elicited only in the corresponding ventral root for instance L2 (or L5) (Hagglund et al., 2013). These studies show that there are distributed networks that are capable of generating rhythmic motor output. It is also known that antagonists of inhibitory neurotransmitters produce synchronized bursting between flexor-extensors as well as left-right sides (Cowley and Schmidt, 1995). Hence it remains to be determined how the glutamatergic and inhibitory systems synergize to produce alternating rhythmic motor out-put.

Genetics to define components of the locomotor CPG Expression of two morphogen gradients imparts dorsoventral positional in-formation in the ventricular zone of the embryonic spinal cord (Jessell, 2000; Goulding, 2009). Signaling molecules that are secreted by cells and act to organize a field of surrounding cells into patterns are called morphogens (Gurdon and Bourillot, 2001). On the ventral side the notochord and the floor plate secrete sonic hedgehog (Shh) (Jessell, 2000) and dorsally bone morphogenetic protein (bmp) is secreted from the epidermis overlying the neural tube and the roof plate (Lee and Jessell, 1999). This results in a gradi-ents of morphogens that result in the creation of progenitor domains from the ventricular zone (11 in chicken and mice) from which different classes of neuronal progenitor cells derive (Goulding, 2009; Arber, 2012). Of the 11 progenitor domains in mice 6 are dorsally derived and 5 are ventrally de-rived (Figure 4). The 5 dorsal progenitor domains (pd1-pd5, Figure 4) give rise to the spinal sensory circuits, and the ventral progenitor domains (pV0-pV3 and pMN, Figure 4) along with the intermediate progenitor domain

22 

(pd6) comprise the spinal locomotor circuitry. Neurons derived from the 5 ventral progenitor domains have been fairly well characterized. One class of ventrally projecting neurons that are derived from a dorsal progenitor do-main (pd6) are currently the substrate for many studies (Dyck et al, 2014; Andersson et al., 2013)

Figure 4. Cross-section through a hemisected mouse spinal cord during embryonic development (adapted from Goulding, 2009). The 11 progenitor domains and their neuronal progeny (pointed by arrows) are indicated.

Interneurons derived from progenitor domains and their roles in locomotion The selective expression of transcription factors during development drives differentiation of interneurons and provides a means for identification of different interneuron populations. Using advanced genetic techniques interneuronal populations have been visualized, manipulated or eliminated, which has resulted in the identification of several interneuronal modules that are part of the locomotor CPG.

V0 population The V0 population derives from the ventral pV0 progenitor domain in the embryonic spinal cord (Figure 4). They are a major source of commissural interneurons in the ventral spinal cord where the networks for locomotor central pattern generation are primarily located (Kjaerulff and Kiehn, 1996). The V0 interneurons are made of 2 distinct populations that have a dorso-ventral distribution. The dorsal population (V0D) make up 2/3s of the total V0 population and comprise inhibitory interneurons. The rest of the V0 pop-ulation is made up of the excitatory ventral V0V interneurons (Talpalar et al., 2013). Ablation of V0 interneurons in mice resulted in a synchronous gait of the forelimbs and hindlimbs at all locomotor frequencies (Tapalar et al., 2013). Selective ablation of the V0D population results in uncoordinated

23

rhythm at low frequencies, while for high frequencies L2 output was alter-nating. Selective ablation of the V0V neurons results in clear alternation at lower frequencies and synchronization at higher frequencies. Thus there seems to be a modular organization of V0 interneurons that are necessary for left-right alternation at different locomotor frequencies. Taken together, the V0 interneuron population is necessary for producing alternation at all locomotor frequencies, and does not affect normal flexor-extensor alterna-tion.

V1 interneurons The V1 population derives from the ventral pV1 progenitor domain in the embryonic spinal cord, and is ipsilaterally projecting (Figure 4). V1 neurons include Renshaw cells that mediate recurrent inhibition, as well as 1a inhibi-tory interneurons mediating reciprocal inhibition of motor neurons. They require the transcription factor Engrailed-1 for their differentiation. Knock-ing out/ablating/acute silencing the V1 population of spinal neurons results in slowing of locomotor rhythm while normal flexor-extensor and left-right alternation remain unaffected (Gosgnach et al., 2006). Thus the V1 spinal interneurons have a role in speed regulation.

V2 interneurons The V2 population derives from the ventral pV2 progenitor domain in the embryonic spinal cord (Figure 4), and is comprised of V2a and V2b subpop-ulations. The V2a interneurons are a ipsilaterally projecting excitatory popu-lation of interneurons that are characterized by the expression of the tran-scription factor Chx10 (Crone et al., 2008). Ablating V2a interneurons re-sults in impaired left-right coordination as seen during fictive locomotion whereas flexor-extensor alternation is not impaired (Crone et al., 2008). Ab-lating V2a interneurons results in normal alternation during treadmill loco-motion at low speeds, however this is replaced by synchrony of both fore- and hindlimbs at higher speeds (Crone et al., 2009). V2a neurons contact V0 as well as other commissural interneurons, hence they may provide the ex-citatory drive necessary for maintaining left-right alternation (Crone et al., 2008). Ablation of V2a interneurons disrupts the fictive locomotor output evoked by reticulospinal and sensory stimulation. Thus the V2a interneurons are necessary for neurally-evoked locomotor activity (Crone et al, 2008). Further intracellular recordings V2a neurons revealed the existence of 2 sub-populations: one that shows a speed dependent increase in recruitment of these neurons, and another group that does not participate in CPG activity (Zhong et al., 2011). Recently it has been shown that the V1 and V2b inter-neuron populations are together responsible for securing flexor-extensor alternation (Zhang et al., 2014).

24 

V3 interneurons The V3 population is an excitatory population of spinal interneurons that derives from the ventral pV3 progenitor domain in the embryonic spinal cord (Figure 4). V3 interneurons make most of their synapses contralaterally as well as a small number of ipsilateral synapses (Zhang et al., 2008). V3 inter-neurons make synapses onto motor neurons and other interneurons- includ-ing 1a inhibitory interneurons and Renshaw cells (Zhang et al., 2008). Selec-tively blocking synaptic transmission from the V3 neurons results in an al-tered locomotor rhythm, that shows a variation in the frequency of bursting. Acutely silencing these neurons in adult mice showed an effect on their locomotor pattern, showing variability in frequency of bursts as well as burst and inter burst periods (Zhang et al., 2008).

Glutamate as the main neurotransmitter for central pattern generation It has been shown that glutamate acting through both NMDA as well as non-NMDA receptors can generate locomotor rhythm (Talpalar and Kiehn, 2010). In lampreys, glutamate acting via NMDA and non-NMDA receptors drive distinct speeds of the locomotor rhythm (Brodin et al., 1985; Alford and Grillner, 1990). However such an organization does not exist in the mammalian locomotor system, and both NMDA and non-NMDA receptors can generate overlapping frequency ranges (Cazalets et al., 1992; Beato et al, 1997). Glutamate has further been shown to be necessary for the generation of different locomotor speeds, acting through inotropic glutamate receptors (Talpalar and Kiehn, 2010). Another study reveals that VGLUT2 is dispen-sable for locomotor rhythm generation (Mackenzie et al., 2006).

Effect of inhibitory neurotransmitters on central pattern generation Application of antagonists for glycine and GABA receptors during drug induced fictive locomotion produce synchronized rhythmic bursting between flexor-extensor and left-right sides (Cowley and Schmidt, 1995; Hinckley et al., 2005). Thus inhibitory neurotransmitters are dispensable for rhythm gen-eration but are necessary for alternation between the flexor-extensors and the left-right sides.

25

Aims of the projects

The overall aim of the thesis is to characterize the activity patterns in popula-tions of genetically identified interneurons as well as motor neurons in the spinal cord. Further to determine how the interneurons may contribute to generation of functional locomotor output.

Project I: To probe the functional connectivity from intersegmental sensory and motor roots to Renshaw cells at a developmental stage when they are known to integrate sensory and motor information.

Project II: Characterizing the activity patterns of Dmrt3 interneurons during fictive locomotion. Removal of inhibitory transmission results in defective locomotor rhythm, functional investigations were carried out on microcir-cuits of motor neurons to determine if there is any effect during fictive lo-comotion.

