MOTOR UNIT PROPERTIES FOLLOWING CROSS-REINNERVATION
OF CAT TRICEPS SURAE MUSCLES.
ROBERT C. FOEHRING
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
I would like to thank the members of my supervisory committee, Drs.
John Munson, Don Stehouwer, George Sypert and Floyd Thompson, for their
advice and interest throughout the course of my dissertation project.
Dr. Janet Zengel was extremely helpful at several stages of this
project. Expert technical assistance was provided by William Ireland,
C.J. Thomas and Audrey Kalehua.
I am especially grateful to Drs. John Munson and George Sypert, who
have been excellent teachers, admirable role models, and great company
for the past few years.
Deb and Christen deserve special thanks for putting up with the life
of a graduate student, and making this all seem worthwhile.
Finally, I thank all those responsible for the microcomputer.
I have been supported as a graduate research associate on an
National Institute for Neurological and Communicative Diseases and
Stroke (NINCDS) grant to Dr. John Munson. The W.L. Gore corporation
donated Gore-Tex for use as nerve-sleeves.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ....... .................................... ......
LIST OF TABLES ..... ............................. ... ........ ......
LIST OF FIGURES...................................................
ABSTRACT ............... ....... ..................................
I. INTRODUCTION .................... ..........................
II. METHODS................... ........................... ..... .
Initial Surgery .........................................
Acute Experiments. .............................. .....
Whole Muscle Twitch.........................................
Motor Unit Studies.. ............ ................. ..........
Muscle Histochemistry .......................................
Statistical Considerations................ .................
III. NORMAL MG AND LONG-TERM REINNERVATION OF M3 ................
Introduction................... ......... .. ......... ......
Results ........................ ...........................
Discussion .... ......................................
IV. AXOTOMY AND THE TIME COURSE OF SELF-REINNERVATION OF MS....
Introduction................ .. .... .......................
Results. ....... ..... ............................... ......
Discussion............................ ..... .......... ......
V. PROPERTIES OF NORMAL LG AND SOLEUS.........................
Results............... .................. ...............
Discussion ............ ................. ...................
VI. CROSS-REINNERVATION OF LO AND SOLEUS MUSCLES BY MG MOTOR
NERVE: SELECTIVITY OF REINNERVATION AND MOTONEURON
INFLUENCE ON MUSCLE........................................ 117
Introduction ............................................. 117
VII. CROSS-REINNERVATION OF LG AND SOLEUS MUSCLES BY MG MOTOR
NERVE: MUSCLE INFLUENCE UPON MOTONEURONS.................... 149
Introduction .................... ..... ................... ... 149
Results............................... ..................... 151
Discussion................. ..... ....... ..... ............ 167
VIII. CONCLUSIONS .... ........................... .oo .... ....... 175
LITERATURE CITED ................ ...................... ......... 181
BIOGRAPHICAL SKETCH.......,........................................ 196
LIST OF TABLES
3-1 Whole Muscle Properties: Normal and Nine Month
Self-Reinnervated MG. ...... ............................. 19
3-2 Muscle Unit Contractile Properties (Long-Caged vs.
Non-Caged; Normal vs. Nine Month Self-Reinnervated)..... 22
3-3 Motoneuron Electrical Properties (Long-Caged vs.
Non-Caged; Normal vs. Nine Month Self-Reinnervated)..... 27
3-4 Differences Between Motor Unit Types: Normal vs. Nine
Month Self-Reinnervated MG.............................. 28
3-5 Data for Non-Contracts (Nine Month Self-Reinnervation
Model) ................ ... ..................... ....... 36
3-6 Motor Unit Types, Muscle Fiber Types and Innervation
Ratios (Normal and Nine Month Self-Reinnervated)........ 38
4-1 Whole Muscle Properties: Time Course of
4-2 Motoneuron Electrical Properties: Axotomy and Time
Course of Self-Reinnervation.............................. 56
4-3 Low-Re Motoneuron Electrical Properties: Contracts vs.
Non-Contract....................... ........... ..... 66
4-4 Motoneuron Electrical Properties of Non-Contracts:
Self-Reinnervation ................................... 66
4-5 Muscle Unit Contractile Properties: Time Course of
Self-Reinnervation ..................................... 68
4-6 Significance of Differences Between Motor Unit Types for
Med-Re Motor Units....................................... 71
4-7 Motor Unit Types, Muscle Fiber Types and Innervation
Ratios: Time Course of Self-Reinnervation.............. 76
5-1 Whole Muscle Properties: Normal MG, LG and Soleus....... 92
5-2 Percent Muscle Fiber Types by Innervation Compartment;
Normal LG.............................. ........ ....... 92
5-3 Muscle Unit Contractile Properties: Normal MG, LG and
Soleus................ ..................... ..... ..... 98
5-4 Muscle Fiber Areas By Innervation Compartment: Normal LO 100
5-5 Motoneuron Electrical Properties: Normal MG, LG and
5-6 Significance of Differences Between Motor Unit Types for
Normal MG ............. ............ .. ............... 101
5-7 Motor Unit Types, Muscle Fiber Types and Innervation
Ratios for Normal M3, LG and Soleus............. .... 110
6-1 Whole Muscle Properties: Cross-Reinnervation ........... 120
6-2 Percent Motor Unit Types................... ............ 120
6-3 Muscle Unit Contractile Properties: Normal vs.
Cross-Reinnervated ... ...... ........................... 127
6-4 Percent Muscle Fiber Types By Innervation Compartment
(Long-X)........ o ... .......... ...... ................ 131
6-5 Percent Muscle Fiber Types: Normal and
Cross-Reinnervated .................................. 137
6-6 Significance of Differences Between Motor Unit Types
(Long-X)......................... ... ................... 137
6-7 Significance of Differences Between Motor Unit Types
(Med-X) .............. ... ... ..... ......... ..... ...... 137
6-8 Motor Unit Types, Muscle Fiber Types and Innervation
Ratios (Long-X) ........................... ...... .... 140
7-1 Data for Non-Contracts (Long-X)........................ 152
7-2 Motoneuron Electrical Properties: Cross-Reinnervation... 154
7-3 Significance of Differences Between Motor Unit Types
7-4 Significance of Differences Between Motor Unit Types
(Med-X).......... ...................................... 164
LIST OF FIGURES
2-1 Cross-Reinnervation surgery............................ 7
3-1 Frequency histograms for muscle unit contractile
properties: normal and long self-reinnervated......... 24
3-2 Frequency histograms for motoneuron electrical
properties: normal and long self-reinnervated......... 30
3-3 Relationships between AHP half-decay time and axonal
conduction velocity: normal and long self-reinnervated
MG ................................................ 32
3-4 Relationships between rheobase and input resistance:
normal and long self-reinnervated MG................... 34
3-5 Photomicrographs of glycogen-depleted motor unit........ 42
4-1 Frequency histograms for motoneuron rheobase and input
resistance: normal vs. axotomy........................ 59
4-2 Frequency histograms for motoneuron AHP half-decay time
and axonal conduction velocity: normal vs. axotomy.... 61
4-3 Relationships between rheobase and input resistance:
axotomy, low-re and med-re............................ 63
4-4 Relationships between AHP half-decay time and axonal
conduction velocity: axotomy, low-re and med-re....... 65
4-5 Time course of recovery of motoneuron electrical
properties: self-reinnervation........................ 90
5-1 Frequency histograms for normal LG and soleus: muscle
unit contractile properties ........................... 96
5-2 Location of 31 motor units in normal LG................. 98
5-3 Frequency histograms for normal LG and soleus:
motoneuron properties................................... 105
5-4 Relationships between rheobase and input resistance
in normal LG and soleus motoneurons.................. 108
5-5 Relationships between AHP half-decay time and axonal
conduction velocity in normal LG and soleus
motoneurons...... ................................... 110
5-6 Photomicrographs of normal LG and soleus muscle
histochemistry ........................ ................. 114
6-1 Frequency histograms for normal LG and longX-LG
muscle unit contractile properties.................... 131
6-2 Frequency histograms for normal soleus and longX
soleus muscle unit contractile properties.............. 135
7-1 Frequency histograms for electrical properties of
self-regenerated MG motoneurons, and MG motoneurons
which innervated longX-LG.................. ............ 158
7-2 Relationship between AHP half-decay time and axonal
conduction velocity in MG motoneurons which innervated
LG or soleus.......... ...... ......... ..... .. ....... 160
7-3 Relationships between rheobase and input resistance in
M3 motoneurons which innervated LG or soleus.......... 162
7-4 Frequency histograms for electrical properties of
self-regenerated MG motoneurons, and MH motoneurons
which innervated longX-soleus............. ............ 164
7-5 Time course of recovery of motoneuron electrical
properties: self and cross-reinnervation............... 174
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
MOTOR UNIT PROPERTIES FOLLOWING CROSS-REINNERVATION
OF CAT TRICEPS SURAE MUSCLES
Robert C. Foehring
Chairman: Dr. John B. Munson
Major Department: Neuroscience
This study showed several aspects of the relationships between
cat triceps surae alpha-motoneurons, and the muscle fibers they
innervate. Relationships between motoneuron electrical properties and
muscle unit contractile properties were investigated for normal medial
gastrocnemius (MG), lateral gastrocnemius (LG) and soleus motor units.
The relationships between motoneuron and muscle fibers were similar in
quality and strength for normal LG and soleus and for normal MS.
Surgical section of the MG nerve, followed by re-anastomosis to MG nerve
(self-reinnervation), or to the combined LG-soleus nerve
(cross-reinnervation), was employed to examine motoneuron influence on
muscle phenotype, and effects of muscle on the expression of motoneuron
electrical properties. Functional connection to muscle was a necessary
condition for expression of normal, mature motoneuron electrical
properties. Long-term self-reinnervation resulted in complete recovery
of motoneuron electrical properties, muscle unit contractile properties,
and the relationships between them. Reinnervation of a foreign muscle
(LG), with similar original muscle fiber type composition to MG, also
resulted in complete recovery of these parameters. Muscle fibers
of soleus, an almost purely slow muscle, resisted the influence of Ho
motoneurons. Since MO motoneurons failed to convert soleus muscle
fibers, MG motoneuron properties were altered, thus maintaining a close
relationship between motoneuron electrical type and muscle unit
contractile type. This relationship was of similar strength to that
between normal and also self-reinnervated MG motoneurons and M3 muscle
fibers, but of different quality In MG motoneurons which innervated
soleus muscle, incomplete recovery from axotomy is a possible
explanation for the altered values for motoneuron rheobase, input
resistance and axonal conduction velocity. Afterhyperpolarization
potential (AHP) half-decay time of MG motoneurons which innervated
soleus muscle was longer than in MG motoneurons which innervated MH
muscle, whereas axotomy resulted in no change in mean AHP half-decay
time. Incomplete recovery cannot explain this finding. Thus type of
muscle innervated influences expression of motoneuron electrical
properties. A retrogradely transported chemical message from muscle is a
likely mediating mechanism.
The mammalian neuromuscular system has been an important model
system for study of how cells influence the expression of
properties of other cells. In particular, the expression of
muscle phenotype has been shown to be under strong neural influence.
Buller, Eccles, and Eccles surgically re-routed the nerve from the
'slow' soleus muscle into the 'fast' flexor hallucis longus muscle
(FHL), and vice versa, in young cats (cross-reinnervation; Buller et al.
1960). This study illustrated two major points. First, the type of
nerve influenced the contractile speed of the whole muscle; the 'fast'
FHL muscle became slow under soleus nerve influence, and the 'slow'
soleus muscle became faster, when innervated by FHL nerve. Second, the
conversion of 'fast' muscle by a 'slow' nerve was more complete than the
reverse. These initial observations have been repeated several times,
and at finer levels of resolution, including single motor unit studies
(see references cited in CHAPTER III). Collectively, these studies
provide strong evidence for neural regulation of muscle phenotype.
The electrical activity pattern of cross-reinnervated muscles is
that of the muscle originally innervated by the nerve (Sperry, 1945;
Cohen, 1978; Brinkman et al. 1983; Mulkey, 1983; O'Donovan et al., in
press ). That is, the activity pattern of the nerve is unchanged when
innervating a foreign muscle. This observation has led most workers to
regard motoneuron properties as uninfluenced by the particular muscle
innervated (e.g. Gordon, 1983). A similar conclusion was reached by Kuno
et al. (1974b), who measured action potential overshoot, axonal
conduction velocity, and afterhyperpolarization (AHP) duration in soleus
motoneurons, up to five months after self- or cross-reinnervation
of flexor digitorum longus muscle (FDL).
Relatively little is known, however, about the expression of
motoneuron electrical properties, and the relationships between them,
after reinnervation of a foreign muscle. A complication in early studies
was that motor units within a muscle were regarded as a single,
homogeneous population. Thus Kuno et al. regarded medial gastrocnemius
(M3) or FDL motoneurons as uniformly 'fast', and soleus motoneurons as
uniformly 'slow'. MG and FDL, although predominately 'fast', are
actually 'mixed' muscles, containing several types of motor unit, and
soleus has occasional fast units (reviewed in Burke, 1981). Changes may
be differential between motor unit types, following reinnervation.
This study utilizes the the cat triceps surae neuromuscular system
as an experimental model of motor unit properties. The triceps surae
includes the soleus, MG and lateral gastrocnemius muscles (LG), all of
which are ankle extensors. The motor unit is defined as an alpha
motoneuron and the muscle fibers it innervates, the muscle portion alone
being referred to as the muscle unit (Burke, 1981). MG motor units can
be classified into four types, based on contractile characteristics of
the muscle unit (Burke, 1981). Motor units can be divided into fast or
slow on the basis of twitch time-to-peak, or the tension profile in an
unfused tetanus ('sag'; Burke et al. 1973). The fast units can be
subdivided on the basis of resistance to fatigue, into fast-fatiguable
(type FF), fast-fatigue-resistant (FR), and fast with intermediate
fatigue-resistance (FI). All slow, type S units are highly
fatigue-resistant. Normal MH contains approximately 45% type FF, 5% type
FI, 25% type FR and 25% type S motor units (Burke, 1981; Fleshman et al.
1981). Normal soleus contains nearly 100% type S units (Burke, 1981).
Normal LG has not previously been studied independent from MG, but is
considered to be similar to MG in motor unit types (from muscle
histochemistry; Ariano et al. 1973; see also Hamant, 1977).
Motoneuron electrical properties have been shown to vary according
to motor unit type, in MG (Fleshman et al. 1981; Zengel et al. 1985).
Parameters investigated include axonal conduction velocity, AHP
half-decay time (related to the ability of the cell to fire
repetitively), rheobase (an inverse indicator of cell excitability), and
input resistance (influenced by cell size and geometry, as well as
specific membrane conductances). Zengel et al. (1985) devised a method
to classify motoneurons on the basis of their electrical
characteristics. This scheme was found to predict motor unit type,
determined independently by contractile properties, with greater than
90% accuracy. In light of the close relationship evident between
motoneuron electrical properties and muscle unit contractile properties
in normal MG, it is of interest whether this is a general property of
normal motor units of other triceps surae muscles, and whether
motoneurons exhibit plasticity in response to experimental manipulation.
