<<

Health Science Campus

FINAL APPROVAL OF MASTER THESIS Master of Science in Biomedical Sciences

Potential mismatches in structural and functional organization in the gracile

Submitted by: Shalini Niranjan

In partial fulfillment of the requirements for the degree of Master of Science in Biomedical Sciences

Examination Committee

Major Advisor: John Wall, Ph.D.

Academic Nicolas Chiaia, Ph.D. Advisory Committee: Richard D. Lane, Ph.D.

Senior Associate Dean College of Graduate Studies Michael S. Bisesi, Ph.D.

Date of Defense: July 25, 2008

Potential mismatches in structural and functional

organization in the gracile nucleus

Shalini Niranjan

Neurosciences and Neurological Disorders

University of Toledo, Health science campus

2008

Acknowledgements

I would like to thank Dr. John Wall, my major advisor, who has supported me all through my research and thesis, with his valuable knowledge and suggestions and by his unending support, while at the same time allowing me room to work in my own way. He has challenged and encouraged my thinking throughout the academic program which has immensely helped me in writing this thesis. My master’s degree and this thesis, too, would not have been completed if not for his patience and continued guidance.

I would also like to thank Dr. Richard D. Lane and Dr. Nicolas L. Chiaia who have guided me through the dissertation process, giving me many invaluable suggestions, and serving on my graduate committee.

I would like to thank Xin Wang for his friendship. He has helped me a lot both with his suggestions and his technical expertise.

Finally I would like to thank all my friends and family, especially my mom and dad for cheering me through the whole program and my husband for his support and my two baby girls for their patience and giving me time to work on my thesis.

ii

TABLE OF CONTENTS

ACKNOWLEDGMENTS………………………………………………………...……..ii

TABLE OF CONTENTS……………………………………………………………….iii

INTRODUCTION…………………………………………………………………….....1

Background for thesis……………………………………………………….…….1

Focus of thesis ………………………………………………………………….....5

LITERATURE REVIEW…………………………………………………………….…7

How the GN contributes to the ……………………….…..7

Features of GN structural and functional organization………………………..…..8

A. A main input to the GN is from primary sensory that

process tactile inputs from the skin……..…………………………….8

B. What is know about the spatial organization of ascending

terminations in the GN…………………………………………………9

a) GN organization of dorsal column inputs from Golgi

material ...……………………………………….……..9

b) GN organization of primary neurons terminations

from labeling of inputs from dorsal root ganglia or nerve

locations ………………………….…………………10

c) GN organization of primary neurons terminations

from labeling of inputs from specific body

locations………………...…………………………….12

C. GN neurons …………...…………………………………………..…13 iii

D. GN connectivity ...…………………………………………………...14

E. GN functional RFs……………………………………………………16

Literature Summary……………………………………………….………….....18

DESIGN………………………………………………………………………………. .20

MANUSCRIPT……………………………………………………… …………...…. .23

DISCUSSION ………………………………………………...………………………...63

Present results……………………………………………………………………63

How the present results supplement understanding of primary

input organization in the GN…...………………………...………………66

How the present results supplement current understanding of GN

functional RFs on the toes……………………………...………………...70

Some broader implications of the findings of this thesis………………………...72

A. Possible mismatches in GN structure and function and

a potential explanation………………………………………………..72

B. Some features of cortical and GN processing may be similar………..74

C. Potential contribution to plasticity of GN RFs……………………...... 75

BIBLIOGRAPHY………………………………………………………………………77

ABSTRACT……………………………………………………………………………..91

iv

Introduction

Background for thesis

The somatosensory system conveys tactile and pain sensation from our body. Transmission of touch information from the body is initiated by peripheral primary sensory neurons that project from endings near the body surface to the central nervous system, including a large ascending central projection to the dorsal column nuclei (DCN), within the (Willis & Coggeshall, 2004; Tracey, 2004).

Peripheral sensory axons from the toes on the foot, for example, enter the and ascend to the brainstem where they terminate and make excitatory synaptic connections on postsynaptic neurons in the gracile nucleus (GN) subdivision of the DCN. GN neurons, in turn, send ascending projection axons to the where connections through a second set of excitatory synapses is made onto postsynaptic thalamic neurons.

Thalamocortical axons of thalamic neurons project to primary somatosensory cortex where subsequent synapses are made onto cortical neurons. In this way, tactile stimulation of a toe on the foot leads to excitation of GN and thalamic neurons which, in turn, initiate cortical activation that contributes to feelings of touch.

Individual somatosensory neurons commonly have an excitatory tactile receptive field (RF). This is an area of skin within which tactile stimuli must be located to cause excitation of the . Thus, individual GN, thalamic, and cortical somatosensory neurons usually have a delimited RF that encompasses the body location where touch can activate that neuron. It is generally thought that our ability to localize touch stimuli to specific body parts like a toe results, in part, because stimulation of RFs 1

on different parts of the body activates restricted clusters of central neurons.

Understanding how ascending projections and related postsynaptic neurons are

structurally organized to process information from different locations on the body is an

important part of understanding normal somatosensory functional organization.

The detailed synaptic arrangements of ascending axonal inputs onto neurons in the GN, thalamus, and cortex remain only partly understood. This is due to uncertainties about the diverse arrangements of connections that are possible. For example, at each level it is possible that terminations from only a few or, alternatively, a large number of ascending axonal projections may converge onto an individual postsynaptic neuron. It is also possible that an individual projection axon may send

divergent terminations to only a few or to a large number of postsynaptic neurons.

Moreover, at each synaptic level the excitatory connections of ascending axonal

projections onto postsynaptic neurons are supplemented by further connections of

inhibitory neurons which inhibit excitatory convergent and divergent connections of input

projection axons. Given these and other connection possibilities, it is difficult to know

how structural convergence of excitatory input terminations versus inhibitory connections

and other factors shape the functional RF properties of a postsynaptic neuron at any

particular time.

Some insight into this issue has come from studies of RF plasticity in the

adult somatosensory cortex. Cortical neurons that normally have RFs on a particular part

of the body have been studied after peripheral inputs from their RFs were lost as a result

of a range of manipulations including injury of nerves, loss of a part of the body (and its

2

nerves) due to amputation or local anesthetic block of inputs. The results show that

cortical neurons that lose their normal inputs and RFs often acquire new RFs in other parts of the body that remain normally innervated. These findings show that cortical neurons can ‘reorganize’ by becoming responsive to new inputs and to shifted RFs on new locations on the body. Cortical RF reorganization has been seen in adult animals

and humans and, thus appears to reflect an important concept of organization. (Jones,

2000; Florence, 2002; Kaas, 2002; Calford, 2002; Wall et al., 2002).

The above RF reorganization of cortical neurons indicates adjustments can

occur in adult ascending connections at some level(s) of the system between the body and cortex. Cortical studies initially focused on RF reorganization after chronic input disruption because it was thought that once developmental changes were complete, adult

somatosensory connections were highly resistant to modification and would require long

times to change. However, much subsequent work has shown that initial cortical

reorganization can occur quite rapidly (Calford, 2002; Wall et al., 2002). For example,

after injury of nerves to the hands of adult monkeys, cortical RFs and maps of noninjured

tactile inputs from the hand undergo changes in organization within minutes to hours.

Similarly, after disruption of peripheral inputs by cutaneous injections of local anesthetics

like lidocaine, RFs of cortical neurons rapidly enlarge within minutes (Byrne and

Calford, 1991; Calford and Tweedale, 1991; Faggin et al., 1997; Krupa et al., 1999;

Nicolelis et al., 1993; Panetsos et al., 1995; Shin et al., 1995).

3

The rapidity of these initial cortical RF changes suggested a broader range of inputs was available to activate a cortical neuron than was normally seen at a particular time, and that some of these available inputs were normally masked (Florence, 2002;

Calford, 2002; Wall et al., 2002). Loss of peripheral inputs, then, led to conditions that rapidly unmasked available inputs.

Two further lines of evidence suggested that masked inputs were at least partly derived from highly convergent/divergent thalamocortical axon projections onto cortical neurons. First, rapid cortical RF reorganization, like that seen within the initial minutes after disruptions of peripheral inputs can also be produced if local cortical inhibition was disrupted by blocking inhibitory effects of cortical interneuron connections with GABA receptor antagonists (Alloway & Burton, 1991;Dykes et al., 1984;

Chowdhury & Rasmusson, 2002). These findings suggested that local cortical inhibition normally decreases RF size by tonically inhibiting functional expression of some subset of structurally existent and broadly convergent/divergent thalamocortical inputs. Second, structural studies, which traced thalamocortical axon arbors as they traveled to synapse on cortical neurons, indicated that individual thalamocortical axons can project across large distances of cortex, and that these distances were broader than cortical distances that contained neurons with RFs on a single body part (Garraghty et al., 1989; Arnold et al.,2001; Rausell & Jones, 1995; Rausell et al. 1998; Garraghty et al. 1990; Churchill et al. 2004; for reviews see Snow and Wilson, 1991). This suggested that thalamocortical axons provide inputs to cortical neurons that have RFs that normally do not reflect all available inputs. 4

Focus of thesis

The above cortical findings point to a simple but important concept about

structural organization of ascending presynaptic inputs and associated RF locations of postsynaptic neurons: i.e., the normal RF location of a neuron at a particular time reflects only a part of a larger range of locations that is possible from available overlapping convergent (and divergent) input axons onto that neuron. This concept has important implications for understanding how postsynaptic functional organization arises from presynaptic structural organization. For example, structurally broad sets of overlapping inputs that are only partially expressed at a particular time could give postsynaptic neurons greater functional adaptability than organization based on narrow inputs.

Cortex is commonly thought to be the most complex part of the nervous system, and adaptation and modifiability of connections is an important intrinsic cortical property. It is possible that this concept may apply to cortex but be less pertinent in other central structures where organization may be less complex or based on different concepts.

For example, presynaptic input organization in subcortical circuits may reflect concepts that support accurate information transmission rather than adaptability, and may be more narrowly defined than presynaptic cortical organization.

Working from the above perspectives, the question addressed in this thesis is: does the concept that RF location reflects only part of an available overlapping convergence (and divergence) of ascending input axons also apply to connections at an 5 earlier level of central touch pathways, and more specifically, at the brainstem level in the GN? To examine this question, this thesis tested whether GN terminations of peripheral primary sensory projections from different locations on the body of adult rats overlap and are organized in convergent fashion or, alternatively, whether these terminations are more discretely organized in nonoverlapping, nonconverging ways. To do this, I studied GN terminations of primary neuron inputs from hindpaw toes, first, because each toe is physically distinct and offers a body location that can be distinguished from other body locations and, second, because high densities of primary neuron inputs originate from the toes. Locations on one or more hindpaw toes were injected with horseradish peroxidase conjugates to permit labeling and reconstruction of

GN terminations of sensory axons from specific toes. Degrees of overlap of input terminations from different toes were assessed from these reconstructions. In addition, patterns of GN terminations from different toes were compared to GN toe RFs. This permitted examination of whether overlap of structural inputs from different toes matched or was different than the numbers of toes that are usually included in normal functional RFs. This approach could also complement other recent studies, including inhibition block studies, which deal with GN RFs but not from the present perspective of

RF relationships to structural inputs (see discussion).

The following Literature Review outlines salient properties of organization in the rat GN that served as a base of understanding for the present GN work.

6

Literature Review

How the GN contributes to the somatosensory system

The GN is a major brainstem processing nucleus in the somatosensory

system that, together with the cuneate nucleus (CN), makes up the dorsal column nuclei

(DCN) (Willis and Coggeshall, 2004; Tracey, 2004). The GN nucleus is located dorsally

in the medulla, is close to the midline and medial to the adjacent CN, and is elongated in

the rostrocaudal direction. In rats, it occupies about 39% of the volume of the DCN and

does not appear to have clear cytoarchitectonic subdivisions (Bermejo et al., 2003).

The DCN form the brainstem processing component of the dorsal column-

medial lemniscal system, a large ascending projection system from the body to cortex

(Tracey, 2004). This system processes touch and information, and the GN subdivision of the DCN is recognized as the main processing complex for touch and proprioception signals from skin areas on the foot, hindlimb, and lower body (Kaas,

2004). There is further evidence that the GN also contributes to processing of pain information (Willis & Westlund, 2004).

It is of interest to know whether concepts of processing that apply at one level of this system, e.g., cortex, apply at other levels, e.g. GN. Present understanding of major structural and functional features of the GN that are pertinent to the question that is addressed in this thesis is outlined next. This comes from studies of the GN as well as the adjacent CN.

7

Features of GN structural and functional organization

A. A main input to the GN is from primary sensory neurons

that process tactile inputs from the skin

Neurons in the GN receive a major part of their inputs from primary

sensory neurons. The distal axons of these primary sensory neurons end in the body and,

via the actions of receptors and voltage-gated ion channels, transduce mechanical and

other stimuli into receptor potentials which, in turn, are converted to action potentials that

are transmitted up the primary neuron’s distal and central axons. The central axons of

primary sensory neurons to the GN enter the spinal cord at thoracic, lumbar, or sacral

levels and ascend in the ipsilateral gracile fascicle of the spinal dorsal column to

terminate and make excitatory synapses on ipsilateral GN neurons (Willis & Coggeshall,

2004).

Each primary has a RF area or volume on the body within

which stimuli must fall to activate that primary neuron. For primary neuron inputs to the

GN, these RFs are commonly on lower body skin locations including, for example,

hindpaw toes. Thus, tactile stimuli activate primary neurons with RFs on a stimulated

skin area which, in turn, activate GN neurons on which activated primary neurons

terminate and synapse. As a consequence, most GN neurons have excitatory tactile RFs

on the skin (Nord, 1967; Berkley & Hubscher, 1995; Panetsos, 1995; Kondo et al., 2002;

Costa-Garcia & Nunez, 2004; Malmierca & Nuñez, 2004). Thus, by nature of its main

8

structural input connections, a major function of the GN is to process primary neuron

inputs from toe and other lower body skin areas.

B. What is known about the spatial organization of ascending

axon terminations in the GN?

A main focus of the present thesis is on how primary neuron terminals

from individual hindpaw toes are distributed in the GN. Of particular interest is whether

ascending input terminations from individual hindpaw toes overlap or, alternatively, occupy nonoverlapping GN locations.

As discussed next, present information about the organization of ascending input terminations in the GN come from: a) Golgi staining of dorsal column inputs, b) labeling of primary neuron inputs from or peripheral nerve locations, and c) labeling of primary neuron inputs from specific skin locations.

a) GN organization of dorsal column inputs from Golgi material

Early studies suggested that Golgi-stained dorsal column axons that

entered the GN had extensively overlapping terminations (Valverde, 1966; Odutola,

1977). Terminal arborizations of individual and small groups of dorsal column axons

formed intertwining glomeruli which surrounded dendrites of GN neurons. Adjacent

glomeruli appeared to form a continuous and relatively uniform plexus of terminations.

