вход по аккаунту


Peripheral and spinal neural mechanisms in arthritis with particular reference to treatment of inflammation and pain.

код для вставкиСкачать
Number 7, July 1994, pp 965-982
8 1994, American College of Rheumatology
Arthritis & Rheumatism
Official Journal of the American College of Rheumatology
In this review, we assess the role of different
types of nerve fibers and neurotransmitters/modulators in arthritic joints, dorsal root ganglia, and the
dorsal horn of the spinal cord, with special reference
to mechanisms relevant to inflammation and pain and
their treatment. Stimulated polymodal type IV/C primary afferent nociceptive (PAN) fibers release substance P (SP) and calcitonin generelated peptide
(CGRP), which synergistically exert proinflammatory
paracrine effects in synovium. Postganglionic sympathetic efferents (PGSN), activated by local stimulation
or via a spinal reflex, release noradrenaline (NA), as
Supported by grants from the Finnish Academy, the E d
Aaltonen Foundation, the Perklen Foundation, the Y j o Jahnsson
Foundation, the Signe and Ane Gyllenberg Foundation, Finland, the
Arthritis and Rheumatism Council, and the Grand Charity of
Freemasons, UK.
Yrjo T. Konttinen, MD, PhD: University of Helsinki,
Helsinki, Finland, and the University of Alberta, Edmonton, Alberta, Canada; Pertti Kemppinen, MSc: University of Helsinki;
Margaretha Segerberg, MD, PhD: University of Helsinki; Mika
Hukkanen, BSc: Royal Postgraduate Medical School, University of
London, London, UK; Richard Rees, MB, BCh, MRCP Charing
Cross Hospital, London, UK; Seppo Santavirta, MD, PhD: Orthopaedic Hospital of the Invalid Foundation, Helsinki, Finland; Tim0
Sorsa, DDS: University of Helsinki; Antti Pertovaara, MD, PhD:
University of Heidelberg, Heidelberg, Germany, and the University
of Helsinki; Julia M. Polak, MD, DSc, FRCPath: Royal Postgraduate Medical School, University of London.
Address reprint requests to Yrjo T. Konttinen, MD, PhD,
Department of Anatomy, P.O. Box 9 (Siltavuorenpenger 20 A),
FIN-00014 University of Helsinki, Helsinki, Finland.
Submitted for publication August 20, 1991; accepted in
revised form February 15, 1994.
well as ATP, adenosine, prostaglandin I, (PGI,),
PGE,, interleukin-1 (IL-1), and neuropeptide Y.
Synthesis, fast axonal transport, and release of
SP and CGRP are enhanced in arthritis, probably due
to stimulation by nerve growth factor (NGF). Interestingly, intramuscular gold and capsaicin, agents selectively neurotoxic to type IV/C fibers, lower SP levels
and have been found clinically useful. Both type III/AG
and IV/C PAN fibers are stimulated by bradykinin
(BK), serotonin, histamine, K+, H+, and movement,
and are sensitized by PGE,, PGE,, PGF,,, PGI,,
8R,lSS-diHETE, BK, H+, and adenosine, Accordingly, nonsteroidal antiinflammatory drug (NSAID)type PGH, synthetase (cyclooxygenase) inhibitors
(diminishing production of prostanoids) and intraarticular corticosteroids (also diminishing production of
8R,ISdiHETE and of tumor necrosis factor a [TNFa],
which drives the IL-l/IL-6/PG and IL-8/sympatheticmediated pain-producing cascades) are effective in the
treatment of inflammation and pain in arthritis.
Stimulation of PANS in arthritis leads to release
of glutamate (Glu) and SP from the central spinal
terminals. Gluhonotropic non-NMDA-type receptor
interaction causes fast excitatory postsynaptic potentials (EPSP) and neurotransmission. This nociceptive
input can be inhibited at the pre- and postsynaptic
levels by stimulation of proprioceptive and tactile type
I/Aa and IYAp fibers, which probably mostly act via
GABAergic-inhibitory intrinsic interneurons. Segmental spinal control is utilized with moderately good
results in physiotherapy, e.g., high-frequency, low-
Table 1. Some characteristicsof 3 common human neuroDeDtides
No. of amino acids
Molecular weight (daltons)
Primary structure
Gene location on chromosome
Exonshessenger RNA species
Substance PIPPT-A gene
* NPY = neuropeptide Y; CFQN = C-flanking peptide of neuropeptide Y.
t Neuropeptide K and substance K, also products of the preprotachykinin A (PFT-A) gene, are not shown. Neuromedin K is a product of a
different gene.
$ In human calcitoningene-related peptide I1 (CGRP-11), coded by the CALC-I1gene located at 1 IplZpter, Asp’ is substituted with Am3, Valz2
with Metzz,and Amz5 with Se?’.
intensity transcutaneous neural stimulation, which
stimulates type IYAp fibers. GldNMDA receptor and
SPheuromedin, (NK,) receptor (or neurokinin A
[NKA]/NK, receptor) interactions lead to the formation of delayed, slow, and prolonged EPSPs, wind-up
and hyperexcitability of the nociceptive-specific (NS)
and wide dynamic-range (WDR) projection neurons.
Stimulation of PANs in arthritis activates cascades of
transcriptionalfactors, such as immediate-early genes.
This leads to an increased synthesis of dynorphin and,
in chronic arthritis, of enkephalin as well; these are
components of the prodynorphidproenkephalidproopiomelanocortin opioid peptide system, which act on
K , 8, and p receptors. Neuropeptide-degrading enzymes such as neutral endopeptidase are at least
temporarily decreased in arthritis and, via synaptic
and volume transmission, slow activity-induced
changes develop and contribute to centralizationof pain.
In coptrast to fibromyalgidfibrositis, in arthritis bulbospinal serotonergic and noradrenergic endogenous analgesia pathways are not defective but are activated.
Peripheral sensitization, neurogenic i d m a t i o n , central processing, hyperexcitability, and neuroplasticity
are probably important pathophysiologic mechanisms
in arthritis.
Paralyzed joints can be spared from arthritis in
patients who have had a cerebrovascular accident,
poliomyelitis, or peripheral nerve lesion. Patients with
reflex sympathetic dystrophy may have synovitis, and
surgical sympathectomy and sympatholytic drugs
have been reported to be beneficial in severe rheumatoid arthritis (RA). Neuropathic Charcot joints may
result from the absence of neurotrophic influences or
from increased strain in the absence of adequate,
protective afferent input. Other important observations relate to the joint predilection and symmetry that
are typical of RA, features that have been connected
with the density of innervation and with cross spinal
reflexes. In palindromic rheumatism, resolution of
symptdms in one joint is often followed by disease in
the contralateral joint (for recent reviews, see refs.
1-3). These clinical observations and new advances in
the understanding of the function of different types of
nerve fibers and neuropeptides/transmitters make it
pertinent to review their role in arthritic pain to
advance understanding of the possible mode of action
of currently available and experimental treatments.
Gene, message, and protein structure
Since von Euler and Gaddum (Karolinska Institutet, Stockholm, Sweden) described substance P (4),
the number of structurally well-characterized neuropeptides has increased to well above 100. SP and
CGRP, which often co-localize in small-diameter, unmyelinated, slowly conducting, nociceptive polymodal
type IV/C PAN fibers, and neuropeptide Y (NPY),
which often co-localizes with cytoplasmic tyrosine hydroxylase and intravesicular dopamine-phydroxylase
in PGSN fibers, are of particular interest (5) (Table 1).
