CLINICAL ORTHOPAEDICS AND RELATED RESEARCH
Number 376, pp. 172-182
Distribution of Neuropeptides in
Synovium of the Knee With Osteoarthritis
Tomoyuki Saito, MD; and Tomihisa Koshino, MD
Synovial tissue was obtained from medial, lat-
eral, and suprapatellar sites of 21 knees (15 pa-
tients) with medial compartmental osteoarthri-
tis at surgery. All patients reported pain around
the medial joint of their knees while walking
and climbing stairs. For investigation of the
synovial innervation, six samples were stained
with modified gold chloride and the others
with an immunohistochernical method using
antisera against neuropeptides. The extent of
synovitis in each part was scored using a new 10-
point scale. The results showed that the syn-
ovium had an extensive neural network in the
somatic and autonomic nervous systems. Neu-
ropeptides were most abundant, with an espe-
cially large number of substance P and calci-
tonin gene related peptide immunoreactive free
nerve endings. Some of the substance P positive
nerve endings were surrounded by monocytes.
Substance P and calcitonin gene related peptide
were found more frequently in the medial than
in the lateral or suprapatellar areas. Substance
From the Department of Orthopaedic Surgery, Yoko-
hama City University School of Medicine, Yokohama,
Japan.
Supported by a Grant-in-Aid for Scientific Research (C)
of The Ministry of Education, Science, Sports and Cul-
ture.
Reprint requests to Tomoyuki Saito, MD, PhD, Depart-
ment of Orthopaedic Surgery, Yokohama City University
School of Medicine, 3-9 Fukuura, panazawa.,ku, Yoke-
hama 2364004, Japan.
Received: July 24, 1998.
Revised: August 23, 1999; December 1, 1999.
Accepted: December 3, 1999.
172
0 2000 Lippincott Williams & Wilkins, Inc.
P positive free nerve endings showed more den-
dritic morphologic features in the medial region
than did those in the lateral and suprapatellar
regions, and small nerves were accompanied by
newly developed vessels in synovial villi. In the
medial region, the synovitis was more remark-
able than in the lateral region. These findings
suggest that free nerve endings containing sub-
stance P may modulate inflammation and the
pain pathway in osteoarthritis.
Progressive failure of articular cartilage in-
duced by biomechanical factors is considered a
main cause of osteoarthritis, which is deemed
a noninflammatory ~ondition,~ but the syn-
ovium shows severe inflammatory change in
some cases.29Osteoarthritic synovitis may be
merely a secondary reaction brought about
by the debris of cartilaginous destruction:
but some patients with mildly destructive
changes seen on radiographsreport spontaneous
pain in their knees from joint swelling and lo-
cal heat. The clinical severity of osteoarthritis
does not always correspond to radiographic
findings. The development and modulation of
synovitis in osteoarthritis have not been ex-
dained.
Patients with arthritis report joint pain,
which induces disabilities in daily life, and
various treatments may be prescribed accord-
ing to the severity of pain. There have been
thorough studies on joint innervation and the
pain sensory mechanism in arthritis.13,26,32
Number 376
July, 2000
However, the provocation of pain in arthritis
remains a matter of controversy.
The role of neurological mechanisms in the
pathophysiologic features of joint diseases has
been discussed in several recent articles. Several
neuropeptides have been identified in the pe-
ripheral and central nervous systems, and phys-
iologic properties of some of these neuropep-
tides in the inflammatory process have been
~ l a r i f i e d . ~ .These
~ ~ . ~ 'various issues prompted
the authors of the current study to determine the
distribution of neural elements and location of