Axons that enter the GN from the dorsal column can potentially originate from the

ascending axons of primary sensory neurons from lower body locations or, alternatively,

9

from ascending axons that originate from central spinal neurons. Thus, early Golgi

studies suggested that GN terminations of ascending dorsal column axons contribute to

glomeruli that appear to be part of a continuous input termination plexus. However, these

studies do not distinguish terminations of axons that are from primary neurons versus

spinal neurons, and do not indicate how primary neuron terminations from specific

locations on the body, like the toes, are organized in the GN.

b) GN organization of primary neuron terminations from labeling of

inputs from dorsal root ganglion or nerve locations

With development of axonal transport methods for labeling primary

sensory neuron axons and their central terminations, a large number of studies focused on

organization of GN terminations of primary neuron inputs that were labeled at dorsal root ganglion or peripheral nerve locations (Leong & Tan, 1987; Ding et al., 1999; Hwang et al., 2001; Novikov, 2001; Robertson & Grant, 1985; LaMotte et al., 1991; Ueyama et al.,

1994; Tong et al., 1999; Shehab et al., 2003). In these studies, horseradish peroxidase conjugates were injected into dorsal root ganglia, which contain cell bodies of primary neurons, or into peripheral nerves, which contain distal axons of primary neurons. The label was taken up, transported, and deposited in the central terminations of primary neurons that originated in the labeled structure. The results show that, regardless of the injection site, virtually all labeled terminations were in the GN ipsilateral to injections.

GN terminations of inputs from different dorsal root ganglia, including lumbar ganglia

3-5, which contain cell bodies of primary neurons from the hindlimb, hindpaw, and toes,

10

were organized as rostrocaudally continuous columns; in addition, columns of terminations from different ganglia overlapped to substantial extents (Rivero-Melian and

Arvidsson, 1992; Prats-Galino et al., 1999; Puigdellivol-Sanchez et al., 2000). Similarly,

GN terminations of inputs from different peripheral nerves, including the sciatic and saphenous nerves, which contain primary neuron axons from the toes, hindpaw, and hindlimb, were also organized as rostrocaudally continuous columns that overlapped to large extents (LaMotte et al., 1991; Ueyama et al., 1994).

These studies clearly advanced understanding beyond the findings of the early Golgi studies by selectively identifying GN terminations that were associated with ascending primary neuron inputs. In addition, they identified GN termination locations of primary neurons associated with specific dorsal root ganglia and peripheral nerves, and established that terminations from different ganglia or nerves overlapped. From the perspective of the present thesis, however, these studies are limited by the fact that ganglia and nerves contain primary neurons with peripheral endings originating in broad body locations; thus, these studies do not have the spatial resolution to determine whether

GN terminations from a delineated body location, like an individual toe, overlap or are separate from terminations from a second delineated body location, like another toe.

11

c) GN organization of primary neuron terminations from labeling of

inputs from specific body locations

In contrast to the above extensive work involving labeling of primary

neuron inputs from ganglia and nerves, there has been little work on GN termination patterns of primary neuron inputs from specific body locations (Maslany et al., 1991;

Maslany et al., 1992). In this work, HRP conjugates were cutaneously injected into toe or hindpaw locations. Terminations from toes were in dorsal locations of the ipsilateral

GN that appeared separated from locations occupied by hindpaw and hindlimb terminations. The results further suggested that toe inputs appeared to be organized in a complex pattern (Maslany et al.,1991) . Specifically, it was shown that rostro-caudal columns of terminations from toes 1 and 4 appeared to reversibly and repeatedly interchange in their relative medial and lateral positions at different rostrocaudal GN locations. Whether medial and lateral positioning of other toe termination columns similarly changed with respect to each other, or whether there was interchanging of termination columns of toes 1 and 4 versus columns of other toes, remained unclear. It also remained unclear whether interchanging termination locations of toes 1 and 4 involved overlapping of terminations from these or other toes. There have been no further studies of GN termination patterns of toe inputs, or inputs from other specific body locations, since this early report. Thus, there presently is no picture of how central terminations of primary neuron inputs from different hindpaw toes are organized in the rat GN and particularly, whether termination areas of different toes overlap or are distinct.

12

C. GN neurons

The DCN, including the GN, contain two type of neurons, projection

neurons and interneurons (Snow and Wilson, 1991; Tracey, 2004) ). These neurons

have the following properties.

Projection neurons have long main axons that project signals from the GN

to other central structures. The main axons of projection neurons commonly also branch within the GN to give rise to collateral axons which terminate and synapse within the

GN. Projection neurons are glutamatergic and the main and collateral axons provide excitatory inputs to neurons they synapse on. The cell bodies of projection neurons can have large dendritic trees that can extend up to a few hundred microns from the cell body

(Kemplay & Webster, 1989; Weinberg et al., 1990; Lue et al., 2000; Bermejo et al.,

2003). Projection neurons to thalamic levels of the appear to make up a large percentage (up to 80%) of DCN projection neurons in rats (Bermejo et al., 2003).

In contrast to projection neurons, interneurons have short axons which terminate locally within the GN. DCN interneurons, including GN interneurons, all appear to be inhibitory. Immunohistochemical analyses indicate interneurons can be

GABAergic, glycinergic, or combined GABA- and glycinergic (Lue et al., 1993; Lue et al., 1996; Lue et al., 2000; Popratiloff et al., 1996a). Interneuronal cell bodies and dendrites are generally smaller than cell bodies and dendritic trees of projection neurons.

Interneurons are believed to make up a smaller percentage (up to about 10-25%) of DCN neurons than projection neurons (Bermejo et al., 2003).

13

D. GN connectivity

Previous work has provided a partial understanding of the synaptic

connections of primary neurons, projection neurons, and interneurons (Willis &

Coggeshall, 2004). This connectivity includes the following main features (Fig.1).

Primary neuron inputs to the GN largely come from large diameter, myelinated afferents that transmit touch and proprioception signals. A further small fraction of primary neuron inputs appears to come from small diameter afferents.

Primary neuron inputs have glutaminergic excitatory synapses on GN projection neurons and interneurons. Inhibitory interneurons synapse on other inhibitory interneurons,

projection neurons, and terminations of primary neurons to mediate postsynaptic and

presynaptic inhibition (Fig.1). Projection neurons receive inputs from primary neurons

and interneurons and, via their axon collaterals, activate interneurons and projection

neurons. Primary neuron inputs comprise a major fraction, i.e. about 15-25%, of DCN

inputs. Other smaller input contributions come from central structures including cortex

(Chimelli et al., 1994; Desbois et al., 1999), spinal cord, (Cliffer and Willis 1994; De

Biasi et al. 1995; French et al. 2002), cerebellum (Asanuma et al., 1983), and other brainstem nuclei ( Weinberg & Rustioni, 1990). From this connectivity, activation of primary neuron inputs from the body can produce a number of effects on GN neurons including: (a) direct excitation of GN interneurons and projection neurons, (b) indirect

inhibition of interneurons, projection neurons, and primary neuron axon terminations through connections on inhibitory interneurons, and (c) more complex GN disinhibition

14

through serial inhibitory interneuron connections (Fig.1). These effects of primary

neuron activation are, in turn, presumably integrated with effects of other inputs.

Figure 1. Primary sensory neuron inputs to GN neurons and GN synaptic connectivity

Primary sensory neurons (S) from the body have excitatory synapses on GN projection neurons (P) and on inhibitory interneurons (I). P axon collaterals have excitatory synapses on I and P (not shown) neurons. I neurons have inhibitory synapses on I and P neurons postsynaptically, and on S neuron terminations presynaptically. GN synaptic activity is dependent on inputs from S neurons and from inputs from central structures. Activation of S neurons can: (a) directly activate I and P neurons, (b) indirectly inhibit I and P neurons and S neuron terminations via I neuron activation, and (c) mediate more complex disinhibition through serial I neuron connections.

15

With regard to output connections, GN and DCN projection neurons

transmit information to a number of central nuclei. A major projection ascends from the

GN to thalamic nuclei, particularly the ventroposterior lateral nucleus (VPL) via the , to form the brainstem tract component of the dorsal column medial lemniscal system. VPL neurons, in turn, project to cortex; consequently, the GN serves as the first level nucleus in the dorsal column medial lemniscal system for processing of

ascending signals to cortex. Other central structures receiving direct GN and DCN projections include spinal cord, cerebellum, brainstem nuclei, and diencephalic nuclei

(Kemplay et al., 1989; Di Biasi et al., 1994; Bermejo et al., 2003).

E. GN functional RFs

The focus of this thesis was on how terminations of primary neuron inputs

from the toes are spatially distributed in the GN. This was done with an eye toward also

examining how structural organization of input terminations from different toes

contributes to GN neuron functional RFs on the toes. Of specific interest was how many

toes are usually included in normally-sized GN RFs on the toes.

It is known that a major percentage of GN neurons have an excitatory

tactile RF on the skin of the toes, hindpaw, hindlimb, or some other lower body location

(Nord, 1967; Berkley & Hubscher, 1995; Panetsos, 1995; Kondo et al., 2002; Costa-

Garcia & Nunez, 2004; Malmierca & Nuñez, 2004). In addition, beside a skin RF, many

GN neurons can have further RF components in visceral locations including, for example,

anal and vaginal locations ( Al-Chaer ED et al., 1996; Hubscher & Berkley, 1994). Most

16

data that pertain to normal sizes of GN RFs on the hindpaw indicate RF size in terms of

areas of skin, but do not indicate size with respect to numbers of toes (McComas, 1963;

Suzuki & Dickenson, 2002; Miki et al., 1998). A small number of GN RFs that include

toes have been illustrated anecdotally in figures of reports on other issues (Nord et al.,

1967; Berkley et al., 1995; Kondo et al., 2002; Costa-Garcia et al., 2004; Wang & Wall,

2005; 2006; Panetsos et al., 1995; Pettit & Schwark, 1993 ). These RFs range from a part

of one toe to up to 5 toes and appear to emerge from summation of smaller RF’s of

primary neurons (Lynn & Carpenter, 1982; Leem et al., 1993; Reinke & Dinse, 1996;

Shortland & Woolf, 1993; Shortland et al., 1989; Sanders & Zimmermann, 1986).

However, the few GN RFs shown in different studies do not provide indications of the

usual number of toes that are included in GN RFs. Other than these observations that

numbers of toes in GN toe RFs can vary to some degree, there is little understanding of

typical degrees of convergence of inputs from different toes.

Consistent with the above findings that primary neuron inputs provide the

major structural input to the GN, disruptions of primary neuron inputs from specific skin

locations, including the toes, causes loss of parts of GN RFs that are innervated by the

disrupted inputs (Pettit & Schwark, 1993; Panetsos et al. 1997; Schwark et al., 1999;

Wang & Wall, 2005; 2006). In contrast, GN RFs, including RFs on the toes, continue to

be generated after disruption of GN inputs from central structures like cortex and

cerebellum (Pettit & Schwark, 1993; Wang & Wall, 2005; 2006). This suggests GN

functional RFs depend to a large degree on primary neurons.

17

Literature Summary

The above literature review points out the following salient points about

what is and is not known about the GN, and about relationships between GN distributions

of primary neuron terminations from hindpaw toes, and GN RFs on these toes.

First, the GN is a major brainstem processing nucleus for information that

reaches cortex via the dorsal column medial lemniscal system. As part of a major cortical

input system, it is of interest to know whether concepts of processing that apply to cortex,

also apply to the GN. This issue has received little attention and is poorly understood.

Second, although cortex and other central structures provide inputs to the

GN, primary neuron inputs and, more specifically, large diameter primary neuron inputs

that process touch and proprioception signals, provide the major structural input to the

GN.

Third, axons from the dorsal column terminate in glomeruli which appear

to be organized in a plexus of terminations in the GN. Axons from dorsal root ganglia or

nerves terminate in rostrocaudal columns, and termination columns associated with

different ganglia or nerves overlap. Limited evidence suggests GN terminations of

primary neurons from toes versus hindpaw and hindlimb locations occupy different GN

locations. GN terminations of primary neurons from toes have been reported to

interchange relative locations at different rostrocaudal levels of the GN. Whether this interchanging involves overlap or some other complex arrangements of termination areas

of inputs from different toes remains unclear.

18

Fourth, primary neurons activate GN inhibitory interneurons and projection neurons. Through the activity of inhibitory interneurons, primary neuron inputs can be inhibited at postsynaptic locations on inhibitory interneurons and projection neurons, and at presynaptic locations on primary neuron axon terminations. Thus, inhibition within the GN may attenuate or cancel excitatory effects of some primary neuron terminations.

Fifth, a major percentage of GN neurons have tactile RFs on the toes, hindlimb, or lower body. These RFs are dependent to a large degree on activity of large diameter primary neuron inputs. Available data do not indicate the number of toes that typically contribute to GN toe RFs.

The above findings provide the foundation for examining the thesis question of whether structural inputs from different toes overlap or are distinct, and whether structural termination patterns of inputs from the toes match or differ from the numbers of toes that are usually included in functional RFs.

19

Design

This thesis examined whether a concept of cortical processing, i.e. that

excitatory RFs reflect functional expression of only part of an available overlapping

larger structural pool of ascending inputs, also applies to the GN.

The design was simple. The main part of the study examined whether GN

termination fields of primary neuron inputs from individual hindpaw toes overlap in a convergent fashion or, alternatively, are discretely organized in a nonoverlapping fashion.

The degree of overlap of structural inputs from different toes was then compared with the number of toes that are typically included in functional RFs on the toes. Inputs from toes were studied because toes are physically distinct (as compared to other continuous skin surfaces) and have higher densities of tactile inputs than more proximal body locations.

Toe RFs that were used for comparison with termination patterns were derived from our recent studies in which normal rat GN tactile RFs were assessed prior to experimental

manipulations (Wang and Wall, 2005; 2006).

To carry out this design, GN termination patterns of inputs from the toes

were reconstructed following toe microinjections of BHRP (cholera toxin subunit B

conjugated to horseradish peroxidase) or a mixture of BHRP and WGA-HRP (wheat germ agglutinin conjugated to horseradish peroxidase). The label was taken up and

transported by distal endings of primary neurons in the injection site. BHRP was used as the primary label because BHRP and the cholera toxin B subunit have been shown to be taken up and transported mainly by large diameter primary neurons that provide the main

20 input to the GN, and that provide activation for GN tactile RFs on the toes (LaMotte et al., 1991; Grant, 1993; Tong et al., 1999; Maslany et al., 1991).

Previous studies suggest that BHRP provides a more comprehensive estimate of total extents and densities of terminations than other labels (e.g., WGA-HRP, free HRP). Use of BHRP was, thus, consistent with the goal of identifying total terminal fields of inputs from individual toes and potential overlap of fields of different toes.