In this study, experimental alteration of the innervation of
triceps surae muscles is used to investigate several aspects of the
relationship between motoneuron and muscle. MH motoneurons were isolated
by intracellular methods, to allow recording of motoneuron electrical
properties. Intracellular current injection was employed to activate the
muscle unit to measure contractile responses. In addition, whole muscle
contractile properties and muscle histochemical properties were
documented. Acute experiments were performed on normal, unoperated-on
animals, animals whose MG nerve had been sectioned (prior to
reinnervation of muscle), and animals whose MG and LG-soleus (combined
nerve) nerves had been sectioned, and allowed to reinnervate either the
original muscle (self-reinnervation), or the opposite muscles)
(cross-reinnervation). Acute experiments for reinnervated muscles were
performed at various times after the initial surgery (from three weeks
to ten months). All studies herein that involved surgical manipulation,
were analyzed for MG motoneurons only.
The questions addressed in this study are as follows:
1) Are the relationships between motoneuron electrical properties
and muscle unit contractile properties of similar quality and strength
in LG and soleus to that observed in MG, (CHAPTER V)?
2) Is functional connection to muscle required for expression of
normal, mature motoneuron electrical properties (CHAPTER IV)?
3) Do motoneuron electrical properties recover to control values
upon reinnervation of the original muscle (CHAPTER III)?
4) What is the time course and pattern of recovery during
self-reinnervation (CHAPTER IV)?
5) Given a 'choice' of end organs, is there selectivity of
reinnervation on the basis of motor unit type (CHAPTER VI)?
6) What are the limits of motoneuron influence upon muscle
properties (CHAPTER VI)?
7) Does the expression of motoneuron electrical properties depend
upon the particular muscle innervated; i.e. is there a regulatory
influence of muscle, upon motoneurons (CHAPTER VII)?
Experiments were performed on 51 adult cats (22 normal cats, 29
experimental). Animals were anesthetized to a depth sufficient to
maintain areflexia, during all surgical and experimental procedures. All
animals were adult at the time of the initial surgery.
Animals were anesthetized with a gaseous mixture of oxygen, nitrous
oxide, and halothane. Under aseptic conditions, the MG and LO-S nerves
of the left hind leg were surgically isolated within the popliteal
fossa, sectioned about 15-25mm proximal to the triceps surae muscles,
and the proximal stumps of each nerve were resutured to the distal stump
leading to the muscles) not innervated by that nerve originally
(X-reinnervation; Fig. 2-1) or to the stump leading to the original
muscle (self-reinnervation). In some experiments this procedure was also
carried out on the right leg. To direct axonal growth, 15mm long sleeves
of Gore-Tex (30um internodal distance) were used (Young et al. 1984;
see their Fig. 2). The two nerve ends were joined and their orientation
fixed with two 9/0 sutures through the epineurium.
Following recovery animals were maintained in pairs in 90x90cm
cages. The animals received care and were exercised daily outside of the
Figure 2-1. Schematic diagram of cross-reinnervation surgery. Location
of Gore-Tex sleeves shown as black cylinders.
The acute experiments utilized techniques and protocol published
previously (Fleshman et al. 1981; Zengel et al. 1985). Operated animals
were studied three weeks to 11 months following the initial surgery.
Animals were anesthetized via intraperitoneal injection of sodium
pentobarbitol (35mg/kg). Leg, body, and spinal cord temperatures were
maintained at 350-370C with a heat lamp and heating pad.
The unoperated, normal M3 population derives from 16 animals (15
females, one male; 2.0 to 6.4 kg, mean 3.5 kg). The normal LG and soleus
data derive from 15 animals (eight animals were the same as for M3; 14
females and one male; 1.9 to 3.8 kg, mean 2.8kg).
Four acute experiments were performed on animals (all females, 2.8
to 3.4 kg, mean 3.0 kg) whose H3 nerve was sectioned and had not yet
reinnervated muscle (no-re; 20-35 days post-operative). Eleven animals
were examined at various times after self-reinnervation of M3 muscle.
Two animals (both female, 3.0 and 3.3 kg) were investigated five to six
weeks following the initial surgery (low-re; these animals had whole
muscle twitch tensions of less than 200g-wt). Two additional animals
(both female, 2.9 and 3.3 kg) were examined nine to ten weeks
post-operative (med-re). These animals had whole muscle twitch tensions
approximately 50% of long self-reinnervated MG. Five animals (four
female, one male, 2.9 to 3.6 kg, mean 3.4 kg) were examined nine months
after self-anastomosis of the M3 nerve (long-re). A final two animals
(both female, 3.5 and 2.9 kg) were examined whose recovery was
intermediate between the low-re and med-re stages .
Cross-anastomosis of the MG and LO-soleus nerves was accomplished
for 14 animals (all females). Four of these were investigated at
nine-to-ten weeks (Afme, 2.0 to 2.8 kg, mean 2.4 kg). The remaining ten
animals were examined nine-to-11 months post-operatively (longX, 2.2 to
3.7 kg, mean 2.9 kg).
Whole Muscle Twitch
Prior to single unit recording, whole muscle twitch contractions
were obtained by stimulating the MG or LG-S nerve, with the MG, LG, or
soleus tendon attached to a force transducer. Each muscle was tested in
response to each of the two nerves, against a passive tension of
100g-wt., and also in response to the nerve eliciting the greater
tension, at the length providing maximal tension. The muscles were
activated tetanically before each measurement to potentiate the twitch
response. In some animals whole muscle twitches were obtained again
following single unit studies. The MG and LG muscle were then completely
separated and the series of twitches recorded again.
Motor Unit Studies
MG motoneurons were identified by antidromic stimulation of the MS
nerve, central to the nerve sleeve (innervates MG in
self-reinnervation, LG and soleus in cross-reinnervation). Motoneurons
were impaled with 3M KCl-filled microelectrodes, with impedances in the
4-12 Mohm range (measured at 1KHz). The muscle innervated by the MG
motoneuron, was differentiated by intracellular current injection, and
locating the muscle response.
Methods for obtaining and analyzing motoneuron and muscle unit data
were as reported previously (Fleshman et al. 1981; Zengel et al. 1985).
We recorded antidromic action potential conduction time, rheobase, input
resistance [by the direct method of measuring voltage deflection by
means of a bridge circuit while injecting a InA current, without
correction for nonlinearities, (Ito and Oshima, 1965)], and
after-hyperpolarization potential (AHP) for each cell. The derived value
rheobase/input resistance was calculated for each motor unit because
that value (in conjunction with AHP) has been shown to precisely
categorize the unit type to which a motoneuron belongs in normal M3
(Zengel et al. 1985). The time for the AHP to decay to half-maximum
amplitude (AHP half-decay time) was measured. Only motoneurons with
action potential amplitude of 60mV, or greater, were used for analysis
of motoneuron electrical properties. Muscle unit contractile data were
obtained from a few units where motoneuron action potentials were less
The contractile responses of motor units were activated by
intracellular current injection at the motoneuron, and measured using
the methods of Burke et al. (1973). Muscle unit contractile responses
were measured against a passive tension of 100g-wt. The following
responses were recorded: unpotentiated twitch; potentiated twitch
(following 100Hz tetanic stimulation); unfused tetanus ('sag'); tetanic
tension (600 ms at 100 Hz for fast units, 1500 ms at 100Hz for slow
units) and the fatigue test (Burke et al. 1973).
Motor units were classified on the basis of potentiated twitch
time-to-peak (fast units had a time-to-peak of potentiated twitch of
40ms or less and exhibited sag, slow units had time-to-peak of > 40ns
and did not sag) and fatiguability (FF units had fatigue index of (
0.25, S and FR units had fatigue indices of > 0.75, and FI units were in
between) as type FF, FI, FR, or S (Burke et al. 1973; Fleshman et al.
1981). Units with twitch time-to-peak in the range 35-45ms were
classified as fast or slow by the presence (fast) or absence (slow) of
"sag" (Burke et al 1973) in an unfused tetanus (inter-stimulus interval
= 1.25X time-to-peak).
Animals were sacrificed by overdose of sodium pentobarbitol. The
left MG, LG, and soleus muscles were excised, cleaned of excess
connective tissue, blotted dry, and weighed to the nearest 0.1g. For MG
and soleus, a 10mm block of tissue was removed from the thickest part of
the muscle belly, cutting perpendicular to muscle fiber direction.
Lateral gastrocnemius muscles were cut into three blocks: proximal,
middle, and distal, each including approximately one third of the muscle
length. The blocks of muscle were fixed to a piece of cork with gum
tragacanth. Muscle orientation was carefully noted. The cork and muscle
were then immersed in isopentane cooled to -1600C by immersion in liquid
nitrogen. The tissue was then placed in a cryostat maintained at -200C.
After a few minutes drying time, the frozen tissue was wrapped in
parafilm, placed in a plastic vial, and stored until cut and stained.
Muscle histochemistry was performed within one week of the
acute experiment. The tissue and cork were mounted on a cryostat chuck
with OCT compound. Ten um thick serial sections were cut from each
block, and stained for myosin ATPase (preincubated at pH 10.3 and 4.2,;
Padykula and Herman, 1955; Guth and Samaha, 1970), Nicotinamide adenine
dinuoleotide diaphorase (NADH-D; Novikoff et ll. 1961) and
A-glycerophosphate dehydrogenase (OPD; Wattenberg and Leong, 1960).
Muscle fiber types were classified using the system of Peter et al.
(1972). This scheme designates fibers as fast or slow, on the basis of
alkaline myosin ATPase (dark= fast; light = slow). Fast fibers are
further subdivided into fast glyoolytic fibers (FG), staining light for
NADH-D and dark for GPD, and fast oxidative glycolytic (FOG), staining
dark for both NADH-D and GPD. Slow (SO) fibers stain dark for NADH-D and
light for GPD.
The distribution of muscle fiber types was determined from
projections of the original slides with a microprojector. For LG,
approximately 600-1000 fibers were analyzed for myosin ATPase and
NADH-D, from each of the four compartments of the muscle (English and
Letbetter, 1982a,b). The LGm compartment was analyzed in the most
medial portion of the proximal section (separated by tendon). The LG2
compartment was located just lateral to the LGm compartment, in the
middle section. LG1 was analyzed from the distal section, near the
lateral border (lateral to tendonous inscription). Finally, LG3 was
located near the medial border of the distal section (see English and
Letbetter, 1982b; their Fig. 3). For soleus, the number of fibers with
"fast" (type II; Engel, 1970) staining characteristics were counted. For
MG approximately 1000 fibers from each of three separate areas (single
cross-section) were sampled. Muscle fiber areas were determined by
planimetry from photographs for 25-50 fibers of each type (FG,FOG,SO)
for each compartment of LG and for each MG or soleus muscle.
Following single unit studies in two long self-reinnervation
experiments in which both M3 nerves had been sectioned and
re-anastomosed, a single type FF motor unit was isolated on the
contralateral side of the animal by intracellular impalement. After
measurement of motor unit properties, the unit was repetitively
activated using the same stimulus regime as for the fatigue test. This
was repeated over approximately one half hour, with one minute rest
between each two minute stimulation period, in order to deplete the
muscle fibers of that motor unit of their stores of glycogen (Edstrom
and Kugleberg, 1968; Kugleberget al. 1970). These muscles were prepared
as above for muscle histochemistry. An additional serial section was
stained for glycogen (PAS stain). The muscle was sectioned until the
section with the maximum number of depleted fibers was located. Serial
sections for myosin ATPase and NADH-D were examined to determine whether
reinnervated motor units were homogeneous with respect to muscle fiber
types and to examine their distribution.
All statistical analysis was performed with programs written in the
language of the Statistical Analysis System (SAS; Helwig and Council,
1979). Mann-Whitney U-test was used to determine differences between
two means. Wilcoxon's multiple range test was used to compare multiple
means. Post-hoc analysis with Tukey's range test was used to determine
which specific means were different. Significant differences between the
proportions of motor unit or muscle fiber types were determined by the
chi square test, and comparison to the hypergeometric distribution. An
alpha value of 0.01 was considered to be significant as different
oontrol samples ooassionally differed at the 0.05 level. SAS programs
were executed with the facilities of the Northeast Regional Date Center
at the University of Florida.
NORMAL MG AND LONG-TERM SELF-REINNERVATION OF MG
Since the initial cross-reinnervation study of Buller et al. (1960),
attention has been focused on the role of the motoneuron in dictating
muscle properties. Numerous studies of contractile properties at the
whole muscle (Close, 1965; Close and Hoh, 1969; Luff, 1975; Prewitt and
Salafsky, 1967) and single unit levels (Bagust et al. 1981; Burke, 1980;
Burke et al. 1979; Chan et al. 1982; Dum et al. 1979; Lewis et al. 1982)
have supported the original hypothesis of neural determination of muscle
properties, and extended its documentation to finer levels of
In addition, experiments in which muscles of mixed fiber type
composition were subjected to chronic electrical stimulation indicate
that the frequency and/or amount of muscle activity are important
factors in the determination of muscle contractile properties (Eerbeek
et al. 1984; Goldring et al. 1981; Hudlicka et al. 1982; Lomo et al.
1974, 1980; Salmons and Vrbova, 1969; Smith, 1978; see additional
references in Pette, 1984; Salmons and Henrikson, 1981).
All of these studies (cross-reinnervation and stimulation) indicate
that it is easier to change a fast muscle to slow than the reverse. The
chronic stimulation experiments suggest that neural activity plays a
major role in determining muscle fiber characteristics, but there appear
to be limits to the extent of this influence.
Only two laboratories have previously related motor unit properties
to motor unit type following reinnervation. The motoneuron properties
investigated were axonal conduction velocity (Burke, 1980; Gordon and
Stein, 1982a,b), AHP duration (Burke, 1980), and extracellular action
potential amplitude (Gordon and Stein, 1982a,b). All studies reported
recovery to control levels by nine months to one year post-operative.
Kuno et al. (1974b) reported that five months following the
initial surgery, conduction velocity, resting membrane potential, ARP
duration, and action potential overshoot were incompletely recovered in
reinnervated cat M3 and soleus motoneurons. In that study all MH
motoneurons were considered 'fast' and soleus motoneurons were
considered 'slow. Czeh et al. (1978) suggested that AHP duration is
regulated by retrograde trophic messages from muscle. There is some
indication of a muscle influence on axonal conduction velocity as well
(Burke, 1980; Lewis et al. 1978). Thus, in addition to neural induction
of muscle properties there may also be retrograde influence upon the
motoneuron from the muscle.