neuropeptides in the synovium of the os-
teoarthritic knee using specific immunohisto-
chemical staining in examining the role of neu-
ropeptides in modulation of synovitis and the
pain pathway.
MATERIALS AND METHODS
Gross Synovial Innervation
To visualize the synovial innervation of the knee
before the immunohistochemical study, the authors
obtained six samples of synovial tissue from three
osteoarthritic knees. These were placed in normal
saline solution and stained in bulk using a modified
gold chloride method.33 After the specimens were
stained, they were rinsed with isopentane for 30
seconds and frozen. The frozen specimens were
sectioned on a sliding microtome at 100 pm. The
floating sections in alcoholic gelatin were mounted
on slides, dehydrated with graded ethanol, cleared
in xylene, and observed with a light microscope.
Immunohistochemical Study
Materials
Synovial samples from 18 knees (11 right, seven
left) of 15 patients (11 women, four men) with me-
dial compartmental osteoarthritis' were obtained at
the time of surgery. High tibial osteotomy was per-
formed in 11 knees, hemiarthroplasty in five, and
total knee replacement in two because of sponta-
neous rupture of the anterior cruciate ligament. The
patients' ages at surgery ranged from 55 to 78
years, with an average of 68 years. Before the op-
erations, all patients had severe spontaneous pain
and tenderness on the medial sides of the knees,
with walking ability limited to less than 500 m be-
cause of the pain. All of the affected knees had
varus deformity, with an average femorotibial an-
Synovial Innervation in Osteoarthritis 173
gle of 186" (6" anatomic varus angulation) (Table
1). The patients' clinical signs and symptoms were
evaluated using a knee score of the Hospital for
Special Surgery.28The average total score was 61
2 9 points (Table 2).
The synovial samples were obtained from three
different parts of the knee. Preoperative permission
was obtained from all patients. Seventeen synovial
samples were from the medial knee compartment,
15 were from the lateral compartment, and 15 were
from the suprapatellar pouch. The samples were
placed immediately in Zamboni's fixative, fixed for
3 days on a shaker at 4" C, washed in 10 mmol
phosphate buffered saline (pH 7.4), and immersed
for 2 hours each in phosphate buffered saline con-
taining 7%, lo%, and 15% sucrose. They were im-
mersed overnight in phosphate buffered saline con-
taining 20% sucrose and frozen in Optimal cutting
temperature compound (Tissue-Tek; Miles Labo-
ratories, Naperville, IL). Antisera used in this study
were substance P (Cambridge Research Biochemi-
cals, Cambridge, UK, 1:1000 dilution), calcitonin
gene related peptide (MILAB, Malmo, Sweden,
1:4800 dilution), neuropeptide Y (MILAB, 1: 1600
dilution), and vasoactive intestinal peptide (MI-
LAB, 1:10,000 dilution).
Immunohistochemical Procedure
Cryostat sections, 15 pm thick, were placed on
poly-L-lysine coated slides,12and these slides were
soaked in methanol containing 0.3%H202solution
for 20 minutes to quench endogenous peroxidase.
The sections were treated with 1.5% normal goat
serum diluted with phosphate buffered saline con-
taining 0.1% bovine serum albumin in a humid
chamber for 20 minutes to block nonspecific bind-
ing. The sections were incubated overnight at 4" C
TABLE 1. Preoperative Radiographic
Findings on AP Radiograph in Stand
Position
Medial Joint Space Number of Knees
~
Narrowing 7
Obliteration 5
Tibia1 defect 6
Mean FTK 185.7 2 4.5
____ ~
FTA (fernorotibial angle): a lateral angle at the intersection of
the anatomic femoral axis and the anatomic tibial axis.
AP = anteroposterior.
 Clinical Orthopaedics
174 Saito and Koshino and Related Research