Electron microscopic studies indicate that label is located in terminals of sensory inputs after cholera toxin B subunit injections in hindlimb nerves (Hwang et al., 2001), and other studies have reported that label is mainly in sensory input terminals in the dorsal horn after BHRP or cholera toxin B subunit injection of hindlimb nerves or skin locations

(LaMotte et al., 1991; Levinsson et al., 2002; Maslany et al., 1992; Takahashi et al.,

2003). Previous reports also indicate BHRP that is taken up by primary neurons is not trans-synaptically transported (Robertson & Grant, 1985; LaMotte et al., 1991; Rivero-

Melian and Arvidsson, 1992). Consistent with previous observations (Maslany et al

1991; Rivero-Melian et al., 1992; Ueyama et al., 1994; Novikov, 2001), labeled axons were virtually all contained in the ipsilateral GN. This permitted microinjections of different toes in the two hindpaws, and cross-side comparisons of termination fields of different combinations of toes.

GN tactile RFs were defined using cutaneous tactile stimulation with hand-held probes and single or multiple unit GN responses in normal rats that were anesthetized with ketamine (80-90 mg/kg, im) and acepromazine (1.7 mg/kg, im). RF

21 locations on the toes were charted on scaled figures of the hindpaw, and counts were made of the numbers of toes per RF.

These structural and functional data provided analyses of degrees of structural overlap of GN terminations from different toes, and degrees of functional expression of toe inputs. These analyses were used to test whether structural overlap or non overlap matched or did not match functional expression. Further details about procedures are given in the following manuscript that summarizes the findings.

22

Manuscript

Title: Spatial relationships of structural and functional inputs to the brainstem gracile nucleus and potential functional underexpression of inputs

Shalini Niranjan, Xin Wang, and John T. Wall

Department of Neurosciences University of Toledo Medical Center Toledo, OH 43614-2598

Keywords: rat gracile nucleus, primary afferents, dorsal column nuclei, receptive fields, transganglionic transport

We gratefully acknowledge support of National Institutes of Health Grant HD 39791.

23

Abstract

The gracile nucleus (GN) is a major brainstem processing center for somatosensory inputs. GN neurons in rats often have excitatory tactile receptive fields

(RFs) on hindpaw toes, but there is little consensus on how the locations and sizes of these RFs emerge from GN substrates. To study this issue, the present experiments assessed relationships between spatial distributions of GN terminations of afferents from the toes, and sizes of GN tactile RFs on toes. Subcutaneous injections and transganglionic transport of tracers were used to reconstruct the distributions of GN termination areas of inputs from the toes, and neural recordings were used to define RFs on the toes. Overlap in termination areas of inputs from different toes was assessed and compared to the numbers of toes that contributed to toe RFs. GN terminations from each toe were distributed in a continuous rostrocaudal column in the GN. There was intermixing of termination columns of at least 3-4 toes across dorsal GN areas which, in cytochrome oxidase stained sections, appeared uniformly stained. Although a small percentage of toe RFs had inputs from up to 4-5 toes, on average RFs involved 2 toes, and most were on 1-2 toes. These results raise the possibility that, to some degree, toe

RFs in the rat GN emerge from sensory afferent substrates that have wider capacities than may be functionally expressed at a particular time.

24

Introduction

Afferent fibers from the hindpaws of rats project to the brainstem where

they terminate and make excitatory synaptic connections on GN neurons (Willis and

Coggeshall, 2004). GN neurons commonly have tactile receptive fields (RFs) on

hindpaw skin which, when stimulated by touch, cause neuronal activation. The locations

and sizes of these RFs, to some degree, reflect available direct input connections;

consequently, the structural distribution of primary afferent terminations from different

hindpaw locations is one factor in generating spatial properties of GN RFs. In addition,

GN input connections are influenced by intrinsic GN mechanisms, including inhibition

mechanisms, that affect GN processing and RF size (Broman, 1994; Willis and

Coggeshall, 2004; Wang and Wall, 2006). How spatial properties of excitatory tactile

RFs emerge from distributions of afferent terminations, GN processing mechanisms, and

other factors remains conjectural.

To study this issue, the present experiments evaluated relationships between the distributions of GN terminations of primary afferents from hindpaw toes and

GN excitatory tactile RFs on hindpaw toes. Tracers were injected into the hindpaw toes

of normal adult rats to permit labeling and reconstruction of the spatial distributions of

GN terminations of toe afferents. Overlaps in termination areas of different toes were

assessed and compared to results from a sample of GN toe RFs that estimated

percentages of RFs involving different numbers of toes. In addition, cytochrome oxidase

staining was used to test for subdivisions in GN areas with toe terminations.

25

Materials and Methods

All experiments were done in accordance with protocols that were approved by

the institutional animal care and use committee. Adult female Sprague Dawley rats were

used.

Toe injections

Animals were anesthetized with ketamine ( 80-90 mg/kg, im) and xylazine (8-12

mg/kg, im), and one or more toes on one or both hindpaws were injected subcutaneously

with cholera toxin subunit B conjugated to horseradish peroxidase (BHRP, 1%) or a mixture of BHRP plus wheat germ agglutinin conjugated to horseradish peroxidase

(BHRP, 1% and WGA-HRP, 14%). BHRP was used as the main tracer because it

effectively labels terminations of large diameter sensory inputs in normal rats (Robertson

and Grant, 1985; LaMotte et al., 1991; Tong et al., 1999; Shehab et al., 2003), and an aim

was to label large diameter sensory neurons that contribute to GN tactile RFs.

Microinjections were made from a microsyringe in one shot (≤8 µl) into the distal

phalanx. This was done to focus delivery in distal parts of toes that are expected to have

high innervation by large diameter cutaneous inputs, and to avoid spread beyond the

injected toe. Due to possible loss of label from uptake by blood, leakage from the

injection site, and other factors, it is difficult to stipulate exact amounts of label that were

transported. A total of 59 toes were injected in 20 rats.

26

Tissue processing

Following a post-injection time of 3 days, animals were given an overdose

of anesthesia and transcardially perfused with heparinated saline, 2% paraformaldehyde

in 0.1 M phosphate buffer, and 2% paraformaldehyde in 0.1 M phosphate buffer with

10% sucrose. The brainstem was removed, placed in the latter solution, and refrigerated

overnight.

The next day, a block of medulla spanning the entire rostrocaudal length of the GN was serially cut into 50 µm transverse sections. Each section was sequentially

placed in an individual well in 0.1 M phosphate buffer. A 5 or 7 in 10 series of these

sections was reacted in successive stages using tetramethylbenzidine (TMB) as a

chromogen (Florence et al., 1994). Given the interest in visualizing total termination

areas, sequential TBM reaction stages were monitored under microscopic view and

repeated to a point where emerging areas of GN labeling progressed no further, and

before background labeling became apparent in the GN and neighboring cuneate nucleus.

The remaining alternate 5 or 3 in 10 series of sections was processed with cytochrome

oxidase (CO) methods (Xu and Wall, 1999; Wang and Wall, 2006). The TMB series and

the CO series were separately mounted in sequence on slides and coverslipped.

In attempts to further document injection sites, in early experiments (n=4)

hindpaws were degloved by surgically dissecting all toe and other hindpaw skin from

underlying hindpaw tissues as one dorsal, and one plantar, skin piece. Dorsal and plantar

pieces, and the underlying hindpaw, were reacted separately using the above TMB

27 procedures. These attempts proved unreliable due to reactions with residual blood in paw tissues.

Reconstructions of GN termination areas

GN label typically appeared as collections of fine particles that were interpreted to involve terminal label (see Discussion). Transported label was clearly distinguishable from label associated with blood vessels or background. Labeled soma were not seen. GN areas containing labeled terminations were systematically reconstructed from transverse sections using dark field illumination and a drawing tube.

In agreement with previous findings (Rivero-Melian and Arvidsson, 1992; Ueyama et al.,

1994; Novikov, 2001), the present procedures resulted in labeling in only the ipsilateral

GN; consequently, it was possible to do cross-side comparisons following injections of toes on the two sides. Cross-side comparisons were done by first defining the left and right GN borders on individual sections. The GN borders on one side were then rotated around the midline axis and matched to the mirror image GN borders on the opposite side. Following matching of borders, reconstructions of labeled termination areas and lines indicating perimeters of labeled termination areas on each side were added at their respective locations within their GN borders. Cross-side matching was done entirely on the basis of GN borders and without reference to labeled terminations. Thus, cross-side analyses were “blinded” to the locations of the termination areas that were compared.

Digital images of TMB and CO sections were taken with a digital camera on a Nikon microscope. Images were collected using Spot basic software that was

28

interfaced with Adobe Photoshop programs. Images were adjusted using only brightness and contrast controls.

Toe receptive fields

Excitatory tactile RFs on hindpaw toes were used as indicators of suprathreshold functional convergence of toe inputs in the GN. A sample of 78 toe RFs

(39 single and 39 multi unit) was derived from studies in this lab in which normal GN tactile RFs were defined prior to subsequent experimental manipulations (Wang and

Wall, 2005; 2006). The sample reflected the range and distribution of sizes of normal toe

RFs defined with our procedures.

Recording, stimulation, and RF definition procedures have been previously described (Wang and Wall, 2005; 2006). Briefly, extracellular single unit

(window discriminated, ≥3:1 signal:noise) or multi unit GN activity was monitored as toe and hindpaw skin areas were tactilely stimulated with hand-held probes or brushes in anesthetized adult rats (ketamine, 80-90 mg/kg, im; acepromazine, 1.7 mg/kg, im). A RF was defined as the total area of skin that elicited neuronal discharges to cutaneous stimulation, and was judged to reflect a maximal excitatory tactile RF. For purposes of the present study, the sizes of GN RFs on toes were characterized in terms of the number of toes that were included in each RF.

29

Results

GN terminations of toe inputs

Injections were made into all five toes of a hindpaw to estimate the total

GN distribution of inputs from all toes. Injections of all 5 toes resulted in label in the ipsilateral, but not contralateral GN (e.g., Fig 1A). Across rostrocaudal levels, the composite termination area for inputs from all toes was roughly centered at mid- mediolateral locations in the more dorsal half to two-thirds of the ipsilateral GN (e.g. Fig.

1A-C). Terminations were continuously and densely distributed within this termination area across sections and extended as a continuous column along the rostrocaudal axis of the nucleus (e.g. Fig. 1B,C).

Injections of an individual toe were used to study terminations of inputs from different toes. After injection of any toe, associated terminations were centered around mid-mediolateral GN locations and, with sparing of label along the dorsal edge of the GN, were located in the dorsal half to two-thirds of the GN (e.g., Figs. 2,3). With

some variations in density, terminations were continuously distributed mediolaterally

and dorsoventrally across sections, and extended as a continuous column along the

rostrocaudal axis of the nucleus (e.g. Figs. 2,3).

A pattern of continuous terminations also occurred following injections of

nonadjacent toes on the same hindpaw. For example, injections of nonadjacent toes 3+5 or 2+4 resulted in largely continuous mediolateral and dorsoventral distributions of

labeled terminations, which extended as a continuous rostrocaudal column (e.g., Fig. 4).

A similar largely continuous pattern of terminations was produced after injections of toes

30

1+5, even though these toes were separated by the remaining three intervening toes 2-4

(e.g., Fig. 4).

Taken together, the above results suggested that the termination areas of

all five toes, any single toe, or nonadjacent toes were distributed in largely continuous

fashion in mediolateral and dorsoventral directions in the ipsilateral GN. Across

rostrocaudal levels, the termination areas of individual toes or combinations of toes

consistently formed continuous elongated columns in the GN.

Spatial relationships of termination columns of different toes

Further analyses were directed at assessing whether termination columns of inputs from different toes overlap. Three lines of evidence suggest that terminations of a given toe overlapped terminations of the other toes to substantial extents.

First, the termination column of each toe spanned similar locations and extents of the dorsal GN, with little apparent somatotopic shifting of termination areas of different toes. These results held when termination columns of different toes were compared across cases, or across sides in the same case. For example, the column of terminations from toe 1 occupied dorsal GN locations that appeared to coextend to some extent with termination columns of toes 2, 3, or 4 from other cases (e.g., compare Fig. 3A and B-D). Similarly, across cases, the columns of terminations from toes 1, 2, 3, or 4, and the termination column of toe 5 occupied overlapping locations (e.g., compare Fig. 3

A-D and E). Analogous statements applied when termination columns were compared across sides within the same case, as indicated for example by cross-side comparisons of

31

the termination columns of toes 2 and 4 (Fig. 3B and D), and the most separated toes 1

and 5 (Fig. 3E).

Second, injections of two nonadjacent toes on the same hindpaw resulted

in a continuous column of labeled terminations (Fig. 4). Given that the combined

terminations of all five toes also occupied a continuous area (e.g. Fig. 1), continuous

termination areas of nonadjacent toes suggest that terminations of inputs from the

injected toes likely overlapped termination areas of inputs from other (e.g., intervening)

toes that were not injected. The continuous termination area seen when the most distant

toes 1+5 were injected (e.g. Fig. 4A), suggests terminations of toes 1 and/or 5 cross areas

of terminations of intervening toes 2-4, thus reflecting overlap of terminations from at

least 3-4 toes. Further consistent with this possibility, cross-side comparisons of the

continuous termination areas seen, for example, after injection of nonadjacent toes 3+5

on one side and toes 2+4 on the other side, also suggested there was substantial overlap in the termination areas of at least 3-4 toes (Fig. 4C).

Finally, other analyses indicated that the column of terminations that was associated with any individual toe occupied substantial extents of the column of terminations associated with combined injections of either the most separated toes 1+5, or all five toes. For example, in different cases the continuous column of terminations that was associated with combined injection of the most separated toes 1+5 on one side colocalized to major degrees with the columns of terminations that were associated with toe 5 (Fig. 5A), toe 3 (Fig. 5B), or toe 4 (Fig. 5C) on the opposite side. Similarly, the column of terminations associated with injection of one toe on one side occupied

32

substantial extents of the column of terminations seen after injections of all five toes on

the other side. For example, across rostrocaudal levels, the termination column of toe 5

inputs spanned major central fractions of the termination column for all toes on the

opposite side, leaving relatively narrow margins for terminations from the remaining 4

toes (e.g., Fig. 5D).

The above results suggest that inputs from each toe terminated in a

continuous rostrocaudally elongated column in dorsal GN locations, and that there was

consistent and substantial overlap in the columns of terminations from at least 3-4, and

most likely all 5, toes. From this it appears that the GN contains areas where terminations

from each toe may intermix to substantial extents with terminations from most or all

other toes.

GN structure from CO stained sections

In CO stained sections, caudal levels of the GN contain densely stained neuropil areas that are disrupted by lightly stained bundles of fibers which are continuous with the gracile fascicle (e.g. Fig. 6A,B, asterisks). At successively more rostral levels, fiber bundles occupy reduced areas and densely stained areas occupy increased space as the main part of the nucleus first enlarges and then recedes (e.g., Fig. 6B-E). No consistent subdivisions were seen in the densely stained areas across the caudorostral axis

(e.g., Fig. 6A-E). On adjacent CO and TMB sections, the termination columns of inputs from the toes avoided fiber areas with light CO staining, and were localized in densely stained areas. Termination areas of toes were located in dorsal GN locations that had

33

relatively uniform CO staining (e.g., compare Figs. 6 and 3). These areas did not contain recognizable CO patches or other signs of subdivision, even though patches were consistently seen in the neighboring cuneate nucleus on the same sections. These results suggested that extents of the GN that contained terminations from different toes colocalize with GN areas with relatively uniform, dense CO staining.