This is because type IV/k and type IIUAS PANs and
PGSNs (Figure 1) seem to play a major role in arthritis
and chronic pain, although large-diameter afferent and
efferent fibers (Table 2) also play a role, for example,
in the modulation of nociceptive input in the dorsal
horn of the spinal gray matter (6,7) and in mediating
reflex responses that are responsible for changes in
muscle tone (8). Differences in the functional effects,
depending on the amount of peptide released, and
significant interactions between neuropeptides and
neurotransmitters make simple straightforward interpretations-from the existence of a certain peptide (or
usually, a combination of peptides) in a neuron to its
function-difficult. There may not be any clear corre-
dorsal root ganglion
Figure 1. A, Primary afferent nociceptive (PAN) type IV C fibers pass through the dorsal root to the spinal cord and terminate mainly in lamina
I1 of the spinal gray matter. Lamina I1 of the dorsal horn of the spinal cord is also called substantia gelatinosa because it is myelin free and
therefore histologically “clear.” The neuronal cell bodies of the small-diameter, unmyelinated C-type afferent fibers are located in the dorsal
root ganglion, where substance P and calcitonin gene-related peptide are synthesized. Type I11 AS PANS are small myelinated fibers, mainly
originating from medium-sized sensory neurons and synapse in lamina I (the marginal layer), as well as in deeper laminae (see text for details).
The gray matter of the spinal dorsal horn can be divided into different layers, so-called rexed laminae, based on their histology, which also
reflects different functions of the different layers. B, A postganglionic sympathetic efferent fiber passing through the gray ramus (gray because
it contains mainly unmyelinated fibers) into the spinal nerve. The neuronal cell body, where neuropeptide Y is synthesized, is located in the
paravertebral sympathetic ganglion, which is part of the sympathetic trunk. A preganglionic sympathetic efferent fiber originating from
the lateral horn of the spinal cord (Tl-L2-3) is shown passing to the sympathetic ganglion at the level of entrance, but it could also pass through
the paravertebral ganglion to the synapse in one of the prevertebral or paravisceral ganglia (broken line).
Table 2.
Classification and characterization of peripheral nerve fibers
Myelin diameter
Spike refractory
Pain, temperature, touch
Proprioception: muscle
spindle, annulospiral
Proprioception: Golgi
tendon organs
Proprioception: muscle
spindle, flower spray
endings; Ruffini
endings; Pacinian
Touch, pressure
Preganglionic autonomic
Postganglionic autonomic
c IV
Pain, itch, temperature
Aa la
Aa Ib
A6 I11
To pressure
To local
To hypoxia
Most sensitive Least
Most sensitive Least
Most sensitive Least
Most sensitive Least
Most sensitive Least
Most sensitive Least
Most sensitive Least
Intermediately Intermediately
Most sensitive
Most sensitive
Most sensitive
Figure 3. Ubiquitin C-terminal hydrolase (PGP 9.5jcontaining
nerves in human tissues. There is rich innervation of the insertions
of ligament and tendon to bone, indicating a potential role in the
pathogenesis of enthesopathies. Neuropeptides may also be involved in the function of specialized proprioceptor end-organs
(original magnification x 280).
lations between the neuropeptide content and the
sensory modality served by that neuron (9).
Neuropeptide nerves in joint tissues and morphologic
evidence of involvement in arthritis
Hormones are secreted by various endocrine
organs and are present at very low concentrations in
the circulating blood. Site-specificity is conferred by
specific, high-affinity receptors. In contrast, neuropeptides are delivered to the site of action in dense
core vesicles in an extravascular, intra-axonal space,
protected from dilution effects and from inactivation
by degradation. At their peripheral, paracrine site of
action, they are released close to their targets by
action potential-induced depolarization and calciumdependent exocytosis of dense core vesicles. After
release, the range of action is restricted both spatially
and time-dependently, due to dilution and degradation.
This means that without the presence of neuropeptidecontaining nerve fibers, there will be no neuropeptide
effects at that particular site. Thus, innervation of
certain tissues and structures by neuropeptidecontaining nerve fibers allows us to draw some con-
clusions about possible functional effects, based on the
signal transduction pathways that are utilized by that
particular neuropeptide and the character of the potential target cells.
Figures 2 and 3 illustrate some examples of
peptide-containinglpeptidergic innervation of organs
and tissues of potential relevance in rheumatology. In
particular, synovial tissue proper and juxtaarticular
bone have been recently shown to be innervated by
SP-, CGRP-, and NPY-containing nerves. Arthritis
alters this pattern of innervation, such that there are
no or only a few peptide-containingnerve fibers visible
in synovium or bone where extensive inflammatory
changes have developed. Neuropeptide depletion,
caused by local release of neuropeptides (lo), is likely
to contribute to this end. Studies using pan-neuronal
markers, such as antisera to protein gene product 9.5
and neurofilament triplet proteins (not affected by
local neuropeptide exocytosis), suggest that nerves are
also to some extent destroyed or “retract” from the
inflamed areas (1 l), although sprouting and collateral
formation can be seen in other areas as an attempt at
regeneration (12). However, it is unlikely that the lack
of peptide-containing or peptidergic nerves would reflect their inability to keep up with inflammatory tissue
proliferation and neovascularization, because the inherent growth potential of SP-, CGRP-, and NPYcontaining nerves is good, and the inflammatory process itself may contribute to neural regeneration by the
production of a range of neuroactive growth factors
and cytokines.
Neuropeptides, paracrine effects, and neurogenic
Excessive local release of SP and CGRP may
lead to so-called “neurogenic inflammation,” whereas
a lack of neuropeptides in other areas could lead to a
depletion of trophic and regulatory effects normally
exerted by neuropeptides (Table 3). The earliest observations of neurogenic inflammation relate to axon
reflexes. Stimulation of type IV/C PANS elicits action
potentials that, once produced, are transmitted from
the periphery to the spinal cord, i.e., in an orthodromic direction. The orthodromically transmitted signals reach branching points of the axon collaterals,
where they are also diverted and are conducted in an
antidromic direction along the axon collaterals to the
nerve terminals in peripheral tissues. There they lead
to co-release of SP and CGRP and a consequent wheal
Table 3.
Examples of paracrine neuropeptide effects*
Substance P
Release of NO,
Histamine release
Inhibition of bone
Mitogenic to
smooth muscle
cells, fibroblasts,
endothelial cells,
and synoviocytes
in vivo
Modulation of
IL-1, IL-6,and
TNFa release
from monocytes
LTB,, PGE,, and
TXB, release
from macrophages
Muscle trophic
Modulation of
Modulation of
release from
sensory and
PGS neurons
Inhibition of
Trophic effects
on nonneuronal cells
Up-regulation of
of endothelial
cells for
Mitogenic to
Stimulation of
Enhancement of
Enhancement of B
lymphocyte IgA
and IgM
Enhancement of
adherence to
endothelial cells
PGE, and MMP-1
release from
fibroblasts and
Mitogenic to
endothelial cells
accumulation and
IL-1 and SP
Chemotactic to T
Enhancement of
adherence to
endothelial cells
Inhibition of natural
killer cell activity
* CGRP = calcitonin gene-related peptide; NPY = neuropeptide Y;
NO = nitric oxide; PGS = postganglionic sympathetic; IL =
interleukin; TNFa = tumor necrosis factor a; LTB, = leukotriene
B4;PGE, = prostaglandin E,; TXB, = thromboxane B2;SP =
substance P; MMP-1 = matrix metalloproteinase 1.
and flare, mediated in part by mast cells activated by
SP acting together with CGRP.