In light of the differences in normal animals between motor unit
types in their motoneuron electrical and muscle unit contractile
properties (c.f. Fleshman et al. 1981; Zengel et al. 1985), it is of
interest whether motoneuron electrical properties, muscle unit
contractile properties, and their interrelationships, are restored
following self-reinnervation. The difference between slow soleus and
the overall M3 motoneuron sample in the studies of Kuno and colleagues
(Kuno, 1984) suggests that fast and slow motor units in a mixed muscle
may show different responses to axotomy and subsequent reinnervation.
A second question in reinnervation of adult skeletal muscle is
whether all motoneurons have an equal ability to reinnervate and
maintain connections with muscle fibers. Lewis et al. (1982) observed a
dramatically increased amount of type I (associated with slow motor
units) muscle fibers in one cat flexor halluois longus (FHL) muscle
three years following self-reinnervation. In combination with the
earlier observation of a reversal in the normal relationship of axonal
conduction velocity to motor unit tension production in soleus muscle
cross-reinnervated by the FDL nerve (slower units larger than fast
units; the reverse of normal FDL; Bagust et al. 1981), this suggested
to these workers that slow axons might have a competitive advantage
over fast axons in reinnervating adult slow muscle fibers.
We here use an intracellular approach to the model of
self-reinnervation of the M1 muscle in the cat, to explore the limits
of regulatory interactions between alpha-motoneurons and the muscle
fibers they innervate. Specifically, three questions are addressed.
First, are motoneuron and muscle properties reestablished (overall and
within motor unit types) following nine months self-reinnervation?
Second, is there evidence for a competitive advantage of any motor unit
type in reestablishing motor unit size or numbers? Finally, to what
extent are muscle unit properties dictated by the innervating
motoneurons and vice ersa; that is, are the normal relationships
between motoneuron and muscle properties restored?
Effects of Cage Time
Since some of the animals in this study were caged for nine-months
prior to the acute experiments, it was important to control for any
effects of cage-time per se. Data were compared from non-caged animals
(11 cats, all female, 2.0 to 6.4 kg, mean weight 3.7 kg) and cats caged
from 151 to 284 days (5 oats, four females and one male, 2.1 to 3.6 kg,
mean weight 2.9 kg). There were no differences in the distribution of
motor unit types between non-caged and long-caged cats.
The only differences in motoneuron properties (p<0.01, Mann-Whitney
U-test) between non-cagcd and long-caged cats were rheobase and
rheobase/input resistance of the population as a whole (across motor
unit types), and input resistance in type S motor units (Table 3-2).
Similarly, contractile parameters did not differ between non-caged and
long-caged cats at the 0.01 alpha level (Table 3-3). We conclude that
the effects of cage time were minimal and the non-caged and long-cage
groups will be considered together as the normal control group for the
remainder of this paper.
Table 3-1 lists values for muscle wet weights and whole muscle
twitches for normal and self-reinnervated MG muscles. Normal MG values
for muscle weight, muscle weight/body weight, and speed and
tension-related parameters were similar to those obtained for cat MG in
other studies (Gardiner et al. 1978; Mayer et al. 1984; Sacks and Roy,
1982, Spector et al. 1980).
Table 3-1. Whole Muscle Properties: NORMAL and Nine Month
Time-To-Peak b ms)
Half-Rise Time b(ms)
Half-Relaxation Time (ms)
Wt./Cat Wt. (g/kg)
Tension/Cat Weight (g-wt./kg)
a. MeansSE (for NORMAL n=14;
b. Twitch at muscle length at
c. = significant difference
for LONG-RE n=4).
which maximum tension was obtained.
from control (p<0.05);
The normal population (non-caged plus caged, see Methods) consisted
of 48% type FF units, 4% type FI, 24% type FR, and 2i% type S units
(Table 3-6). This is similar to that reported by Burke et al. (1973),
and Fleshman et al. (1981).
Normal MG muscle contained 56% type FG, 23% type FOG, and 21% type
SO fibers (Table 3-6; see also Ariano et al. 1973; Burke and Tsairis,
1973). There were no differences between non-caged and long-caged cats
with respect to muscle fiber type distribution.
In cat MG, type FF muscle units are comprised of type FG muscle
fibers (or IIb), type FR units of type FOG fibers (IIa), and type S
units of type SO fibers (I; Burke and Tsairis, 1973). Table 3-6
estimates relative innervation ratios for each motor unit type.
In general, normal values for motoneuron electrical properties
(Table 3-3) and muscle unit contractile properties (Table 3-2) were
similar to those reported previously (Burke et al. 1973; Fleshman et al.
1981; Zengel et al. 1985).
Control vs. Reinnervated
Acute physiological experiments were performed on four animals
(three female, one male; 2.9 to 3.6 kg, mean 3.4 kg) nine months after
the initial surgery. An additional animal (female, 2.7 kg) was used for
muscle histochemistry only. The normal distribution of motor unit types
was reestablished, confirming results obtained by Burke and colleagues
(Burke, 1980; Burke I 3aL 1979; Dum et al. 1979, in press b) in
self-reinnervated cat FDL, and by Gordon and Stein (1982a) in
self-reinnervated cat M3 (Table 3-6).
Following nine months recovery from the initial surgery, HO muscle
weights were about 70% of unoperated normal MD. Whole muscle twitch
time-to-peak as well as half-rise and half-relaxation times were normal
(Table 3-1; Bagust and Lewis, 1974; Buller et al. 1960; Burke, 1980;
Burke et al. 1979; Chan et al. 1982; Dum et al. 1979, 1985b; Gordon and
Stein, 1982b). In the operated animals muscle weight was significantly
reduced as was muscle weight/body weight. Twitch tension was also lower
in reinnervated animals (p<0.05), but tension /muscle weight was
unchanged (Table 3-1). These results are similar to those of Chan et al.
(1982), Bagust et al. (1981), and Lewis et al. (1982), but Gordon and
Stein (1982b) and Burke (1980) saw complete recovery of muscle weight
and twitch tension.
Muscle Unit Contractile Properties.
There were several differences in contractile properties between
normal and self-reinnervated animals (Table 3-2). Overall, across motor
unit types, mean maximum tetanus was lower in the reinnervated group (23
vs. 35g-wt.; Table 3-2). This trend was still present when maximum
tetanus was normalized by body weight, although no longer statistically
significant (p>0.05). This overall effect reflects the significantly
lower maximum tetanic contractions of type FF units (also significant
for normalized values). Types FR and S unit tensions were normal.
Consistent with the mean values, there were few units with large maximum
tetanic contractions in the reinnervated muscles (Fig. 3-10). We saw no
Table 3-2. Muscle Unit Contractile Properties. (Long-caged vs. Non-caged;
Normal vs. Nine month self-reinnervated). abc
FF Fl FR S ALL
TWITCH TIME-TO-PEAK (ms) d
TWITCH TENSION (g-wt.)
H CTIWT HALF-RELAXAT N
MAXIMUM TETANIC TENSION (g-wt.)
a. = Significant difference from NON-CAGE at 0.05 level; *" 0.1 level
b. + = Significant difference from NORMAL at 0.05 level: ++ 0.01 level
c. Means SE (number of units).
d. Potentiated Twitch.
e. NORMAL = NON-CAGE + LONG-CAGE.
TWTTCH BALF- RE7-AXATTON-~
TIME (ms) d
evidence of greatly enlarged motor units with reinnervation (Chan et al.
1982; Gordon and Stein, 1982a,b; but see Rosenfalck and Buchthal, 1970;
Yahr et al. 1950). Most studies of reinnervated muscle have reported
decreased mean tension with increased variance [Bagust and Lewis, 1974;
Burke, 1980 (oat flexor digitorum longus: FDL); Chan et al. 1982 (oat
flexor hallucis longus: FHL)]. Distributions of twitch and tetanic
tension in this study were similar to those of normals in range for all
types except type FF where distributions were shifted to lower values
Mean potentiated twitch amplitude was smaller in self-reinnervated
MH, but not significantly so (Table 3-2). The overall distribution was
similar to that of normals, except there were few large twitch units
within type FF (Fig. 3-1A). Gordon and Stein (1982b) reported recovery
of mean twitch tension, with reinnervated units showing higher variance
than normals. Bagust and Lewis (1974), Bagust et al. (1981), and Lewis
et al. (1982) found little change in mean twitch tension, although the
distribution included a few very large units and many very small units.
Part of the difference between their results and ours may be due to
their expression of motor unit tension relative to whole muscle tension.
Absolute motor unit sizes were not abnormally large compared to normal
In general, type FF units produced less tension and type S units
more tension than normals (Table 3-2; also true if normalized by cat
weight). The increased type S unit twitch amplitude, without increased
tetanus resulted in significantly increased twitch/tetanus ratio. We saw
no difference in twitch/tetanus ratio between normal and reinnervated
5t. --.._ .
C T iL
I n s
a- T ** -DEER TI ME (s 0
I .1 1.1 I.I I2. 1 3.
INPUT RESISTANCE lMobsh
o I .N it
RXONAL COOLCTICN VELOCTT WI/s)
Figure 3-1. Frequency histograms for normal (unfilled) and nine
month self-reinnervated (filled) MG motoneuron properties.
(A) Rheobase (B) Input resistance (C) AHP half-decay time (D)
axonal conduction velocity. See text for explanation.
There were no differences between reinnervated and normal muscle
units in twitch time-to-peak (Bagust and Lewis, 1974b; Burke, 1980; Chan
et al. 1982; Gordon and Stein, 1982a) or half-relaxation time (Bagust
and Lewis, 1974b). There were no differences in the frequency
distributions for the reinnervated units' twitch time-to-peak or fatigue
index (Fig. 3-1B,D) compared to normals. There was a significant
increase in mean fatigue index in type FF units of reinnervated animals
(0.06 vs. 0.03), confirming a general impression of a "residual" tension
following the fatigue test in these animals (unpublished observations).
Reinnervated motor units were similar to normal units in the pattern
of differences between motor unit types for contractile properties
(Table 3-4). The exceptions were that reinnervated types FR and S unit
fatigue indices did not differ, and the only difference for
twitch/tetanus was between types S and FR units.
In reinnervated units the overall correlation between twitch
time-to-peak and tetanic tension was -0.47 (p< 0.0001), similar to
normals (-.40, p< 0.0001), and confirming Bagust and Lewis (1974). The
correlation was -0.53 within reinnervated type S units (p< 0.02; not
significant in normals). Types FF and FR units show no correlation
between these properties in normals or reinnervated animals.
The overall correlation between twitch time-to-peak and twitch
amplitude was weak (-0.32, p<0.004) for self-reinnervated units (vs.
-.29, p< 0.003 in normals) and nonsignificant within any motor unit type
in either population. Gordon and Stein (1982a) reported a restoration
of the normal negative correlation between twitch time-to-peak and
twitch amplitude with reinnervation.
Speed- and fatigue-related properties were normal in mean values and
distribution nine months' following section and re-suture of the MH
nerve. The relationships between time-to-peak and tension were also
normal. The exception to complete restoration of normal properties was
that type FF units did not fully recover ability to generate tension.
Motoneuron Electrical Properties
There were no significant differences in motoneuron properties
between operated animals and normals (p<0.01). At the 0.05 alpha level
only rheobase in FF units, and rheobase/input resistance in S units
reached significance (Table 3-3). Significant differences between motor
unit types in reinnervated motor units were similar to normals (Table
3-4). Exceptions include input resistance, where FR units were not
significantly different from FF or S units, and rheobase, where FR and S
units differ at p<0.05 only.
There was no significant difference in axonal conduction velocity
between regenerated and normal MG motoneurons (Table 3-3). This was also
seen by Gordon and Stein (1982a) for MH self-reinnervated for 9 months.
Kuno et al. (1974b) reported that after 4 months, regenerated MG
motoneurons' axonal conduction velocity had not quite recovered to
non-caged levels. This difference of results may be due to the shorter
recovery time in the Kuno et al. (1974b) study. Consistent with this,
Lewiset al. (1978) reported lower than non-caged conduction velocity in
self-reinnervated FDL axons at 6 months but not at two years following
the initial surgery.
Table 3-3. Motoneuron Electrical Properties. (Long-caged vs. Non-caged; Normal
vs. Nine month self-reinnervated). a, ,8
NORMAL MG d
INPUT RESISTANCE (Mohms)
(15) 0.90 (67)
(13)** 1.20 (50)
(28) 1.0+0 (117)+
(11) 1.00 (45)+
RHEOBASE/INPUT RESISTANCE (nA/Mohms)
NORMAL MG d
AHP HALF-DECAY TIME (ms)
NORMAL MG d
AXONAL CONDUCTION VELOCITY (ms)
NORMAL MG d
* = Significance at 0.05 level from NON-CAGE; ## 0.01 level.
+ = Significance at 0.05 level from NORMAL; ++ 0.01 level.
Means SE (number of units).
NORMAL MG = NON-CAGE + LONG-CAGE.
Table 3-4. Results of Tukey's Studentized Range Test: Significance of
Differences Between Motor Unit Types: NORMAL VS. Nine Month
Self-reinnervated MG (LONG-RE).
Axonal Conduction Velocity
AHP Half-Decay Time
Twitch Amplitude a,b
Twitch Time-To-Peak b
Twitch HRT b,c
Maximum Tetanic Tension a
a) The same result was obtained
b) Potentiated twitch.
c) HRT = half-relaxation time.
with raw data and data normalized for body
d) Twitch/tetanus = unpotentiated twitch/maximum tetanus.
All motor unit types had ranges of AHP half-decay times similar to
normal (Fig. 3-2). Kuno et al. (1974b) found that regenerated Mi
motoneurons had AHP durations that were not significantly longer than
normal (82 vs 75ms). In contrast, regenerated soleus motoneurons
(virtually all slow in normals) had AHP durations that were
significantly shorter than normal (123 vs 150ms). Thus the type S
motoneurons of the mixed muscle MG appear to behave differently in this
respect than soleus type S motor units following regeneration.
Kuno et al. (1974b) found that 45-60% of five-month regenerated MG
motoneurons had values for AHP duration/axonal conduction velocity
within the normal range. We found that 96% (44/46) of regenerated MG
motoneurons were in the normal range with respect to AHP half-decay time
and axonal conduction velocity (Fig. 3-3) and there was little overlap
between the distribution of fast and slow units (35/36= 97% of
regenerated fast motoneurons in the normal fast range; 13/18 = 72% of
regenerated slow motoneurons in the normal slow range). Axonal
conduction velocity decreased as AHP half-decay time increased in
regenerated motoneurons (r= -0.42, p<0.0001; vs. -0.52, p<0.0001 in
normals). Within reinnervated type S units the correlation was -0.69
(p<0.0001; -0.51, p<0.03 in normals), and there was no relationship for
types FF and FR units as in normals.
The distribution of values for rheobase was indistinguishable from
that of normals (Fig. 3-2A). The same was true for input resistance,
with the possible exception of a lack of large input resistance (>1.5
Mohm) cells in S units of the operated animals (Fig. 3-2B). This may
represent an artifact of sampling.