TABLE 2. Preoperative Clinical TABLE 3. Ten-Point Scoring System

Evaluation of Evaluation for Synovitis of
Osteoarthritic Knee
Clinical Evaluation HSS Score* (range)
Histologic Findings Score
Pain 17 2 ? 3 6 (10-20)
Function *
12 3 3 7 (8-22) Synovial lining cell hyperplasia
Range of motion 13 9 t 2 2 (10-18) 1-2 layers of cells 0
Muscle strength 7 8 i 1 6 (2-10) More than 3 layers 1
Flexion deformity 3 4 2 3 0 (0-10) Vascular changes
Instability 6 5 + 2 4 (0-10) None 0
Edema and vascular congestion 1
Total score 61 1 2 8 8
~
Vascular proliferation 2
*Kneescore by the Hospital for Special Surgery Cellular infiltration
None 0
Scattered cells 1
Moderate; cell cluster 2
Perivascular cell aggregation 3
Marked cellular infiltration with 4
in a humid chamber in the primary antibodies, lymphoid follicle
which were diluted to suitable concentrations with Villous hyperplasia
phosphate buffered saline containing 0.1% bovine Smooth surface 0
serum albumin and 0.25% Triton-X. After the sec- Irregular 1
tions were incubated, they were treated with bi- Partially fingerlike hyperplasia 2
otinylated goat antirabbit immunoglobulin G di- Diffuse fingerlike hyperplasia 3
luted in phosphate buffered saline containing 0.1% Total score 10
bovine serum albumin for 30 minutes at room tem-
perature, and incubated with avidin-biotin-peroxi-
dase complex for 30 minutes at room temperature
(Elite ABC kit, Vector Laboratories, Inc, Burlingame, RESULTS
CA). The sections were incubated for 3 minutes in
0.05 mol Tris hydrochloride buffered solution (pH Gross Synovial Innervation
7.6) containing 0.05% 3,3-diaminobenzidine and
0.01% H,O,. After that, the slides were counter- Modified gold chloride staining showed an ex-
stained in hematoxylin or methyl green. The inci- tensive neural network in the synovial tissue.
dence of neuropeptides appearing was calculated as In the stained section neural elements,collagen
the number of positive slides per each total sample bundles, and the fatty tissue were identified. In
x 100. the deep layer adjacent to the fatty tissue there
was a large nerve fiber bundle connected to
Histologic Evaluation medium size nerve fiber bundles having a net-
Evaluation of the extent of synovial inflammation like appearance (Fig 1). In the subsynovial tis-
was made in the medial compartment, the lateral sue, nerve fibers branched at almost right an-
compartment, and the suprapatellar pouch using a gles and proceeded laterally (Fig 2). Some of
10-point scoring system consisting of four parts. the nerve fibers in the subsynovial tissue were
One point was allotted for synovial hyperplasia,
directed toward the intimal layer, and the oth-
two for vascular changes, four for cellular infiltra-
tion, and three for villous hyperplasia (Table 3). ers were located parallel to the surface, form-
ing a network of nerve fibers just under the
Statistical Analysis synovial cell lining (Fig 3). Terminals of pe-
For statistical analysis, Student’s two-tailed t test ripheral nerve fibers also were found in the
and a chi square test were used to verify the differ- subsynovial tissue. An abundant nerve supply
ence in the incidence of synovial changes for each was found around the wall of blood vessels;
neuropeptide between two sites. Probability values these small nerve fibers wound around the wall
less than 0.05 were considered significant. in a spiral manner. The authors failed to find
Number 376
Julv. 2000 Svnovial Innervationin Osteoarthritis 175

Fig 1. Photomicrograph show-

ing a network of medium diame-
ter nerve fiber bundles branch-
ing out from large nerve fiber
bundles (Stain, modified gold
chloride; original magnification,
x 20).

mechanoreceptors such as Ruffini end organs
or Pacinian corpuscles in the synovium in this
study. Synovial innervation of the knee could
be divided into three types: large and medium
nerve fiber bundles, free nerve endings, and
fibers and perivascular neural networks. The
distribution pattern of the neuropeptides was
investigated by immunohistochemical staining
according to this classification.
Location of Neuropeptides
in the Synovium of the Knee
In the immunohistochemical study, the syno-
vium was composed of the synovial cell lin-
ing, the subsynovial tissue, and a layer formed
of dense bands of collagen fibers. There was a
large neurovascular network between the sub-
lining tissue and the layer of collagen fiber
bundles. The axons, which were immunoreac-
tive for antisera to neuropeptides, appeared as
brown strands in longitudinal sections of a
nerve fiber bundle and as brown spots in trans-
verse sections. Fine immunoreactive nerve
fibers resembled brown beaded strands, indi-
cating axoplasmic transport of neuropeptides.
Nerve fiber bundles with more than five
strands were defined as large, those with two
to four strands as medium, single strands as

Fig 2. Photomicrograph show-

ing small diameter nerve fibers
branching at a right angle from a
medium nerve fiber bundle and
proceeding in a lateral direction
(Stain, modified gold chloride;
original magnification, x 100).
 Clinical Orthopaedics
176 Saito and Koshino and Related Research

Fig 3. Photomicrograph show-

ing small diameter nerve fibers
running toward the synovial lin-
ing in the subsynovial tissue
(Stain, modified gold chloride;
original magnification, x 100).