In articular tissue, PGSNs may to a large extent
mediate these effects initiated by stimulation of type
IV/C PANs (13), although direct effects of sympathetic
mediators are also possible. Antidromic stimulation
and release of neuropeptides can be induced by electrical stimulation of the nerve at some point along
its path, or, perhaps more importantly, direct local
stimulation of the sensitized nerve terminals can also
cause release of neurotransmitters/modulators. Subsequently, many other proinflammatory paracrine effects, in addition to those relevant to wheal and flare,
have been attributed to SP and CGRP (Table 3). These
effects emphasize the efferent role of the “sensory
fibers. ”
PGSNs have also been implicated in neurogenic
inflammation. In contrast to PANs, PGSNs are spontaneously active and respond to electrical or noxious
type IIVAG and type IV/C nerve fiber stimulation.
Strong sympathetic stimulation causes a release, not
only of NA, but also of NPY. Some paracrine NPY
effects (Table 3) would be compatible with an involvement of NPY in neurogenic inflammation.
Neuropeptides in arthritis
Although early clinical observations of the
modulating effect of lesions in the central or peripheral
nervous system on arthritis implicated neural involvement (see above), it was not clear what component
was involved. Capsaicin (truns-8-methyl-N-vanillyl-6nonenamide) is the pungent ingredient of hot paprika
or chili peppers. It initially produces pain by a receptormediated activation of polymodal PAN fibers (by opening a nonselective cation channel, which can be
blocked by ruthenium red), which is followed by
desensitization and/or degeneration of the polymodal
PANs. Capsaicin treatment has been reported to alleviate experimental inflammatory arthritis (14)and RA
(15). These observations specifically implicate the
polymodal PANs in the pathogenesis of arthritis. Interestingly, gold sodium thiomalate, another neurotoxin, has been shown to produce a significant and
selective decrease in the numbers of small unmyelinated axons, which is associated with a decreased SP
content in the peripheral nerves and an elevated
nociceptive threshold (13).
Polymodal PANs may also contain CGRP,
NKA, NPK (the N-terminal-extended form of NKA),
vasoactive intestinal peptide (VIP), peptide histidineisoleucine (PHI), somatostatin (SOM), galanin (GAL),
bombesin, and excitatory amino acids (EAAs). The
peptide content and combinations thereof can vary
according to the target tissue innervated and depending on various other circumstances, including inflammation in arthritis (16,17) and neuropathic lesions.
Some of the many possible neuropeptide interactions
are antinociceptive and/or antiarthritic, rather than
proinflammatory. For example, SOM inhibits SP actions at the periphery and in the spinal cord and has
97 1
been used in a clinical trial in the treatment of psoriatic
arthritis, the preliminary results of which were favorable (18).
Aggravation of arthritis following intraarticular
infusion of SP (19) and alleviation of arthritis by its
antagonists (20) more specifically indicate the influence of SP. Immunoneutralization studies utilizing
immunization or systemic administration of specific
SP and/or CGRP antibodies have shown the alleviation
of both neurogenic inflammation and arthritis. There
are as yet no reports on the effect of a newly invented
potent nonpeptide antagonist of the SP NK, receptor
in arthritis.
NPY-containing PGSNs have been found in
synovial tissues, and the sympathetic influence has
been implicated in reflex neurogenic inflammation in
the limb contralateral to the affected limb. The active
involvement of PGSNs is suggested by release of NPY
into the synovial fluid in RA. Furthermore, chemical
sympathectomy by guanethidine (thought to cause
immune-mediated destruction of PGSNs), reserpine
(which blocks uptake of catecholamines from the
cytoplasm to the vesicle, thus causing NA depletion),
and selective p,-adrenergic receptor-blocking drugs
has been shown to alleviate adjuvant-induced arthritis
(13). In patients with RA, a 2-week randomized,
double-blind study of regional intravenous guanethidine demonstrated diminished pain and increased
pinch strength (21). This effect was interpreted to
result from the diminution of the reflex sympathetic
efferent discharge. Consistent with this, it has been
reported that adjuvant-induced arthritis is worse in
spontaneously hypertensive rats, and that adrenaline
and &-agonists exacerbate arthritis (13).
These observations, however, indicate NA
rather than NPY as an inflammatory modulator. The
inhibitory action of NPY on NA release from PGSNs
and on tachykinin release from type IVIC PANs via a
prejunctional action implicates a role for N/PY. Recent evidence suggests that NPY may suppress neurogenic inflammation via Y1 and Y2 receptors. At
present, there is no direct evidence specifically linking
NPY to arthritis. Furthermore, as with type IVIC
PANs, the situation is complicated by cotransmittersl
modulators produced and released by PGSNs, including PGE,, PGI,, ATP, adenosine, and IL-1. It seems
clear, however, that noxious stimulation of PANs and
local inflammatory mediators cause a reflectory discharge and activation of PGSNs, which by direct and
indirect mechanisms further stimulate PANs in a kind
of vicious circle, which can lead to sympathetically
mediated and maintained inflammation and pain.
PANs and PGSNs are not the only neural
elements contributing to artlhtis and pain. Dorsal
rhizotomy causes exacerbation of arthritis, although a
more selective depletion of type IV/C PANs leads to
alleviation. Because dorsal rhizotomy also affects the
large-diameter, myelinated type IYAp and type YAa
fibers, it can remove an inhibitory influence on preganglionic sympathetic fibers, and thus lead to an increase
in efferent PGSN activity. Although not the subject of
this review, the hypothalamus-pituitary-adrenal axis
and stress responses probably also play important
roles in arthritis (22).
Involvement of nociceptors in pain mechanisms in
Although polymodal type IVIC PANs, due to
their neuropeptide content, seem to be of particular
relevance for neurogenic inflammation, both type IVIC
and type IIYAS high-threshold PAN fibers are involved in nociception, i.e., in responses to potentially
harmful and usually painful stimuli. Most type IVIC
nerve fibers are classified as polymodal fibers, because
they often respond to 3 different stimulus modalities:
mechanical, thermal, and chemical. In contrast, most
of the high-threshold myelinated type IIYAS PANs
respond only to mechanical or to mechanical and
thermal stimuli. Recently published neurophysiologic
single-fiber monitoring studies show that both type
IVIC and type IIYAS PANs are chemosensitive, although this chemosensitivity may be restricted to
relatively slow-conducting fibers in the C and AS fiber
range. The substances that have been considered to be
able to directly stimulate chemosensitive PANs are
BK, serotonin (5-hydroxytryptamine [S-HT]), and histamine, which act on PANs via specific B, (or B,),
5-HT2 (or 5-HT3), and histamine receptors, respectively, and potassium and protons, which in sufficient
concentrations can act nonspecifically to depolarize
the PANs (Table 4). These compounds probably often
act in concert, and interestingly, SP and inflammation
enhance the excitation caused by these compounds.