Is M as s ?S
TWITCH TIME-TO-PEAK (as)
TETANIC TENSION tg-wtl
ih- n. ,
TWITCH TENSION tg-wtl
inL -.. ..
v a 1 A-Ns
FATIGUE INDEXX (TII 7 E S
TWITCH TENSION (g-vt)
FRTISUE :cE3x (TICS IlU)
Figure 3-2. Frequency histograms for normal (unfilled) and nine
month self-reinnervated (filled) M3 muscle unit contractile
properties. (A) potentiated twitch amplitude (B) potentiated
twitch time-to-peak (C) tetanic tension (D) fatigue index. See
text for explanation.
Figure 3-4A shows the relationship between rheobase and input
resistance in regenerated motoneurons (3-4B) as compared to normals
(3-4A). Figure 3-4B illustrates that, as in normals, motoneurons
segregated according to type on the basis of rheobase: input resistance
ratio. The correlation between log rheobase and input resistance was
-0.63 (normal, -0.60, p<0.0001; see also Fleshman et al. 1981; Zengel
et al. 1985). No significant correlations were found within any motor
unit type in normal or reinnervated units.
The criteria used by Zengel et al. (1985) were applied to these data
to estimate motoneuron type (AHP half-decay time < 30ms = fast, AHP
half-decay time > 30ms = slow; rheobase/input resistance <7 = S,
rheobase/input resistance > 18 = FF, with FR motoneurons having
rheobase/input resistance between 7 and 18). Based on these criteria
there was an 86% agreement in normal MG, between motoneuron type and
motor unit type (determined by contractile properties). When these same
criteria were applied to the reinnervated motor units, there was an 841
agreement between motor unit type contractilee) and motoneuron type
Thus, after nine months, regenerated MG motoneurons had normal mean
values and frequency distributions for membrane electrical properties.
Relationships between motoneuron electrical properties were normal, and
motoneuron type (defined by electrical properties) accurately predicted
motor unit type (defined by contractile properties).
XON CO CION VELOCITY
XONIL CONOJICTION VELOCITY (/.s)
I I I
I I I I I
RXONL COXOJCTION VELOCITY (./s)
Figure 3-3. Relationships between AHP half-decay time and axonal
conduction velocity in normal and nine month self-reinnervated MG
motoneurons. (A) Normal M3. Note overlapping but largely separate
distributions for slow (S) and fast (F=FF, R=FR, I=FI) motoneurons.
Solid lines outline the distribution of slow (upper) and fast
motoneurons. (B) Nine months self-reinnervated MR. Note precise
reestablishment of distributions for slow (S) and fast (F=FF, R=FR,
I=FI) motoneurons. Most non-contracts (N) fall outside the normal range.
Solid lines outline the distribution of slow (upper) and fast (lower)
motoneurons in controls.
I II I
Unlike those in the normal population, there were a few motoneurons
(8/81) in the self-reinnervated animals which, when stimulated, did not
elicit measurable muscle contractions. We have termed these cells
"non-contracts" as we cannot rule out their being .-fusimotor in nature
(Gregory et al. 1982). It is unlikely that these are gamma motoneurons
as they received monosynaptic input from the LG-S nerve (Table 3-5). A
subpopulation within the regenerated motoneurons which did not elicit
muscle contraction has been observed previously (Bagust and Lewis, 1974;
Gordon and Stein, 1982b; Kuno et al, 1974b; Dum et al. in press).
Mean values and raw data for individual non-contracts' electrical
properties are seen in Table 3-5. In axotomized motoneurons, mean axonal
conduction velocity and the mean and range for rheobase decreased, while
mean input resistance was increased (Chapter 4; Gustafsson, 1974;
Gustafsson and Pinter, 1984; Kuno et al. 1974a). Mean AHP half-decay
time is little changed with axotomy, but the range was compressed from
both the high and low ends (Chapter IV; Gustafsson, 1974; Gustafsson and
Pinter, 1984; Kuno et al. 1974a). Some of the non-contracts at nine
months' self-reinnervation fit this pattern (cells 6,7,and 8) while
others show various combinations of properties within and outside the
range for normal motoneurons. Six of the eight non-contracts fell
outside the normal range of values for the ratio AFP half-decay time:
I I I I
I.S I 1. 2 2.1 S
INPUT REISTFCE (Moh*sJ
gI I l I I
I8 .5 1.I 2 2.1 I 1. 4
INPLT RESISTANCE IMoh.%)
Figure 3-4. Relationships between rheobase and input resistance in
normal and nine month self-reinnervated MG motoneurons. F= type FF
units, I=type FI, R=type FR, S=type S units, N=non-contract. (A) Normal
M3. NoLE segregation by motor unit type. (B) Nine months
self-reinnervated MG. Segregation by motor unit type is re-established.
The lack of high input resistance type S units may reflect sampling
I ~~ I
conduction velocity (Fig. 3-4). It.is possible that some of the cells
with relatively normal properties were injured in dissection or have
made contact with intrafusal muscle fibers (Gregory et al. 1982).
Relationships Between Motoneuron and Muscle Unit Properties
In normal animals, the strongest correlation between a motoneuron
electrical property and muscle unit contractile property was between AHP
half-decay time and twitch time-to-peak (r= 0.74, p<0.0001 overall; 73).
This relationship was also present in reinnervated units (0.42,
p<0.0001). We saw no significant relationships within motor unit types
in operated or normal animals, in contrast to Zengel etal. (1985), who
found a significant correlation within S units of MG. Huiszar et al,
(1977) also reported a significant correlation between AHP duration and
twitch time-to-peak for soleus motor units.
In operated animals, as in normals, the correlations between axonal
conduction velocity and twitch (r = 0.23, p<0.04) or tetanic tension
(0.37, p<0.001), normalized by cat weight and expressed as log tension,
were weak ( 0.41, p<0.0001 for twitch; 0.44, p<0.0001 for tetanus;
2,29). There were no significant correlations between these variables
within any motor unit type in reinnervated units. In normals the
relationships were significant within type S units only (0.58 for
twitch, ns for tetanus).
The correlation between axonal conduction velocity and twitch
time-to-peak was -0.50 overall (p<0.0001) and -0.53 for type S units
(p<0.02). This compares to -0.52 and ns in normals. Gordon and Stein
(1982a) also showed restoration of the normal negative correlation
Table 3-5. Data
for Non-Contracts (Nine Month Self-Reinnervation
RNb RHEO/RNc HALFTIMEd
MEAN VALUES +SE (NUMBER OF UNITS)
RN b (Mohm)
RHEO/RN c (nA/Mohm)
AXONAL C.V. e(m/s)
RHEO = rheobase
RN = input resistance
RHEO/RN = rheobase/input resistance
HALFTIME = AHP half-decay time
C.V. = conduction velocity
PSP = composite monosynaptic Ia EPSP from LG-S nerve
between axonal conduction velocity and twitch contraction time in
self-reinnervated cat H3. In contrast, Bagust and Lewis (1974) reported
that the normal inverse relationship between time-to-peak and axonal
conduction velocity was lost in reinnervated muscles. While it is not
clear what can account for these differences, the present surgeries were
all performed on adult cats whereas Bagust and Lewis (1974) operated on
young cats of around one kg in weight.
In summary, the relationships between electrical properties of
self-regenerated motoneurons and contractile properties of
self-reinnervated muscle units were similar to those which exist between
normal motoneurons and muscle units.
Muscle Fiber Histochemistrv and Fiber Areas
Mean cross-sectional areas of nine months' self-reinnervated MS
muscle fibers were 3098um2 for type FG fibers (3873um2 normal), 2355um2
for type FOG (2264um2 normal), and 2296um2 for type SO fibers (1972um2
normal), with an overall average of 2738um2 (88% of normal = 3104um2;
Table 3-6). The areas of types FOG and SO fibers were similar to normal
values but type FG fiber areas were significantly below normal levels.
The reduced overall mean area was due to this lack of type FG recovery.
Burke (1980) found that after one year, the normal distribution of
muscle fiber types was present in FDL, although "type-grouping"
(Dubowitz, 1967; Romanul and Van Der Meulen, 1967) was evident (Burke,
1980). We found 54% type FG fibers, 15% type FOG fibers, and 31% type
SO fibers (56;23;21 in normals) and fiber "type-grouping" was evident
(Fig. 3-5). This overall distribution was significantly different from
normal (chi-square test, p<0.05). There was an increase in the
Table 3-6. Motor Unit Types, Muscle Fiber Types, and Innervation
Ratios (Normal and Nine Month Self-Reinnervated).
MUSCLE FIBER TYPES
SO n/N a
Observed Mean Fiber Area (um2)
Est. Number of Muscle Fibers.
68560 28160 25710
58320 16200 33480
EQUIVALENT MOTOR UNIT TYPE
% Of Population
Est. Number in Pool
Relative IR d
FF+FI FR S
Number of Cells/Number of Animals.
(Physiological Cross-Sectional Area)/Mean Fiber Area.
Estimated as 280-[(280) (% Non-Contracts)].
Relative Innervation Ratio = % Motor Unit Type/% Muscle
proportion of type SO muscle fibers, and a decrease in fast muscle
fibers as a group (types FOO+FO). This alteration of proportions of
muscle fiber types was not as extensive as that reported by Lewis t
al. (1982), however, who presented histochemical data for one
self-reinnervated FHL, after over two years' recovery. In that muscle
80$ of the fibers were type I (equivalent to type SO) compared to the
normal FHL which is 86% type II (Ariano et al. 1973).
Physiological Cross-Section and Innervation Ratios
The physiological cross-section (CSA) of the whole MG muscle was
estimated as follows (ref. Dum et al. 1982):
CSA= (muscle weight)(cosA)
(muscle density)(muscle fiber length)
Muscle density was assumed to be 1.06g cm'3 (Mendez and Keys, 1960).
Values for A (angle of pennation) and muscle fiber length were assumed
to be 210 and 21mm, respectively, for both normal and reinnervated
muscles (Sacks and Roy, 1982). For the reinnervated muscle we estimate
physiological cross-section as 3.0cm2 (3.8cm2 in normals).
The number of muscle fibers in WM was estimated as physiological
cross-section/mean muscle fiber area (Burke and Tsairis, 1973). Mean
muscle fiber area was estimated as the sum of the area for each muscle
fiber type, multiplied by the frequency of that type in the population.
For normals, we estimated the number of muscle fibers in MS as 122,400.
This is lower than the 170,000 estimated by Burke and Tsairis (1973) for
cat M3, presumably they obtained higher values for A and fiber length,
as their value for mean fiber area was higher (3426um2 calculated from
Burke and Tsairis, 1973, vs. 3104um2 in this study).
The overall average innervation ratio (mean number of muscle fibers
innervated by an individual motoneuron; Eccles and Sherrington, 1930) is
calculated as the ratio of the number of muscle fibers in MG (estimated
as 122,400) to the number of MG motoneurons (280: Boyd and Davey, 1968;
Burke et al. 1977). In this study average innervation ratio for normal
MG was 436 (vs. 607: Burke and Tsairis, 1973). For reinnervated MG we
estimate average innervation ratio to be 429 (108,000/252; estimated
numbers of both muscle fibers and innervating axons were decreased;
Table 3-6). The relative innervation ratio for each type is the
frequency of a given muscle fiber type divided by the frequency of the
corresponding motor unit type (Dum et al. 1982; their eq.2).
Relative innervation ratios for reinnervated types FF+FI motor units
were 1.1 (vs. 1.1 normals), 0.6 for type FR units (vs. 1.0, although
Burke and Tsairis, (1973) reported 0.7 in normal MG), and 1.2 in type S
units [vs. 0.9 in normals; 0.9 in Burke and Tsairis, 1973). This may
reflect some competitive advantage of S units in capturing and/or
maintaining contacts with muscle fibers, although not to the extent
reported by Lewis et al. (1982).
Mean motor unit tetanic tensions (Table 3-2) recovered to normal
levels in type FR units despite reduced innervation ratios (see below)
suggesting increased specific tension (Table 3-6). Bagust et al. (1981)
reported altered muscle specific tension following reinnervation and
cite Hob (1974) as showing myofibril packing density was influenced by
innervation. Type S units recover to 67% of normal tension levels
despite increased numbers of fibers and unchanged area (suggesting lower
specific tension, Table 3-6). Type FF units recovered to only 64% of
normal tetanic tension (Table 3-2), primarily due to decreased muscle
fiber area (no change in innervation ratio).
Fig. 3-5 shows a single glycogen-depleted type FF motor unit from a
nine month self-reinnervated M3. Note the "clumped" nature of the
distribution and the homogeneity of muscle fiber type. There were 141
fibers counted in this unit, in the section with the maximal number of
depleted fibers. All were FG fibers, confirming previous studies (Burke,
1980; Gauthier et al. 1983; Kugelberg et al. 1970), in that muscle fiber
type was homogeneous within a single motor unit following long-term
reinnervation. Burke and Tsairis (1973) estimated that counting depleted
fibers from a single section of normal MG sampled 50%-75% of the whole
unit. From this relation we estimate that the unit in Fig. 3-5 contained
212-284 muscle fibers. This unit had a motoneuron rbeobase of 32, muscle
twitch time-to-peak of 30ms, 4.4 g-wt. twitch tension, 25 g-wt. tetanic
tension, exhibited 'sag', and had a fatigue index of 0.07. A second
glycogen depleted type FF unit had 365 fibers in the section with
maximal number (estimate of 548-730 muscle fibers). Rheobase was 16 and
twitch time-to-peak was 20ms. Tension data are unavailable. Again all
fibers were type FG. This unit also exhibited type-grouping, but not to
the extent of the unit in Fig. 3-5. The innervation ratios of these
depleted FF units fall within the range reported by Burke and Tsairis
(1973) for normal MG.
Since calculated innervation ratios were increased for type S motor
units (Table 3-6), type S motoneurons may be at advantage in innervating
and maintaining connections with muscle fibers following reinnervation
of a mixed muscle. Gycogen-depleted type FF units (n=2) fell within the
Figure 3-5. Photomicrographs of a single glycogen-depleted type FF motor
unit from a nine month self-reinnervated MG. (A) The entire unit (PAS
stain for glycogen, 70X) (B) PAS stain (100X) (C) myosin ATPase, pH 4.3
(100X) (D) NADH-D (10OX).
normal range for innervation ratios of unoperated normal type FF motor
In this study we tested whether motoneuron electrical properties,
muscle unit contractile properties, and the normal relationships between
them, would be restored nine months following section and re-anastomosis
of the MG muscle nerve in oats.
We have shown that the normal proportions of motor unit types are
re-established in self-reinnervated muscles (Burke, 1980; Gordon and
Stein, 1982a). Motoneuron electrical properties (rheobase, input
resistance, AHP half-decay time, and axonal conduction velocity), as
well as correspondence between motoneuron (electrical) type and motor
unit contractilee) type, are normal following nine months reinnervation.