GN excitatory receptive fields on the toes

Excitatory RFs on the toes were studied to assess functional convergence

of toe inputs in the GN. A sample of 78 toe RFs (39 single and 39 multi unit) was used to

analyze numbers of toes in RFs. Given the observed overlap of terminations of inputs

from the toes, RFs involving at least 3-4 toes might be expected to be common. In this

direction, GN RFs could involve excitatory inputs from up to 5 toes (e.g., Fig. 7).

However, further results suggest that most RFs did not reflect this expectation, including

that: (a) substantial percentages of toe receptive fields were restricted to one toe (single

unit = 46%; multi unit = 28%), (b) the majority of fields involved 1 or 2 toes (single unit

= 85%; multi unit = 56%), (c) few toe receptive fields involved 4 or 5 toes (single unit =

0%; multi unit = 11%), and (d) the overall mean number of toes per RF was 2 (single unit

= 1.7, multi unit = 2.3). These results suggest that although GN terminations of each toe

appeared to commonly intermix with terminations from most or all other toes, most RFs

did not functionally reflect this intermixing. These findings raise the possibility of

common mismatching in functional and structural organization where numbers of toes in

functional RFs underexpress existing structural overlaps in terminations of toe afferents.

34

Discussion

1. Present results

The present study suggests three main results. First, GN termination areas of toe inputs were organized as continuous rostrocaudal columns, and the column of terminations associated with a given toe overlapped the columns of other toes to large extents. As a consequence, there appeared to be intermixing of terminations from 3-4 or all toes along the GN rostrocaudal axis. Second, as assessed with CO staining, the termination areas of the toes avoided fiber areas with light CO staining, and were located in densely stained dorsal GN locations that did not appear to be subdivided. Finally, a small percentage of GN excitatory tactile RFs functionally reflected inputs from up to 4-5 toes, however, toe RFs averaged 2 toes, and the majority were on 1-2 toes. Taken together, these findings raise the possibility that there is some structural-functional mismatching in the form of common overlapping of afferent terminations from 3 or more toes, but common functional expression of inputs from 1-2 toes.

There is little understanding of how structural distributions of toe afferent terminations and functional RFs of toe inputs are related in the rat GN. The present results supplement current understanding in the following ways.

2. Peripheral sensory inputs to the rat GN

Extensive GN studies have been done in rats on sensory input projections from nerves (Leong and Tan, 1987; LaMotte et al., 1991; Ueyama et al., 1994; Shortland et al., 1998; Ding et al., 1999; Tong et al., 1999; Hwang et al., 2001; Shehab et al., 2003),

35

and dorsal root ganglia (Robertson and Grant, 1985; Rivero-Melian and Arvidsson, 1992;

Novikov, 2001). Although these studies clearly define GN termination patterns of

primary sensory inputs from the injected nerves or ganglia, they do not distinguish

termination patterns of toe inputs.

GN termination patterns of sensory inputs from hindpaw toes have been

examined in one previous study in which label was injected into toes (Maslany et al.,

1991). Results from that study, which were also based on transport of HRP conjugates,

resemble the present results in that terminations from an individual toe were distributed

as a rostrocaudal column in the ipsilateral dorsal GN. The present findings and this study

further agree that GN terminations from different toes are not distributed in a simple somatotopically shifted pattern. In this respect, the previous study suggested that the

termination columns of toes 1 and 4 reversibly and repeatedly interchanged their medial

and lateral positions with respect to each other across the rostrocaudal GN axis. How the

columns of these and other toes were related as this mediolateral interchanging occurred,

remained unclear; however, it seems possible that with mediolateral interchanging, the

termination columns of toes 1 and 4 may have overlapped the termination columns of

intervening toes. In the previous study, injections were designed to label limited inputs

from small areas on the plantar side of a toe (Maslany et al., 1991). In the present study,

the distal injections presumably involved most distal plantar and dorsal inputs of a toe;

consequently, termination areas of toes 1 and 4 in the present study appeared larger and

overlapping across the rostrocaudal axis. The present study also tested input pattern

relationships that weren’t analyzed in the previous report, including termination patterns

36

after combined injections of different multiple toes on a hindpaw, and termination

patterns of individual toes versus combinations of toes. These additional analyses provide

new evidence for substantial overlap in termination columns of different toes.

Overlap in GN terminations of toe afferents resembles the overlap in other terminations in the rat GN. For example, substantial overlap is seen in GN termination areas of afferents from the sciatic and saphenous nerves (LaMotte et al., 1991; Ueyama et al., 1994). Overlap is also seen in the GN termination areas of afferents from different lumbar dorsal root ganglia, with substantial overlap in termination areas of lumbar ganglia 3-5 which contain afferents from the hindpaw (Rivero-Melian and Arvidsson,

1992).

The present results are based on injections of BHRP or a combination of

BHRP and WGA-HRP. In normal rats, the cholera toxin B subunit and BHRP are taken up and transported mainly by large diameter sensory inputs (Robertson and Grant, 1985;

LaMotte et al., 1991; Tong et al., 1999; Shehab et al., 2003). A main component of large diameter inputs from the toes originates from cutaneous tactile inputs. Consistent with the present results, previous studies suggest hindlimb nerve inputs predominantly terminate in the ipsilateral GN, with no or only a few fibers projecting to the contralateral

GN (Rivero-Melian and Arvidsson, 1992; Ueyama et al., 1994; Novikov, 2001).

Previous work suggests that BHRP also provides a more complete estimate of total extents and densities of GN terminations than WGA-HRP (LaMotte et al., 1991; Maslany et al., 1991).

37

In the present studies, label was located in GN areas that appeared

uniformly stained in CO sections, avoided GN fascicles, and was not seen in soma. This

suggests that GN label mainly reflected terminal or preterminal, rather than axonal or

soma, labeling. This is consistent with electron microscopic findings of label in GN terminals of sensory afferents after cholera toxin B subunit injections in hindlimb nerves

(Hwang et al., 2001), with findings that label is mainly in sensory afferent terminals in the dorsal horn after BHRP or cholera toxin B subunit injection of hindlimb nerves or skin sites (LaMotte et al., 1991; Levinsson et al., 2002), and with findings that BHRP uptake by primary neurons is not trans-synaptically transported (Robertson and Grant,

1985; LaMotte et al., 1991; Rivero-Melian and Arvidsson, 1992).

Other pertinent findings come from Golgi studies of the rat GN. Terminal fields of Golgi stained dorsal column fibers have been reported to overlap extensively

(Odutola, 1977) and to be organized in a largely continuous fashion (Valverde, 1966).

Terminal arborizations of individual and small groups of dorsal column inputs have been described as intertwining to form glomeruli, within which dendrites of neurons receive axo-dendritic synapses. Glomeruli that occupy adjacent locations also fused to form a relatively uniform dense plexus of terminal arbors (Valverde, 1966). Although dorsal column inputs include spinal as well as primary afferents, the terminal labeling and CO staining patterns in the present studies could be consistent with this organization if primary afferents from different toes are organized in an overlapping plexus.

Postsynaptically, dendritic fields of GN neurons, including fields of projection neurons, appear to also extensively overlap (Valverde, 1966; Odutola, 1977).

38

We suggest that label in the present study is associated with afferents from injected toes, rather than with nonspecific uptake or transport by other inputs.

Nonspecific uptake may occur by peripheral spread of injected label into adjacent toes or other hindpaw or hindlimb locations, or by spread of label from labeled toe axons to adjacent axons in projection systems between the toes and GN. Our attempts to directly assess injection sites, although inconclusive (see Methods), did not indicate spread of label was a problem. Other considerations are also consistent with this. Spread of label across peripheral nerve or dorsal column axons is rarely discussed in transganglionic studies but, when examined, has not been seen (Ding et al., 1999). Primary afferents from the toes project to the GN via the sciatic and saphenous nerves and the dorsal column. These projection systems also contain axons that innervate hindlimb locations that terminate in parts of the GN where inputs from the toes did not terminate in the present study (LaMotte et al., 1991; Maslany et al., 1991; Ueyama et al., 1994). This suggests that spread of label, and particularly spread from afferents of an injected toe to afferents of another toe, would have had to occur without spread to other afferents.

GN overlap of toe terminations appears to be a further reflection of overlapping organization of toe projections at sub-brainstem levels in rats. For example, at peripheral levels, primary sensory neurons innervating the five hindpaw toes are located mainly in lumbar ganglia L3-6, with neurons from each toe occupying two or three ganglia in a coarsely shifted and overlapping manner (Prats-Galino et al., 1999;

Puigdellivol-Sanchez et al., 2000). Individual primary sensory neurons that respond to touch stimulation of the toes usually have a restricted receptive field spanning a small

39

fraction of one toe (Lynn and Carpenter, 1982; Sanders and Zimmermann, 1986;

Shortland et al., 1989; Leem et al., 1993; Shortland and Woolf, 1993; Reinke and Dinse,

1996). Interestingly, a few axons with tactile receptive fields on two toes have also been seen (Lynn and Carpenter, 1982; Woolf, 1987), as have a small number of double- or triple-labeled dorsal root ganglion cells after injection of multiple toes (Puigdellivol-

Sanchez et al., 2000). These findings suggest that some peripheral neurons, including mechanosensory neurons, have “overlapping” toe inputs at the peripheral axon level.

Overlap has also been reported in spinal dorsal horn terminations of toe inputs. Injections of individual toes label terminations in the medial dorsal horn at L3-5 levels in rats and, in both superficial and deeper dorsal horn layers, terminations from adjacent and up to 3-

4 toes can overlap (Molander and Grant, 1985; Maslany et al., 1992; Levinsson et al.,

2002; Takahashi et al., 2003). Injection and labeling of individual sensory fibers from toes, including mechanosensory fibers, suggest that these fibers span long rostrocaudal extents in the medial dorsal horn, and that the termination arbors of individual fibers from one toe are partially offset from, but also overlap with, arbors of fibers from other toes

(Shortland et al., 1989; Shortland and Woolf, 1993). From these findings, overlap of toe terminations in the GN appears to be a property that is shared with hindpaw toe projection systems below the brainstem.

3. CO and other structure in the rat GN

CO stained material from our injection cases suggested that toe termination areas in the dorsal GN were uniformly stained and did not appear to be

40

subdivided. CO staining in other brainstem nuclei in rats reveals distinct, densely stained

patches that colocalize with sensory input termination areas from specific body parts,

including, for example, CO patches that colocalize with main termination areas of each

forepaw toe in the other dorsal column nucleus, the cuneate nucleus (Crockett et al.,

1993b). The present failure to see CO patches in the GN was not attributable to technical

failures in staining because CO dense patches were consistently seen in the neighboring

CN on the same sections where uniform staining was seen in the GN toe area.

The present findings that GN toe termination areas do not involve obvious subdivisions are consistent with other observations from the rat GN. A stereological evaluation briefly noted vertical ‘slabs’ of dense CO staining in the rostrodorsal GN, but indicated staining was not sufficiently prominent or consistent to justify subdivisions

(Bermejo et al., 2003). Another report indicated the dense and uniform CO staining pattern in the GN contrasted with the parcellated pattern of CO dense patches in the CN

(Crockett et al., 1993b). Finally, no distinct dorsal GN subdivisions have been recognized in rat with other methods that reveal presynaptic and/or postsynaptic structural organization, including: Nissl staining (Valverde, 1966; Bermejo et al., 2003), p75NGFR-like immunostaining (Crockett et al., 1993a), neuropeptide immunoreactivity

(Conti et al., 1990; Noguchi et al., 1995; Ma and Bisby, 1999; 2000; Ossipov et al., 2002;

Ma et al., 2003), neurotrophin expression (Li et al., 1999; Ha et al., 2000), synaptic

vesicle associated protein immunostaining (Crockett et al., 1996), substance P receptor

binding (Schwark et al., 1998), GABA and glycine immunostaining (Popratiloff et al.,

1996a), calcium binding protein immunoactivity (Crockett et al., 1996; Magnusson et al.,

41

1996), or retrograde labeling of GN projection neurons from GN projection targets

(Kosinski et al., 1988; Kemplay and Webster, 1989; Magnusson et al., 1996; Li and

Mizuno, 1997; Bermejo et al., 2003; Del Cano et al., 2004; Wree et al., 2005).

4. Comparison of toe RFs with primary afferent organization

The distribution of afferent terminations from the toes presumably affects

GN toe RFs, but how terminal organization contributes to the location and size of these

RFs remains unclear. In the present study, some GN excitatory receptive fields had functionally expressed inputs from up to 4-5 toes, but the majority had inputs from 1-2 toes. Considered with the termination findings, these results raise the possibility that some existing terminations may be functionally underexpressed in toe RFs.

Available data on normal sizes of GN tactile RFs on rat hindpaws are usually expressed as areas of skin (e.g., mm2) and do not distinguish toe areas, or

numbers of toes in RFs (McComas, 1963; Miki et al., 1998; Suzuki and Dickenson, 2002;

Kitagawa et al., 2005). RF data in rats that are pertinent to the present results are limited

to a few toe RFs that are incidentally illustrated in reports on other issues (Nord, 1967;

Berkley and Hubscher, 1995; Panetsos et al., 1995; 1997; Kondo et al., 2002; Costa-

Garcia and Nunez, 2004). Illustrated toe RFs range from part of one toe to up to 5 toes,

but typical number of toes per RF cannot be gauged from the few available examples.

Comparisons of GN toe RFs and toe RFs of primary afferents suggest that

spatial convergence of afferent inputs contributes to GN toe RF size and location, i.e.,

whereas GN RFs commonly involve 1-2 toes, most RFs of primary afferents span a small

42

fraction of one toe (Lynn and Carpenter, 1982; Sanders and Zimmermann, 1986;

Shortland et al., 1989; Leem et al., 1993; Shortland and Woolf, 1993; Reinke and Dinse,

1996). However, unless one assumes that input convergence is the sole determinant of

GN RF organization, it is difficult to use relative sizes of GN and primary afferent RFs to suggest how GN RF size and location are determined.

The present findings raise the possibility that toe RFs may not always reflect all available structural inputs from the toes. Other findings seem consistent with this possibility. Toe RFs in the rat GN have capacities to rapidly enlarge under some conditions. For example, disruption of peripheral input activity by cutaneous injection of the local anesthetic lidocaine can cause immediate enlargements of hindpaw GN RFs, including enlargements involving one or more toes (Panetsos et al., 1995; 1997; Wang

and Wall, 2005). In addition, disruption of local inhibition in the rat GN with the

GABAA receptor antagonist bicuculline methiodide commonly causes immediate RF

enlargements involving larger areas on individual toes and increased numbers of toes

(Wang and Wall, 2006). The lidocaine findings are consistent with an availability of functionally unexpressed cutaneous inputs from toes, and the bicuculline results suggest structural substrates of unexpressed toe inputs are located within the GN itself and are regulated by local inhibition. The present results provide new evidence for the possibility that these substrates may, to some degree, involve overlapping terminations of sensory afferent fibers.