Some other compounds (PGE,, PGE,, PGF,,,
PGI,, 8R, 15s-diHETE, adenosine, 5-HT, H f ) (Table
4) are able to sensitize type IVIC and IIYAS PANs to
subsequent chemical, mechanical, or thermal stimulation. Although PGE2 and PGI, have been considered
to be sensitizing, rather than directly stimulating,
compounds, they seem to be able to do both, and they
Table 4. Substances able to directly stimulate and/or sensitize
primary afferent nociceptive fibers*
8R- i5S-diHETE
Probably many other
* PGE,
prostaglandin E,; LTB, = leukotriene B,; X I , =
prostacyclin; 8R,lSS-diHETE = the 8R,15S stereoisomer of dihydroxyeicosatetraenoic acid; TNFa = tumor necrosis factor a;
NGF-OP = the amino-terminal octapeptide of nerve growth factor.
t Has been suggested to act via the postganglionic sympathetic
nerve fibers. Bradykinin, however, has also been suggested to act
without involvement of the sympathetic nervous system.
also act on both type IV/C and type IW6 PANs.
However, it must be recognized that although prostanoids can activate nociceptors, this effect is markedly less pronounced than their ability to sensitize
nociceptors. PGE,, PGI,, 8R,15s-diHETE, adenosine,
and 5-HT act directly on their specific prostaglandin E,
receptor, prostacyclin I, receptor, 8R,lSS-diHETE, A,,
and 5-HT2/5-HT, receptors, respectively, and therefore with a short latency of onset, whereas many more
stimuli have an indirect mode of action and a long
latency of onset. These include BK (which thus has a
dual effect), NA, and leukotriene B4 (LTB,). It has
been suggested that BK and NA act on PGSNs to
induce the production of PGE, and PGI,, respectively,
which then sensitize PANs (13). More recent observations, however, suggest that nociceptor sensitization
by BK may not be dependent on sympathetic neurons.
Interpretation of the effects of BK in arthritis is
further confounded because in acute situations, BK
affects B, receptors on PANs and PGSNs and, accordingly, selective B, receptor antagonists are effective
analgesic and antiinflammatory agents. When inflammation is prolonged, B, receptors, which are not
expressed to a significant degree in healthy tissues, are
up-regulated, so that in arthritis, B, receptor antagonists reverse the hyperalgesia, whereas B, receptor
antagonists are relatively inactive. IL-1p/IL-6/
PG-mediated and IL-8/sympathetic-mediated sensitizing pathways have also been described, with TNFa
playing a pivotal role in both cascades (23). Protons
selectively induce long-lasting excitation and sensitization to mechanical stimulation of PANs.
In neurophysiologicterms, peripheral sensitization in arthritis implies increased spontaneous activity,
decreased activation threshold of the nociceptive fibers, increased responsiveness to suprathreshold stimuli, increased local release of neuropeptides upon
stimulation, and recruitment of the so-called “silent”
nociceptive fibers, which are usually unresponsive to
even maximal stimulation (24) (Table 5). All this
contributes to pain on motion and tenderness to pressure, which are typical symptoms in arthritis patients.
NSAIDs inhibit PGH, synthetase enzymes
(also known as cyclooxygenase) and therefore diminish the formation of PGE,, PGE,, PGF,,, and PGI,
from arachidonic acid. Corticosteroids prevent the action of phospholipase A,, which could otherwise cleave
arachidonic acid from the cell membrane phospholipids and provide substrate for both cyclooxygenase/
prostanoid and lipoxygenaseheukotriene (including
gR,lSS-diHETE as an intermediate) pathways (7). In
high doses, corticosteroids also inhibit IL-1 and TNFa
synthesis and release, and down-regulate PGH, synthetase (25). These may be the peripheral mechanisms
of action of NSAIDs and corticosteroids in the treatment of inflammatory pain.
Alterations in the neuropeptide synthesis and axonal
transport in arthritis
Although neuropeptides are stored in and released from the peripheral nerve terminals, they are
synthesized in the perikarya of PANs or PGSNs by
conventional ribosomal protein synthesis (rather than
by synthetase enzymes) in the dorsal root or sympathetic ganglia, well away from their peripheral storage
and release site. The final products would often be too
small to be synthesized by ribosomes or to be
“tagged” for sorting and packaging for regulated secretory pathways. Therefore, many neuropeptides are
synthesized as prepropolyproteins, which yield several similar or different neuropeptides by proteolytic
processing by endo- and exopeptidases in the Golgi
complex and mature vesicles.
Arthritis and chronic pain are associated with
activation of preprotachykinin A and CGRP genes in
the dorsal root ganglion cells (17). This increase in
messenger RNA (mRNA) expression is associated
with an increase in the content of the corresponding
neuropeptides in the PANs (16). SOM increases only
slightly, whereas the synthesis of VIP and GAL does
not seem to increase in arthritis, although they are
up-regulated in response to peripheral nerve injury.
Interestingly, NPY (not normally expressed in sensory
neurons) is induced after peripheral axotomy, at least
in myelinated type IIIAp and III/AG axons terminating
in laminae 111-V of the dorsal horn of the spinal gray
matter. Under normal conditions, SP and CGRP are
likely to be neuromodulators, responsible in part for
the central hyperexcitability, and may even be neurotransmitters in the first afferent synapse, but VIP
seems to replace them after peripheral nerve injury.
GAL is able to antagonize the activity of SP, CGRP,
and VIP. In contrast to arthritis, peripheral axotomy
causes a decrease in SP and CGRP levels as a result of
decreased synthesis.
The regulatory stimulus responsible for the
increased reactive SP and CGRP synthesis in arthritis
may be nerve growth factor. Anti-NGF antiserum
prevents the increase, whereas local injection of NGF
enhances it (26). The reason for the differential regulation of SP, CGRP, and SOM compared with VIP,
GAL, and NPY in inflammation and in peripheral
nerve injury is not known, because the stimuli that
up-regulate (or lead to disinhibition of) VIP, GAL, and
NPY in peripheral axotomy are not known.
Mature, peptide-containing, dense core vesicles
are probably transported along microtubules by an
ATP-dependent process mediated by kinesin (in contrast to fast retrograde transport, which is mediated by
dynein-like compounds), from the perikarya of the
nerve cell toward the axon terminus and preterminal
varicosities, and centripetally into the dorsal horn of
the spinal cord. Many neurons in the dorsal root
ganglion may contain such low levels of some of the
“sensory” neuropeptides SP, CGRP, NKA, VIP,
PHI, SOM, GAL, and bombesin that they can be
reliably demonstrated only after colchicine treatment
or nerve ligation, both of which block axonal transport. The increased content of SP and CGRP in dorsal
root ganglia seems also to be associated with increased
axonal transport (26). In arthritis, the increased synthesis of SP and CGRP in the dorsal root ganglion
neurons and the increased axonal transport (5-40
cdday, compared with 0.2-5 mdday slow axonal
transport or axoplasmic flow of microtubules, neurofilaments, actin filaments, etc.) are unable to prevent
peripheral (11) and spinal (27) depletion of SP and
CGRP caused by local release (10). In contrast to
neuropeptides, the classic neurotransmitters acetylcholine and NA are synthesized by enzymes, which
are also found in axon terminals, and are thus better
able to keep pace with the required functional demand.