As in controls, self-regenerated MG motoneurons segregate by motor unit
type with respect to the ratio rheobase:input resistance. This ratio, in
conjunction with AHP half-decay time, predicts motor unit types in
self-reinnervated as well as control MG (Zengel et al. 1985).
We found that axonal conduction velocity recovers to normal levels
following reinnervation (Burke, 1980; Burke et al. 1979; Gordon and
Stein, 1982a; Kuno et al. 1974b; Lewis et al. 1978). Kuno et al. (1974b)
showed that after five months self-reinnervation, action potential
overshoot, resting membrane potential, and conduction velocity were
normal or nearly normal in M5, while AHP duration had not recovered
completely. Our study confirms these earlier reports, and extends the
documentation of recovered motoneuron properties to motoneuron rheobase,
input resistance, and their relationship. Previous works have not
distinguished between motor unit types with respect to motoneuron
properties [except axonal conduction velocity (Burke, 1980; Gordon and
Stein, 1982a) and AHP duration (Burke, 1980)]. We also found motonearon
properties within each motor unit type in self-reinnervated motor units
to be similar to properties in control motor units of the same type.
Besides the above-mentioned relationship between rheobase and input
resistance, other relationships between motoneuron properties in
long-term self-reinnervated motor units are similar to normal MG.
Overall negative correlations were seen between rheobase and input
resistance, and between axonal conduction velocity and input resistance
or AHP half-decay time. As reported by Kuno et al. (1974a,b), fast and
slow motoneurons tend to segregate on plots of conduction velocity vs.
AHP half-decay time (AHP duration in Kuno et al. 1974a,b), in both
control and regenerated motoneurons (Fig. 3-4). Kuno et al. (1974a,b)
defined 'fast' as M3, and 'slow' as soleus motoneurons. We extend this
observation to fast and slow motor unit types within one mixed muscle,
M3, and document the complete recovery of this relationship after nine
months reinnervation. The lack of correlation between these variables
within motor unit types suggests that overall correlations may be a
consequence of the presence of different motor unit types, each with
characteristic properties, as in normal motoneurons (Fleshman a.t al
Self-reinnervated muscle units did not differ from normal with
respect to speed (isometric twitch time-to-peak, half-relaxation time)
or fatigue resistance (although type FF units were slightly less
fatigable). The exception to complete recovery of overall mean values
for motor unit properties was an overall decline in mean maximum
tetanic tension, especially for type FF units [also seen
in reinnervated FDL (Burke, 1980) and in flexor hallucis longus (Chan &t
Al. 1982)]. This failure to recover tension was due to lack of recovery
of muscle fiber area (type FG fibers). We found no evidence for greatly
enlarged motor units in operated animals (Chan et al. 1982; Dum et al.
1979; Gordon and Stein, 1982a,b). These data suggest that neural
influence may be more effective in regulating muscle twitch speed and
fatigue resistance than muscle unit tension during self-reinnervation
in adult cats.
Twitch speed and fatigue resistance are the parameters which define
motor unit type (Burke et al. 1973). These parameters are also the focus
of most chronic stimulation experiments (reviewed in Pette, 1984;
Salmons and Henrikson, 1981), although Lomo et al., (1980) reported
influence of frequency of activation on force generation by chronically
denervated and stimulated rat soleus muscles. Spector (1984) compared
denervated rat muscles to muscles completely inactivated by TTI, and
reported that while changes in speed of shortening could be accounted
for by activity alone, muscle weight, force generation, and specific
tension appeared to be regulated by release of a trophic substance as
well. Our data are consistent with differential regulation of force and
speed-related muscle properties by the motoneuron.
Relationships between contractile properties after nine months
self-reinnervation were also similar to normal. Motor unit twitch
time-to-peak was negatively correlated with twitch amplitude and with
maximum tetanic tension across the whole population of reinnervated
units, similar to controls (Bagust and Lewis, 1974). We found no
correlations between contractile properties within motor unit types
(except twitch time-to-peak vs. tetanic tension in type S units).
The similarity between normal and self-regenerated motor units in
motoneuron electrical properties, and in motor unit type distribution,
argues against a selective advantage for any motor unit type in terms of
contacting muscle fibers and surviving, although the number of fibers
innervated by a given unit type may vary (see below). Thus although 10%
of the axons made no functional reconnection with extrafusal muscle
fibers, there was no systematic loss of motoneurons of any type, and the
original proportions of each motor unit type were restored by nine
The close correspondence between motoneuron "type" and motor unit
type, described above, suggests that proper match-ups between motoneuron
and muscle properties are reestablished at the level of single units,
following long-term self-reinnervation of MG. This can also be seen in
the overall correlations between axonal conduction velocity and twitch
or tetanic tension, axonal conduction velocity and twitch time-to-peak,
and between AHP half-decay time and twitch time-to-peak. This latter
relationship was the strongest between any motoneuron and muscle
property in both normal and operated animals. These correlations are
weak or absent within motor unit types, suggesting that the overall
correlations can be explained by the presence of different types of
motor units in self-reinnervated as well as normal MG.
The AHP half-decay time (or duration) is closely related to the
motoneurons' steady-state firing frequency (Kernell, 1965;
Oustafsson, 1974). The correlation between AHP half-decay time and
twitch time-to-peak may reflect the control of muscle contractile speed
by pattern and quantity of activation (reviewed in Salmons and
Henrikson, 1981; Pette, 1984). The question still remains as to whether
this match-up is dictated solely by the motoneuron to the muscle fibers
(as suggested by fiber type grouping of units and results from
experiments using chronic electrical stimulation of muscle), or whether
there are also retrograde influences by the muscle on the motoneuron
(Czeh et al. 1978).
Muscle fiber types were present in nearly normal proportions after
long-term self-reinnervation of MS, although there was an increase in
type SO muscle fibers and a corresponding decrease in the proportion of
fast (FG+FOG) muscle fibers in reinnervated MG. "Type grouping" of fiber
types was evident (Dubowitz, 1967; Romanul and Van Der Meulen, 1967).
These data are generally consistent with Chan et al. (1982), Burke
(1980) and Gauthier et al. (1983), who reported restoration of normal
proportions of muscle fiber types with reinnervation. In contrast, Lewis
et al. (1982) presented histochemical data for one cat FHL muscle,
reinnervated for three years, which was 80% type I (equivalent to type
SO) fibers, as compared to a normal 85% type II (FG and FOG). This
result, in combination with their finding of slow units being larger
than fast units in soleus cross-reinnervated by the mixed FDL nerve, led
these authors to propose that slow axons may have a competitive
advantage in reinnervation. Our histochemical data show a slight trend
in the direction of increased numbers of SO fibers, but not of the
magnitude reported by Lewis et al. (1982) and perhaps not outside of
normal variation between animals (real and due to sampling).
The combined data on innervation ratios, muscle fiber type
proportions, and motor unit tension production also suggest that there
could be a slight competitive advantage for type S axons in
reinnervation, as suggested by Lewis et al. (1982). Perhaps within the
type S population, recovery of neuromuscular transmission occurs more
rapidly when the proper match-up between type S motoneuron and SO fiber
is achieved, while fast axons are at a disadvantage unless they
innervate an originally fast muscle fiber. This was suggested by Ip and
Vrbova (1983) for reinnervation of soleus muscle in rats and could
account for the small alteration in motor unit size in the present
study. Ip and Vrbova (1983) showed that either soleus' own nerve or a
foreign nerve were eventually able to innervate soleus muscle
successfully. At early times following the initial surgery, however,
soleus' own nerve induced more tension in the muscle, formed more
synapses with the original endplates, and could follow high-frequency
electrical activation more securely. They noted that the original
post-synaptic specialization remains for some time following
denervation, and may retain its original properties. Finally,
transmission might initially be most effective when the presynaptic
terminal is of the same type as the post-synaptic specialization.
Nine months following self-reinnervation, MN motor units have
properties remarkably similar to normal adult M3. In addition, the
relationships between motoneuron properties and muscle unit properties
are normal, Slow units may have some advantage in making or retaining
contacts with muscle fibers, although there is no strong evidence for
this conclusion. Fiber type grouping and the restoration of motor unit
properties and relationships suggest a major role for the motoneuron in
dictating muscle unit properties. The motoneuron may regulate muscle
unit speed- and fatigue-related properties differently than muscle unit
force generation. The latter property may indicate a limit to
motoneuron regulatory influence. Information concerning the nature of
neural influence on muscle force is of interest with respect to the
observed recruitment of motor units in order of increasing force (Stuart
and Enoka, 1983; Zajac and Faden, 1985) and controversies over the
neural substrate for such a recruitment pattern (Henneman, 1980a; Sypert
and Munson, 1981; Stuart and Enoka, 1983).
This study cannot directly assess whether or not the muscle also
effects motoneuron properties, or whether one type of motoneuron or
another is better able to make early contact with muscle, or recover its
properties more rapidly. Subsequent chapters will address these
questions by utilizing surgical cross-reinnervation of muscles with
different muscle fiber type distributions, and by looking at the time
course of recovery of motoneuron and muscle unit properties during self-
AXOTOMY AND THE TIME COURSE OF SELF-REINNERVATION OF MG
CHAPTER III described virtually complete recovery of cat MG
motoneuron electrical properties, muscle unit contractile properties,
and the relationships between them, following nine months
self-reinnervation. One exception was that type FF motor units failed to
regain control tensions, apparently due to a failure of type FG muscle
fiber areas to recover. Based upon proportions of muscle fiber types and
calculated innervation ratios, it was suggested that type S motor units
might have an advantage over type FR units in the ability to make
and/or maintain functional connections with muscle fibers.
Very little is known about motor unit properties during the process
of reinnervation of skeletal muscle because most studies have focused
upon the end-point, when all changes would presumably be completed.
Gordon and Stein (1982a,b) used a chronic paradigm to measure motor unit
twitch tension and axonal action potential amplitude extracellularr)
for cat MG at various times after nerve section and repair. They found
that, at the earliest stages of reinnervation, twitch tensions were low,
a sizable number of axons did not elicit muscle tension, and the normal
relationships between muscle speed and tension, and axonal action
potential amplitude and twitch tension, were lost. When whole muscle
tension reached about 50% of the value eventually attained, the
relationships between twitch time-to-peak and tetanic tension returned
(3 mos. post-operative). After this time they saw no further increase
in the estimated number of motor units and suggested that the remaining
increase in whole muscle tension was due to increased muscle fiber
areas. The relationship between axonal action potential amplitude and
twitch time-to-peak returned also (at about three months, although not
at control levels until six months). Gordon and Stein (1982a,b) did not
attempt to classify units by motor unit type until the final acute
experiment, at about one year post-surgery.
Kuno et al. (1974b), studying cat M3 and soleus motoneurons,
reported that at early stages of reinnervation many motoneurons did not
elicit muscle contraction. They reported that these cells did not differ
in electrical properties from motoneurons which did elicit muscle
tension. They suggested that functional reconnection was neither a
necessary nor a sufficient condition for restoration of motoneuron
electrical properties. In contrast, Gordon and Stein (1982b) also
reported that a group of axons did not make functional reconnection with
the muscle, but that in these axons, axonal action potential amplitude
did not recover from axotomized levels.
On the basis of histochemistry of self-reinnervated flexor hallucis
longus (FHL) and a reversed relationship between axonal conduction
velocity and twitch tension in soleus muscle cross-reinnervated by
flexor digitorum longus (FDL) nerve, Lewis et al. (1982) hypothesized
that slow motor units might have a selective advantage in reinnervation
of muscle (see also Ip and Vrbova, 1983).
The purpose of the present study was to examine the time course of
self-reinnervation of cat M3 to determine: (1) What happens to
motoneuron properties following axotomy? From what level must recovery
occur? (2) Is functional reoonnection with muscle necessary and/or
sufficient for recovery of motoneuron electrical properties? (3) At what
stage are motor unit types first recognizable? How does this relate to
the degree of recovery of motoneuron properties, muscle contractile
properties, and their relationships? (4) Until what stage are new motor
units being formed? (5) Is any motor unit type at advantage in
reinnervating MG? (6) Are there parallels between the time couse of
reinnervation and motor unit ontogeny?
Table 4-1 lists values for muscle wet weights and whole muscle
twitches for all treatment groups where muscle twitch tension could be
elicited by nerve stimulation. After the M3 nerve was severed there was
muscle atrophy. At the earliest stage of reinnervation examined (low-re;
two cats), muscle weight and tension were less than half of control
values. At this stage isometric twitch time-to-peak was prolcnrfed, as
was half-rise time and half-relaxation time.
The medium self-reinnnervation (med-re; two cats) category was
defined in terms of recovery of tension. These MG muscles had maximum
twitch tensions of over one half of nine month self-reinnervated
Table 4-1. Whole Muscle Properties: Time Course of Self-Reinnervation. a
NORMAL LOW-RE ED-RE LONG-RE
Twitch Time-To-Peak b (ms) 321 46114 340 333
Twitch Half-Rise Time b (ms) 11+0 193 130 100
Twitch Half-Relaxation Time b (ms) 22-2 702 388 252
Twitch Tension b (g-wt.) 2123151 7737 89311 1609276
Muscle Weight (g) 9.01 5.00 4.4.1 7.01
Muscle Wt./Cat Wt. (g/kg) 3.00 1.60 1.50 2.30
Twitch Tension/Cat Weight (g-wt./kg) 74265 24+11 30219 48866
a. MeansSE (for NORMAL n=14; for LOW-RE n=2; for MED-RE n= 2;
for LONG-RE n=4).
b. Twitch at muscle length at which maximum tension was obtained.
(long-re) values (maximum twitch >1000g-wt tension). This group was also
characterized by time after the initial surgery (63 and 71 days). By
this stage values for contraction time (time-to-peak, half-rise,
half-relaxation) had recovered to normal or near normal levels. Muscle
weights were about 60% of long-re (significant at aLPHA = 0.001). Data
for the nine months reinnervated and normal MG are from CHAPTER III.
One animal (FE17), for which we only have whole muscle contractile
and histochemical data, was intermediate between the low-re and med-re
groups in terms of recovery of tension. This animal was sacrificed at 52
days post-surgery. Maximum whole muscle twitch was 281 g-wt.and
time-to-peak was 6 ms. Muscle weight was 5.5g and estimated
physiological cross-section was 2.1 cm2.
A second animal (FE12) also exhibited recovery to a level
intermediate between the low-re and med-re stages after only 34 days.
This animal's whole muscle twitch was closer to the med-re level
(maximum twitch of 653 g-wt., 30 ms time-to-peak). The mass of M3 was
7.4g and physiological cross-section was estimated at 3.1 cm2. Motor
unit data were obtained from this animal.
These two animals indicate that recovery was a continuous process,
with considerable variability in the rate of recovery for different
animals. Data obtained from the other animals were generally consistent
at a given defined stage.