free nerve fibers, and single strands with a spi-

der shape as free nerve endings. The incidence
of nerve fibers immunoreactive for each anti-
serum at these sites was recorded and com-
pared with that of the medial, lateral, and
suprapatellar synovium.
Substance P immunoreactive nerve fibers
were distributed widely, and substance P was
found in free nerve endings and fibers (Fig 4),
nerve fiber bundles accompanied with or with-
out blood vessels, and in the perivascular
neural network in all three parts of the joint.
The free nerve fibers and endings showed sub-
stance P immunoreactivity in I5 of 17 (88%)
medial samples, in seven of 15 (47%) lateral
samples (p < 0.05), and three of 15 (20%)
suprapatellar samples (p < 0.005). The inci-
dence was most predominant on the medial
side. Substance P positive free nerve endings
in the medial synovium showed marked ax-
onal branching, and some free nerve endings
containing substance P were surrounded by
several monocytes forming a cluster (Fig 5).
The distribution of the nerves containing
calcitonin gene related peptide showed almost
the same tendency as those of substance P. Fig 4. Photomicrograph showing substance P
The calcitonin gene related peptide im- immunoreactive free nerve fibers in the subsyn-
ovial tissue (arrows). The fine nerve fibers are
munoreactivity of large or medium nerve fiber
made up of a single strand of substance P posi-
was found in lo Of l7 (59%) tive axon that runs parallel to the synovial lining
of synovium from the medial compartment, in (Stain, methyl green counterstain; original mag-
10 of 15 (67%) samples of synovium from the nification, XIOO).
Number 376
Julv. 2000 Synovial Innervation in Osteoarthritis 177

Fig 5. Photomicrograph show-

ing substance P immunoreactive
free nerve endings, which are
forming a spiderlike appearance
and are surrounded by several
monocytes (arrow) (Stain, hema-
toxylin counterstain; original mag-
nification, x 200).

lateral compartment, and in five of 15 (33%)
samples of synovium from the suprapatellar
pouch. The main location of calcitonin gene
related peptide was in large or medium nerve
fiber bundles. The incidence of calcitonin
gene related peptide immunoreactive free
nerve fibers or endings or both was most
prevalent in synovium from the medial com-
partment (47%), followed by the lateral com-
partment (20%), and the suprapatellar pouch
(0%) (p < 0.01) (Table 4).
The localization of neuropeptide Y and va-
soactive intestinal peptide revealed a different
manner from those of substance P and calci-
tonin gene related peptide. The immunoreac-
tivity for neuropeptide Y showed the most
characteristic modality of the distribution of
neuropeptides among the four neuropeptides.
The positive nerve fibers were located pre-
dominantly in the perivascular area (Fig 6),
and the immunoreactivity could not be de-
tected in free nerve fibers or endings. Neu-
ropeptide Y immunoreactive perivascular
fibers were found in 12 of 17 (71%) medial
samples, six of 15 (40%) lateral samples, and
six of 15 (40%) suprapatellar samples. The in-
cidence of the neuropeptide Y positive nerves
was the highest in synovium from the medial
compartment. The large nerve fiber bundles
containing strands of neuropeptide Y were ac-
companied with blood vessels located be-
tween the thick collagen bundle layer and the
sublining tissue.
Vasoactive intestinal peptide antiserum

TABLE 4. Frequency and Location of Substance P and Calcitonic Gene Related

Peptide in Synovium
Substance P Calcitonin Gene Related Peptide
Neuropeptide
Location M (n = 17) L (n = 15) S (n = 15) M (n = 17) L (n = 15) S ( n = 15)

Free nerve fibers and 15* 7 3 8 3 0

endings
Medium or large nerve 9 8 6 10 10 5
fiber bundles
Perivascular neural network 8 6 5 5 5 3
~

M = medial synovium; L = lateral synovium; S = suprapatellar pouch.