Nociceptive transmission in the first afferent synapse
in the spinal dorsal horn
Many peptides (for example, SP, CGRP, NKA,
VIP, PHI, SOM, GAL, bombesin) have been found in
type IV/C PANS, with the spinal terminals concentrated in the superficial dorsal horn. Type IV/C fibers
terminate in lamina I1 and in lamina I (marginal zone),
whereas type III/AG fibers project to laminae I, 11,, V,
and X, but there is also a good degree of variation
depending on the tissue of origin (cutaneous, muscular, articular) of the sensory neurons. Dorsal rhizotomy leads to gross depletion of the sensory peptides in
the dorsal horn of the spinal cord, which suggests that
these compounds are mainly located in the presynaptic
nerve terminals and that they could therefore be
involved in sensory transmission; segmental spinal
interneurons would not be affected by rhizotomy (or
spinal transection).
It is possible that some SP can be released from
the spinal interneurons, but inflammation/arthritis in
combination with mechanical stimulation of the diseased joint releases SP and CGRP, probably mainly
from the central terminals of the type IV/C PANS(28).
As in the peripheral tissue, arthritis leads to an initial
depletion of SP and CGRP in the primary sensory
fibers that terminate in the spinal gray matter (24,27),
probably because the replenishment occurs via transport of neuropeptides synthesized in the dorsal root
ganglia, and is therefore slow. Although there is a
corelease of glutamate (Glu) with SP and CGRP, Glu
staining is not compromised: Glu is taken up by the
glial cells, where it is converted into glutamine (Gln),
released, and taken up by PANS to be converted back
to Glu (27), which means that Glu replenishment is fast.
Exogenous administration of SP into the rat
spinal cord produces hyperalgesia, as is evidenced by
behavioral features. However, depending on the dose
and route of administration, antinociceptive effects
and direct or indirect (via interneuron) inhibition of
dorsal horn neurons can also be evoked by SP. Although patients with a reduced population of SPcontaining neurons (i.e., type IV/C fibers) naturally
have diminished pain sensitivity, it is interesting that
SP levels in cerebrospinalfluid of patients with chronic
pain syndromes are low rather than high (29). There is
good correlation between the localization of SPcontaining nerve terminals and the NK, receptor sites,
but amphiphilic SP is also able to exert receptorindependent actions. Therefore, the ovepll function of
spinally released SP is still not clear, although it seems
to contribute to the development of central hyperexcitability in arthritis (see below).
CGRP potentiates SP release at the spinal cord
level and delays its degradation, therefore also potentiating the action of SP. Noxious stimuli in arthritis
also lead to an early release of the EAA Glu and
aspartate (Asp), followed by the inhibitory amino
acids (IAA) glycine and serine (30), after which the
release of SP and CGRP ensues.
Spinal type III/AG nerve fiber terminals mainly
contain small electron-translucent synaptic vesicles
that are thought to contain EAAs, whereas type IV/C
terminals also contain large dense-core vesicles containing neuropeptides. Glu, via its action on ionotropic
non-NMDA receptors, has emerged as a primary
neurotransmitter candidate possibly responsible for
relaying nociceptive information from both type IV/C
and III/A6 PANS in the form of fast excitatory
postsynaptic potentials (EPSP) to dorsal horn neurons. It is of interest that both SP and CGRP enhance
the release of EAAs and that SP induces delayed,
slow, and prolonged EPSPs, which contribute to the
development of central hyperexcitability (see below).
Accordingly, in arthritis there is a parallel enhancement of the response of spinothalamic-tract projection
neurons to mechanical stimulation as well as of the
EAAs acting at non-NMDA receptors (31). Recent
data also emphasize a possible role for NKA and its
receptor NK2 in nociception.
NPY in the intrinsic spinal interneurons or in
the descending noradrenergic bulbospinal system
could be involved in antinociception. Spinally administered NPY has antinociceptive effects and inhibits
stimulus-evoked SP release in the superficial dorsal
horn. Interestingly, NPY in the spinal cord often
co-localizes with the IAA GABA in local interneurons
in the superficial dorsal horn.
Segmental spinal processing and diffuse noxious
inhibitory control of nociceptive input in arthritis
Transmission of nociceptive information from
the inflamed joint to the somatosensory cortex does
not occur along “a labeled line’’ in a one-to-one
manner. The nociceptive input is first processed at the
segmental spinal level. Type IV/C and IIYAG PANS
relay nociceptive information to the second-order neurons (32). The spinal dorsal horn neurons display
different response properties, with some cells being
excited solely by PANS (nociceptive-specificneurons)
and other cells responding to non-nociceptive proprioceptive and tactile stimuli, but further increasing
their response rate when a noxious stimulus is applied
(wide dynamic range neurons).
One approach to modulating nociceptive signal
transmission at a segmental level in the spinal cord is
to stimulate type IIIAp and/or type IIAa nonnociceptive primary afferent neurons; such stimulation
via local inhibitory interneurons leads to presynaptic
(PAN) and/or postsynaptic (second-order projection
neuron) inhibition of nociceptive transmission, with
pain relief as a result (6). This afferent inhibition is
clinically utilized in physiotherapy, in which various
stimuli such as surface or deep heat, massage, and
range-of-motion exercises are used to stimulate type
IJIAp and/or I/Aa non-nociceptive nerves and to suppress nociceptive transmission. It is possible to obtain
good results in the treatment of pain by selective
high-frequency and low-intensity transcutaneous
nerve stimulation of type IYAp large-diameter fibers,
which have the lowest electrical threshold (33).
Segmental spinal inhibition is mediated by nonnociceptive neurons, which mainly terminate in laminae I11 and IV (nucleus proprius). According to the
original gate-control theory, these type IVAp and I/Aa
non-nociceptive nerve fibers also stimulate local inhibitory interneurons located abundantly in lamina I1 (or
substantia gelatinosa), where many type IV/C PANS
also terminate. Type IV/C PANS inhibit, whereas
thick myelinated type IYAP and YAa non-nociceptive
nerve terminals excite, inhibitory interneurons. These
local inhibitory interneurons, which do not project to
supraspinal sites, modulate nociceptive signal transmission at the segmental spinal level.
In neurochemical terms, GABA and endogenous opioids have been implicated in segmental spinal
control. The exact neuroanatomical organization of
the segmental spinal control system is not yet known.
In particular, axo-axonic synapses on the PAN terminals, thought to be the anatomical substrate of presynaptic inhibition, are rare in the superficial laminae;
those which do exist are often the wrong way round:
afferent fibers presynaptic to dendrites of the substaotia gelatinosa neurons. However, opioid and GABA
receptors have been found on some terminal-type
PAN fibers. It has therefore been suggested that
dendroaxonic synapses or nonsynaptic interactions
(see below) could be involved in presynaptic inhibition. In contrast, endogenous opioid and GABA syn-
apses have been shown on dorsal horn projection
It is interesting that in contrast to high-intensity
stimulation of type IVIC and IIIIAG nociceptive fibers
(e.g., “electroacupuncture,” which leads to a release
of endogenous opioids), the effect of selective lowintensity stimulation of type IYAp fibers is not generally reversed by naloxone. This suggests that analgesia
from transcutaneous electrical nerve stimulation at the
spinal cord level is mediated by a nonopioid inhibitory
neuromodulator, such as GABA. In addition to its
effects at the spinal cord level, transcutaneous electrical nerve stimulation may block afferent fibers.