Project III: Rhythmic locomotor output is the sum of the individual motor neuronal activity. In this study we set out to characterize the activity patterns of individual motor neurons generating rhythmic locomotor output.

Project IV: Neuronal networks are distributed in three-dimensional (3D) space, and most cellular phenomena occur with sub-second resolution. Hence the project involved implementation of user-configurable fast 3D imaging.

26 

Materials and Methods

Mice All animal procedures were approved by and performed following the guide-lines of the appropriate local Swedish Ethical Committee (Ethical permit C248/11, C135/14, Uppsala Animal Ethics Committee, Jordbruksverket). In vitro spinal cord preparation

Mice were anesthetized with isofluorane (Baxter) and decapitated using sharp scissors. Incision was made on either side of the ventral thorax and abdomen with scissors, and the mice was then eviscerated. Further dissection of the spinal column was performed in a Sylgard (Dow Corning) coated dis-section chamber and the tissue was immobilized by using insect pins. The spinal column was then rinsed in ice-cold dissection buffer containing (in mM): 128 NaCl, 4.69 KCl, 25 , 1.18 , 3.5, 0.25 and 22 D-Glucose; equili-brated with 95% Oxygen and 5% Carbondioxide. A longitudinal ventral laminectomy was performed by cutting the bony spinal column using fine scissors . The spinal cord was then removed by disconnecting the nerve roots on both sides of the spinal cord and cutting the meninges. Dissected spinal cords were pinned down with the ventral side facing up and perfused with artificial cerebrospinal fluid (aCSF) (containing in mM: 128 NaCl, 4.69 KCl, 25 , 1.18 , 1.25, 2.5 and 22 D-Glucose; equilibrated with 95% Oxygen and 5% Carbondioxide). Flow-rate of aCSF was 6ml/min and the spinal cord was incubated at room temperature for an hour for good dye loading. Reagents and chemicals were, if not otherwise stated purchased from Sigma Aldrich.

Two-photon microscopy We built a custom designed two-photon laser scanning microscope (2PLSM), in collaboration with the Helmchen Group, University of Zurich, based on the 2PLSM described previously (Gobel et al., 2007). Briefly, a mode locked Ti:Sapphire laser system (COHERENT) was used to deliver laser pulses at 100-fs with a repetition rate of 80 MHz at 870 nm wave-length. A telescope was used to focus the beam size and the laser intensity was modulated by a Pockel’s cell (Conoptics). Two galvanometric scan mir-rors (model 6210; Cambridge Technology) were used to perform 2-D frame scanning (x-y scanning). In order to perform 3D imaging, a third piezoelec-

27

tric device (PIFOC) was used for scanning along the z-axis. The PIFOC is coupled to the objective and hence directly displacing the objective along the z-axis (Gobel et al, 2006). A 40X water-immersion objective with a numeri-cal aperture (NA) of 1 was used to acquire fluorescence signals (Plan Apochromat; 1 numerical aperture (NA); ZEISS). Emitted fluorescence sig-nals were passed through bandpass filters for the red (610/75 nm) and green (535/50 nm) channels respectively (AHF Analysentechnik). The signals were amplified through two photomultiplier tubes (PMT) (R6357; Hamama-tsu), sampled at 10 MHz and digitally integrated over pixel dwell times. Laser power at the focal spot never exceeded 50 mW. The microscope was controlled by the software HelioScan (Langer et al., 2013). Functional imag-es were acquired at at a rate of 6 Hz with 128 × 128; 256 x 256 pixel resolu-tion.

Electrophysiology Stimulations of ventral or dorsal roots were performed by applying glass-suction electrodes. Stimuli were generated using a constant-current stimula-tor (A365, World Precision Instruments, Sarasota, FL) and delivered to the chosen roots. Motor neuron activity was measured from lumbar ventral root by connecting the roots to suction electrodes. The signals were amplified 10,000 times and band pass filtered 100 Hz to 1 kHz (Model EX4-400, Da-gan Corporation). Signals were converted from analog-to-digital using Digidata 1440A AD-converter (Axon CNS, Molecular Devices, Sunnyvale, CA) and recorded with Clampex 10.2 (Molecular Devices). Drug induced fictive locomotion and change in frequency of rhythm

Spinal cords from neonatal mice (0-4 days old) were used to induce locomotor-like activity by bath application neuroactive drugs. Most experi-ments were performed at a single frequency of locomotor rhythm, elicited by bath application of N-methyl-d-aspartate (NMDA; 5 μM), dopamine (DA; 50 μM) and serotonin (5HT; 10 μM). Different frequencies of locomotor rhythm were generated by keeping serotonin and dopamine constant: 50 μM and 10 μM, and varying N-methyl-d-aspartate (NMDA; 1–10 μM) (Talpalar and Kiehn, 2010).

Conditioning of monosynaptic reflex response Stimulations of ventral or dorsal roots were performed by applying glass-suction electrodes. Stimuli were generated using a constant-current stimula-tor (A365, World Precision Instruments, Sarasota, FL) and delivered to the chosen roots. Motor neuron activity was measured from lumbar segment 1 ventral root (VRrL1) by connecting glass-suction electrodes. Experiments on conditioning of monosynaptic reflex response were adapted after Machacek

28 

et al., 2006. A test reflex response (TRR) is a monosynaptic reflex response and is elicited by dorsal root stimulation. In our experiments we chose the right lumbar segment 1 (rL1) for measurement of monosynaptic reflex re-sponse. The stimulus for eliciting a TRR is termed test stimulus (TS). Its amplitude was fixed at twice the threshold for evoking a monosynaptic re-flex response, typical values were in the range 10-50 μA. The monosynaptic reflex response has a latency onset of 6-8 ms. and it is quantified by measur-ing the amplitude in the following 3 ms (Machacek et al., 2007). The TRR output was measured at the corresponding ventral root (VRrL1). When a conditioning stimulus (CS) is applied to an adjacent root preceding the test stimulus, the resulting response measured at the VRrL1 is defined as condi-tioned reflex response (CRR). It has been observed that conditioning results in either suppression of monosynaptic reflex response. The CS consisted of a pulse train with an amplitude of 300 uA, a duration of 2 seconds, a frequen-cy of 20 Hz and a pulse width of 200 μs, ending 50 ms before the delivery of test stimulus (TS). The conditioning-test interval was 50 ms. The condition-ing stimulus and the conditioning-test interval were empirically chosen as they elicited the maximum suppression of monosynaptic reflex response. Each conditioning of monosynaptic reflex response experiment involved TRR and CRR measurements in alternation to avoid sequential effects. Each single experiment was preceded by a rest phase of at least two minutes. Am-plitude suppression (%) was defined as ((TRRCRR )/TRR )*100, where TRR is the amplitude of TRR and CRR is the amplitude of CRR. At least 6 repetitions (for both TRR and CRR) were performed for each experiment. The amplitudes of response was determined as the mean value of all the repetitions.

Calcium indicator loading Two complementary strategies were used to label neurons with calcium in-dicators for two-photon imaging: Bulk loading of dye Oregon green BAPTA

Activity was visualized using a calcium indicator dye Oregon Green BAPTA-1 AM (OGB-1 AM; Molecular Probes, Life technologies). OGB-1 AM was dissolved in DMSO/20% Pluronic F-127 and diluted in Ringer so-lution to a concentration of 1 mM (Johannssen and Helmchen, 2013). Multi-cell bolus loading of calcium indicator dye (Stosiek et al., 2003) Populations of neurons were loaded in intact, hemisected or dorsal horn removed prepa-rations. A glass micropipette filled with calcium indicator dye was posi-tioned close to the surface of the spinal cord at the desired region under vid-eo guidance and a 10X air objective (Plan Apochromat; 0.45 numerical aper-ture (NA); ZEISS). In two-photon imaging mode (40X objective), a positive pressure was applied to the dye and advanced into the tissue at a shallow

29

angle. The pipette tip was positioned in the tissue where dye loading is de-sired. The dye was pressure-ejected (150–700 mbar for several minutes) at an approximate depth, ranging from 60-140 μm from the surface of the cord. This was repeated at 2 or 3 more spots in the regions close to the first inject-ed spot. The spinal cord was incubated with the dye for one hour for good loading.