*Number of samples in which antiserum immunoreactive nerve fibers were found.
 Clinical Orthopaedics
178 Saito and Koshino and Related Research

Fig 6. Photomicrograph show-

ing neuropeptide Y containing
nerve fibers accompanied with
a blood vessel. The fibers are
tortuous and form a helical
structure (arrows) (Stain, methyl
green counterstain; original mag-
nification, x200).

showed the decreased intensity of staining and
the lowest incidence among the four neu-
ropeptides in this study. The nerves that were
immunoreactive for vasoactive intestinal pep-
tide were found in large nerve fiber bundles
between a tight and a loose collagen fiber bun-
dle layer in deep areas of the specimens and
around the wall of the blood vessels. There
seemed to be no difference in the incidence be-
tween the medial and the lateral parts of the
joint. In the samples of the suprapatellar
pouch, it was hard to find vasoactive intestinal
peptide positive nerve fibers (Table 5 ) (Fig 7).
Synovial Inflammation
The average total score of synovitis was 5.5 ?
2.3 points in the medial compartment, 3.3 5
2.5 in the lateral compartment, and 4.9 ? 2.4
in the suprapatellar pouch. There was a sig-
nificant difference in the severity between
synovium from the medial and lateral com-
partment (p < 0.05) with the Student’s t test.
In the medial synovium, newly developed vas-
cularity and cell infiltration in the subsynovial
tissue were observed more often (Table 6).
DISCUSSION
Synovial Innervation
Early literature about the innervation of the
knee suggested that nerve fibers were widely
distributed in the synovium and were accom-
panied mainly by a vascular tree.13 Recent re-
ports discuss the existence of a plexus of non-

TABLE 5. Frequency and Location of Neuropeptide Y and Vasoactive Intestinal

Peptide in Synovium
Neuropeptide Y Vasoactive Intestinal Peptide
Neuropeptide
Location M (n = 17) L (n = 15) S (n = 15) M (n = 17) L (n = 15) S ( n = 15)
Free nerve fibers and O* 0 0 1 1 1
endings
Medium or large nerve 6 2 2 4 3 0
fiber bundles
Perivascularneural network 12 6 6 3 1 0

M = medial synovium; L = lateral synovium; S = suprapatellar pouch.

‘Number of samples in which antiserum immunoreactive nerve fibers were found
Number 376
July, 2000 Synovial Innervation in Osteoarthritis 179

Synovial lining cells

F m
fibers:
en
Subsynovial
tissue / "".."
Free nerve endings: SP, CGRP

Layer of neural
collagen network:
bundles SP.CGRP.

L Vessel

Fig 7. Schema of synovial innervation and location of neuropeptides in the synovium of osteoarthritis
of the knee. The schema shows the distribution of neural elements and the location of neuropeptides
in synovial tissue of medial compartmental osteoarthritis of the knee. Substance P (SP)and calcitonin
gene related peptide (CGRP) are distributed widely in most nerve fibers and especially in main loca-
tions that are peripheralnerve terminals in the subsynovialtissue. The location of neuropeptideY (NPY)
is closely related to vascular trees. Vasoactive intestinal peptide (VIP) exists in the neural network
around a large vessel.

medullated nerves in addition to the perivas-
cular ones.14,27,30Recent advances in immuno-
histochemical staining and neurohistologic
analysis have shown numerous free nerve
fibers and endings in the synovium.15 How-
ever, to clarify the exact modality of nerve dis-
tribution in connection with the anatomy of the
synovium, it seems to be crucial to use a stain-
ing method with a high specificityfor neural el-
ements and to use thicker sections than usual.
The modified gold chloride method is said to
show free nerve endings and several mech-
anoreceptors well.I4 However, when this stain
is used on thick sections and the specimens are
treated with strong acid, identifying the
anatomic location of neural elements is labori-
ous. In immunohistochemical staining, the
avidin-biotin-peroxidase complex method is a
precise technique." Thus, this study on the
synovial innervation of the human knee was
performed using both of these methods.
The results of this study showed there may
be two different neural systems in the syn-
ovium of a human knee, one somatic and the
other autonomic. In the somatic neural system
there seem to be nerve fiber bundles of a large
diameter accompanied by a large blood vessel.
These are connected by medium size nerves
and form a neural network between the sub-
synovial tissue and a layer of dense collagen

TABLE 6. Pathologic Abnormalities in Medial, Lateral, and Suprapatellar Synovium

of the Osteoarthritic Knee Evaluated with a 10-point Scoring System
~ ______~ _ _ _ ~

Synovial Changes Medial (n = 17) Lateral (n = 15) Suprapatellar (n = 15)