Acupuncture may involve an element of counterirritation, in which a heterotopic, noxious stimulus
causes a relatively generalized inhibition of dorsal
horn WDR neurons, with excitatory fields in other
body areas. This has the dual function of 1) enhancing
the stimulus-to-noise ratio of the most recently applied
noxious stimulus, and 2) giving pain relief to other
injured or painful sites-at the cost of the pain associated with counterirritation (34). This phenomenon,
called diffuse noxious inhibitory control (DNIC),
seems particularly relevant to rheumatologists, because much of the visceral, deep dorsal neck muscle
and joint input signals are distributed via WDR neurons (8); DNIC does not affect the NS neurons relaying nociceptive input from superficial cutaneous structures. DNIC involves a supraspinal brainstem loop,
with spinoreticular tract and dorsolateral funiculus as
the ascending and descending tracts, respectively. In
freely moving individuals, there is probably a continuous, relatively random, activation of the lowthreshold, rapidly-adapting proprioceptive and tactile
type IIAa and IIIAp fibers signaling to dorsal horn
WDR neurons as a form of “somatosensory noise.”
As a result of DNIC type arrangement, a strong
noxious stimulus will not only send a positive signal to
the central nervous system, but it will also reduce the
background (34).
In healthy volunteers, the effect of large afferent fiber Stimulation on dorsal horn neurons lasts for
only seconds, but in patients with chronic pain, including those with musculoskeletal diseases and RA, relief
lasts for extended periods of time (32,33). It can
therefore be assumed that the segmental control system is also subject to activity-induced changes in its
function, as suggested in the original gate-control
theory (6).
Table 5. Characteristics of sensitization
Spontaneous activity
Decreased pain threshold
Recruitment of so-called “silent” nociceptors
Increased responsiveness to suprathreshold stimuli
Expansion of the excitatory receptive field*
Additional receptive fields*
Increased stimulus-evoked release of neuropeptides
Disturbed sympathetic (and other) reflexeslresponses
* Acts mainly or only at the central nervous system level.
Central hyperexcitability in arthritis
In arthritis, activity in the PANs leads to considerable changes in the properties of pain-relay neurons of the spinal dorsal horn in the form of enhancement of the responses of the ascending tract cells
(Table 5) (24). Such changes in synaptic efficacy can
last for several minutes (short term) or for extended
periods of time (long term). Short-term sensitization
has been explained by modulation of ion-channel
and/or neurotransmitter (or modulator) receptor function, whereas long-term facilitation requires changes
in gene regulatory events and/or synaptic reorganization. This section deals with short-term sensitization;
the next section is about long-term sensitization.
Exogenous administration of SP into the spinal
cord produces delayed, slow, and prolonged EPSPs on
second-order dorsal horn neurons. SP and CGRP
enhance the release of EAAs and lead to a long-lasting
potentiation of NMDA receptor-mediated EAA responses (35). The delayed, slow, and prolonged EPSPs
induced by SP in dorsal horn projection neurons may
be mediated by an endogenous mediator (such as Glu
or Asp) acting at the NMDA receptor, compatible with
the suggestion that SP may be a neuromodulator rather
than a neurotransmitter. SP is synthesized and stored
in type IV/C PANs, released into the synaptic extracellular space, and bound by postsynaptic NK, receptors, and thus fulfills many of the criteria of a candidate for primary afferent transmitter. However, its
time course of action is slow compared with nociceptive activity, and its action may be prolonged in
arthritis due to diminished degradation.
It seems that central hyperexcitability to nociceptive stimulation is mediated mainly via two mechanisms: by EAA action on NMDA-type receptor Ca2’
channels after removal of the voltage-sensitive Mg2+
block of that receptor, and by SP action on NK,
receptors. The latter interaction stimulates phospholipase C to hydrolyze phosphatidylinositol-4,5bisphosphate into the intracellular messengers inositol
trisphosphate (and derivatives) and diacyglycerol(36).
Stimulation of diacylglycerol formation together with
an increase in intracellular calcium leads to protein
kinase C translocation and activation. Protein kinase C
(and other protein kinases activated due to stimulation
of the postsynaptic projection neurons) may alter
membrane excitability via phosphorylation of integral
membrane proteins, such as ion channels. It has been
recently shown that protein kinase C removes the
voltage-sensitive Mg2+ block from NMDA receptors
(35); therefore, many other neurotransmitters and neuromodulators, which contribute to protein kinase C
activation, can also switch Glu action to NMDA
receptors. According to more recent findings, NMDA
may also stimulate the nitric oxide/cGMP pathway and
thus alter nociceptive processing. Interestingly, excessive GWNMDA action leads to excitotoxicity, i.e.,
nerve cell damage and death.
Together with NMDA receptor agonists (such
as Glu), SP causes NK, receptor-mediated formation
of delayed, slow, and prolonged EPSPs in dorsal horn
neurons after mechanical and electrical stimulation in
the cat, but capsaicin-induced formation of slow
EPSPs in the rat seems to be a combined effect of
NMDA receptor agonists acting together with N K N
NK,. However, EAALNMDA and SP/NK, systems
are usually considered to act in concert on WDR and
NS spinothalamic-tract neurons, leading to the formation of delayed, slow, and prolonged EPSPs, sensitization, and wind-up (Table 5 ) (35). “Wind-up” describes an augmented response of nociceptive dorsal
horn neurons to repetitive type IV/C fiber stimulation
caused by temporal summation of depolarizations.
Accordingly, NMDA and NK, receptor antagonists
attenuate sensitization of the dorsal horn neurons, and
intrathecal antibodies to SP and CGRP inhibit production of hyperalgesia.
Prostaglandins and PGH, synthetase may be
involved in the EAA/NMDA receptor and SP/NK,
receptor signal-transduction pathways, because intrathecal PGE, produces hyperalgesia, and hyperalgesia
mediated by Glu or SP is blocked by spinal cyclooxygenase inhibition (37). This latter feature also dissociates the spinal analgesic effect of NSAIDs from their
peripheral antiinflammatory actions (see above).
NSAIDs also have analgesic effects on supraspinal
structures, which may be partly mediated by inhibition
of central prostaglandin formation induced in inflammatory and noninflammatory pain states. These central
analgesic effects are perhaps linked to prostaglandininduced inhibition of the descending monoaminergic
pathways (i.e., NA and 5-HT release from the spinal
terminals), which also contributes to down-regulation
of nociception at the spinal level (see below) (38).
These studies have established a role, aside
from the actions of EAAs on ionotropic NMDA receptors, of the SP/NK, receptor system as an important
contributor to central sensitization. It is worth emphasizing that these stimuli are required only for the
development of enhanced synaptic transmission and
sensitization of the spinal and supraspinal pain-relay
neurons. Once the central sensitization has been established, ongoing activity of these systems is not a
prerequisite for ongoing pain (or provoked pain, arising as a result of “pain memory”), which may continue even after deafferentation or as “phantom pain,”
without further input from the initially painful area
(36). This centralization of pain may depend on a
cascade of slow changes sweeping over the central
nervous system in chronic inflammation (see below).