Retrograde labeling of motor neurons by crystals of Calcium Green The second technique involved retrograde labeling of motor neurons by ap-plying suitable tracers at cut ventral roots (Kosumacic et al., 2010). Crystals of Calcium Green-1-conjugated dextran amine (CGDA, 3000 MW, Molecu-lar Probes) were applied at the cut end of the lumbar segment 2 (L2) ventral root as previously described (Rabe et al., 2009). The L2 ventral roots are largely innervated by flexor related motor neurons and hence our studies are concerned with flexor burst generation. The L2 ventral root was cut first before the other roots in order to prevent leakage into neighboring roots. Further they were cut close to the spinal cord in order to minimize labelling time. The spinal cord was incubated in the dark at room temperature for 3 hours (Kosumacic et al., 2010).

Imaging

After incubation with Oregon green Bapta dye (OGB) or calcium green-1 conjugated dextran amine salts (CGDA), a spot is chosen for imaging. The spot is chosen such that there are a good number of neurons loaded with the calcium indicator. As an additional step, stimulations of segmental ipsilateral ventral or dorsal roots are performed and activity responses in neurons ob-served. This helps in determining that the loaded cells were functional and capable of producing activity. Each experiment comprised of 6-30 loaded neurons in one field of view. The region of interest (ROI) for each imaged neuron was manually outlined to extract only the somatic activity profiles. Only neurons with a clearly visible soma in the field of view were included for analysis.

Combined ventral root electrophysiology and imaging During root stimulations Experiments involving activity imaging of neurons in response to stimula-tions involved the following steps. The electrophysiological and imaging signals are acquired by different systems and need to be temporally aligned in order to extract meaningful information. Hence the shutter opening and closing signals from the imaging system were recorded along with the elec-

30 

trophysiological measurements to indicate the start- and endpoint of the im-aging session. The stimulation artifact was recorded to indicate the onset of stimulation during the imaging period. The stimulus consisted of a pulse train with an amplitude of 300 uA, a duration of 2 seconds, a frequency of 20 Hz and a pulse width of 200 μs. Quantification and normalization of activity responses

Calcium imaging was performed on neurons in response to dorsal, ventral or simultaneous dual root stimulations. Each imaging experiment comprised of stimulations of single or simultaneous dual roots in alternation with right lumbar segment 1 ventral root (VRrL1) stimulations. Calcium response in neurons was quantified by measuring the peak amplitude (PA). The differ-ence between the peak value of fluorescence intensities (peak) after the onset of stimulation and the mean of the baseline fluorescence intensities (base-line) pre-stimulation was defined as the peak amplitude. Peak amplitudes (PA) were considered significant and included in further analysis if it was greater than 2 standard deviations from the mean of baseline intensities. Ac-tivity response in neurons to different root stimulations were normalized to VRrL1. Normalized peak amplitude is defined as [( / )*100], where is the peak amplitude in response to single or simultaneous dual root stimulation and is the peak amplitude in response to VRrL1 stimulation. Peak amplitudes for individual cells in the field of view as well as the population were determined. The mean responses in a population of cells, imaged in a field of view, were de-termined and used as a representative population response. The mean popu-lation responses were also normalized to VRrL1. Imaging during fictive locomotion

After incubating for calcium indicator uptake, fictive locomotion is elicited by bath application of drugs. A spot is selected with a good number of load-ed neurons at the L2 segment for calcium imaging. The neurons were either OGB loaded interneurons or motor neurons that are retrogradely labeled with calcium indicator dye. We defined the baseline ( ) as the 10th percen-tile value of the fluorescent intensities for the respective neuron.

Imaging data analysis Time-lapse imaging data was acquired and saved as multi-page TIFF files. A mean projection image from the time lapse imaging data was generated us-ing the image processing program ImageJ. The mean projection image has sharper quality and aids in better identification of somata for demarcation of the region of interest. After background subtraction, the mean of fluores-cence intensities per ROI /frame was extracted for the entire time-lapse im-aging series. The fluorescence time-series were extracted for all the ROI’s in

31

the field of view. Analysis of imaging data was carried out using the pro-grams IGOR Pro (Wavemetrics Inc.) and Excel (Microsoft). Fluorescence time- series extracted from the neurons in the field of view were smoothened using box smoothing in IGOR Pro. The region of interest (ROI) for each imaged neuron were manually outlined to extract only the somatic activity profiles. Only neurons with clearly visible somata in the field of view were included for analysis. The activity response for each neuron was quantified as the percentage change in fluorescence (∆F) over the baseline fluorescence ( ), (∆F/ (%) = ((F – )/ )*100). Imaging experiments were per-formed and data acquired with the software HelioScan (Langer et al., 2013).

32 

Results and Discussions

Paper 1 Renshaw cell mediated recurrent inhibition of motor neurons has been of great interest in locomotor function. After their discovery about 7 decades ago many studies have been carried out to elucidate their functional signifi-cance. While a lot is known about their anatomical, functional, molecular characteristics and the neural systems they interact with, their overall contri-bution to locomotor function remains a challenge (Alvarez et al., 2007. This study extends the body of knowledge about the Renshaw cells, and deals particularly with the ability of Renshaw cells to integrate sensory and motor information early in development (Alvarez et al., 2007; Mentis et al., 2006). Renshaw cells can be genetically targeted through their selective expression of the α2 subunit of nicotinic acetylcholine receptor (Chrna2; Perry et al., 2015). The Chrna2:Cre mouse has provided a tool to carry out investigations on populations of genetically identified Renshaw cells. The Chrna2 positive cells have been verified to label Renshaw cells selectively in the ventral part of the spinal cord (Perry et al., 2015).

We used two-photon calcium imaging in order to visualize activity pat-terns in Renshaw cells in response to inputs from segmental and intersegmental sensory and motor roots. Two-photon calcium imaging ena-bles sampling of activity from populations of Renshaw cells with cellular resolution. It is known that at neonatal stages, Renshaw cells receive sensory inputs in addition to motor inputs (Mentis et al., 2006). We confirmed and extended these observations on Renshaw cell populations in spinal cords of P0 mice at lumbar segment 1. All the Renshaw cells simultaneously imaged show response to segmental sensory and motor inputs across experiments. Further, individual Renshaw cells show variability of responses and greater numbers showed larger response to segmental dorsal (75%) compared to ventral root stimulations.

Renshaw cell identification has been based on morphology, location and electrophysiological patterns of response, thus limiting studies to individual-ly identified Renshaw cells (Alvarez and Fyffe, 2007; Mentis et al., 2006). Identification of Renshaw cells based on genetic labeling of fluorescent pro-teins enables measurement of activity from populations of Renshaw cells (Perry et al., 2015). Thus population activity was investigated at lumbar segment 1 in response to intersegmental inputs. Five intersegmental dorsal and ventral roots were chosen located rostral and caudal to lumbar segment

33

1, of which two were ipsilateral (rT11 and rL5) and three were contralateral (lT11, lL1 and lL5). Analysis of intersegmental root stimulations showed that Renshaw cells were responding to all the roots tested to different extents. The mean population response amplitudes showed a trend depending on the root stimulated for both ventral and dorsal root stimulations.

The rostral root stimulations elicited significantly larger population re-sponse compared to caudal inputs, irrespective of dorsal or ventral root test-ed, whereas contralateral inputs did not differ significantly from ipsilateral inputs. The development of the nervous system proceeds along the rostrocaudal axis (Ensini et al., 1998; Jessell, 2000) and this could be reflect-ed in the rostrocaudal graded influence onto Renshaw cells. Interestingly, dorsal inputs were more efficient in producing a response in Renshaw cells compared to ventral root inputs. This may indicate a profound influence on the formation of functional circuitry involving Renshaw cells at these devel-opmental stages.