To examine the physiological effects of axotomy upon HG motoneurons,
four animals were examined three to five weeks (25, 26, 33 and 35 days)
following nerve section. No tension could be elicited from the MG muscle
by MG nerve stimulation, nor were electomyographic responses seen.
Although it is not known whether there were connections which were too
weak to elicit tension, structural regeneration of the nerve terminals
is not necessarily immediately accompanied by functional neural
transmission (Birks et al. 1960; but see Carmignoto et al. 1983).
The mean values for motoneuron electrical properties are seen in
Table 4-2. Frequency histograms for motoneuron electrical properties
following axotomy are seen in Figures 4-1 and 4-2. There was a marked
decrease in axonal conduction velocity in axotomized motoneurons (Ecoles
et al. 1958; Kuno et al. 1974a,b; Hoffer et al. 1979; Milner and Stein
1981; Cragg and Thomas, 1961; Mendell et al. 1976; Kiraly and Krnjevic,
1959; Gallego et al. 1979b). The range of values for axonal conduction
velocity was similar to control, although shifted to lower values (Fig.
4-2B). This decreased conduction velocity is believed to be due
primarily to decreased axonal diameter (Gutman and Sanders, 1943;
Sanders and Young, 1946), as well as shorter internodal distance and
altered ion channel distribution (Beeryet al. 1944; Cragg and Thomas,
1964; Ritchie, 1982; Kocsis et al. 1982; Kocsis and Waxman, 1983).
Myelin thickness does not change markedly with axotomy (Gillespie and
Table 4-2. Motoneuron Electrical Properties: Axotomy and Time Course of
FF FI FR S ALL
161 (6) 111 (37)
162 (9) 112 (13)
183 (3) 111 (18)
50 (34) 141 (147)+
31 (8) 101 (33)'
61 (19) 151 (73)+
INPUT RESISTANCE ( :hms)
1.11 (30) 1.50 (28)
1.70 (12)' 1.90 (7)
1.20 (10) 1.30 (11)
RHEOBASE/INPUT RESISTANCE (nA/Mohms)
41t (28) 212 (117)+
42 (7) 123 (29)'+
61 (11) 233 (45)+
AHF HALF-tDF'IY T"T? (ur-
192 (5) 251 (31)
161 (9) 232 (12)
231 (3) 242 (18)
AXO'IAL CC;PU'TIQON VELO:iTY fm1 -
971 (55) 1004 (7)
801 (14)0 751 (9)'
962 (18) 812 (19)
a. Means + SE (number of units).
b. = Significant difference from NORMAL (p<0.01);
c. + = Significant difference from NO-RE (p<0.01);
There was no change in mean AHP half-decay time following axotomy
(Gustafsson, 1979; Kuno et al. 1974b), but the range was compressed from
both extremes (Fig. 4-2A). In contrast, Gustafsson and Pinter (1984b)
reported a slight but significant increase in AHP duration for
lumbosacral motoneurons (16 of which were triceps surae). Kuno et al.
(1974a) also showed a slight increase in MG AHP durations following
axotomy. This later study also examined the slow soleus motoneurons,
which exhibited markedly decreased AHP durations following axotomy. This
finding, in conjunction with the altered distribution of AHP durations,
suggests that the difference between studies reflects the relative
composition of fast and slow motor units in the motoneuron pool
examined. The presence of approximately 25% slow units in MG may offset
the change in AHP duration of fast units (Kuno et al. 1974a; Gustafsson
and Pinter, 1984). Gustafsson and Pinter (1984) suggested that the
compressed range for AHP duration with axotomy indicated that fast
motoneurons increase AHP duration, while slow motoneurons decrease AHP
duration. Alteration in AHP half-decay time could reflect changes in the
kinetics of the Ca++-dependent K+ conductance thought to underlie the
AHP (Barrett et al. 1980; Krnjevick et al. 1978) or changes in the
magnitude of membrane nonlinearities (Ito and Oshima, 1965; Gustafsson
and Pinter, 1985).
Most studies on spinal motoneurons have reported an increase in
input resistance with axotomy (Gustafsson, 1979; Gustafsson and Pinter
1984). In contrast, Kuno and Llinas (1970) reported no change of input
resistance following nerve section. The discrepancy may be due in part,
to different criteria for recognizing axotomized cells.
S t; 5E :.2a;
:h; ES S'L; .Y *
Figure 4-1. Frequency histograms for MG motoneuron properties following
section of the MG nerve (filled), compared to normal MG motoneurons
(unfilled). (A) Rheobase, note decreased range and mean vs. controls.
(B) Input resistance, note similar range, but shifted to higher values
than controls. (3) Ahp half-decay time, note range compressed from both
extremes, no change in means. (D) axonal conduction velocity, range is
decreased and mean is lower.
Kuno and Llinas (1970) included only those cells exhibiting a 'partial'
active dendritic response to afferent input, in their axotomized
population. Also, in that study motoneurons were axotomized by ventral
root section, as opposed to peripheral nerve section in the present work
(also Gustaffson, 1979; Gustafsson and Pinter, 1984). The axon reaction
is known to be dependent upon the length of the intact proximal stump
(Kuwada and Wine; 1981, Lieberman, 1971; Mendell et al. 1976). The
range of values for input resistance was similar to that of normal MG,
although shifted to higher values (Fig. 4-1B). Mechanisms thought
responsible for alterations of input resistance with axotomy are
discussed by Gustafsson and Pinter (1984).
Rheobase was decreased in axotomized motoneurons, in agreement with
earlier reports (Eccles et al. 1958; Kuno and Llinas, 1970; and
Gustafsson, 1979). The range of values seen for rheobase was compressed
as well as shifted to lower values (Fig. 4-1A). Presumably the
decreased rheobase indicates altered excitability of the initial segment
and the soma-dendritic membrane (Eccles et al. 1958; Faber, 1984).
The relationship between rheobase and input resistance is altered by
axotomy (Fig. 4-3A). The range of values was compressed. The mean value
for rheobase/input resistance was lower than normal (Table 4-2).
When compared to motoneurons from normal type S units, axotomized MG
motoneurons had significantly higher input resistance, slower axonal
conduction velocity and lower rheobase/input resistance. Thus, while
some axotomized cells fell within the type S range for some parameters,
they were not identical in properties to normal type S motoneurons.
kn 3S 35
a Ii--l "^ -
I- 2S -
0 15 0 is
29 48 6 8 18N 5s 73 9 lit 1 IN
RHP HALF-DECAY TIME (ms) RXONRL CONDUCTION VELOCITY (m/s)
Figure 4-2. Frequency histograms for MG motoneuron properties following
section of the MG nerve (filled), compared to normal MG motoneurons
(unfilled). (A) AHP half-decay time, note range compressed from both
extremes, no change in means. (B) axonal conduction velocity, range is
decreased and mean is lower.
In the no-re population, the correlation seen in control motoneurons
between rheobase and input resistance was lost (Fig. 4-3A), as was that
between AHP half-decay time and axonal conduction velocity (Fig. 4-4A).
There was no segregation of motoneurons with respect to these
properties. For AHP half-decay time vs. axonal conduction velocity, 85%
(51/60) of motoneurons sampled were outside the normal range for either
fast or slow motoneurons. Motoneuron types based on the ratio rheobase:
input resistance and AHP half-decay time (Zengel et al. 1985) could not
be recognized in the no-re population.
These data are consistent with a dedifferentiationn' of motoneuron
properties from the normal adult state (Kuno et al. 1974a; Gustafsson
and Pinter, 1984). While some properties of axotomized motoneurons are
similar to those of normal type S motoneurons (rheobase, input
resistance, rheobase/input resistance), AHP half-decay time is shorter,
and axonal conduction velocity is slower than in normal type S units.
Thus, the dedifferentiated state is not simply an 'S' or 'super S'
motoneuron type (Gustafsson and Pinter, 1984).
We examined two animals (39 and 40 days post-operative: low-re)
which had MG muscles which generated 41 and 115 g-wt. twitch tension,
respectively, in response to MG nerve stimulation. Previous studies
indicate that a minimum of about four weeks is required before muscle
contraction can be elicted by nerve stimulation (Eccles et al. 1962;
Gordon and Stein, 1982b).
a i r i 1CF:
r Ii I, a ar i
* Li U aa L I 1 I I
Figure 4-3. Relationships between rheobase and itput resistance
following section of the MG nerve, and at early and intermediate times
of reinnervation. (A) axotomy, note uniformity of the sample and
altered distribution (N= non-contract). (B) low-re, distribution is
similar to no-re stage; no differences between contracts (C) and
non-contracts (N). (C) med-re, note recstablishment of sr-egation by
motor unit type, although distribution is not at normal levels.
Of 37 motoneurons sampled, 20 elicited muscle unit contraction
(57%). Assuming 280 motoneurons in the normal M1 motor nucleus (Boyd and
Davey, 1968; Burke et al. 1977), this suggests that approximately 160
axons innervated MG muscle.
Motoneuron electrical properties. Low-re motoneurons had overall
mean values which were significantly different from controls for all
motoneuron electrical properties except AHP half-decay time (Table
4-2). In most cases the overall means were indistinguishable from the
no-re population. There were no significant correlations between any
motoneuron parameters at the low-re stage, unlike controls, but similar
to axotomy (Figs. 4-3B, 4-4B).
Motoneurons at the low reinnervation stage were divided into two
groups on the basis of whether they elicited muscle contraction
(contracts) or not (non-contracts). The only differences in mean values
between these two groups were rheobase and input resistance (both lower
in contracts; Table 4-3). The similarity between properties of contracts
and non-contracts could be interpreted in two ways: either making
functional reconnection is not sufficient for re-establishment of normal
motoneuron properties, or additional time is required following
reinnervation for normal electrical properties to be expressed.
Kuno t1 aL (1974b) found two MG motoneurons (five months
self-reinnervated muscle) which did not elicit muscle tension, but did
not differ from connected cells in AHP duration, axonal conduction
S( *p. Cf [
axA ON L t.
B CO )iDK lD'. t.IXXlt1. 1
4*- J* p. ---- 0 0 40.__~.i.
Figure 4-4. Relationships between AHP half-decay time and axonal
conduction velocity after nerve section, and at early and intermediate
stages of reinnervation. (A) no-re. Virtually all motoneurons fall
outside the normal range. (B) low-re. Virtually all motoneurons fall
outside the normal range. There was no difference between contracts (C)
and non-contracts (N). (C) med-re. Many contracts (F,I,R,or S) approach
or enter the range of control motoneurons, while all non-contracts (N)
fall outside the normal range. There was segregation of fast and slow
motoneurons. Solid lines in all graphs outline the distribution of slow
(upper) and fast motoneurons in controls.
velocity, or action potential amplitude. They preferred the explanation
that functional reoonnection is neither necessary nor sufficient for
recovery of motoneuron properties.
That functional reconnection is necessary for re-establishment of
normal electrical properties is suggested by the similarity between
non-contracts at all times following axotomy (Table 4-4). At the med-re
and long-re stages non-contracts had mean values significantly different
from the overall sample at the same stage. Mean values for rheobase and
rheobase/input resistance (med-re only), input resistance, and axonal
conduction velocity also differed from type S units at the same stage of
recovery. AHP half-decay times were short, as in axotomy, and different
from control type S units. These data are consistent with suggestions,
based on recording extracellular action potentials of single units, that
the effects of peripheral axotomy are not continuous, but reach a
stable, plateau value (Davis et al. 1978). This suggests that motoneuron
survival following peripheral axotomy is not dependent upon connection
to muscle in regenerating adult motoneurons, although expression of
normal electrical properties is. All cells at the long-re or med-re
stages which made functional reconnection with the MG muscle showed
recovery of motoneuron properties, suggesting that reconnection, with
sufficient recovery time, is sufficient for expression of mature
There were some non-contracts in the long-re and med-re populations
with properties in the normal range for one or several motoneuron
properties. These cells could have been injured in the final dissection
or perhaps represent alpha-fusimotor innervation (Gregory et al. 1982).
Table 4-3. LOW-RE Motoneuron Electrical Properties: Contract vs.
RN b (Mohms)
AHP HALF-DECAY TIME (ms)
AXONAL C.V. (m/s)
a. Means SE (number of cells).
b. RN = input resistance.
c. = significant difference fr
(alpha = 0.01).
d. C.V. = conduction velocity.
om NON-CONTRACT (alpha = 0.05);
Table 4-4. Motoneuron Electrical Properties of Non-contracts: Self-
reinnervation Model. a,c
RN b (Mohms)
AHP BALF-DECAY TIME
AIONAL C.V. (m/s)
a. Means SE (number of cells).
b. RN = Input resistance.
c. = significant difference from NO-RE (alpha = 0.05);
0" (alpha = 0.01).
d. C.V. = conduction velocity.
--- "" ""~ "'
At the long-re stage the vast majority of non-contracts were outside of
the normal range for AHP half-decay time vs. axonal conduction velocity
(o.f. Fig. 3-4).
Muscle unit contractile properties. All speed-related properties of
the whole muscle twitch were prolonged at the low-re stage (Table 4-1;
twitch time-to-peak, half-relaxation time, half-rise time). Prolonged
isometric time-to-peak has been reported to occur with denervation (Kean
t al. 1974; Spector, 1984), with relaxation prolonged more than twitch
rise-time. This has been attributed to alterations in the sarooplasmic
reticulum, troponins, and myosin in the denervated muscle fibers (Cecchi
et al. 1984; Spector, 1984). At the low-re stage, motor units did not
fit into adult motor unit types and so were treated as a single
population. The mean value for low-re units' potentiated twitch
time-to-peak was not significantly different from control (Table 4-5).
Half-relaxation time was prolonged in low-re units. Amplitudes of
potentiated twitch and maximum tetanus were greatly reduced in low-re
units. All low-re units were fatiguable, but most showed residual
tension at the end of the fatigue test.
We found no significant correlations between motor unit contractile
properties in the low-re population. The relationship between twitch
time-to-peak and maximum tetanus (-0.54, p<0.07)) approached
These results are in agreement with Gordon and Stein (1982a,b) in
that initially after reinnervation mean unit tensions were smaller,
with no correlations between motor unit speed and tension.
Table 4-5. Muscle Unit Contractile Properties: Time Course of
FF FE FR S ALL
TWITCH TIME-TO-PEAK d(ms)
282 (7) 261 (30)
TWITCH TENSION d(g-wt.)
41 (7) 10 (28) 0.20 (22)
62 (10) 10 (13)
2.6 (3) 10 (17)
TWITCH HALF-RELAXATION TIME d(ms)
272 (7) 261 (29)
20+3 (7) 13i1 (45)
71 (39) 352 (175)
4.+1 (9) 163 (36)*
6-1 (16) 232 (67)*
0.00 (82) 0.50 (7)
0.1+0 (3) 0.50 (10)
0.10 (33)* 0.50 (3)
0.7 0 (34)*
Means + SE (number of units).
* = Significant difference from NOERAL MG (p<0.01).