There is recognition of the principle that spatial properties of central RFs can emerge from structural substrates that have wider capacities than are functionally

43 expressed at a particular time. However, identifying substrates where structural and functional convergence can be compared has been difficult. In the dorsal column medial lemniscal projection system to somatosensory thalamus and cortex, thalamic afferent substrates to cortex appear to work by this principle. “Masked” substrates between the skin and GN have been long discussed, but there is little data on relationships between structurally identified afferents, and the size and location of functional RFs. The present findings raise the possibility that this principle may apply to peripheral afferent substrate connections in the GN.

Literature Cited

Berkley KJ, Hubscher CH. 1995. Are there separate central nervous

system pathways for touch and pain? Nature Medicine 1:766-773.

Bermejo PE, Jimenez CE, Torres CV, Avendano C. 2003. Quantitative

stereological evaluation of the gracile and cuneate nuclei and their projection

neurons in the rat. J Comp Neurol 463:419-433.

Broman J. 1994. Neurotransmitters in subcortical somatosensory

pathways. Anat Embryol (Berl) 189:181-214.

Conti F, De Biasi S, Giuffrida R, Rustioni A. 1990. Substance P-

containing projections in the dorsal columns of rats and cats. Neuroscience

34:607-621.

44

Costa-Garcia M, Nunez A. 2004. Nociceptive stimuli induce changes in

somatosensory responses of rat dorsal column nuclei neurons. Brain Res

1025:169-176.

Crockett DP, Foschini DR, Girgis WS, Egger MD. 1993a.

Immunocytochemical localization of the low-affinity nerve growth factor receptor

(p75NGFR) in the cuneate nucleus of the rat and its relationship to cytochrome-

oxidase activity. Brain Res 603:324-327.

Crockett DP, Maslany S, Harris SL, Egger MD. 1993b. Enhanced

cytochrome-oxidase staining of the cuneate nucleus in the rat reveals a modifiable

somatotopic map. Brain Res 612:41-55.

Crockett DP, Maslany S, Egger MD. 1996. Synaptophysin

immunoreactivity and distributions of calcium-binding proteins highlight the functional organization of the rat's dorsal column nuclei. Brain Res 707:31-46.

Del Cano GG, Uria I, Gerrikagoitia I, Martinez-Millan L. 2004.

Connection from the dorsal column nuclei to the superior colliculus in the rat: topographical organization and somatotopic specific plasticity in response to neonatal enucleation. J Comp Neurol 468:410-424.

Ding Y-Q, Shi J, Wang D-S, Xu J-Q, Li J-L, Ju G. 1999. Primary afferent fibers of the pelvic nerve terminate in the gracile nucleus of the rat. Neurosci Lett

272:211-214.

45

Florence SL, Garraghty PE, Wall JT, Kaas JH. 1994. Sensory afferent projections and area 3b somatotopy following median nerve cut and repair in macaque monkeys. Cereb Cortex 4:391-407.

Ha SO, Kim JK, Hong HS, Kim DS, Cho HJ. 2000. Expression of brain- derived neurotrophic factor in rat dorsal root ganglia, spinal cord, and gracile nuclei in experimental models of neuropathic pain. Neuroscience 107:301-309.

Hwang SJ, Rustioni A, Valtschanoff JG. 2001. Kainate receptors in primary afferents to the rat gracile nucleus. Neurosci Lett 312:137-140.

Kemplay S, Webster KE. 1989. A quantitative study of the projections of the gracile, cuneate, and trigeminal nuclei and of the medullary to the thalamus in the rat. Neuroscience 32:153-167.

Kitagawa J, Tsuboi Y, Ogawa A, Ren K, Hitomi S, Saitoh K, Takahashi

O, Masuda Y, Harada T, Hanzawa N, Kanda K, Iwata K. 2005. Involvement of dorsal column nucleus neurons in nociceptive transmission in aged rats. J

Neurophysiol 94:4178-4187.

Kondo E, Iwata K, Ogawa A, Tashiro A, Tsuboi Y, Fukuoka T, Yamanaka

H, Dai Y, Morimoto T, Noguchi K. 2002. Involvement of glutamate receptors on hyperexcitability of wide dynamic range neurons in the gracile nucleus of the rats with experimental mononeuropathy. Pain 95:153-163.

Kosinski RJ, Lee HS, Mihailoff GA. 1988. A double retrograde fluorescent tracing analysis of dorsal column nuclear projections to the basilar

46

pontine nuclei, thalamus, and superior colliculus in the rat. Neurosci Lett 85:40-

46.

LaMotte CC, Kapadia SE, Shapiro CM. 1991. Central projections of the sciatic, saphenous, median, and ulnar nerves of the rat demonstrated by transganglionic transport of choleragenoid-HRP (B-HRP) and wheat germ agglutinin-HRP (WGA-HRP). J Comp Neurol 311:546-562.

Leem JW, Willis WD, Chung JM. 1993. Cutaneous sensory receptors in the rat foot. J Neurophysiol 69:1684-1699.

Leong SK, Tan CK. 1987. Central projection of rat sciatic nerve fibres as revealed by Ricinus communis agglutinin and horseradish peroxidase tracers. J

Anat 154:15-26.

Levinsson A, Holmberg H, Broman J, Zhang M, Schouenborg J. 2002.

Spinal sensorimotor transformation: relation between cutaneous somatotopy and a reflex network. J Neurosci 22:8170-8182.

Li H, Mizuno N. 1997. Collateral projections from single neurons in the dorsal column nuclei to the inferior colliculus and the ventrobasal thalamus: a retrograde double-labeling study in the rat. Neurosci Lett 225:21-24.

Li W-P, Xian C, Rush RA, Zhou X-F. 1999. Upregulation of brain-derived neurotrophic factor and neuropeptide Y in the dorsal ascending sensory pathway following sciatic nerve injury in rat. Neurosci Lett 260:49-52.

Lynn B, Carpenter SE. 1982. Primary afferent units from the hairy skin on the rat hindlimb. Brain Res 238:29-43.

47

Ma W, Bisby MA. 1999. Ultrastructural localization of increased

neuropeptide immunoreactivity in the axons and cells of the gracile nucleus following chronic constriction injury of the sciatic nerve. Neuroscience 93:335-

348.

Ma W, Bisby MA. 2000. Partial sciatic nerve ligation induced more dramatic increase of neuropeptide Y immunoreactive axonal fibers in the gracile nucleus of middle-aged rats than in young adult rats. J Neurosci Res 60:520-530.

Ma W, Chabot JG, Powell KJ, Jhamandas K, Dickerson IM, Quirion R.

2003. Localization and modulation of calcitonin gene-related peptide-receptor component protein-immunoreactive cells in the rat central and peripheral nervous systems. Neuroscience 120:677-694.

Magnusson A, Dahlfors G, Blomqvist A. 1996. Differential distribution of calcium-binding proteins in the dorsal column nuclei of rats: a combined immunohistochemical and retrograde tract tracting study. Neuroscience 73:497-

508.

Maslany S, Crockett DP, Egger MD. 1991. Somatotopic organization of

the dorsal column nuclei in the rat: transganglionic labelling with B-HRP and

WGA-HRP. Brain Res 564:56-65.

Maslany S, Crockett DP, Egger MD. 1992. Organization of cutaneous

primary afferent fibers projecting to the dorsal horn in the rat: WGA-HRP versus

B-HRP. Brain Res 569:123-135.

48

McComas AJ. 1963. Responses of the rat dorsal column system to

mechanical stimulation of the hind paw. J Physiol 166:435-448.

Miki K, Iwata K, Tsuboi Y, Sumino R, Fukuoka T, Tachibana T,

Tokunaga A, Noguchi K. 1998. Responses of dorsal column nuclei neurons in rats

with experimental mononeuropathy. Pain 76:407-415.

Molander C, Grant G. 1985. Cutaneous projections from the rat hindlimb

foot to the substantia gelatinosa of the spinal cord studied by transganglionic

transport of WGA-HRP conjugate. J Comp Neurol 237:476-484.

Noguchi K, Kawai Y, Fukuoka T, Senba E, Miki K. 1995. Substance P

induced by peripheral nerve injury in primary afferent sensory neurons and its effect on dorsal column nucleus neurons. J Neurosci 15:7633-7643.

Nord SG. 1967. Somatotopic organization in the spinal trigeminal nucleus, the dorsal column nuclei and related structures in the rat. J Comp Neurol 130:343-

356.

Novikov LN. 2001. Labeling of central projections of primary afferents in adult rats: a comparison between biotinylated dextran amine, neurobiotin, and

Phaseolus vulgaris-leucoagglutinin. J Neurosci Methods 112:145-154.

Odutola AB. 1977. Patterns and fields of dorsal column fiber terminals in

the cuneo-gracile nuclei of the rat. Experimental Neurology 57:112-120.

Ossipov MH, Zhang E-T, Carvajal C, Gardell L, Quirion R, Dumont Y,

Lai J, Porreca F. 2002. Selective mediation of nerve injury-induced tactile

hypersensitivity by neuropeptide Y. J Neurosci 15:9858-9867.

49

Panetsos F, Nunez A, Avendano C. 1995. Local anaesthesia induces immediate receptive field changes in nucleus gracilis and cortex. NeuroReport

7:150-152.

Panetsos F, Nunez A, Avendano C. 1997. Electrophysiological effects of temporary deafferentation on two characterized cell types in the nucleus gracilis of the rat. Eur J Neurosci 9:563-572.

Popratiloff A, Valtschanoff JG, Rustioni A, Weinberg RJ. 1996.

Colocalization of GABA and glycine in the rat dorsal column nuclei. Brain Res

706:308-312.

Prats-Galino A, Puigdellivol-Sanchez A, Ruano-Gil D, Molander C. 1999.

Representations of hindlimb digits in rat dorsal root ganglia. J Comp Neurol

408:137-145.

Puigdellivol-Sanchez A, Forcada-Calvet P, Prats-Galino A, Molander C.

2000. Contribution of femoral and proximal sciatic nerve branches to the sensory innervation of hindlimb digits in the rat. Anat Rec 260:180-188.

Reinke H, Dinse HR. 1996. Functional characterization of cutaneous properties in aged rats. Neurosci Lett 216:171-174.

Rivero-Melian C, Arvidsson J. 1992. Brain stem projections of rat lumbar dorsal root ganglia studied with choleragenoid conjugated horseradish peroxidase.

Exp Brain Res 91:12-20.

Robertson B, Grant G. 1985. A comparison between wheat germ agglutinin and choleragenoid horseradish peroxidase as anterogradely transported

50

markers in central branches of primary sensory neurones in the rat with some

observations in the cat. Neuroscience 14:895-905.

Sanders KH, Zimmermann M. 1986. in rat glabrous

skin: redevelopment of function after nerve crush. J Neurophysiol 55:644-659.

Schwark HD, Petit MJ, Fuchs JL. 1998. Distribution of substance P receptor binding in dorsal column nuclei of rat, cat, monkey, and human. Brain

Res 786:259-262.

Shehab SAS, Spike RC, Todd AJ. 2003. Evidence against cholera toxin B subunit as a reliable tracer for sprouting of primary afferents following peripheral nerve injury. Brain Res 964:218-227.

Shortland P, Woolf CJ. 1993. Morphology and somatotopy of the central arborizations of rapidly adapting glabrous skin afferents in the rat lumbar spinal cord. J Comp Neurol 329:491-511.

Shortland P, Woolf CJ, Fitzgerald M. 1989. Morphology and somatotopic

organization of the central terminals of hindlimb hair follicle afferents in the rat

lumbar spinal cord. J Comp Neurol 289:416-433.

Shortland PJ, Wang HF, Molander C. 1998. Transganglionic transport of the lectin Soybean agglutinin (Glycine max) following injection into the sciatic nerve of the adult rat. J Neurocytol 27:233-245.

Suzuki R, Dickenson AH. 2002. Nerve injury-induced changes in opioid modulation of wide dynamic range dorsal column nuclei neurones. Neuroscience

111:215-228.

51

Takahashi Y, Chiba T, Kurokawa M, Aoki Y. 2003. Dermatomes and the central organization of dermatomes and body surface regions in the spinal cord dorsal horn in rats. J Comp Neurol 462:29-41.

Tong Y-G, Wang HF, Ju G, Grant G, Hokfelt T, Zhang X. 1999. Increased uptake and transport of cholera toxin B-subunit in dorsal root ganglion neurons after peripheral axotomy: possible implications for sensory sprouting. J Comp

Neurol 404:143-158.

Ueyama T, Houtani T, Ikeda M, Sato K, Sugimoto T, Mizuno N. 1994.

Distribution of primary afferent fibers projecting from hindlimb cutaneous nerves to the in the cat and rat. J Comp Neurol 341:145-158.

Valverde F. 1966. The pyramical tract in rodents. A study of its relations with the posterior column nuclei, dorsolateral reticular formation of the medulla oblongata, and cervical spinal cord. Zeitschrift fur Zellforschung 71:297-363.

Wang X, Wall JT. 2005. Cortical influences on sizes and rapid plasticity of tactile receptive fields in the dorsal column nuclei. J Comp Neurol 489:241-

248.

Wang X, Wall JT. 2006. Cortical influences on rapid brainstem plasticity.

Brain Res 1095:73-84.

Willis WD, Coggeshall RE. 2004. Sensory mechanisms of the spinal cord.

New York: Kluwer Academic. 962 p.

Woolf CJ. 1987. Central terminations of cutaneous mechanoreceptive afferents in the rat lumbar spinal cord. J Comp Neurol 261:105-119.

52

Wree A, Itzev DE, Schmitt O, Usunoff KG. 2005. Neurons in the dorsal

column nuclei of the rat emit a moderate projection to the ipsilateral ventrobasal

thalamus. Anat Embryol (Berl) 210:155-162.

Xu J, Wall JT. 1999. Functional organization of tactile inputs from the

hand in the cuneate nucleus and its relationship to organization in the

somatosensory cortex. J Comp Neurol 411:369-389.

Figure legends

Figure 1. GN label following combined injections of all 5 toes in one hindpaw and no injection in the other hindpaw. A: Darkfield photomicrograph of GN label after unilateral injection of all 5 toes on the left side, and no injection on the right side. There was a continuous area of label in the indicated area in the left GN, and no label in the right GN (GN borders indicated on each side). B,C: Reconstruction of the caudorostrally

(respective lower to higher numbered sections) continuous column of label that was restricted to the ipsilateral dorsal GN after injection of all 5 toes in two cases. Midline and GN borders are indicated. Dorsal is up in all sections.

Figure 2. Examples of GN label after injection of different individual toes in each hindpaw. An injection was made into toe 5 (T5) on the left side and toe 1 (T1) on the right side. Low power photomics illustrating label on each side at indicated transverse planes. Reconstructions illustrate label across the caudorostral (respective lower to higher numbered sections) axis of the GN. Inputs from each toe were distributed as a

53 continuous caudorostral column spanning similar locations in the dorsal GN. Midline and

GN borders are indicated and dorsal is up in all sections.