Intersegmental inputs are known to influence the functional motor output, this has been extensively characterized in the context of 1a afferent mediated monosynaptic reflex response (Machacek et al., 2007; Talpalar et al., 2011; Rudomin and Schmidt, 1999). This phenomenon is termed conditioning. We have shown that intersegmental dorsal and ventral root inputs elicit response in populations of Renshaw cells. Further we have shown that dorsal inputs seem to have a profound influence on the formation of functional circuitry involving Renshaw cells. Hence we wanted to investigate intersegmental sensory input mediated conditioning monosynaptic reflex response at lumbar segment 1. All intersegmental dorsal root inputs resulted in significant sup-pression of monosynaptic reflex response. Further, caudal inputs resulted in significantly greater suppression compared to rostral inputs.

Thus, intersegmental sensory inputs resulted in a rostrocaudal gradient of activation of Renshaw cells as well as conditioning mediated suppression of the monosynaptic reflex response, but the two are anti-correlated. While greater activation in Renshaw cells should mediate greater suppression of the monosynaptic reflex response, what caused the anticorrelation? It is known that early in development Gaba and Glycine mediate depolarization and that there is a developmental shift towards hyperpolarization.

There is variability of response in individual Renshaw cells that are im-aged at a spot. Further the same Renshaw cells show variability in response to different root stimulations indicating a heterogeneity in the inputs from different roots. Based on the number of cholinergic inputs it is estimated that on average, 75 motor neurons synapse onto a Renshaw cell (Moore et al., 2015). Our data begin to reveal that the inputs may not be limited to segmen-tal levels and that there could be considerable cross coupling between motor pools. We were not able to determine the extent of cross coupling, but used a limited paradigm of simultaneous inputs to Renshaw cells from dual roots in order to estimate this effect. The resulting responses in Renshaw cells due to

34 

simultaneous dual root inputs fell in 5 categories, which were either larger or smaller or similar to the response compared to individual root stimulations. Remarkably, the largest response was obtained by simultaneous stimulation of two dorsal roots, and it was greater than the sum of individual root stimu-lation responses. This is additional support for the importance of sensory inputs early in development. Thus, our study begins to shed light on the abil-ity of Renshaw cells to integrate sensory and motor information early in development.

Paper II The transcription factor doublesex and mab-3 related transcription factor 3 (Dmrt3) is necessary for normal differentiation of the dI6 population of spi-nal interneurons (Andersson et al., 2012). Mutations in the gene affects gat-ing patterns in horses and effects locomotor central pattern generator output (Andersson et al., 2012). This study aimed to characterize anatomical and physiological properties of Dmrt3 expressing neurons. Further the aim was to determine if there is an effect on locomotor output due to loss of inhibito-ry neurotransmission by conditional deletion of Viaat from the Dmrt3 posi-tive neurons.

The dI6 population of spinal interneurons, of which the Dmrt3 neurons are a subset, are derived from the progenitor domain p6 that is located in the dorsal half of the ventricular zone during embryonic development (Jessell, 2000; Goulding, 2009). The dI6 population is the only dorsally derived and ventrally migrating population of spinal interneurons found so far. They are also the least characterized for their role in locomotion as compared to other ventrally derived spinal neuronal populations. The ventrally located neurons in the spinal cord are necessary and sufficient for generation of locomotor rhythm that is alternating between the flexor-extensor and left-right sides (Kjaerulff and Kiehn, 1996).

A Cre mouse was generated that was bred with a tomato reporter mouse GT(ROSA)26Sortm14(CAG-tdTomato)Hze (tdT) to create a Dmrt3Cre;R26tdTomato (D3;tdT). Using immunohistochemistry it was determined that there was 99.6% co-localization between the DMRT3 immunolabeled cells and Dmrt3Cre;R26tdTomato neurons. There was 74% overlap between Dmrt3Cre;R26tdTomato neurons and DMRT3 immunolabeled cells. The Dmrt3Cre;tdT labeled the corticospinal tract and this tract appears around P7, hence there is no labeling of the tract embryonically. The Dmrt3Cre;R26tdTomato neurons settle in laminae VII and VIII, with a small percentage settling dorsal of the central canal. The Dmrt3 neurons are com-missural and send ascending, descending and bifurcating projections. It re-mains to be determined what the functional significance of these ascending and descending commissural projections are.

35

It is also known that these neurons send projections to ipsilateral and con-tralateral motor neurons (Andersson et al., 2012). Based on their involve-ment in locomotor rhythm generation, earlier studies on dI6 populations have reported the existence of two functional subpopulations. One subpopu-lation is implicated in rhythm while the other is implicated in pattern genera-tion (Dyck et al., 2012). In this study, both single cell electrophysiology as well as two-photon activity imaging on populations of Dmrt3Cre;R26tdTomato neurons have been carried out in the presence of fictive locomotor drugs. Accessibility to these neurons is a challenge in intact spinal cord prepara-tions because of the more medial location of these interneurons both along the dorso-ventral as well as the medio-lateral axis. Hence, to gain access to these neurons we first used a hemisected cord preparation and performed activity imaging of these neurons during fictive locomotion. We confirmed that Dmrt3Cre;R26tdTomato neurons are rhythmically active during fictive lo-comotion. In addition, and surprisingly we found that individual Dmrt3Cre;R26tdTomato neurons showed rhythmic oscillations that were differ-ent and distinct in relation to the stable and constant ipsilateral ventral root rhythm. Hemisecting the cord transects the axons of commissural neurons and Dmrt3Cre;R26tdTomato neurons being commissural, their physiology could be altered by hemisecting the cord. To further examine this possibility, we instead used a dorsal horn shaved preparation to image Dmrt3Cre;R26tdTomato neurons during fictive locomotion. This further corroborated the existence of different and distinct oscillation frequencies in individual Dmrt3Cre;R26tdTomato neurons. Further single cell electrophysiology also cor-roborated the rhythmic oscillations in these neurons in the presence of fictive locomotion inducing drugs in hemisected cord preparations.

In Dmrt3Cre;ViaatKO animals, inhibitory neurotransmission from Dmrt3Cre neurons is conditionally deleted. Behavioral analysis on Dmrt3Cre;ViaatKO mice revealed a disturbed CPG output that was particularly affected when frequency of locomotion increased. This was further corroborated by fictive locomotion experiments involving spinal cords form neonatal mice, confirm-ing that Dmrt3 neurons are required for normal CPG function in mice. This behavioral phenotype in the Dmrt3Cre;ViaatKO animals prompted us to ask what happens at the level of motor neuron microcircuits. We chose to look at the activity in motor neurons innervating the flexor related lumbar segment 2 during drug induced fictive locomotion. In order to visualize activity in mo-tor neurons we retrogradely applied the calcium indicator dye Calcium green. Dmrt3Cre;ViaatKO animals showed a clear effect on motor neuronal activity patterns in relation to littermate controls. Their activity was uncou-pled with the ventral root rhythm as well as between motor neurons. Thus, we uncovered that variable patterns of activity in motor neurons explained the disturbed ventral root output.

36 

Paper III Individual motor neurons imaged during fictive locomotion were found to oscillate with distinct frequencies. Some motor neurons showed slower while other showed faster frequencies in relation to the ipsilateral segmental ventral root rhythm (iVR). A small percentage (20%) of motor neurons also showed the same frequency as the ipsilateral ventral root rhythm. Such pat-tern of activity was found across experiments (20 of 21 experiments, involv-ing 248 motor neurons). The different frequencies of oscillation seen in an ensemble of motor neurons were independent of the motor neuron labeling technique or calcium indicator used (OGB or CGDA). Amplitude modulated outputs from the ventral roots have been obtained and are believed to be due to overlapping outputs from different pools of motor neurons (Tresch and Kiehn, 2000). Motor neurons in our experiments could belong to different pools that innervate the lumbar segment 2 ventral root (L2VR). Each imag-ing experiment was performed from a field of view (FOV) of 200um x 200um and in such a small region we have recorded up to 15 different oscil-lation frequencies. There is a definite organization of pools in longitudinal columns, with a dorsoventral and mediolateral arrangement along the spinal cord (Surmeli et al., 2011). We conclude that motor neurons showing multi-ple rhythmicities may not all belong to different pools, as the possibility of 15 pools existing in a small area 200um x 200um is unlikely. Motor neuron size impacts physiology as well as it forms a basis for classification between alpha and gamma motor neurons (Henneman et al., 1965; Friese, 2009; Zampieri, 2014). We found that there was no dependence of the motor neu-ron oscillation frequency on the size of the motor neuron somata. Hence, the different oscillation frequencies in individual motor neurons may arise across motor neuron subtypes.