Muscle histoohemistry. Low-re MO contained 50% type FG muscle
fibers, 28% type FOG, and 23% type SO, similar to controls (Table 4-7).
It is uncertain to what extent this reflects the original innervation of
the muscle fibers, as fiber type grouping (Dubowitz, 1967; Romanul and
Van Der Meulen, 1967; Karpati and Engel, 1968a) was not seen at this
stage. Mean muscle fiber area was decreased at the low-re stage, with
fast fibers (types FG and FOG) affected more than type SO fibers (Table
The muscles from the animals intermediate in recovery between low-re
and med-re stages contained 55% type FG, 17% type FOG, and 28% type SO
fibers (FE12) and 44% type FG, 32% type FOG, and 25% type SO fibers
(FE17). Of 12 motor units sampled in cat FE12, six were type FF, one
was type FR, three were type S motor units, and two were non-contracts.
No motor unit data were obtained from FE17.
Relationships between motoneuron electrical properties and muscle unit
contractile properties.. We found no significant correlation between
any combination of motoneuron and muscle properties at the low-re stage
(Gordon and Stein, 1982b). At two months following the initial surgery,
Gordon and Stein (1982a,b) reported no correlation between action
potential amplitude (correlated with axonal conduction velocity) and
twitch tension, or between twitch contraction time and twitch tension.
Medium Reinnervation Stage
Two animals were examined which generated greater than 50S of
long-re whole muscle twitch tension (63 and 71 days; med-re). At this
stage motor unit types could be distinguished by contractile criteria.
The proportion of fast and slow units at the med-re stage was similar to
normal MG (75% fast, 25% slow vs. 76:24 in normal HM; 69:31 in long-re
MG). There were 8% type FF, 28% type FI, 39% type FR and 25% type S
units (controls: 48:4:24:24; long-re: 50:4:15:31; Table 7). The
increased proportion of type FI units appears to be at the expense of
type FF units. This may indicate a general increase in fatigue
resistance of fast units at this stage, perhaps due to previous
polyneuronal innervation, increased recruitment due to lowered overall
force-generating ability of the muscle, or a general increase in
oxidative enzymes as a function of reinnervation.
Statistical differences between means for motor unit parameters at
the med-re stage were similar to normal and long-re motor units (Table
Motoneuron electrical properties. The overall means for med-re
motoneuron electrical properties show a definite shift towards those of
the long-re and normal populations (Table 4-2), although axonal
conduction velocity, rheobase/input resistance and rheobase were still
lower, and input resistance was still higher in med-re than long-re or
normal motoneurons (Table 4-2). Mean AHP half-decay time was
essentially the same at all stages.
Table 4-6. Results of Tukey's Studentized Range Test: Significance of
Differences Between Motor Unit Types for MED-RE Motor Units.
Axonal Conduction Velocity
AHP Half-Decay Tipe
Twitch Tension a'
Twitch Time-To-Peak b
Twitch HRT b',
Maximum Tetanic Tension a
a. The same result was obtained for raw
b. Potentiated twitch.
c. HRT = half-relaxation time.
data and data normalized for
Mean values by motor unit type, for motoneuron electrical
properties, were generally similar to properties of the same type of
motoneuron in normal or long-re MG motoneurons. Only input resistance
(type FR) and axonal conduction velocity (types FF and FR) differed
between long-re and med-re motoneurons of the same type (Table 4-2).
Fig. 4-3C shows the relationship between rheobase and input
resistance for med-re motoneurons. Note that segregation by motor unit
type is re-established at this stage, although lower rheobases and higher
input resistances tend to alter the distribution from normal. The
correlation between log rheobase and input resistance was -0.80
(p<0.0001). Within unit types log rheobase and input resistance were
correlated for type S units only (-0.80, p<0.03)).
AHP half-decay time was negatively correlated with axonal conduction
velocity across all units (r= -0.50, p<0.0001; Fig. 4-4C), as in normal
and long-re motoneurons (Fig. 3-2). This relationship was significant
within type FR units (-0.79, p<0.002), but not type FF or type S units.
There was some segregation into fast and slow groups by the relationship
AHP half-decay time: axonal conduction velocity (Fig. 4-4B), although
this relationship was not as in normal or long-re motoneurons (Fig 3-4).
Five of eight type S motoneurons were within the normal range for this
relationship (63%), as were 10 of 22 fast motoneurons (45%). All
non-contracts (12/12) fell outside the normal range for either fast or
Motoneuron type was determined according to the criteria of Zengel
et al. (1985; AHP half-decay time >30ms = fast, AHP <30ms = slow; type
FF units have rheobase/input resistance >18, type S units <7, and type
FR units between 7 and 18). At the med-re stage, there was an 82%
agreement (18/22) between electrically determined motoneuron type, and
motor unit type determined from contractile properties. AHP half-decay
time predicted fast and slow motor units in 91% of cases (29/32).
Muscle unit contractile Droperties. At this stage certain
contractile properties have recovered more than others (Table 4-5). The
overall means for twitch time-to-peak, half-rise time, and
half-relaxation time were similar to normal. Twitch tension (raw data
and normalized by body weight) recovered nearly to the level shown by
long-re units (smaller than control values). This was also true of
maximum tetanus (raw and normalized). Overall fatigue index was higher
than for normal or long-re units, reflecting the increased proportion of
types FR and FI units. All units displayed some residual tension after
the fatigue test.
Across the whole population, twitch time-to-peak was significantly
correlated with the log of twitch amplitude/body weight (r= -0.32,
p<0.04) and tetanic tension/body weight (r= -0.58, p<0.0001), as in
normal and long-re units (CHAPTER III). These relationships did not
hold within any motor unit type.
Relationships between motoneuron electrical properties and muscle
urlt sc. tr.ctile tLrorertie As noted above, there was good agreement
between motoneuron type and motor unit type at the med-re stage. Most
correlations found between motoneuron and muscle properties in normal
and lonr self-reinnervated MG had recovered by the med-re stage. Across
all units, AHP half-decay time was positively correlated with twitch
time-to-peak (0.82, p< 0.0001). This relationship was not significant
within motor unit types. Axonal conduction velocity was correlated with
twitch time-to-peak (-0.51, p<0.0001), the log of twitch tension/body
weight (0.43, p<0.01), and log maximum unit tetanic tension/body weight
(0.63, p<0.0001). The relationship with time-to-peak was only
significant within FR units (-0.73, p<0.03) and the relationship with
log tetanic tension/body weight within FR units (0.59, p<0.02). There
were no significant correlations between axonal conduction velocity and
log twitch tension/body weight within motor unit types.
Gordon and Stein (1982a,b) noted a parallel recovery of axonal
action potential amplitude and motor unit twitch tension. In their
study, units recovered normal tension five to six months following
surgery (nerve-nerve suture). The normal positive correlation across the
whole population of units, for action potential amplitude and twitch
tension, was lost in early reinnervation, and did not recover until the
sixth postoperative month (Gordon and Stein, 1982b). The normal negative
correlation between twitch contraction time and tension recovered
earlier, at three months. These authors felt that the early lack of
correlation between axonal action potential amplitude and twitch tension
may be explained by motor units being heterogeneous with respect to
metabolic enzymes at early stages. They suggested that recovery of
normal relationships between nerve and muscle depends upon time and
extent of whole muscle recovery, with whole muscle recovery having the
best predictive value.
Muscle histochemistry, muscle fiber areas. and innervation ratios.
Histochemistry did not indicate the same magnitude of shift towards
fatigue-resistant units as did the motor unit type distribution. At this
stage 43% of the muscle fibers were type FO, 24% type FOG, and 33% type
SO (Table 4-7). The med-re stage was the first one in which fiber-type
grouping (Dubowitz, 1967) of muscle fibers was seen. The overall
distribution was significantly different from controls (Chi-square, 95%
confidence level). Comparison to expectations for an assumed
hypergeometric distribution showed no significant differences for any
individual muscle fiber type. Oxidative fibers (FOG+SO) made up 57% of
the population, as compared to 44% in controls (31% in long
self-reinnervated). This change is in the direction suggested by the
motor unit type distribution. Perhaps the qualitative comparison of NADH
staining intensity is insufficiently sensitive to detect the degree of
changes in enzyme levels, or the relationship between NADH staining and
fatigue resistance is not particularly strong (Baldwin et al. 1984).
We estimate physiological cross-section at 2.4cm2. Mean fiber areas
and relative innervation ratios are found in Table 4-7. Mean fiber areas
show recovery towards normal values for types FG and FOG fibers, while
type SO fibers are smaller than controls. (see CHAPTER III).
With denervation, types FG and FOG fibers atrophy to a greater
extent than type SO fibers (low-re stage; see also Lowrie and Vrbova,
1984). With time, type SO muscle fibers decrease in size until the
med-re stage, then recover to normal levels. Type FOG fibers recover to
normal size earlier, at the med-re stage. Type FG fibers, on the other
hand, never recover to control levels, although type FG fibers remain
the largest. Motor unit tension recovery follows the pattern of
recovery of muscle fiber area (see below).
Table 4-7. Motor Unit Types, Musole Fiber Types, and Innervation Ratios:
Time Course of Self-Reinnervation. 6
MUSCLE FIBER TYPES
I Histochemical Composition
FG FOG SO n/N a
NORMAL MG 56 23 21 4200/7
LOW-RE 50 28 23 900/2
MED-RE 43 24 33 900/2
LONG-RE 54 15 31 3000/5
Observed Mean Fiber Area (um2)
NORMAL MG 3873 2264 1972 3104
LOW-RE 1439 1274 1803 1491
MED-RE 2125 1634 1211 1714
LONG-RE 3098 2355 2296 2738
EQUIVALENT MOTOR UNIT TYPE
I Of Population
FF+FI FR S n/N a
NORMAL MG 48+4 24 24 176/16
MED-RE 8+28 39 25 36/2
LONG-RE 46+4 24 26 72/4
Est. Number in Pool d
NORMAL MG 134+11 67 67 280
LOW-RE X X X 160
MED-RE 2+59 82 53 210
LONG-RE 116+10 61 66 252
FF+FI FR S
Relative IR c
NORMAL MS 1.1 1.0 0.9
MED-RE 1.2 0.6 1.3
LONG-RE 1.1 0.6 1.2
a. Number of Cells/Number of Animals.
b. [280 (280)(% non-contracts)]
c. Relative Innervation Ratio = %Muscle Fiber Type/% Motor Unit Type.
d. Estimated as (Total number of motoneurons)(% motor unit type).
e. Motor unit types could not be distinguished at the LOW-RE stage.
Relative innervation ratios indicate a possible disadvantage of type
FR motoneurons and advantage of type S motoneurons in capturing muscle
fibers. At the long-re stage there is an even greater trend towards low
relative innervation ratio of type FR units and increased ratio in
type S units than observed at the long-re stage (CHAPTER III).
Gordon and Stein (1982a,b) used chronic recording techniques to
follow the time course of self-reinnervation of cat MG. They calculated
that the estimated number of units did not change from early
reinnervation to the plateau of recovery although whole muscle and
single unit tensions increased (see also Kuno et al. 1974a). These
authors speculated that the primary change was an increase in fiber
diameter. Our data support the notion that a large proportion of the
increase in muscle tension from the med-re stage to the long
self-reinnervated stage is due to increase in fiber area (160$ increase
in mean fiber area, 120% increase in whole muscle twitch tension, 129%
increase in mean motor unit tetanic tension). After nine months
reinnervation, all three of these parameters were still below normal
values (88%, 78%, 63% of normal, respectively).
We saw a larger proportion of non-contracts at the med-re
stage, relative to nine-month reinnervated MG (25% vs. 9%). We estimated
the number of innervating axons from the percentage of non-contracts
(12/48 = 25%) as 210. This indicates that a proportion of axons, which
eventually form functional connections with muscle fibers, had not yet
done so at 63 or 71 days. Alternatively, a greater percentage of units
at this stage which are connected do not elicit measurable levels of
At the med-re stage we found significant correlations between
contractile parameters, as well as between axonal conduction velocity or
AHP half-decay time and contractile parameters. It appeared however,
that some axons destined to do so had not made functional connections at
this time (63-71 days). In general, recovery of muscle and motor unit
properties occurred in parallel.
This study examined the influence of functional connection with
muscle on expression of motoneuron electrical properties. Also examined
was the time course of recovery of motor unit properties during
self-reinnervation of the cat MG muscle. The purpose was to determine
the level from which motoneurons must recover, and the progression of
recovery, in order to place the long-term reinnervation results in
perspective. In addition, such data were required for interpretation of
ongoing studies of reinnervation of foreign muscles by motoneurons.
Following nine months recovery from nerve section and self-reunion,
most motoneuron and muscle unit properties recover to control levels,
and normal relationships between motoneuron and muscle properties are
re-established (Eagust et al. 1981; Lewis et al. 1982; Chan et al. 1982;
Burke, 1980; CHAPTER III). In this study we examined motoneuron
electrical properties and the contractile properties of the respective
muscle units, after intermediate periods of recovery from nerve section.
Following section of the M3 nerve, MG motoneurons become more
excitable to somatic current injection (decreased rheobase), depolarize
more in response to a given current input (increased input resistance),
and M3 axons conduct impulses at a lower rate. Axotomized motoneurons
did not segregate on the basis of either the ratio rheobase: input
resistance or the relationship of AHP half-decay time and axonal
conduction velocity. All of these data support the notion that following
axotomy, motoneurons 'dedifferentiate' (Kuno et al. 1974a; Gustafsson
and Pinter, 1984). Dedifferentiation can be regarded as a shift to a
growth state for the motoneuron (Watson, 1976).
Huiszar et al. (1975) noted the resemblance between axotomized
motoneurons and those of kittens, and suggested that this
dedifferentiation returns adult motoneurons to a state similar to early
ontogeny. Gustafsson and Pinter (1984a) suggested that axotomized
motoneurons dedifferentiate to a level similar to that of adult slow
motor units. The present data show that while many electrical properties
of axotomized motoneurons overlap the range of type S motoneurons, in
most cases the mean values for axotomy are different, and certain
parameters (AHP half-decay time, axonal conduction velocity) are very
different in axotomized motoneurons than in normal type S motoneurons.
In addition, the relationship between AHP half-decay time and axonal
conduction velocity was different for axotomized motoneurons than for
type S motoneurons. Thus type S motoneurons, as well as fast
motoneurons, represent a differentiatiated state from axotomized or
Functional connection to muscle and expression of motoneuron electrical
One issue we addressed was whether functional reconnection of
motoneuron to muscle fibers was necessary or sufficient for recovery of
motoneuron electrical properties. Kuno et al. (1974b) measured axonal
conduction velocity, action potential overshoot amplitude, resting
membrane potential, and duration of the AHP at various times (up to five
months) following self-reinnervation of cat MG and soleus. At early
times, with incomplete reinnervation, they reported no difference
between properties of those motoneurons which elicted muscular
contraction, and those that did not (both groups appeared axotomized).