Central neuroplasticity
Perception of pain in arthritis does not simply
involve a moment-to-moment analysis of afferent nociceptive input in the central nervous system. It also
involves dynamic activity-dependent changes, such as
segmental gate control and alterations of the excitability as discussed above. Activity-dependent changes
also comprise structural and functional changes of
dorsal horn neurons and other higher order projection
neurons, local spinal interneurons, and long descending neurons (for recent reviews, see refs. 36 and 39).
It has been suggested that the induction of
immediate-early genes and their proto-oncogene products, such as c-fos, Fos B, Jun, Jun B, Jun D,
NGF1-A, NGF1-B, and SRF, is triggered by a chain of
events initiated by the action of EAAs on ionotropic
non-NMDA-type AMPA or kainate receptors and
mGlu receptors, of SP on NK, receptors, and finally,
of EAAs acting on NMDA receptors (36). Considering
the post-receptor interactions at the second messenger
level, in particular protein kinase C-mediated removal
of the Mg2+block from the NMDA receptors, it seems
likely that many other neuropeptides and mediators
released by PANS and other convergent fibers can also
contribute to this end.
The conclusion that these changes are indeed
driven by the nociceptive transmitter/modulator
substances/input associated with arthritis and pain is
further suggested by the following two facts. At the
spinal level in monoarthritis, these changes are first
restricted to the exact area of pain relay from the
involved joint, and are thus restricted both unilaterally
and segmentally. When the arthritis becomes chronic,
the changes extend to the contralateral side as well as
in a rostrocaudal direction in the spinal cord, and thus
become more widespread. Second, systemic morphine
suppresses noxious stimulus-evoked fos proteinlike immunoreactivity.
The products of the immediate-early genes act
as transcription factors in the initiation of mRNA
synthesis, considered to be the primary control point
in the regulation of eukaryotic gene expression. This is
accomplished by the interplay of short defined DNA
sequences within and near the transcribing gene and
various DNA binding proteins. These gene products
act as third messenger cascades and couple transsynaptic stimuli with altered neuronal gene expression
(referred to here as central neuroplasticity). It is
possible that in addition to the above-mentioned messenger systems, still other messengers, including the
retrograde messengers nitric oxide, carbon monoxide,
and arachidonic acid (and their metabolites), participate in the maintenance of the central sensitization.
Because of central hyperexcitability and central
neuroplasticity, referred pain and/or hyperalgesia can
be felt in areas remote from the initial trigger site (such
as monoarthritis). Similarly, once established, even
non-noxious stimulation of areas remote from the
inflamed joint can stimulate dorsal horn, thalamic, and
cortical neurons, which show enhanced excitability.
These functional changes are associated with an increased expression of proto-oncogenes in the higherorder projection neurons, pathways, nuclei, and somatosensory cortex, which demonstrates that the
structural changes may also spread trans-synaptically
(36). These changes cannot be explained either by
peripheral sensitization or by convergence of, for
example, cutaneous and articular nerves, on the same
second-order projection neuron. Thus, there is further
support for the existence of central neuroplasticity .
Unilateral inflammation leads to changes in the
excitability of withdrawal reflexes on the contralateral
side as well. These changes are maintained even after
inputs from the injured site are blocked by local
anesthesia or deafferentation. It therefore seems that
once induced, central hyperexcitability is maintained
without further input from the periphery (36). The
therapeutic significance lies in the fact that this component of the pain, which is probably very significant
in chronic inflammatory conditions such as arthritis, is
not affected by the interventions aimed at moderating
inflammatory changes in the diseased joints. Centralization of pain may also make it diacult to assess the
efficiency of short-term treatment with analgesic drugs
in patients with chronic pain because even years after
the original injury, a peripheral trigger can reactivate
the past pain. This shows that the activity-induced
changes and the central memory trace they leave (pain
memory) can, unfortunately, be very long-lasting (36).
Supraspinal NA- and 5-HT-mediated influences in
The endogenous analgesia system involved in
stimulation-produced and opioid-activated analgesia
also modulates nociceptive input at the spinal level in
arthritis. Activation of the nociceptive pathways in
arthritis is counteracted by an increase in tonic descending inhibition, as shown by temporary spinalization with cold block (40). Concomitantly with the
development of arthritis, there is increased NA (41)
and 5-HT (42) synthesis and release from the descending inhibitory pathways passing in the dorsolateral
funiculus. These monoamines seem to mediate an antinociceptive effect via q-adrenoceptors and 5-HTIA/
5-HT3receptors, respectively. However, in chronic RA,
levels of metabolites of only NA (not those of 5-HT)
are high in the cerebrospinal fluid (43). Although not a
subject of this review, it is interesting that neurons
containing p-endorphin (an endogenous opioid of the
pro-opiomelanocortin family), which are relatively restricted in the central nervous system (mainly to the
hypothalamus), send projections to the noradrenergic
nuclei in the brain stem, and that both the synthesis
and secretion of pendorphin are increased in polyarthritis, reflecting a chronic stress state. This would
be compatible with the hypothesis that supraspinal
opioids produce their effects by activating descending
inhibitory controls.
NA and 5-HT inhibit nociceptive signal transmission by pre- and postsynaptic actions (32). This is
partly made possible by the predominance of nonsynaptic organization of the NA- and 5-HT-containing
varicosities, which enable volume transmission (see
also below). In chronic diffuse pain syndromes such as
fibromyalgidfibrositis, cerebrospinal fluid monoamine
metabolites are low (43). Low doses of amitriptyline,
which not only block the high-affinity reuptake of
biogenic amines, but also increase the sensitivity of
the postsynaptic 5-HT receptors and decrease the
presynaptic receptors involved in autoinhibition of NA
release, can be used to treat such patients. It would
Table 6. Characteristicsof oDioid receptors
p receptors
6 receptors
Fend > dyn A pend = leu> met-enk
enk = met> leu-enk
enk > dyn
dyn A >> Fend
>> leu-enk =
Inhibit adenylate
cyclase; close
appear that mixed-type antidepressants (trimipramine
and, particularly, amitriptyline), which inhibit reuptake of both NA and 5-HT, are better than noradrenergic antidepressants (desipramine and maprotiline). The use of serotonergic drugs (clomipramine,
citalopram, fluvoxamine) in the management of
chronic pain has been questioned.
Some of the effects of the noradrenergic and
serotonergic bulbospinal pathways may be mediated
by local interneurons, in particular, those containing
endogenous opioids (32). Endogenous opioids are antagonized by the endogenous “anti-opioid” cholecystokinin (CCK) (see also below). Systemic, spinal (intrathecal or epidural), and peripheral application of
morphine-like drugs can be used to treat labor, postoperative, myocardial, post-arthroscopic, and cancer
pain. Other transmitters/modulators such as IAAs
(glycine, serine, taurine) and neuropeptides (NPY,
neuropeptide FF, GAL, neurotensin, thyrotropinreleasing hormone) found in primary afFerents, spinal
interneurons, or in the long descending (or ascending)
tract neurons also participate in the modulation of
nociceptive input at the spinal level. Utilization of
these systems in the management of inflammation,
arthritis, and pain is still at an experimental stage.