Figure 3. Examples of GN label in caudorostral series (respective lower to higher numbered sections) following injections of different individual toes. Reconstructions are for terminations of toe 1 (A), toe 2 (B), toe 3 (C), and toe 4 (D). B and D are from the two sides of one case. E shows cross-side comparisons of areal extents of label for toe 5

(grey area) on one side and toe 1 (black area and dashed lines) on the other side in another case. To facilitate comparisons, the midline is to the right and dorsal is up in all cases.

Figure 4. Examples of GN label following injections of nonadjacent toes on the same hindpaw. A: Reconstructed terminations after injections of toes 1+5. B:

Reconstructed terminations after injections of toes 3+5 on the left side and 2+4 on the right side in one case. C: Cross-side comparison of terminations following injections of toes 3+5 (black area and dashed lines) and 2+4 (grey area) for a series of sections from the case shown in B. Midline and GN borders are indicated and dorsal is up in all sections.

Figure 5. Cross-side comparisons of GN label following injections of combinations of toes on one side, and an individual toe on the other side. A-C: Cross- side comparisons for injections of the most separated toes 1+5 on one side and of toe 5

(A), toe 3 (B), or toe 4 (C), on the other side in 3 cases. D: Cross-side comparison for injections of all 5 toes (1-5) on one side and toe 5 on the other side in another case. To facilitate comparisons, all cases are illustrated with the midline to the right and dorsal up.

54

Figure 6. Caudorostral sequence (respective lower to higher numbers) of

CO stained sections through the GN nucleus. At more caudal levels (A,B), densely stained neuropil areas are interrupted by lightly stained fiber bundles (e.g., indicated by

*). At more rostral levels (C-E), densely stained neuropil areas become more prevalent.

Dorsal GN areas where toe inputs terminated were relatively uniformly stained and did not contain recognizable subdivisions. Medial is right and dorsal is up in all sections.

Figure 7. Examples of GN excitatory tactile RFs involving toes. The shaded area on each hindpaw figure indicates a RF, and the number indicates number of toes in that RF. Lines link parts of a RF that involved plantar and dorsal surfaces.

55

Figure - 1

56

Figure - 2

57

Figure – 3

58

Figure - 4

59

Figure – 5

60

Figure – 6

61

Figure – 7

62

Discussion

Present results

This thesis examined GN structural organization of primary neuron

terminations from the hindpaw toes, and GN excitatory RFs on toes, in an attempt to

understand how distributions of input terminations from different toes and RFs on toes are related. At the beginning of this thesis, the GN termination patterns of primary neuron inputs from toes and the numbers of toes that usually contribute to normal

excitatory toe RFs were not well understood.

The GN is a major player in the dorsal column medial lemniscal system

which, through related thalamocortical projections, is a main input projection system to

primary somatosensory cortex. At cortical levels of this system, there was substantial

evidence that RFs of cortical neurons can commonly reflect incomplete functional expression of available inputs; i.e., there is structural overlap of thalamocortical terminations from a much wider range of body locations than is reflected in the excitatory

RF at a given time (Garraghty et al., 1989; Arnold et al.,2001; Rausell & Jones, 1995;

Rausell et al. 1998; Garraghty et al. 1990; Churchill et al. 2004; for earlier reviews see

Snow and Wilson, 1991). This thesis asked whether this concept might also apply at brainstem levels of the ascending projection system.

No previous attempts had been made to examine GN structural-functional relationships between primary neuron termination patterns and RFs, and different arrangements seem possible. One alternative, for example, is that GN terminations from each toe occupy a discrete GN area that does not overlap the termination areas of other 63

toes and that this structural organization, in turn, is associated with a “matched” GN RF organization where excitatory RFs are mainly restricted to an individual toe (e.g. Fig 2,

RF neuron A). An alternative “matching” possibility is that GN terminations of different toes overlap to some degree and that this is associated with a “matched” GN RF organization where excitatory RFs reflect these overlapping toe inputs (e.g. Fig 2, RFs of

neuron B & C). “Mismatched” structural-functional arrangements also seem possible.

For example, there may be discrete nonoverlapping primary sensory terminations from

individual toes, but excitatory RFs on multiple or most toes may result from excitatory

driving from spinal or other inputs that add to primary sensory inputs (e.g. Fig 2, RF of

neuron D) , An alternative mismatch could result if GN terminations from each toe

overlap termination areas of most or all other toes, but excitatory GN RFs involve only one or two toes (e.g. Fig 2, RFs of E & F). This last possibility could involve, e.g. GN inhibition and would be consistent with how thalamocortical inputs and RFs can be organized in somatosensory cortex.

64

Figure 2. Potential relationships between structural distribution patterns of primary neuron terminations from the toes and functional RFs on the toes.

Left: Structural patterns of termination may “match” functional RF expression. For example, GN terminations from each toe may occupy discrete, non-overlapping GN areas and this structural organization may “match” RF organization where excitatory RFs are restricted to one toe (GN neuron A and its RF). Another possibility is that GN terminations from different toes overlap to some degree and this overlapping is associated with “matched” RFs from multiple toes (GN neurons B and C and their RFs). Right: There may also be “mismatches” in structural and functional organization. For example, GN terminations from each toe may occupy discrete, non-overlapping GN areas but excitatory RFs may be on multiple toes due to contributions that are made by inputs other than primary neuron inputs (GN neuron D and its RF). Another “mismatch” possibility is that GN terminations from each toe overlap terminations of most or all other toes but, due to GN inhibition, RFs involve only one or two toes (GN neurons E and F and their RFs). The present results, taken together with findings on GN inhibition effects on RFs, provide evidence for “mismatches” like that seen in neurons E and F. This evidence does not preclude that some “matches”, e.g., like that seen in neuron B, may also occur.

65

The present results provide evidence that GN terminations of primary neuron

inputs from individual toes were organized as continuous rostrocaudal columns in the

ipsilateral GN, and that columns of different toes overlapped. In assessing the degree of

overlap, results from injections of each individual toe, different combinations of some or

all toes on one side, and different combinations of toes on the two sides in individual

cases, suggest there is substantial overlap in the termination columns of at least 3-4, and

likely all 5, toes. The termination columns of toe inputs were located in dorsal extents of

the GN which, in CO stained sections, were densely stained and did not appear to contain

subdivisions. This pattern of CO staining could also be consistent with an overlapping,

nonparcellated organization of structural inputs from the toes. Finally, a small fraction of

GN excitatory RFs expressed inputs from up to 4-5 toes, however, the mean number of

toes in RFs was 2, and a majority of RFs involved 1-2 toes. Taken together, these results

raise the possibility of functional and structural organization whereby numbers of toes in

excitatory RFs underexpress existing structurally available primary neuron inputs. While

not precluding some “matches” in organization (e.g. Fig 2, RF of neuron B), this is

consistent with “mismatching” in structural-functional organization (e.g. Fig 2, RFs of

neurons E & F).

How the present results supplement understanding of primary input

organization in the GN

Prior to the present work, early Golgi studies indicated that ascending axons from the dorsal column terminate in the GN in glomeruli around GN neuron

66

dendrites, and form an apparently continuous plexus of GN terminations (Valverde,

1966; Odutola, 1977). Although some stained axons were likely from primary neurons,

Golgi staining does not permit distinguishing of axons of primary neurons and spinal neurons, both of which ascend in the dorsal column. Subsequent axonal transport studies identified distributions of GN terminations that were labeled after injections of dorsal root ganglia or peripheral nerves that contain primary neurons from hindquarter and hindlimb areas. These studies clearly distinguished ascending inputs of primary neurons from inputs from spinal neurons, and indicated that there was considerable overlap in the

GN termination areas of primary neurons from different ganglia or nerves (Rivero-

Melian and Arvidsson, 1992; Prats-Galino et al., 1999; Puigdellivol-Sanchez et al.,

2000); however, from injections of these structures it was not possible to resolve how primary inputs from a specific location on the body were distributed, or whether GN termination areas of inputs from different body locations, like different toes, overlap.

One previous study used axonal transport following subcutaneous injections of label into toe and other locations on the rat hindpaw and hindlimb and found

that terminations from toes were organized in rostrocaudal columns in the dorsal GN

(Maslany et al., 1991). In addition, terminations of primary neuron inputs from toes 1

and 4 repeatedly and reversibly interchanged medial and lateral GN positions at different

rostrocaudal levels. It was not clear from these results whether this medial-lateral

interchanging was associated with overlap of GN terminations of these and/or other toe

inputs. Thus, whether or not GN terminations from each toe overlap terminations of

other toes was not resolved.

67

The present study used individual toe injections and a range of combined

toe injections in one or both hindpaws to examine termination patterns of toe inputs in

detail. The findings agree with previous findings regarding GN toe terminations in

suggesting that primary neuron inputs from individual toes occupy rostrocaudally

continuous columns in dorsal GN locations. The present findings further agree with previous results in suggesting that GN terminations from the toes are not arranged in a

simple somatotopic pattern. The present results supplement previous understanding by

suggesting that GN termination columns of primary neuron inputs from individual toes

overlap, and that terminations from most or all toes appear to overlap to significant

degrees.

The CO results further suggest there was no apparent parcellation or subdivision in dorsal GN areas where toe inputs terminated. Consistent with this,

previous studies that have examined rat GN using CO staining (Crockett et al., 1993b), as

well as a range of other pre- and postsynaptic structural stains and markers like Nissl

staining (Valverde, 1966; Bermejo et al., 2003), Golgi staining(Valverde, 1966;Gulley,

1973), p75NGFR-like immunostaining (Crockett et al., 1993a) , neuropeptide

immunoreactivity (Conti et al., 1990; Noguchi et al., 1995; Ma and Bisby, 1999; 2000;

Ma et al., 2003; Ossipov et al., 2002), neurotrophin expression (Li et al., 1999; Ha et al.,

2000) , synaptic vesicle associated protein immunostaining (Crockett et al., 1996) ,

substance P receptor binding (Schwark et al., 1998) , GABA and glycine immunostaining

(Popratiloff et al., 1996a) , calcium binding protein immunoactivity (Crockett et al.,

1996; Magnusson et al., 1996) , or retrograde labeling (Kosinski et al., 1988; Kemplay

68

and Webster, 1989; Magnusson et al., 1996; Li and Mizuno, 1997; Bermejo et al., 2003;

Del Cano et al., 2004; Wree et al., 2005) of GN projection neurons from GN projection

targets , have not observed subdivisions in the dorsal GN. Although all these approaches

are not all specific for presynaptic termination organization, they appear consistent with

the possibility that neuropil in the rat dorsal GN where the toes terminate does not

contain structurally defined subdivisions. These GN findings differ from the distinct

parcellated patterns of CO patches that are associated with input terminations from

different forepaw toes in the CN subdivision of the DCN (Crockett et al., 1993b), and that

were seen in the present study on the same CO sections where the GN was uniformly

stained. Similarly, the present GN findings also differ from the CO patches and

parcellated patterns of input terminations from different vibrissae in some parts of the

adjacent brainstem trigeminal nuclei (Chiaia et al.,1992a; Chiaia et al., 1992b; Jacquin et al., 1993c; Chiaia et al., 1994; Florence & Lakshman, 1995).

The present finding of overlapping organization of toe primary inputs in the GN appears to be a further extension of overlapping organization of primary neuron inputs from toes at pre-brainstem locations. For example, there is evidence that small numbers of primary neurons, including touch input neurons, have distal axons that branch

to end in more than one toe (Lynn and Carpenter, 1982; Woolf, 1987). The cell bodies of

primary neurons that have distal axons that innervate different hindpaw toes appear to be

arranged in a partly overlapping manner in the lumbar ganglia (Puigdellivol-Sanchez et

al., 2000). Injections and transport of label from toes have shown that terminations from

up to 3-4 toes can overlap in the spinal dorsal horn (Maslany et al., 1992; Levinsson et

69 al., 2002; Takahashi et al., 2003). Finally, studies where individual primary sensory fibers were labeled suggest that individual axons span long rostrocaudal extents in the dorsal horn and that terminal arbors of axons from different toes can overlap (Shortland et al., 1989; Shortland & Woolf, 1993). Taken together, these results suggest primary neuron inputs from different toes on the rat hindpaw have features of overlapping organization at peripheral and spinal levels. The present results supplement this understanding by suggesting that primary neuron inputs from different toes have further overlapping features of organization at a higher central level in the GN.

How the present results supplement current understanding of GN functional RFs on the toes

A second interest of this thesis was to identify numbers of toes that are normally included GN excitatory RFs in rats. As the thesis was begun there was little understanding of this issue. There is a longstanding recognition that many GN neurons have an excitatory tactile RF that is located on some hindquarter area, including toe areas.

Previous work on sizes of GN excitatory RFs in rats have expressed RF size in terms of skin area measurements (McComas, 1963; Suzuki & Dickenson, 2002; Miki et al., 1998;

Pettit & Schwark, 1993; Wang & Wall, 2005), but there have been no studies directed at estimating numbers of toes that usually contribute to GN excitatory RFs. Anecdotal illustrations of a limited number of rat GN RFs on toes have been included in reports on other issues (Nord, 1967; Berkley and Hubscher, 1995; Panetsos et al., 1995; 1997;

Kondo et al., 2002; Costa-Garcia and Nunez, 2004). Illustrated RFs include RFs that

70 involve a fraction of one toe to RFs that involve parts of up to 5 toes, thus, suggesting excitatory RFs on the toes can vary in size. Because only a few RFs have been shown, and because available RFs come from different studies, these results do not provide estimates or conclusions about how many toes are included in most GN RFs on the toes.

The sample of excitatory RFs in the present thesis was derived from studies in this lab that defined normal hindpaw RFs from single and multiple unit responses as part of obtaining control data prior to doing subsequent experimental manipulations (Wang & Wall, 2005; 2006). Many of the RFs from these studies involved the toes, and it was these RFs which served as the present sample. A relevant note of interest is that because these RFs were defined as part of other studies, they were defined

“blindly” with respect to their use in this thesis, i.e. to count number of toes included in

RFs. Thus, the numbers of toes per RF in the present sample are not “biased” by a prior recognition of the structural results of this thesis. The resulting sample of GN RFs suggests that a small percentage of RFs can include up to 4-5 toes, but the majority involve 1-2 toes, and the mean number of toes was 2. These findings are consistent with illustrated RFs from previous studies, which suggested RFs could involve different numbers of toes. These results supplement current understanding by indicating that a majority of GN neurons normally appear to be activated by 1-2 toes at any particular time. Additionally, a smaller fraction of GN neurons or, alternatively, some neurons at a particular time, may be activated by inputs from a larger number of toes.

71

Some broader implications of the findings of this thesis

As discussed above, the findings of this thesis supplement current

understanding of how terminations of primary neurons from the toes are normally

structurally organized in the GN, and of the number of toes that normally contribute to

GN functional RFs. Some implications of these results are discussed next.

A. Possible mismatches in GN structure and function and a potential explanation

The present study provides new evidence that terminations of primary

neuron inputs from 3-4 or more toes overlap extensively, but that functional RFs usually reflect inputs from 1-2 toes. This raises the possibility of structural-functional mismatching whereby availability of primary neuron inputs from toes may often exceed normal functional expression of inputs in toe RFs at a particular time.