After application of drugs there is a period of tonic baseline activity, that we term pre rhythm, and subsequently the rhythm stabilizes after 5-10 minutes. The baseline activity continues even after a rhythm is established. We have imaged activity in motor neurons during the pre rhythm period, as well as during a stable rhythm in the ventral roots. Whenever possible, we have followed the same set of motor neurons in a field of view through the pre rhythm and stable rhythm periods. Multiple rhythmicities were observed both during the pre rhythm as well as during the rhythm period of drug in-duced fictive locomotion. We find it likely that motor neurons with oscilla-tion frequencies outside the frequency of locomotor rhythm underlie baseline activity.

With increase in the speed of locomotor rhythm, we found a correspond-ing increase in the mean value of the motor neuron oscillation frequencies on a population level. But not all the underlying motor neurons show a corre-sponding increase in their oscillation frequencies, some show a decrease in oscillation frequencies. Almost all of the motor neurons imaged showed different oscillation frequencies with change in speed of the iVR rhythm.

37

Entirely different or slightly overlapping groups of motor neurons tuned to the rhythm with change in speed of the locomotor rhythm. There exists a topographical organization in the recruitment of motor neurons with increas-ing swimming speeds in zebrafish (Gabriel, 2011). From our data, we could not determine any particular order of motor neuron recruitment, during a rhythm as well as with change in speed of the rhythm, and it seems that this property is emerging with progressing development of the network. Motor neurons that tune-in to a ventral root rhythm have a certain degree of free-dom in following the ventral root rhythm, while some follow the rhythm robustly others may show variations with respect to the rhythm.

Paper IV Neuronal networks are distributed in the three-dimensional (3D) space, whereas most imaging experiments involve sampling activity from a two-dimensional (2D) plane. This has an inherent drawback: in order to sample activity from 3D space, several 2D images are acquired resulting in slowing down the entire process. Most cellular phenomena occur with sub-second resolution. Acquisition of stacks of 2D images to sample activity from neu-ronal networks in 3D space is infeasible. Hence, in order to improve the temporal resolution, activity from neurons is not sampled uniformly but along a predefined user configured linescan trajectory. The linescan trajecto-ry defines the path along which the laser scans the tissue, and emissions are only restricted to the scanned regions.

There are several advantages of the 3D linescan imaging to sample activi-ty from neuronal networks. It enables fast scanning through the tissue, thereby making it possible to resolve action potential driven calcium transi-ents. The use of linescans limit excitation and consequently reduce photobleaching and toxicity. This would also increase the ability to deter-mine the activity patterns of neurons and identify them by their location. Information on anatomy-physiology-location of populations of neurons can the more readily be studied in relation to functional output. This would ulti-mately help delineate the neural substrates for generation of functional out-put.

38 

Acknowledgements

Thank you Professor Klas Kullander for providing me the opportunity to work on two-photon calcium imaging of neuronal networks and supporting me through my studies. My co-supervisors Christiane Peuckert and Svante Winberg thank you for your support. Ted Ebendal thank you for accepting to be my PhD Examiner. Gilad Silberberg from the Department of Neuro-science, Karolinska Institute, I extend my warm regards to you for agreeing to be the faculty opponent. Hakan Aldskogius thank you for kindly accept-ing to be the Chair person of the thesis committee. Christian Broberger, Robert Fredriksson and Sebastian Barg, I am greatful to you for accepting to be in the thesis evaluation committee.

I joined the lab in 2009 as a master’s thesis student to work on implemen-tation of 3D imaging for two-photon microscopy. From the University of Zurich, I thank Dominik Langer for guidance with the implementation of 3D imaging, and Fritjof Helmchen for hosting me during my visit to Zur-ich. Christiane Peuckert for introducing me to and guidance with setting up of the two-photon microscope. Henrik Gezelius and Helge Johannssen thank you for very exciting and scintillating pilot experiments combining two-photon calcium imaging of Renshaw cells with ventral root electrophys-iology. I would like to extend BIG thanks to Nadine Rabe, Anders Enjin, Karin Nordenankar, Fatima Memic, Casey Smith, Kasia Rogoz, past students that created a great work environment.

I would like to extend my thanks to Cecilia E, Cecilia Y, Mariana, Emma A, Berit, Neil, Lena K from the administration. Your cheerful readi-ness whenever I have approached for help, queries or paperwork is very heartening and I am greatful to you.

Thanks to the groupleaders Klas Kullander, Malin Lagerström, Rich-ardson Leao and Åsa Mackenzie for creating a dynamic and collaborative work environment. Kasia Rogoz it was great working with you on the dorsal horn imaging project, your enthusiasm and drive was inimitable. Martina Blunder a big thank you for help with the mice! Also you were one of the most enthusiastic member of the Beer and Burgers club! Kali bhai you are an artist and there is a touch of art in everything you do! And together with the third musketeer you were the founding members of the Beer and Burg-ers club, fun officemates, partners during beach volleyball games. You two made my time in the lab a lot of fun! You are also an awesome cook and a great host together with Sulena bhabhi! I wish i could multitask the way you do Sharn Perry! Its been fun to work with you and I wish you the best!

39

Fabio Caxieta it has been fun to work with you! Group meetings are incom-plete without you :-), I wish you a great time in the lab! Richardson and Sanja an ideal guide-student duo, limitless supply of ideas and the ability to pursue science untiringly, you guys are awesome! Sanja no description of yours is complete without mention of Pavol, you partner for life and in sci-ence as well! And adding a new dimension to your life is Lea, your bundle of joy. I wish you all the best in everything! Fabio Batista you are very fo-cused and hardworking it was fun collaborating with you. I enjoyed your company during beach volleyball games, bbqs and beer! Samer you are the latest member of the Kullanderlab and managed to strike a chord with eve-ryone from day one. Your supervisors and coworkers were impressed with you, I wish you a great scientific career! Kia, Markus, Malte, Hermany, Stephano, Julia, Ernesto, Maanvee, Bejan, Kate, Amalia, Rik, Jorgen, Henrik B, Soren, Swati and Raja thank you for sharing scientific ideas and all the fun-filled times. Many thanks to Siv for all the help and support and ensuring smooth functioning of the lab, and to Brigitta for ensuring we nev-er ran out of clean labware.

For making my stay away from home pleasant by creating a home away from home. Harisha, we met for the first time in EBC in 2008 and you have remained as friendly as ever. What strikes me is your ability to think fast on your feet, innovative ideas, and I have enjoyed all our discussions scientific and otherwise. Now you are joined by Shruthi who complements you in all respects, I wish you two the best! Naveen you have been a great friend, sup-portive, patient, focused, these are just a few of your outstanding qualities! Together with Divya things are only getting better! Raghuveer also known as chinnappa :-), you are one of the most genuine persons I have come across, and I am glad to have a friend like you. You and Madhu have been great hosts and I wish you the best! Nikhil we are neighbours, in BMC just as in india! We started our innings in Uppsala by making chapati and cutting onions in Dobelnsgatan. I have enjoyed all the fun times, conversations, dinners and outings together with you and Lakshmi! I also thank you Lak-shmi for being a great host and the best cook Iv known! Amol and Priya thanks for all the fun times! Amol you are the only guy that gave me sleep-less nights! Because of the tasty spicy food that you fed us!! I also owe my gratitude for the care, affection and friendship of Shrikanth and Ann, Harsha and Shreya, Smitha and Nimesh, Shreedharan uncle and aunty, Navya and Sandeep, Narendra and Sujata bhabhi, Anirudha and Megha, Samarth, Rohit Menon, Sundar, Shridharan, Vikram, Vineeth, Srikanth, Vinay, Praghna and Swaroop, Jayanth and Vinutha, Rashmi and Sarosh, Devi prasad uncle and Mukta aunty, Suraj and Harshitha.