They interpreted this as evidence that functional reconnection is not
sufficient for recovery of normal motoneuron properties. Kuno et al.
(1974b) reported two of 25 M3 motoneurons which did not elicit muscle
tension following five months recovery, but had AHP duration and axonal
conduction velocity in the normal range. Their interpretation was that
functional reconnection was not necessary for recovery of motoneuron
properties. One potential complication bearing on this interpretation
is that properties of axotomized motoneurons overlap somewhat with
normal type S motor units in terms of electrical properties (Gustafsson
and Pinter, 1984; this study). In the Kuno et al. (1974b) study, all MG
motoneurons were regarded as 'fast', and all soleus motoneurons as
'slow'. Normal MG contains approximately 25% type S motor units (Burke
et al. 1973; Burke, 1981; Fleshman et al. 1981; CHAPTER III).
Our data show that motoneurons which do not make functional
reconnection with muscle display membrane electrical properties similar
to axotomized motoneurons. At the low-re stage, electrical properties of
motoneurons with functional reconnections, and those of non-contracts
were not different from one another, and neither differed from the
axotomized condition, in agreement with Kuno et al. (1974b). This could
indicate that functional reinnervation is not sufficient for recovery of
motoneuron electrical properties or that time is required for these
parameters to reach levels permitted by reconnection.
At the med-re and long-re stages, motoneurons without functional
reconnection to muscle (non-contracts) have electrical properties
different from those motoneurons which elicit muscle contraction. In
particular, the relationships between rheobase and input resistance, and
between AHP half-decay time and axonal conduction velocity, are
different in these cells. At the long-re stage, all motoneurons with
functional connections had electrical properties similar to normal MG
motoneurons innervating the same type of motor unit. At the med-re
stage, motoneuron properties were not fully recovered, but all cells
which elicited muscular contraction showed electrical properties altered
significantly from axotomy (in the direction of controls), and
segregated according to motor unit type with respect to rheobase, input
resistance, and AHP half-decay time. Collectively, these data suggest
that functional reconnection is necessary for re-establishment of
motoneuron electrical properties. In agreement with our results are
those of Gordon and Stein (1982b) who reported a subpopulation of axons
which did not elicit muscle contraction, and which did not recover
normal action potential amplitude extracellularr; a function of axonal
diameter; see also Gordon, 1983; Gutman and Sanders, 1943). All axons
eliciting muscle contraction in their study recovered to normal levels
of action potential amplitude by nine months to one year
Thus, our data suggest that functional reconnection is necessary and
perhaps a sufficient condition for recovery of motoneuron electrical
properties. A similar situation occurs in the goldfish Mauthner cell,
where membrane electrical properties do not recover in the absence of
connection to an end organ (Faber, 1984). In contrast, Kuwada and Wine
(1981) found that the effects of axotomy on somatic excitability of
crayfish central neurons was a transient one, regardless of whether the
axon regenerated and formed new connections. There appears to be
variability in the degree to which the periphery influences neural
properties (Gordon, 1983).
Price (1974) suggested that the motoneuron soma receives two types
of signal from its periphery, an axon-to-soma signal concerning the
state of the axon, and a muscle-to-nerve-to-soma signal, dealing with
the relationship between muscle and nerve. Consistent with this, many
changes seen with axotomy in sympathetic ganglion cells can be induced
by blockage of axonal transport by colchicine (Purves, 1976). Recent
evidence suggests that muscle-derived trophic factors exist (reviewed in
Slack et al. 1983), and that uptake of chemical signals from muscle is
activity-dependent (Watson, 1969; Brown and Ironton, 1977; Duchen and
Strich, 1968; Czeh et al. 1978). Block of action potential conduction
with tetrodotoxin (TTX) has been shown to alter AHP duration in a way
similar to axotomy in cat soleus motoneurons (Czeh et al. 1978), as well
as AHP half-decay time, axonal conduction velocity, rheobase, and input
resistance in cat MG motoneurons (Munson et al. 1985). Electrical
stimulation distal to, but not proximal to the TTX cuff was observed to
reverse the change in AHP duration in soleus motoneurons
(Czeh et al. 1978). Further experiments (Gallego et al. 1979) suggest
that the metabolic state of the innervated muscle is the important
variable, rather than activity per se. Thus the muscle exerts influence
on the expression of motoneuron electrical properties.
Recognition of motor unit types and recovery of properties.
A second goal was to determine at what stage motor unit types are
first recognizable, and how this relates to the level of recovery of
motoneuron properties. We found that motor unit types became
recognizable using contractile criteria at the med-re stage (8-10
weeks), before mean motoneuron electrical properties and mean muscle
unit contractile properties had reached mature levels.
Although mean values for motoneuron electrical properties were not
at normal levels, motoneurons segregated into types with respect to
rheobase, input resistance, and AHP half-decay time at the med-re stage.
At this stage most axons had made functional connections with muscle
fibers, although apparently some axons destined to make connections had
not yet done so. Relationships between motoneuron properties, between
muscle unit speed and tension, and between motoneuron and contractile
parameters, were relatively normal at the med-re stage.
Most of the increase in muscle and motor unit tension beyond the
med-re stage can be accounted for by increase in muscle fiber area,
although there may be a small increase in number of axons with
functional reconnection. Motor unit tension recovery follows the pattern
of muscle fiber area recovery. Muscle fiber area increased 160% from the
med-re to the long-re stages, whole muscle twitch tension increased
120%, and mean motor unit tetanic tension increased 129%. Type FG muscle
fibers did not fully recover cross-sectional area, thus type FF motor
unit tetanic tension was reduced. This failure of recovery could reflect
altered activity (Lapointe and Oardiner, 1984; Spector, 1984), altered
trophic relationships between nerve and muscle (Spector, 1984; Lapointe
and Oardiner, 1984; but see Cangiano and Lutzemberger, 1980), or a
mismatch between motoneuron firing rate and the muscle's ability to
respond (Lowrie and Vrbova, 1984). In general, these data confirm the
suggestions made by Gordon and Stein (1982b) for self-reinnervation of
At the med-re stage a greater than normal proportion of the fast
units were fatigue-resistant, as indicated by the reduced proportion of
type FF, and increased frequency of types FI and FR motor units. It is
likely that some motoneurons, which originally innervated type FF units,
innervated more fatigue-resistant types FI and FR units at this stage.
Lowrie and Vrbova (1984) reported increased fatigue resistance in rat
muscle after crush injury to the nerve. Perhaps the increased resistance
to fatigue reflects increased activation of type FF units due to the
lowered tension of all motor units (see also Baldwin et al. 1984).
Increased fatigue resistance could also reflect increased activation
during the stage of poly-neuronal innervation seen with reinnervation
(McArdle, 1975; Letinsky et al. 1976; Rothshenker and McMahan, 1976), or
be a general feature of the metabolism of differentiating muscle.
Increased fatigue resistance of fast muscle is also a transient
phenomenon in developing rat muscle (Lowrie and *-bova, 1984). By nine
months, fatigue resistance of self-regenerated cat MG units was as in
normal MG (CHAPTER III; Chan et al. 1982; Burke, 1980).
There were no correlations between muscle tension and speed, axonal
conduction velocity and twitch time-to-peak, or axonal conduction
velocity and muscle tension at the low-re stage, but all these
relationships had recovered by the med-re stage, when mean whole muscle
tension was about 60% of that at nine months self-reinnervation. Gordon
and Stein (1982b) reported that the normal relationship between unit
contractile speed and twitch tension was lost in newly reinnervated MG,
but recovered when whole muscle tension was about 50% of that eventually
attained (3 months post-operative). In their study the relationship
between tension and axonal action potential amplitude recovered more
slowly (six months).
Gordon and Stein (1982b) suggested that the initial lack of
correlations was due to the early-reinnervated units being heterogeneous
in nature, since axons do not return to their original muscle fibers
(Karpati and Engel, 1968; Kugleberg at al. 1970). In support of this
notion, we find that recovery of relationships between motoneuron and
muscle properties is coincident with the re-establishment of recognizable
motor unit types. Also, muscle fiber type grouping was first evident at
the med-re stage, suggesting that motoneurons had induced myosin and
mitochondrial enzyme properties of their muscle unit by that time.
Motoneuron properties and muscle unit properties were re-established
with a similar time course.
Relative reinnervation of motor unit types. A third question was
whether any motor unit type might be favored in the process of
self-reinnervation of a mixed muscle. From the earliest stage that motor
unit types could be recognized (med-re), there was no difference in the
relative proportions of fast (FF+FI+FR) vs. slow (S) motor units,
suggesting no advantage for survival of particular motoneuron types.
We found a slight increase in the proportion of type SO muscle
fibers at the med-re stage. Together with the unchanged proportion of
type S motor units, this leads to an estimate of increased mean
innervation ratio for type S units. This appeared to be at the expense
of type FR units (type FOG muscle fibers). We observed a similar
situation after nine months self-reinnervation (CHAPTER III). Since
motor unit types could not be recognized before the med-re stage, it is
not possible to determine whether this slight advantage was due to
different rates of axonal growth (e.g. type S axons reaching the muscle
first) or more successful competition after contacts were made. These
data could be evidence for a slight advantage of type S motoneurons in
competition for muscle fibers during reinnervation as suggested by Lewis
et al. (1982; see also Ip and Vrbova, 1984).
Thus there is no strong evidence for an advantage in reinnervation
by any motor unit type, although type S units may be more effective in
competing for muscle fibers (see also CHAPTER III).
Comparison with ontoRenv. There were some parallels between the
recovery of motor unit properties following self-reinnervation, and the
progression of normal ontogeny. The above mentioned transient phase of
high fatigue resistance was one similarity (Lowrie and Vrbova, 1984).
The resemblance between axotomized adult motoneurons and motoneurons of
young kittens with respect to action potential overshoot, AHP duration,
and axonal conduction velocity, as well as the relationship between AHP
duration and conduction velocity led Huiszar et al. (1975) to suggest
that axotomy results in a dedifferentiation of motoneuron properties to
a level similar to early development.
Most motoneuron and motor unit properties that have been
investigated in kittens differentiate to the adult pattern by the sixth
to tenth post-natal weeks (Kellerth et al. 1971; Bagust et al. 1973;
Hammarberg and Kellerth, 1971; Hammarberg, 1974; Nystrom, 1968). This
corresponds to the time at which kittens acquire the ability to run, and
to the time of weaning (5-6 weeks; Peters, 1984).
During ontogeny, motoneuron soma size (Sato et al. 1977; Mellstrom
and Skoglund, 1969), dendritic morphology (Scheibel and Scheibel, 1970),
and axon myelinization (Berthold et al 1983) continue to mature well
beyond this 6-10 week period. Relative to these morphological changes,
axonal conduction velocity does not reach adult values until late in
development (although fast vs slow axons are differentiated from birth,
or close to that time; Ridge, 1967; Huiszar et al. 1975; Bagust et al.
1973). Also, although the difference between AHP duration of fast and
slow motoneurons is established by ten weeks, adult values are not
attained until late in development (Hammarberg and Kellerth, 1975),
perhaps related to the increase in dendritic tree development. This is
especially true for the slow, soleus motoneurons (Euiszar et al. 1975).
Following axotomy, motoneuron properties dedifferentiate. If
reinnervation of muscle is allowed, motor unit types are again
recognizable by 8-10 weeks. Motoneuron and muscle unit properties mature
in parallel, not reaching control levels until long after motor unit
types are recognizable. The pattern of recovery of motoneuron electrical
properties is shown in Fig. 4-5.
- C 0r -S
WI -.C S-t
a1 ------.---- -----------*
-* *u o t -> -S( i'rt- t
ril cw St *.-i*-'t.,
j ~ ^ '
-t ftt WI) it
*~Y )t li'~*r Y() C
>*0 4 -
' SC i^ t -, **-C ON,:-Wf
t...( .t *St*M.E*t i.
Figure 4-5. The time course of recovery of overall means for motoneuron
electrical properties. (A) Rheobase decreases with axotomy, then
gradulaly recovers to normal levels with time after reinnervation. (B)
Input resistance increases with axotomy, then gradually recovers to
normal levels with time after reinnervation. (C) AHP half-decay time.
The mean values do not change with axotomy of reinnervation. (D) Axonal
conduction velocity decreases with axotomy, then gradually recovers to
control levels by nine-months post-operative.
Huiszar et al. (1975) found no clear relationship between
alteration of particular motoneuron properties and changes in whole
muscle twitch parameters, during postnatal development in oat MG and
soleus. Similarly, we found no clear relationship between alteration of
single motoneuron properties and changes in muscle unit properties
during self-reinnervation of MG in adult cats. Rather, the overall
motoneuron character (as indicated by motoneuron type) was strongly
related to muscle unit type.
While there were many parallels with ontogeny, self-reinnervation of
adult muscle is not a strict recapitulation of events occurring in
development. In regeneration there are additional connective tissue
barriers to axonal growth, and cues for axonal guidance are likely to be
different (reviewed in Gordon, 1983; Vrbova et al. 1978). For example
the basal lamina adjacent to the former endplate region is known to
attract regenerating axons (Sanes, 1983). In regenerating adult muscles
all muscle fibers are present before axonal ingrowth, and these muscle
fibers have already reached a state of differentiation which was lost to
an uncertain degree following loss of innervation. Although the results
of reinnervation studies such as this one suggest that muscle fibers
possess considerable plasticity, it is not certain whether this ability
to alter properties is complete. The increased proportion of type SO
muscle fibers following reinnervation could represent a type difference
in muscle fiber or motoneuron plasticity.
PROPERTIES OF NORMAL LG AND SOLEUS
The cat lateral gastrocnemius (LG) muscle has been an important
model in studies of anatomical and histochemical compartmentation of
muscle (English and Letbetter, 1982a,b; English and Weeks, 1984; Weeks
and English, 1983, 1985), sensory partitioning within muscle (Van den
Noven et al. 1983), differential usage of muscle compartments (English,
1984; Russel et al.1984; Rushmer et al. 1984), and motor reinnervation
of muscle (Gordon and Stein, 1982a). Despite its importance, there are no
detailed reports of motor unit contractile properties or distribution of
motor unit types in this muscle, and even less information about
motoneuron electrical properties. A few LG units are included in studies
by Burke (1967; 14 units), Burke et al.(1973; 12 units), Burke et al.
(1982; 12 units), Pinter et al. (1983, unknown number of units) and
Hammarberg and Kellerth (1974; 36 units). These motor units appeared
similar to those in medial gastrocnemius (MG), and were combined with
data for that muscle in these reports. Gordon and Stein (1982a) present
data for motor unit type distribution and unit tetanic tension from LG
and soleus units combined. Hammant (1977) studied axonal conduction
velocity and contractile properties for 76 LG units, and suggested that
LG may differ from MG in possesing a higher proportion of fatiguable
units, fewer type S units, and smaller tetanic tensions.