How can these potential mismatches be explained? GN inhibitory interneurons inhibit projection neurons, other GN inhibitory interneurons, and terminations of primary neuron inputs (Lue et al., 1993; Lue et al,. 1994; Lue et al., 1996;

Aguilar et al., 2002) (Fig.1). Thus, from known structural connectivity, primary neuron

terminations and synapses on GN neurons are inhibited to some degree by pre- and post-

synaptic actions of GN inhibitory interneurons. GN interneurons, in turn, receive inputs

from primary neurons and from central structures including cortex (Lue et al., 1997; Lue

et al., 2001; Willis and Coggeshall 2004; Popratiloff et al., 1996a; Malmierca et al.,

2004). Many GN interneurons are GABAergic, and pharmacological blockade studies of 72

GN inhibition with the GABAA receptor antagonist bicuculline methiodide indicate GN hindpaw RFs, including RFs on the toes, immediately and reversibly enlarge across larger numbers of toes and other hindpaw areas following local microinjection of bicuculline methiodide around recording sites where GN RFs were being recorded (Schwark et al.,

1999; Wang et al., 2006). This suggests that the spatial extents of excitatory RFs of GN neurons are tonically controlled by intrinsic GN inhibition and that decreases in this inhibition can lead to immediate functional expression of inputs from a broader area of the hindpaw, including a larger number of toes. Tonic GN inhibition in turn, may in part be driven by direct peripheral inputs (e.g. tonically active proprioceptive inputs that signal body position and movement) or peripheral inputs that relay in spinal circuits.

Overall, known connectivity and pharmacological evidence suggests that local GN inhibition plays an important role in confining normal spatial extents of GN toe RFs.

Axon terminations of touch primary neurons provide a major source of input to the GN in rats. Primary neurons synapse on and directly activate GN inhibitory interneurons and projection neurons (Willis & Coggeshall 2004). These major, direct connections appear to be a main substrate for generation of GN excitatory RFs on the toes because acute disruption of primary neuron transmission from parts of a toe or hindpaw RF as a consequence of local anesthetic blockade causes immediate loss of parts of the RF innervated by disrupted inputs (Panetsos et al., 1997; Wang & Wall, 2005).

Another direct input to the rat GN originates from cortex but, in contrast to effects of disruptions of peripheral inputs, sizes of GN RFs on the toes and hindpaw remain unchanged before and after acute removal of cortical inputs to the GN (Wang & Wall,

73

2005). These and other findings have led to suggestions that GN excitatory RFs on the toes and hindpaw are dominantly determined by ascending primary neuron projections, rather than descending central projections (Wang & Wall, 2005, 2006).

Taken with the above findings on GN connectivity, the immediate enlargements of GN toe RFs with pharmacological blockades of GN GABA inhibition, and the major excitatory input influence that primary neurons provide to the GN, the present results raise the possibility that tonic GN inhibition of functional expression of some fraction of structurally available toe inputs may normally contribute to mismatches between structurally available, and functionally expressed, toe inputs.

B. Some features of cortical and GN processing may be similar

The GN nucleus is a main brainstem processing center for touch and other inputs within the dorsal column medial lemniscal ascending system. Following GN processing, information in this system is subsequently sent to thalamic neurons in the ventroposterior lateral nucleus which, in turn, sends thalamocortical projections to somatosensory cortex. Previous studies have indicated that thalamocortical arbors can transverse large distances across cortex. When the cortical distances of thalamocortical arbors were compared to cortical spaces in which neurons had functional RFs on a specific body location, like a digit, it appeared that functional RFs normally reflected inputs from more delimited areas of the body than reflected by the expansive thalamocortical arbor organization (Garraghty et al., 1989; Arnold et al.,2001; Rausell &

Jones, 1995; Rausell et al. 1998; Garraghty et al. 1990; Churchill et al. 2004; for reviews

74

see Snow and Wilson 1991). Consistent with this mismatch in thalamocortical input

availability and spatial extent of cortical RFs, pharmacological blockades of local cortical

inhibition lead to immediate enlargements of cortical RFs (Alloway & Burton, 1991;

Dykes et al., 1984; Chowdhury & Rasmusson, 2002). These and other findings have led

to suggestions that spatial extents of normal cortical receptive fields are limited by tonic

cortical inhibition and that disruption of this inhibition causes immediate RF

enlargement, in part, via drive of normally suppressed thalamocortical inputs (Calford,

2002; Snow & Wilson, 1991)

An implication of the present findings is that the GN organization of

broader ascending structural inputs than are expressed in functional RFs appears similar

to cortical organization. In concept, the present findings raise the possibility that GN organization resembles cortical organization in this respect.

C. Potential contribution to plasticity of GN RFs

GN tactile RFs on the hindpaw and toes of adult rats enlarge within

minutes following disruptions of primary neuron inputs by subcutaneous local anesthetic blockades of inputs (Panetsos et al., 1997; Wang & Wall, 2005). Further studies have shown that enlargement of an individual GN RF on the toes or hindpaw following these blockades spatially resembles the enlargement that is produced in the same RF following acute pharmacological blockade of local GN GABA inhibition with bicuculline methiodide (Wang & Wall, 2006). Immediate changes in GN RFs after peripheral anesthetic blockades of inputs are paralleled by similar concurrent changes in cortical

75

RFs (Byrne et al., 1991; Calford et al., 1991; Panetsos et al., 1995). These cortical

changes are not required for GN changes because acute removal of cortical descending

inputs to the GN does not block GN RF changes (Wang & Wall, 2005). These findings have led to proposals that immediate enlargements in GN RFs after acute disruptions of

peripheral inputs may be caused by decreases in local GN inhibition that do not require

descending cortical inputs (Panetsos et al., 1997; Pettit & Schwark, 1993; Schwark et al.,

1999; Wang & Wall, 2005, 2006).

GN RF enlargements after disruptions of peripheral inputs or GN GABA

inhibition occur within minutes and are clearly dependent on existing structural

connections. Thus a final implication of the present findings is that terminal organization

of primary neuron inputs from toes may serve as a structural substrate that contributes to

immediate plasticity of normal toe RFs after these disruptions, and that provides the GN

with an ability to rapidly adapt RFs to input conditions.

76

Bibliography

Aguilar, J., Canedo, A., Rivadulla, C., and Soto, C. (2002). The lemniscal-cuneate

recurrent excitation is suppressed by strychnine and enhanced by gabaa

antagonists in the anaesthetized cat. Eur. J. Neurosci 16, 1697-1704.

AlChaer, E., Lawand, N.B., Westlund, K.N., and Willis, WD., (1996 ). Pelvic visceral

input into the nucleus gracilis is largely mediated by the postsynaptic dorsal

column pathway. Journal Of Neurophysiology 76, 2675-2690.

Arnold, P., Li, C.X., and Waters, RS. (2001). Thalamocortical arbors extend beyond

single cortical barrels, an in vivo intracellular tracing study in rat. Exp Brain Res

136, 152-68.

Asanuma , C., W.T. Thatch, and E.G. Jones (1983). Brainstem and spinal projections of

the deep cerebellar nuclei in the monkey, with observations on the brainstem

projections of the dorsal column nuclei. Brain research 5, 299-322.

Berkley, K., and Hubscher, C.H. (1995). Are there separate central nervous system

pathways for touch and pain?. Nature Medicine 1, 766-773.

Bermejo, P., Jiménez, C.E., Torres, C.V., and Avendaño, C. (2003). Quantitative

stereological evaluation of the gracile and cuneate nuclei and their projection

neurons in the rat. J Comp Neurol 463, 419-33.

Burton, H., and Alloway, K. D. (1991). Differential effects of GABA and bicuculline on

rapidly- and slowly- adapting neurons in primary somatosensory cortex of

primates. . Exp Brain Res 85, 598-610.

77

Byrne, J., and Calford, M.B. (1991). Short-term expansion of receptive fields in rat

primary somatosensory cortex after hindpaw digit denervation. Brain Research

565, 218-24.

Calford, M. ( 2002). Dynamic representational plasticity in sensory cortex. Neuroscience.

111, 709-38. .

Calford, M., and Tweedale, R. (1991). Immediate expansion of receptive fields of

neurons in area 3b of macaque monkeys after digit denervation. Somatosens Mot

Res 8, 249-60.

Chiaia, N., Bennett-Clarke, C.A., Eck, M., White, F.A., Crissman, R.S., and Rhoades,

R.W. (1992a). Evidence for Prenatal Competition among the Central Arbors of

Trigeminal Primary Afferent Neurons. The Journal of Neuroscience 12, 62-76.

Chiaia, N., Bennett-Clarke, C.A., and Rhoades, R.W. (1992b). Differential-effects of

peripheral damage on vibrissa-related patterns in trigeminal nucleus-principalis,

subnucleus interpolaris, and subnucleus caudalis. Neuroscience. 49, 141-56.

Chiaia, N.L., Zhang, S., King, T.D., Rhoades, R.W.(1994). Evidence for prenatal

competition among the central arbors of trigeminal primary afferent neurons:

single axon analysis.J Comp Neurol., 345, 303-13.

Chimelli, L., Bello, A.R., and Scaravilli, F. (1994). Development of cortical fibres into

the gracile nucleus of the rat, an ultrastructural study. Dev Neurosci 16, 260-6.

Cliffer, K.D. and Willis, W.D. (1994). Distribution of the postsynaptic dorsal column

projection in the cuneate nucleus of monkeys. The Journal of Comparative

Neurology, 345, 84-93.

78

Churchill, J., Tharp, J.A., Wellman, C.L., Sengelaub, D.R., and Garraghty, P.E. (2004).

Morphological correlates of injury-induced reorganization in primate

somatosensory cortex. BMC Neurosci 5, 43.

Conti, F., De Biasi, S., Giuffrida, R., Rustioni, A. (1990). Substance P-containing

projections in the dorsal columns of rats and cats. Neuroscience 34:607-621.

Costa-Garcia, M. and. Nunez, A. (2004). Nociceptive stimuli induce changes in

somatosensory responses of rat dorsal column nuclei neurons. Brain Research,

1025, 169-176.

Crockett, D.P., Foschini, D.R., Girgis, W.S., Egger, M.D. (1993a). Immunocytochemical

localization of the low-affinity nerve growth factor receptor (p75NGFR) in the

cuneate nucleus of the rat and its relationship to cytochrome-oxidase activity.

Brain Res, 603, 324-327.

Crockett DP, Foschini DR, Girgis WS, Egger MD. 1993a. Immunocytochemical

localization of the low-affinity nerve growth factor receptor (p75NGFR) in the

cuneate nucleus of the rat and its relationship to cytochrome-oxidase activity.

Brain Res 603:324-327.

Crockett, D.P., Maslany, S., Egger, M.D. (1996). Synaptophysin immunoreactivity and

distributions of calcium-binding proteins highlight the functional organization of

the rat's dorsal column nuclei. Brain Res, 707, 31-46.

De Biasi, S., Vitellaro-Zuccarello, L., Bernardi, P., Valtschanoff, J.G. and Weinberg, R.J.

(1995). Ultrastructural and immunocytochemical characterization of terminals of

79

postsynaptic ascending dorsal column fibers in the rat cuneate nucleus. J. Comp.

Neurol. 353,109-118.

Del Cano, G.G., Uria, I., Gerrikagoitia, I., Martinez-Millan, L. (2004). Connection from

the dorsal column nuclei to the superior colliculus in the rat: topographical

organization and somatotopic specific plasticity in response to neonatal

enucleation. J Comp Neurol, 468,410-424.

Desbois, C., Le Bars, D., and Villanueva, L. (1999). Organization of cortical projections

to the medullary subnucleus reticularis dorsalis, A retrograde and anterograde

tracing study in the rat. J Comp Neurol. 410, 178-96.

Ding, Y., Shi, J., Wang, D.S., Xu ,J.Q., Li, J.L., and Ju, G. ( 1999). Primary afferent

fibers of the pelvic nerve terminate in the gracile nucleus of the rat. Neurosci Lett.

272, 211-4.

Dykes, R., Landry, P., Metherate, R., and Hicks, T.P. (1984). Functional role of GABAB

in Cat primary somatosensory cortex, Shaping receptive fields of cortical neurons.

Journal of Neurophysiology 52, 1066-1093.

Faggin, B., Nguyen, K.T., and Nicolelis, M.A.L. (1997). Immediate and simultaneous

sensory reorganization at cortical and subcortical levels of the somatosensory

system. Proc Natl Acad Sci USA 94, 9428-9433.

Florence (2002). The Changeful Mind: Plasticity in the Somatosensory System. In:

Nelson, R J., ed., The Somatosensory System. Boca Raton, CRC Press, pp. 335-

366.

80

Florence, S., and Lakshman, S. (1995). Topography of primary afferent projections in the

trigeminal sensory nuclei of rats. Acta Neurobiol Exp (Wars) 55, 193-200.

French, A.S., Dick, S.H. and Rasmusson, D.D. (2002). Postsynaptic dorsal column and

cuneate correlations in the raccoon: A re-evaluation by parallel-cascade analysis.

J. Neurophysiol. 88, 3372-3376.

Garraghty, P., Pons, T.P., Sur, M., and Kaas, J.H. (1989). The arbors of axons

terminating in middle cortical layers of somatosensory area 3b in Owl Monkeys.

Somatosensory and motor Research 6, 401-411.

Garraghty, P.E., and Sur, M. (1990). Morphology of single intracellularly stained axons

terminating in area 3b of macaque monkeys. J Comp Neurol 294(4), 583-93.

Grant, G. (1993). Projection patterns of primary sensory neurons studied by

transganglionic methods, somatotopy and target-related organization. Brain Res

Bull 30, 199-208.

Gulley, R.L. (1973). Golgi studies of the nucleus gracilis in the rat. Anat Rec, 177, 325-

42.

Ha, S.O., Kim, J.K., Hong, H.S., Kim, D.S., Cho, H.J. (2000). Expression of brain-

derived neurotrophic factor in rat dorsal root ganglia, spinal cord, and gracile

nuclei in experimental models of neuropathic pain. Neuroscience, 107, 301-309.

Hwang, S., Rustioni, A., and Valtschanoff, J.G. (2001). Kainate receptors in primary

afferents to the rat gracile nucleus. Neurosci Lett. 312, 137-40.

81

Jacquin, M.F., Renehan, W.E., Rhoades, R.W., Panneton, W.M.(1993). Morphology and

topography of identified primary afferents in trigeminal subnuclei principalis and

oralis. J Neurophysiol, 70, 1911-36.

Jones, E. (2000). Cortical and subcortical contributions to activity-dependent plasticity in

primate somatosensory cortex. Annual review of Neuroscience 23, 1-37.

Kaas, J.H. (2002.). Functional implications of plasticity and reorganizations in the

somatosensory and motor systems of developing and adult primates. In: Nelson,

RJ., ed., The Somatosensory System. Boca Raton, CRC Press, pp. 367-381.