Thanks Reda and Inger for all the fun conversations over coffee and Inger for being my unofficial swedish teacher- du är en fantastisk lärare!

40 

I dedicate this PhD to you baba, mummy and Chandan for all the love, guidance, support and encouragement. Thank you Shwetha, you have added a whole new dimension to my life.

41

References

Ackman, J.B., Burbridge, T.J., and Crair, M.C. (2012). Retinal waves coordinate patterned activity throughout the developing visual system. Nature 490, 219-225.

Alvarez, F.J., Dewey, D.E., Harrington, D.A., and Fyffe, R.E.W. (1997). Cell-type specific organization of glycine receptor clusters in the mammalian spinal cord. The Journal of Comparative Neurology 379, 150-169.

Alvarez, F.J., Dewey, D.E., McMillin, P., and Fyffe, R.E.W. (1999). Distribution of cholinergic contacts on Renshaw cells in the rat spinal cord: a light microscopic study. The Journal of physiology 515, 787-797.

Alvarez, F.J., and Fyffe, R.E. (2007). The continuing case for the Renshaw cell. The Journal of physiology 584, 31-45.

Alvarez, F.J., Jonas, P.C., Sapir, T., Hartley, R., Berrocal, M.C., Geiman, E.J., Todd, A.J., and Goulding, M. (2005). Postnatal phenotype and localization of spinal cord V1 derived interneurons. The Journal of Comparative Neurology 493, 177-192.

Andersson, L.S., Larhammar, M., Memic, F., Wootz, H., Schwochow, D., Rubin, C.-J., Patra, K., Arnason, T., Wellbring, L., Hjalm, G., et al. (2012). Mutations in DMRT3 affect locomotion in horses and spinal circuit function in mice. Na-ture 488, 642-646.

Bellingham, M.C. (1998). DRIVING RESPIRATION: THE RESPIRATORY CENTRAL PATTERN GENERATOR. Clinical and Experimental Pharmacolo-gy and Physiology 25, 847-856.

Benito-Gonzalez, A., and Alvarez, F.J. (2012). Renshaw Cells and Ia Inhibitory Interneurons Are Generated at Different Times from p1 Progenitors and Differ-entiate Shortly after Exiting the Cell Cycle. The Journal of Neuroscience 32, 1156-1170.

Benninger, R.K.P., and Piston, D.W. (2013). Two-Photon Excitation Microscopy for the Study of Living Cells and Tissues. Current protocols in cell biology / edito-rial board, Juan S Bonifacino [et al] 0 4, Unit-4.1124.

Berridge, M.J., Lipp, P., and Bootman, M.D. (2000). The versatility and universality of calcium signalling. Nat Rev Mol Cell Biol 1, 11-21.

Bessou, P., Emonet-Dénand, F., and Laporte, Y. (1965). Motor fibres innervating extrafusal and intrafusal muscle fibres in the cat. The Journal of physiology 180, 649-672.

Bhumbra, G.S., Bannatyne, B.A., Watanabe, M., Todd, A.J., Maxwell, D.J., and Beato, M. (2014). The Recurrent Case for the Renshaw Cell. The Journal of Neuroscience 34, 12919-12932.

Blankenship, A.G., and Feller, M.B. (2010). Mechanisms underlying spontaneous patterned activity in developing neural circuits. Nat Rev Neurosci 11, 18-29.

42 

Bouvier, J., Caggiano, V., Leiras, R., Caldeira, V., Bellardita, C., Balueva, K., Fuchs, A., and Kiehn, O. (2015). Descending Command Neurons in the Brain-stem that Halt Locomotion. Cell 163, 1191-1203.

Brown, T.G. (1911). The Intrinsic Factors in the Act of Progression in the Mammal. Proceedings of the Royal Society of London Series B, Containing Papers of a Biological Character 84, 308-319.

Bui, T.V., Cushing, S., Dewey, D., Fyffe, R.E., and Rose, P.K. (2003). Comparison of the Morphological and Electrotonic Properties of Renshaw Cells, Ia Inhibito-ry Interneurons, and Motoneurons in the Cat. Journal of neurophysiology 90, 2900-2918.

Burke, R.E. (2001). Spinal Reflexes. In eLS (John Wiley & Sons, Ltd). Callaway, E.M., and Luo, L. (2015). Monosynaptic Circuit Tracing with Glycopro-

tein-Deleted Rabies Viruses. The Journal of Neuroscience 35, 8979-8985. Carp, J.S., and Wolpaw, J.R. (2001). Motor Neurons and Spinal Control of Move-

ment. In eLS (John Wiley & Sons, Ltd). Carr, P.A., Alvarez, F.J., Leman, E.A., and W. Fyffe, R.E. (1998). Calbindin D28k

expression in immunohistochemically identified Renshaw cells. NeuroReport 9. Cheng, H., Lederer, W.J., and Cannell, M.B. (1993). Calcium sparks: elementary

events underlying excitation-contraction coupling in heart muscle. Science 262, 740-744.

Cina, C., and Hochman, S. (2000). Diffuse distribution of sulforhodamine-labeled neurons during serotonin-evoked locomotion in the neonatal rat thoracolumbar spinal cord. The Journal of Comparative Neurology 423, 590-602.

Cowley, K.C., and Schmidt, B.J. (1994). A comparison of motor patterns induced by N-methyl-d-aspartate, acetylcholine and serotonin in the in vitro neonatal rat spinal cord. Neuroscience Letters 171, 147-150.

Crone, S.A., Quinlan, K.A., Zagoraiou, L., Droho, S., Restrepo, C.E., Lundfald, L., Endo, T., Setlak, J., Jessell, T.M., Kiehn, O., et al. (2008). Genetic Ablation of V2a Ipsilateral Interneurons Disrupts Left-Right Locomotor Coordination in Mammalian Spinal Cord. Neuron 60, 70-83.

Dai, X., Noga, B.R., Douglas, J.R., and Jordan, L.M. (2005). Localization of Spinal Neurons Activated During Locomotion Using the c-fos Immunohistochemical Method. Journal of neurophysiology 93, 3442-3452.

Denk, W., Strickler, J.H., and Webb, W.W. (1990). Two-photon laser scanning fluorescence microscopy. Science 248, 73-76.

Devanandan, M.S., Eccles, R.M., and Yokota, T. (1965). Muscle stretch and the presynaptic inhibition of the group Ia pathway to motoneurones. The Journal of physiology 179, 430-441.

Diaspro, A., Bianchini, P., Vicidomini, G., Faretta, M., Ramoino, P., and Usai, C. (2006). Multi-photon excitation microscopy. BioMedical Engineering OnLine 5, 1-14.

Donoghue, J.P., and Sanes, J.N. (2001). Motor System Organization. In eLS (John Wiley & Sons, Ltd).

Dougherty, K.J., and Kiehn, O. (2010). Firing and Cellular Properties of V2a Inter-neurons in the Rodent Spinal Cord. The Journal of Neuroscience 30, 24-37.

Eccles, J.C., Eccles, R.M., Iggo, A., and Lundberg, A. (1961a). Electrophysiological investigations on Renshaw cells. The Journal of physiology 159, 461-478.

43

Eccles, J.C., Eccles, R.M., and Magni, F. (1961b). Central inhibitory action attribut-able to presynaptic depolarization produced by muscle afferent volleys. The Journal of physiology 159, 147-166.

Eccles, J.C., Fatt, P., and Koketsu, K. (1954). Cholinergic and inhibitory synapses in a pathway from motor-axon collaterals to motoneurones. The Journal of physi-ology 126, 524-562.

Eccles, J.C., Schmidt, R., and Willis, W.D. (1963). Pharmacological studies on pre-synaptic inhibition. The Journal of physiology 168, 500-530.

Eccles, J.C., and Sherrington, C.S. (1930). Numbers and Contraction-Values of Individual Motor-Units Examined in some Muscles of the Limb. Proceedings of the Royal Society of London B: Biological Sciences 106, 326-357.