Lining cell hyperplasia 0.2 i 0.4 0.1 2 0.3 0.3 t- 0.5
Vascular changes 1.6 i 0.6 12 0.8 1.1 -t 0.6
Cellular infiltration 1.8 i 1.2 1% 0.9 1.5 i 1.2
Villous hyperpasia 1.8 t 0.9 1.1 2 1.1 1.9 t 1.2
Total score 5.4 ? 2.2 3.3 i 2.5* 4.9 2 2.4

p < 0.05, there is a significant difference in the total score between synovium from the medial and the lateral compartment.
 Clinical Orthopaedics
180 Saito and Koshino and Related Research

fibers. In the subsynovial tissue, nerve fibers of study these four kinds of neuropeptides, while
small diameter run in longitudinal and trans- investigating the modality of nerve supply to the
verse directions immediately underneath the synovium of the human knee and the mecha-
intimal layer. These nerve fibers also form a nism that transmits pain in inflammation of the
neural network. The distribution of autonomic osteoarthritic knee.
nerves may be strongly associated with that of In comparing the synovium from three dif-
blood vessels, and they are found around large ferent compartments of knees affected by me-
and medium size blood vessels, forming a web dial compartmental osteoarthritis, the current
of fine nerve fibers. Small autonomic nerve authors' most important observation was that
fibers located along small vessels extend to- neuropeptides were more abundant in the
ward the synovial lining. Thus, the authors synovium from the medial compartment than
think the synovium of the knee may be noci- in the other two compartments. Free nerve
ceptive and insensitive to pressure or tension. fibers and endings containing substance P and
calcitonin gene related peptide were found
Location and Role of Neuropeptides most frequently in the medial compartment.
in the Synovium Free nerve fibers have been recognized as pain
With the advance of neurophysiology, antisera receptors of Type 4 using the classification of
to numerous neural elements have been pro- Freeman and Wyke.8 The abundance of sub-
duced, and today's techniques have made it stance P and calcitonin gene related peptide in
possible to distinguish the qualities of each free nerve fibers suggests the medial sy-
nerve fiber. Among the neuropeptides purified novium may be sensitive to noxious mechani-
from the gastrointestinal tract, the hypothala- cal stimuli, corresponding to the location of
mus, substance P, calcitonin gene related pep- spontaneous pain and tender pain that patients
tide, vasoactive intestinal peptide, neuropeptide with medial compartmental osteoarthritis of
Y, galanin, enkephalins, and growth hormone the knee report most frequently.
releasing factor have been reported in periph- Neuropeptide Y immunoreactive nerve
eral nerve fibers.25 fibers were found in significantly high number,
Substance P in particular is suggested as one and the autonomic nervous system seemed to
of the most important neuropeptides in the mod- be activated in the synovium of the medial
ulation of the inflammatory process of arthritis. side. From the results of this study, the current
Animal studies showed that the intraarticular in- authors were unable to explain fully why va-
jection of substance P increased the severity of soactive intestinal peptide was detected in such
and infusion of substanceP antagonist
arth~itis,'~ minute amounts; however, it may be attribut-
reduced the exaggerated inflammati~n.'~ Sub- able to depletion or original scarcity.
stance P is not only a transmitter of pain signal, Because patients with medial compartmen-
but it also stimulates macrophages, neutrophils, tal osteoarthritis have a varus limb alignment
and endothelial cells to induce phagocytosis, with tightness of the medial soft tissues, show
chemotaxis, and extraplasmavasation?J8 This a lateral thrust (adduction movement) while
activity reportedly is augmented in the presence beginning the stance phase in walking, and
of calcitonin gene related peptide.2JoSubstance have protruded osteophytes around the medial
P and calcitonin gene related peptide were con- femoral and tibia1 condyles, a repeated com-
sidered markers of sensoy nerve fibers, and pressive force is exerted on the medial com-
neuropeptide Y was said to be found in most pe- partment of the knee. Peripheral mechanical
ripheral noradrenergic neurons.ls Vasoactive stimuli are said to increase the production of
intestinal peptide reportedly lowered blood substance P in the dorsal horn of the spinal
pressure by inducing vasodilation and moving cord. 16,24 Synovitis has been recognized in the
circulating cells to the region of inflammati~n.~ animal knee treated by anterior cruciate liga-
For these reasons the current authors chose to ment transaction in the experimental model of
Number 376
July, 2000
0steoarthritis,2~and the instability is consid-
ered a primary cause of neuropeptide accumu-
lation in the medial synovium.
Substance P immunoreactive free nerve