Kaas, J. (2004). Somatosensory system. In: Paxinos, G., ed., The Human Nervous

System. San Diego, Elsevier Academic Press, pp. 1059-92.

Kemplay, S., and Webster, K.E. (1989). A quantitative study of the projections of the

gracile, cuneate and trigeminal nuclei and of the medullary reticular formation to

the thalamus in the rat. Neuroscience. 32, 153-67.

Kondo, E., Iwata, K., Ogawa, A., Tashiro, A., Tsuboi, Y., Fukuoka, T., Yamanaka, H.,

Dai, Y., Morimoto, T., and Noguchi, K. (2002). Involvement of glutamate

receptors on hyperexcitability of wide dynamic range neurons in the gracile

nucleus of the rats with experimental mononeuropathy. Pain 95, 153-63.

Kosinski, R.J., Lee, H.S., Mihailoff, G.A. (1988). A double retrograde fluorescent tracing

analysis of dorsal column nuclear projections to the basilar pontine nuclei,

thalamus, and superior colliculus in the rat. Neurosci Lett, 85, 40-46.

82

Krupa, D., Ghazanfar, A.A., and Nicolelis, M.A.L. (1999 ). Immediate thalamic sensory

plasticity depends on corticothalamic feedback. Proc Natl Acad Sci U S A 96,

8200-8205.

LaMotte, C., Kapadia, S.E., and Shapiro, C.M. (1991). Central projections of the sciatic,

saphenous, median, and ulnar nerves of the rat demonstrated by transganglionic

transport of choleragenoid-HRP (B-HRP) and wheat germ agglutinin-HRP

(WGA-HRP). J Comp Neurol 311, 546-62.

Leem, J., Willis, W.D., and Chung, J.M. (1993). Cutaneous sensory receptors in the rat

foot. J Neurophysiol 69, 1684-99.

Leong, S., and Tan, C.K (1987). Central projection of rat sciatic nerve fibres as revealed

by Ricinus communis agglutinin and horseradish peroxidase tracers. J Anat 154,

15-26.

Levinsson, A., Holmberg, Hans., Broman, Jonas., Zhang, Mengliang., and Schouenborg,

Jens. (2002). Spinal Sensorimotor Transformation, Relation between Cutaneous

Somatotopy and a Reflex Network. The Journal of Neuroscience 22, 8170–8182.

Li, W-P., Xian, C., Rush, R.A., Zhou, X-F. (1999). Upregulation of brain-derived

neurotrophic factor and neuropeptide Y in the dorsal ascending sensory pathway

following sciatic nerve injury in rat. Neurosci Lett, 260, 49-52.

Lue, J., Chen, K.N., Shieh, J.Y., and Wen, C.Y. (1993). Synaptic relationships between

GABA-immunoreactive boutons and primary afferent terminals in the rat cuneate

nucleus. Neuroscience. 56, 973-9.

83

Lue, J., Jiang-Shieh, Y.F., Shieh, J.Y., and Wen, C.Y. (1996). The synaptic

interrelationships between primary afferent terminals, cuneothalamic relay

neurons and GABA-immunoreactive boutons in the rat cuneate nucleus. Neurosci

Res 24, 363-71.

Lue, J., Chen, S.H., Shieh, J.Y., and Wen, C.Y. (2000). Cuneothalamic relay neurons are

postsynaptic to glycine-immunoreactive terminals in the rat cuneate nucleus.

Synapse 37, 222-31.

Lynn, B., and Carpenter S.E (1982). Primary afferent units from the hairy skin of the rat

hind limb. Brain Research 238, 29-43.

Ma, W., Bisby, M.A. (1999). Ultrastructural localization of increased neuropeptide

immunoreactivity in the axons and cells of the gracile nucleus following chronic

constriction injury of the sciatic nerve. Neuroscience, 93, 335-348.

Ma, W., Bisby, M.A. (2000). Partial sciatic nerve ligation induced more dramatic

increase of neuropeptide Y immunoreactive axonal fibers in the gracile nucleus of

middle-aged rats than in young adult rats. J Neurosci Res, 60, 520-530.

Ma, W., Chabot, J.G., Powell, K.J., Jhamandas, K., Dickerson, I.M., Quirion, R. (2003).

Localization and modulation of calcitonin gene-related peptide-receptor

component protein-immunoreactive cells in the rat central and peripheral nervous

systems. Neuroscience, 120, 677-694.

Magnusson, A., Dahlfors, G., Blomqvist, A. (1996). Differential distribution of calcium-

binding proteins in the dorsal column nuclei of rats: a combined

84

immunohistochemical and retrograde tract tracting study. Neuroscience, 73, 497-

508.

Malmierca, E., and Nuñez, A. (2004). Primary somatosensory cortex modulation of

tactile responses in nucleus gracilis cells of rats. Eur J Neurosci 19, 1572-80.

Maslany, S., Crockett, D.P., and Egger, M.D. (1991). Somatotopic organization of the

dorsal column nuclei in the rat, transganglionic labelling with B-HRP and WGA-

HRP. Brain Research 564, 56-65.

Maslany, S., Crockett, D.P., and Egger, M.D. (1992). Organization of cutaneous primary

afferent fibers projecting to the dorsal horn in the rat, WGA-HRP versus B-HRP.

Brain Research 569, 123-35.

McComas, A. J. (1963). Responses of the rat dorsal column system to mechanical

stimulation of the hind paw. J Physiol. 166, 435–448.

Miki, K., Iwata, K., Tsuboi, Y., Sumino, R., Fukuoka, T., Tachibana, T., Tokunaga, A.,

and Noguch,i K. (1998). Responses of dorsal column nuclei neurons in rats with

experimental mononeuropathy. Pain 76, 407-15.

Nicolelis, M., Lin, R.C.S., Woodward, D.J., and Chapin, J.K. (1993). Induction of

immediate spatiotemporal changes in thalamic networks by peripheral block of

ascending cutaneous information. Nature 361, 533-536.

Noguchi, K., Kawai, Y., Fukuoka, T., Senba, E., Miki, K. (1995). Substance P induced

by peripheral nerve injury in primary afferent sensory neurons and its effect on

dorsal column nucleus neurons. J Neurosci, 15, 7633-7643.

85

Nord, S. G. (1967). Somatosensory organization in the spinal trigeminal nucleus, the

dorsal column nuclei and related structures in the rat. J Comp Neurol 130, 343-

356.

Novikov LN. 2001. Labeling of central projections of primary afferents in adult rats: a

comparison between biotinylated dextran amine, neurobiotin, and Phaseolus

vulgaris-leucoagglutinin. J Neurosci Methods 112:145-154.

Ossipov, M.H., Zhang, E-T., Carvajal,C., Gardell, L., Quirion, R., Dumont ,Y., Lai, J.,

Porreca, F. (2002). Selective mediation of nerve injury-induced tactile

hypersensitivity by neuropeptide Y. J Neurosci .,15, 9858-9867.

Panetsos, F., Avendaño, C., and Nunez, A. (1995). Local anaesthesia induces immediate

receptive field changes in nucleus gracilis and cortex. Neuroreport 7, 150-152.

Panetsos, F., Avendaño, C., and Nunez, A.. (1997). Electrophysiological effects of

temporary deafferentation on two characterized cell types in the nucleus gracilis

of the rat. European Journal of Neuroscience 9, 563-572.

Pettit, M., and Schwark, H.D. (1993 ). Receptive field reorganization in dorsal column

nuclei during temporary denervation. Science 262, 2054-2056.

Popratiloff, A., Valtschanoff, J.G., Rustioni, A. and Weinberg, R.J (1996a).

Colocalization of gaba and glycine in the rat dorsal column nuclei. Brain

Research 706, 308-312.

Prats-Galino, A., Puigdellívol-Sánchez, A., Ruano-Gil, D., and Molander, C. (1999).

Representations of hindlimb digits in rat dorsal root ganglia. J Comp Neurol 408,

137-45.

86

Puigdellívol-Sánchez, A., Forcada-Calvet, P., Prats-Galino, A., and Molander, C. (2000).

Contribution of femoral and proximal sciatic nerve branches to the sensory

innervation of hindlimb digits in the rat. Anat Rec 260, 180-8.

Rasmusson, D. D., and Chowdhury S.A. (2002). Effect of GABA receptor blockade on

receptive fields of raccoon somatosensory cortical neurons during reorganization.

Exp Brain Res 145, 150-157.

Rausell, E., Bickford, L., Manger, P.R., Woods, T.M., and Jones, E.G. (1998). Extensive

divergence and convergence in the thalamocortical projection to monkey

somatosensory cortex. J Neurosci 18, 4216-32.

Rausell, E., and Jones, E.G (1995). Extent of intracortical arborization of thalamocortical

axons as a determinant of representational plasticity in monkey somatic sensory

cortex Journal of Neuroscience 15, 4270-4288.

Reinke, H., and Dinse, H.R. (1996). Functional characterization of cutaneous

mechanoreceptor properties in aged rats. Neurosci Lett. 216, 171-4.

Rivero-Melián, C., and Arvidsson, J. (1992). Brain stem projections of rat lumbar dorsal

root ganglia studied with choleragenoid conjugated horseradish peroxidase. Exp

Brain Res 91, 12-20.

Robertson, B., and Grant, G. (1985). A comparison between wheat germ agglutinin and

choleragenoid-horseradish peroxidase as anterogradely transported markers in

central branches of primary sensory neurons in the rat with some observations in

the cat. Neuroscience. 14, 895-905.

87

Sanders, K., and Zimmermann, M. (1986). Mechanoreceptors in rat glabrous skin,

redevelopment of function after nerve crush. J Neurophysiol 55, 644-59.

Schwark, H.D., Petit, M.J., Fuchs, J.L. (1998). Distribution of substance P receptor

binding in dorsal column nuclei of rat, cat, monkey, and human. Brain Res 786,

259-262.

Schwark, H.D., Tennison, C.F., Ilyinsky, O.B., Fuchs, J.L. (1999). Inhibitory influences on receptive field size in the dorsal column nuclei. Exp Brain Res 126, 439-42.

Shehab, S., Spike, RC., and Todd, AJ. (2003). Evidence against cholera toxin B subunit

as a reliable tracer for sprouting of primary afferents following peripheral nerve

injury. Brain Research 964, 218-27.

Shin, H.-C., Park, S., Son, J., and Sohn, J-H. (1995). Responses from new receptive fields

of VPL neurones following deafferentation. Neuroreport 7, 33-36.

Shortland, P., Fitzgerald, M., and Woolf, CJ. (1989). Morphology and somatotopic

organization of the central terminals of hindlimb hair follicle afferents in the rat

lumbar spinal cord. J Comp Neurol. 289, 416-33.

Shortland, P., and Woolf, CJ. (1993). Morphology and somatotopy of the central

arborizations of rapidly adapting glabrous skin afferents in the rat lumbar spinal

cord. J Comp Neurol. 329, 491-511.

Snow, P.J., and Wilson, P. (1991). Plasticity in the somatosensory system of developing

and mature mammals- The effects of injury to the central and peripheral nervous

88

system. In: Autrum, H., Ottoson, D., Perl, ER., Schmidt, RF., Shimazu, H., Willis,

WD., eds., Sensory Physiology. Berlin, Springer-Verlag, pp. 225-285 & 403-482

Suzuki, R., and Dickenson, A.H. (2002). Nerve injury-induced changes in opioid

modulation of wide dynamic range dorsal column nuclei neurones. Neuroscience.

111, 215-28.

Takahashi, Y., Aoki Y., Chiba, T., and Kurokawa, M. (2003). Dermatomes and the

Central Organization of Dermatomes and Body Surface Regions in the Spinal

Cord Dorsal Horn in Rats. The Journal Of Comparative Neurology 462, 29–41.

Tong, Y., Wang, H.F., Ju, G., Grant, G., Hökfelt, T., and Zhang, X. (1999). Increased

uptake and transport of cholera toxin B-subunit in dorsal root ganglion neurons

after peripheral axotomy, possible implications for sensory sprouting. J Comp

Neurol 404, 143-58.

Tracey, D. (2004). Somatosensory system. In: Paxinos, G., ed., The rat Nervous system.

San Diego, Elsevier Academic Press, pp. 797-815.

Ueyama, T., Houtani, T., Ikeda, M., Sato, K., Sugimoto, T., and Mizuno, N. (1994).

Distribution of primary afferent fibers projecting from hindlimb cutaneous nerves

to the medulla oblongata in the cat and rat. J Comp Neurol. 341, 145-58.

Valverde, F. (1966). The pyramidal tract in Rats. A study of its relations with the

posterior column nuclei, dorsolateral reticular formation of the medulla oblongata,

and cervical spinal cord. Z Zellforsch Mikrosk Anat 71, 298-363.

Wall, J., Xu, J., and Wang, X. (2002a). plasticity, an emerging view of the

multiple substrates and mechanisms that cause cortical changes and related

89

sensory dysfunctions after injuries of sensory inputs from the body Brain

Research reviews 39, 181-215

Wang, X. and Wall, J.T. (2005). Cortical influences on sizes and rapid plasticity of tactile

receptive fields in the dorsal column nuclei. Journal of Comparative Neurology.

489, 241-248.

Wang, X., Wall, J.T.(2006). Cortical influences on rapid brainstem plasticity.Brain Res.,

1095, 73-84.

Weinberg, R., Pierce, J.P., and Rustioni, A. (1990). Single fiber studies of ascending

input to the cuneate nucleus of cats: I. Morphometry of primary afferent fibers. J

Comp Neurol 300, 113-33.

Willis, W., and Coggeshall, R.E. (2004). Sensory mechanisms of the spinal cord. New

York, Kluwer Academic.

Willis, WD., and Westlund, K.N. (2004). Pain system. In: Paxinos, G., ed., The Human

Nervous System. San Diego, Elsevier Academic Press, pp. 1125-1170.

Woolf, C. (1987). Central terminations of cutaneous mechanoreceptive afferents in the rat

lumbar spinal cord. J Comp Neurol 261, 105-19.

Wree, A., Itzev, D.E., Schmitt, O., Usunoff, K.G. (2005). Neurons in the dorsal column

nuclei of the rat emit a moderate projection to the ipsilateral ventrobasal thalamus.

Anat Embryol (Berl) 210, 155-162.

90

Abstract

The gracile nucleus (GN) is a major brainstem processing center for

somatosensory inputs. GN neurons often have functional receptive fields (RFs) on the

toes, but there is little consensus on how these RFs are related to primary sensory inputs from toes. This project compared structural overlap in GN terminations of afferents from different toes, and numbers of toes in GN functional RFs. The results suggest that

structural convergence of afferent terminations from at least 3-4 toes is common; in

contrast, although some GN RFs reflected functional inputs from up to 4-5 toes, most

RFs reflected inputs from 1-2 toes. These findings raise the possibility that structural

convergence of sensory afferents from the toes is broader than the afferent convergence

that is usually reflected in functional RFs. This type of potential mismatch has been

previously described in somatosensory cortex and may partly explain rapid GN RF

changes after disruptions of afferents.

91