Friedman, W.A., Sypert, G.W., Munson, J.B., and Fleshman, J.W. (1981). Recurrent inhibition in type-identified motoneurons. Journal of neurophysiology 46, 1349-1359.

Friese, A., Kaltschmidt, J.A., Ladle, D.R., Sigrist, M., Jessell, T.M., and Arber, S. (2009). Gamma and alpha motor neurons distinguished by expression of tran-scription factor Err3. Proceedings of the National Academy of Sciences 106, 13588-13593.

Fyffe, R.E. (1991). Spatial distribution of recurrent inhibitory synapses on spinal motoneurons in the cat. Journal of neurophysiology 65, 1134-1149.

Gabriel, J.P., Mahmood, R., Walter, A.M., Kyriakatos, A., Hauptmann, G., Calabre-se, R.L., and El Manira, A. (2008). Locomotor Pattern in the Adult Zebrafish Spinal Cord In Vitro. Journal of neurophysiology 99, 37-48.

Garaschuk, O., Linn, J., Eilers, J., and Konnerth, A. (2000). Large-scale oscillatory calcium waves in the immature cortex. Nature neuroscience 3, 452-459.

Geiman, E.J., Knox, M.C., and Alvarez, F.J. (2000). Postnatal maturation of gephyrin/glycine receptor clusters on developing Renshaw cells. The Journal of Comparative Neurology 426, 130-142.

Geiman, E.J., Zheng, W., Fritschy, J.-M., and Alvarez, F.J. (2002). Glycine and GABAA receptor subunits on Renshaw cells: Relationship with presynaptic neurotransmitters and postsynaptic gephyrin clusters. The Journal of Compara-tive Neurology 444, 275-289.

Gobel, W., Kampa, B.M., and Helmchen, F. (2007). Imaging cellular network dy-namics in three dimensions using fast 3D laser scanning. Nat Meth 4, 73-79.

González-Forero, D., and Alvarez, F.J. (2005). Differential Postnatal Maturation of GABAA, Glycine Receptor, and Mixed Synaptic Currents in Renshaw Cells and Ventral Spinal Interneurons. The Journal of Neuroscience 25, 2010-2023.

Gonzalez-Forero, D., Pastor, A.M., Geiman, E.J., Benítez-Temiño, B., and Alvarez, F.J. (2005). Regulation of Gephyrin Cluster Size and Inhibitory Synaptic Cur-rents on Renshaw Cells by Motor Axon Excitatory Inputs. The Journal of Neu-roscience 25, 417-429.

Gosgnach, S., Lanuza, G.M., Butt, S.J.B., Saueressig, H., Zhang, Y., Velasquez, T., Riethmacher, D., Callaway, E.M., Kiehn, O., and Goulding, M. (2006). V1 spi-nal neurons regulate the speed of vertebrate locomotor outputs. Nature 440, 215-219.

Goulding, M. (2009). Circuits controlling vertebrate locomotion: moving in a new direction. Nat Rev Neurosci 10, 507-518.

44 

Granit, R., Pascoe, J.E., and Steg, G. (1957). The behaviour of tonic α and β motoneurones during stimulation of recurrent collaterals. The Journal of physi-ology 138, 381-400.

Grewe, B.F., Langer, D., Kasper, H., Kampa, B.M., and Helmchen, F. (2010). High-speed in vivo calcium imaging reveals neuronal network activity with near-millisecond precision. Nat Meth 7, 399-405.

Grienberger, C., and Konnerth, A. (2012). Imaging Calcium in Neurons. Neuron 73, 862-885.

Grillner, S. (2003). The motor infrastructure: from ion channels to neuronal net-works. Nat Rev Neurosci 4, 573-586.

Guertin, P.A. (2009). The mammalian central pattern generator for locomotion. Brain Research Reviews 62, 45-56.

Göppert-Mayer, M. (1931). Über Elementarakte mit zwei Quantensprüngen. Annalen der Physik 401, 273-294.

Hanson, M.G., and Landmesser, L.T. (2006). Increasing the Frequency of Spontane-ous Rhythmic Activity Disrupts Pool-Specific Axon Fasciculation and Pathfinding of Embryonic Spinal Motoneurons. The Journal of Neuroscience 26, 12769-12780.

Helmchen, F., and Denk, W. (2005). Deep tissue two-photon microscopy. Nat Meth 2, 932-940.

Hinckley, Christopher A., Alaynick, William A., Gallarda, Benjamin W., Hayashi, M., Hilde, Kathryn L., Driscoll, Shawn P., Dekker, Joseph D., Tucker, Ha-ley O., Sharpee, Tatyana O., and Pfaff, Samuel L. (2015). Spinal Locomotor Circuits Develop Using Hierarchical Rules Based on Motorneuron Position and Identity. Neuron 87, 1008-1021.

Hinckley, C.A., and Pfaff, S.L. (2013). Imaging spinal neuron ensembles active during locomotion with genetically-encoded calcium indicators. Annals of the New York Academy of Sciences 1279, 71-79.

Hultborn, H., Jankowska, E., Lindström, S., and Roberts, W. (1971). Neuronal pathway of the recurrent facilitation of motoneurones. The Journal of physiolo-gy 218, 495-514.

Hultborn, H., and Pierrot-Deseilligny, E. (1979). Input-output relations in the path-way of recurrent inhibition to motoneurones in the cat. The Journal of physiolo-gy 297, 267-287.

Hunt, C.C. (1951). The reflex activity of mammalian small-nerve fibres. The Journal of physiology 115, 456-469.

Hägglund, M., Dougherty, K.J., Borgius, L., Itohara, S., Iwasato, T., and Kiehn, O. (2013). Optogenetic dissection reveals multiple rhythmogenic modules underly-ing locomotion. Proceedings of the National Academy of Sciences 110, 11589-11594.

Jessell, T.M. (2000). Neuronal specification in the spinal cord: inductive signals and transcriptional codes. Nat Rev Genet 1, 20-29.

Kirkby, Lowry A., Sack, Georgeann S., Firl, A., and Feller, Marla B. (2013). A Role for Correlated Spontaneous Activity in the Assembly of Neural Circuits. Neuron 80, 1129-1144.

Kjaerulff, O., and Kiehn, O. (1996). Distribution of Networks Generating and Coor-dinating Locomotor Activity in the Neonatal Rat Spinal Cord In Vitro : A Le-sion Study. The Journal of Neuroscience 16, 5777-5794.

45

Kriellaars, D.J., Brownstone, R.M., Noga, B.R., and Jordan, L.M. (1994). Mechani-cal entrainment of fictive locomotion in the decerebrate cat. Journal of neuro-physiology 71, 2074-2086.

Kudo, N., and Yamada, T. (1987). Morphological and physiological studies of de-velopment of the monosynaptic reflex pathway in the rat lumbar spinal cord. The Journal of physiology 389, 441-459.

Kuffler, S.W., Hunt, C.C., and Quilliam, J.P. (1951). FUNCTION OF MEDULLATED SMALL-NERVE FIBERS IN MAMMALIAN VENTRAL ROOTS: EFFERENT MUSCLE SPINDLE INNERVATION. Journal of neuro-physiology 14, 29-54.

Lindsay, A.D., and Binder, M.D. (1991). Distribution of effective synaptic currents underlying recurrent inhibition in cat triceps surae motoneurons. Journal of neu-rophysiology 65, 168-177.

Llewellyn, M.E., Thompson, K.R., Deisseroth, K., and Delp, S.L. (2010). Orderly recruitment of motor units under optical control in vivo. Nat Med 16, 1161-1165.

Machacek, D.W., and Hochman, S. (2006). Noradrenaline Unmasks Novel Self-Reinforcing Motor Circuits within the Mammalian Spinal Cord. The Journal of Neuroscience 26, 5920-5928.

Machado, Timothy A., Pnevmatikakis, E., Paninski, L., Jessell, Thomas M., and Miri, A. (2015). Primacy of Flexor Locomotor Pattern Revealed by Ancestral Reversion of Motor Neuron Identity. Cell 162, 338-350.