endings were found to encompass monocytes.
This indicates that monocytes do not react to
free substance P, which is released from free
nerve endings into the surrounding tissue, but
that these cells receive input from the nervous
systems terminal more directly and actively.
Thus, free nerve endings may function to trans-
mit information from the nervous system to
cells. This strongly supports the view that the
nervous system immunologically modulates
the inflammatory process in osteoarthritis.
Accordingly, in the process of inflammation
and destruction in medial compartmental os-
teoarthritis of the knee, the authors think the in-
stability of the knee may greatly activate nerve
terminals, and accumulated neuropeptides in
peripheral nerves may stimulate inflammatory
cells, such as mast cells and monocytes. Hista-
mine may be released from connective tissue
mast cells found at inflammatory sites having
substance P recept01-s.~~ This induces edema of
the synovium and fluid accumulation in the
knee. The monocytes stimulated by substance
P release interleukin-1, tumor necrosis factor,
and other cytokines,22and these soluble in-
flammatory mediators activate fibroblasts and
chondrocytes to secrete collagenase and pro-
tease, which may result in destruction of the
~artilage.~ However, the articular surface of the
lateral compartment of the knee always is
maintained, even under inflammatory condi-
tions. This suggests the superficial layer of the
articular cartilage may act as a defensive guard
against the attack of soluble inflammatory me-
diators and inflammatory cells, but that once
the superficial layer is damaged and the matrix
of the cartilage is exposed by mechanical
stress, the destruction of the cartilage may
progress rapidly to its final stage.
References
I . Ahlbdck S: Osteoarthritis of the knee. A radiographic
investigation. Acta Radio1 277(Suppl):1-72, 1968.
2. Ahmed M, Bjurholm A, Schultzberg M, et al: In-
creased levels of substance P and calcitonin gene-re-
Synovial Innervation in Osteoarthritis 181
lated peptide in rat adjuvant arthritis. Arthritis
Rheum 38:699-709, 1995.
3. Arend WP, Dayer JM: Cytokines and cytokine in-
hibitors or antagonists in rheumatoid arthritis. Ar-
thritis Rheum 33:305-315, 1990.
4. Bar-Shavit Z, Goldman R, Stabinsky Y, et al: En-
hancement of phagocytosis: A newly found activity
of substance P residing in its N-terminal tetrapep-
tides sequence. Biochem Biophys Res Commun
94: 1445-1 45 1, 1980.
5. Bole GG: Osteoarthritis. In Cohen A, Bennet JC
(eds). Rheumatology and Immunology. Orlando,
Grune and Stratton 345-3.53, 1986.
6. Christensen SB: Osteoarthritis. I. Pathogenesis of os-
teoarthrosis. Acta Orthop Scand 56(SuppI 214):2-9,
1985.
7. Danek A, O’Dorisio MS, O’Dorisio TM, et al: Spe-
cific binding sites for vasoactive intestinal polypep-
tide on nonadherent peripheral blood lymphocytes. J
Immunol 131:1173-1177, 1983.
8. Freeman MAR, Wyke B: The innervation of the knee
joint, an anatomical and histological study in cat. J
Anat 101:505-532, 1967.
9. Gronblad M, Konttinen YT, Korkara 0, et al: Neu-
ropeptides in synovium of patients with rheumatoid
arthritis and osteoarthritis. J Rheumatol 15:1807-
1810, 1988.
10. Gulbenkian S, Merighi A, Wharton J, et al: Ultra-
structural evidence for the coexistence of calcitonin
gene-related peptide and substance P in secretory
vesicles of peripheral nerves in the guinea pig. J Neu-
rocytol 15:535-542, 1986.
11. Hsu SM, Raine L, Fanger H: Use of avidin-biotin-
peroxidase complex (ABC) in immunoperoxidase
techniques: A comparison between ABC and unla-
beled antibody (PAP) procedure. J Histochem Cy-
tochem 29577-580, 1981.
12. Huang WM, Gibson SJ, Facer P, et al: Improved sec-
tion adhesion for immunocytochemistory using high
molecular weight polymers of L-lysine as a slide
coating. Histochemistry 77:275-279, 1983.
13. Kellgren JH, Samuel EP: The sensitivity and inner-
vation of the articular capsule. J Bone Joint Surg
32B:84-92, 1950.
14. Kennedy JC, Alexander IJ, Hayes KC: Nerve supply
of the human knee and its function importance. Am
J Sports Med 10:329-335, 1982.
15. Kidd BL, Mapp PI, Blake DR, et al: Neurogenic in-
fluences in arthritis. Ann Rheum Dis 49:649-652,
1990.
16. Kuraishi Y: Evidence that substance P and somato-
statin transmit information related to pain in the
spinal horn. Brain Res 325:294-298, 1985.
17. Lam FY, Ferrell WR: Inhibition of carrageenan in-
duced inflammation in the rat knee joint by substance
P antagonist. Ann Rheum Dis 48:928-932, 1989.
18 Lembeck F, Holzer P: Substance P as neurogenic
mediator of antidromic vasodilation and neurogenic
plasma extravasation. Arch Pharmacol 310: 175-
183, 1979.
19. Levine JD, Clark R, Devor M, et al: Intraneural sub-
stance P contributes to the severity of experimental
arthritis. Science 226:547-549, 1984.
20 Levine JD, Collier DH, Basbaum AI, et al: Hypoth-