Marder, E., and Bucher, D. (2001). Central pattern generators and the control of rhythmic movements. Current Biology 11, R986-R996.

Mears, S.C., and Frank, E. (1997). Formation of Specific Monosynaptic Connec-tions between Muscle Spindle Afferents and Motoneurons in the Mouse. The Journal of Neuroscience 17, 3128-3135.

Mentis, G.Z., Siembab, V.C., Zerda, R., O'Donovan, M.J., and Alvarez, F.J. (2006). Primary afferent synapses on developing and adult Renshaw cells. The Journal of neuroscience : the official journal of the Society for Neuroscience 26, 13297-13310.

Miyawaki, A., Llopis, J., Heim, R., McCaffery, J.M., Adams, J.A., Ikura, M., and Tsien, R.Y. (1997). Fluorescent indicators for Ca2+based on green fluorescent proteins and calmodulin. Nature 388, 882-887.

Morquette, P., Lavoie, R., Fhima, M.-D., Lamoureux, X., Verdier, D., and Kolta, A. (2012). Generation of the masticatory central pattern and its modulation by sen-sory feedback. Progress in Neurobiology 96, 340-355.

Nakamura, Y., and Katakura, N. (1995). Generation of masticatory rhythm in the brainstem. Neuroscience Research 23, 1-19.

Nishimaru, H., Koganezawa, T., Kakizaki, M., Ebihara, T., and Yanagawa, Y. (2010). Inhibitory Synaptic Modulation of Renshaw Cell Activity in the Lumbar Spinal Cord of Neonatal Mice. Journal of neurophysiology 103, 3437-3447.

Nishimaru, H., and Kudo, N. (2000). Formation of the central pattern generator for locomotion in the rat and mouse. Brain Research Bulletin 53, 661-669.

Noga, B.R., Kriellaars, D.J., Brownstone, R.M., and Jordan, L.M. (2003). Mecha-nism for Activation of Locomotor Centers in the Spinal Cord by Stimulation of the Mesencephalic Locomotor Region. Journal of neurophysiology 90, 1464-1478.

46 

Perry, S., Gezelius, H., Larhammar, M., Hilscher, M.M., Lamotte d'Incamps, B., Leao, K.E., and Kullander, K. (2015). Firing properties of Renshaw cells de-fined by Chrna2 are modulated by hyperpolarizing and small conductance ion currents Ih and ISK. European Journal of Neuroscience 41, 889-900.

Piercey, M.F., and Goldfarb, J. (1974). Discharge patterns of Renshaw cells evoked by volleys in ipsilateral cutaneous and high-threshold muscle afferents and their relationship to reflexes recorded in ventral roots. Journal of neurophysiology 37, 294-302.

Renshaw, B. (1941). INFLUENCE OF DISCHARGE OF MOTONEURONS UPON EXCITATION OF NEIGHBORING MOTONEURONS. Journal of neurophys-iology 4, 167-183.

Renshaw, B. (1946). CENTRAL EFFECTS OF CENTRIPETAL IMPULSES IN AXONS OF SPINAL VENTRAL ROOTS. Journal of neurophysiology 9, 191-204.

Romanes, G.J. (1951). The motor cell columns of the lumbo-sacral spinal cord of the cat. The Journal of Comparative Neurology 94, 313-363.

Rosenberg, S.S., and Spitzer, N.C. (2011). Calcium Signaling in Neuronal Devel-opment. Cold Spring Harbor Perspectives in Biology 3.

Rudomin, P., and Dutton, H. (1969). Effects of conditioning afferent volleys on variability of monosynaptic responses of extensor motoneurons. Journal of neu-rophysiology 32, 140-157.

Ryall, R.W., Piercey, M.F., and Polosa, C. (1971). Intersegmental and intrasegmental distribution of mutual inhibition of Renshaw cells. Journal of neurophysiology 34, 700-707.

Shneider, N., Brown, M., Smith, C., Pickel, J., and Alvarez, F. (2009). Gamma mo-tor neurons express distinct genetic markers at birth and require muscle spindle-derived GDNF for postnatal survival. Neural Development 4, 42.

Smith, J.C., Abdala, A.P.L., Borgmann, A., Rybak, I.A., and Paton, J.F.R. (2013). Brainstem respiratory networks: building blocks and microcircuits. Trends in Neurosciences 36, 152-162.

Song, J., Ampatzis, K., Björnfors, E.R., and El Manira, A. (2016). Motor neurons control locomotor circuit function retrogradely via gap junctions. Nature 529, 399-402.

Stifani, N. (2014). Motor neurons and the generation of spinal motor neurons diver-sity. Frontiers in Cellular Neuroscience 8.

Stosiek, C., Garaschuk, O., Holthoff, K., and Konnerth, A. (2003). In vivo two-photon calcium imaging of neuronal networks. Proceedings of the National Academy of Sciences 100, 7319-7324.

Svoboda, K., and Yasuda, R. (2006). Principles of Two-Photon Excitation Micros-copy and Its Applications to Neuroscience. Neuron 50, 823-839.

Sürmeli, G., Akay, T., Ippolito, Gregory C., Tucker, Philip W., and Jessell, Thom-as M. (2011). Patterns of Spinal Sensory-Motor Connectivity Prescribed by a Dorsoventral Positional Template. Cell 147, 653-665.

Talpalar, A.E., Bouvier, J., Borgius, L., Fortin, G., Pierani, A., and Kiehn, O. (2013). Dual-mode operation of neuronal networks involved in left-right alter-nation. Nature 500, 85-88.

Taylor, C.W. (1998). Inositol trisphosphate receptors: Ca2+-modulated intracellular Ca2+ channels. Biochimica et Biophysica Acta (BBA) - Molecular and Cell Bi-ology of Lipids 1436, 19-33.

47

Thomas, D., Lipp, P., Berridge, M.J., and Bootman, M.D. (1998). Hormone-evoked Elementary Ca2+ Signals Are Not Stereotypic, but Reflect Activation of Differ-ent Size Channel Clusters and Variable Recruitment of Channels within a Clus-ter. J Biol Chem 273, 27130-27136.

Tripodi, M., Stepien, A.E., and Arber, S. (2011). Motor antagonism exposed by spatial segregation and timing of neurogenesis. Nature 479, 61-66.

Tsien, R.W., and Tsien, R.Y. (1990). Calcium Channels, Stores, and Oscillations. Annual Review of Cell Biology 6, 715-760.

Tyagarajan, S.K., and Fritschy, J.-M. (2014). Gephyrin: a master regulator of neu-ronal function? Nat Rev Neurosci 15, 141-156.

Wallén-Mackenzie, Å., Gezelius, H., Thoby-Brisson, M., Nygård, A., Enjin, A., Fujiyama, F., Fortin, G., and Kullander, K. (2006). Vesicular Glutamate Trans-porter 2 Is Required for Central Respiratory Rhythm Generation But Not for Locomotor Central Pattern Generation. The Journal of Neuroscience 26, 12294-12307.

Wallén, P., and Williams, T.L. (1984). Fictive locomotion in the lamprey spinal cord in vitro compared with swimming in the intact and spinal animal. The Journal of physiology 347, 225-239.

Vanderhorst, V.G.J.M., and Holstege, G. (1997). Organization of lumbosacral motoneuronal cell groups innervating hindlimb, pelvic floor, and axial muscles in the cat. The Journal of Comparative Neurology 382, 46-76.

Whelan, P., Bonnot, A., and O'Donovan, M.J. (2000). Properties of Rhythmic Activ-ity Generated by the Isolated Spinal Cord of the Neonatal Mouse. Journal of neurophysiology 84, 2821-2833.

Whelan, P.J. (2010). Shining light into the black box of spinal locomotor networks. Philosophical Transactions of the Royal Society of London B: Biological Sci-ences 365, 2383-2395.

Zhong, G., Masino, M.A., and Harris-Warrick, R.M. (2007). Persistent Sodium Currents Participate in Fictive Locomotion Generation in Neonatal Mouse Spi-nal Cord. The Journal of Neuroscience 27, 4507-4518.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1190

Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofMedicine. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-280062

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2016