182 Saito and Koshino
esis: The nervous system may contribute to the
pathophysiology of rheumatoid arthritis. J Rheuma-
to1 12:406-411, 1985.
21. Levine JD, Moskowitz MA, Basbaum AI: The con-
tribution of neurogenic inflammation in experimen-
tal arthritis. J Immunol 1352343-847, 1985.
22. Lotz M, Vanghan JH, Carson DA: Effect of neu-
ropeptides on production of inflammatory cytokines
by human monocytes. Science 241:1218-1220,1988.
23. Myers SL, Branat KD, O’Connor BL, et al: Synovi-
tis and osteoarthritic changes in canine articular car-
tilage after anterior cruciate ligament transaction. Ef-
fect of surgical hemostasis. Arthritis Rheum
33: 1406-1 4 15, 1990.
24. Oku R: Calcitonin gene-related peptide promotes
mechanical nociception by potentiating release of
substance P from the spinal dorsal horn in rats. Brain
Res 403:350-354, 1987.
25. Pol& JM, Bloom SR: Immunocytochemistry. In Po-
l& JM, VanNoorden S (eds). Immunocytochem-
istry. Practical Applications in Pathology and Biol-
ogy. Bristol, Wright PSG 184-211, 1983.
26. Rabinowicz T, Jacqueline F Pathology of the cap-
sular and synovial hip nerves in chronic hip diseases.
Pathol Res Pract 186:283-292, 1990.
Clinical Orthopaedics
and Related Research
27. Ralston I1 HJ, Miller MR, Kasahara M: Nerve endings
in human fascia, tendons, ligaments. periosteum and
synovial membrane. Anat Rec 36: 137-147, 1960.
28. Ranawat CS, Shine JJ: Duo-Condylar total knee
arthroplasty. Clin Orthop 94:185-195,1973.
29. Revel1 PA, Mayston V, Lalor P, Mapp P: The syno-
vial membrane in osteoarthritis, a histological study
including the characterisation of the cellular infiltrate
present in inflammatory osteoarthritis using mono-
clonal antibodies. Ann Rheum Dis 47:300-307,
1988.
30. Samuel EP: The autonomic and somatic innerva-
tion of the articular capsule. Anat Rec 113934-92,
1952.
3 1. Wassennan SI: Mast Cells and Eosinophils. In Kel-
ley WN, Harris ED, Ruddy S, Sledge CB (eds). Text-
book of Rheumatology. Philadelphia, WB Saunders
367-375,1989.
32. Wojtys EM, Beaman DN, Glover RA, Janda D: In-
nervation of the human knee joint by substance-P
fibers. Arthroscopy 6:254-263, 1990.
33. Zimny ML, Ouge MS, Shutte M: A modified gold
chloride method of the demonstration of nerve end-
ings in frozen sections. Stain Techno1 60:305-306,
1985.