Opioid peptides in arthritis
One class of neuropeptides which deserves
particular attention in arthritis is the opioid peptides of
the POMC, proenkephalin, and prodynorphin families,
which act on at least 3 different opioid receptors (Table
6). Under normal circumstances, enkephalin mRNA
and peptide levels in the spinal cord are much higher
than those of dynorphin. Enkephalins, found in spinal
interneurons and long-tract neurons, act mainly on 6
and p-opioid receptors (Table 6) and may be important
in the evoked pre- and postsynaptic inhibition of pain
pathways (32).
In hyperalgesia and acute arthritis, levels of
preprodynorphin mRNA and dynorphin peptide are
rapidly and greatly increased in the primary afferent,
local spinal cord, and projection neurons in areas that
receive sensory input from the afFected limb (12), in
arthritis in laminae I, 11, IV, and V, and dorsolateral to
the central canal, probably as a result of induction of
the third-messenger immediate-early genes (36,39).
Interestingly, these changes as well as the associated
behavioral changes are initially restricted unilaterally
and segmentally, but later spread to the contralateral
side and in a rostrocaudal direction, such that the
spinal changes start to resemble those seen in chronic
polyarthritis. It therefore seems clear that inflammation and arthritis, via the local spinal release of
transmitters/mediators, lead to modulation of nuclear
transcription factors, and thus to slow alterations of
the phenotype and function of the target cells via
gene-regulatory events.
In the spinal cord, the opioid-containing neurons are in synaptic contact mainly with projection
neurons and occassionally also with PANs. However,
due to the presence of p, 6, and K receptors on spinal
PAN terminals and due to the well-documented presynaptic effects of local opioids, it seems that nonsynaptic opioid actions on PANs should also be considered (see below).
Although dynorphin inhibits hyperalgesia, it
can also contribute to hyperexcitability , excessive
depolarization, and excitotoxicity (39). It is therefore
of interest that dynorphin immunoreactivity in chronic
polyarthritis and in long-lasting monoarthritis can be
observed throughout the cord, whereas in these longstanding cases, leu-enkephalin and met-enkephalin
also increase, but only in the dorsal horn. This would
be compatible with an antinociceptive role of enkephalin opioids in arthritis, whereas dynorphin has also
been indicated in neurotoxicity, possibly mediated
indirectly via NMDA receptors (36).
Opioid binding in the dorsal horn is not altered
in arthritis, but spinal sensitivity to the antinociceptive
effects of specific p- and Sopioid agonists is enhanced
in arthritis due to increased sensitivity of adenylate
cyclase to their inhibitory effects. Dynorphin, first
described as a specific endogenous ligand of the Kopioid receptor, excerts antinociceptive effects in the
spinal cord in arthritis. This might be due to its
substantial avidity for the p-opioid receptor, but more
likely reflects dynorphin actions on a K-receptor subtype. Similarly, arthritic rats are hypersensitive to systemic opioids and naloxone, in which sensitivity to pand particularly K-opioid receptors seems to play a role.
Opiate-induced analgesia is counteracted by an
endogenous anti-opioid peptide cholecystokinin
(CCK). Injury or signals for impending danger activate
endogenous opiate analgesia, which can be regarded
as an adaptive behavior because under such stress, it
is possible to ignore the pain in a situation potentially
critical to survival. Once the stress has passed, safety
cues activate the antianalgesia system, probably the
CCK/CCK-B receptor system, to allow perception of
pain and recuperative behaviors. This system can be
utilized to modulate nociceptive input at the spinal
level: CCK agonists block, and antagonists potentiate,
opioid analgesia.
In addition to their spinal and supraspinal sites
of action, opioids participate in the modulation of
nociceptive transmission in the peripheral tissues via
p-, 6,and K-opioid receptors. In peripheral tissues,
p-opioid agonists inhibit the discharges of PANS from
inflamedjoints in experimental arthritis, and according
to more recent observations, intraarticular opioids can
be utilized to prevent pain after arthroscopic knee
surgery (44). Agonists of S and K-opioids inhibit
PGSN-dependent BK-induced hyperalgesia and
plasma extravasation. Local application of opioids
can, theoretically, be expected to be very effective in
the prevention of peripheral sensitization: NSAIDs
and corticosteroids only inhibit the production of
eicosanoids, which leaves BK, SHT, and adenosine
free to act (Table 4). Because the receptors of all these
sensitizing compounds are connected to a stimulatory
G protein that activates the adenylate cyclasemediated increase in CAMP,opioid agonists can counteract all of the compounds by inhibiting CAMP formation (Table 6) (9,45).
Interestingly, opioids of the POMC and proenkephalin families have been found in increased
concentrations in inflamed joint tissues and fluids (46).
Immunohistochemicaland in situ hybridization studies
show that these endogenous opioids (47) as well as
other important neuropeptides, such as corticotropinreleasing hormone (48), are produced by non-neural
inflammatory cells.
The presence of p, 6 and K receptors at the
peripheral, spinal, and supraspinal sites, and dosedependent actions and interactions of more than 20
known opioid peptides make it difficult to assess the
exact site and mechanism of action of systemic opioid
antagonists and agonists (or of enzymes inhibiting
degradation of endogenous opioids). Characterizing
the acti6ty of large populations of neurons in awake,
freely moving animals, based on third messenger expression, may facilitate evaluation of these compli-
cated interactions in dserent settings, including acute
and chronic arthritis.
Neuropeptide degradation and volume transmission
One aspect of importance for both the supraspinal and segmental pain control system and
arthritis-induced changes in the spinal cord relates to
neuropeptide degradation. In addition to the synaptic
(transmission across a synaptic cleft, the “wiring”
connection) and ephaptic (conduction of nerve impulses across a point of lateral contact, or “crosstalk”) organization, it is also possible for neuropeptides to be released and to diffuse widely in the dorsal
horn, where they can act as neuromodulators (“open”
volume transmission) (49). This has been clearly demonstrated for NKA, which in acute arthritis is rapidly
released at the dorsal surface of the spinal cord and in
the superficial dorsal horn. Within 1 hour, NKA is
detectable throughout the dorsal horn and adjacent
white matter (50).
In contrast to wired synaptic transmission, the
speed and segregation (or “safety”) of volume transmission are low, but the degree of divergence and
plasticity are very high because of the many possible
combinations and interactions of neuropeptides and
other transmitters at the pre- and postreceptor levels
(49). More importantly, volume transmission exerts
long-term actions, which affect relatively wide areas in
the dorsal horn (49) and may explain how.short-term
activity can lead to long-term changes in #he response
properties of the spinal dorsal horn.
A “mismatch” problem is a lack of correspondence between the localization of neuropepide release
sites and their receptors; this was first noticed for the
enkephalinergic system. It seemed logical to speculate
that locally released neuromodulators might have access to their distant receptor sites via transport in the
extracellular fluid.
Volume transmission could explain the discrepancy between studies showing a presynaptic action of
a peptide/amine but no anatomical substrate for such
an interaction. One example is the presynaptic inhibition of PAN by enkephalin and GABA in the absence
of axoaxonic connections from enkephalinergic or
GABAergic neurons to PAN fibers (32). That this
indeed may be relevant is suggested by an elegant
study in which prevention of degradation of endogenous enkephalins in the spinal cord inhibited nociception (51). Similarly, it was shown that substance P
degradation can also be inhibited by peptidase inhibi-
Table 7. Some enzymes involved in neuropeptide degradation
Removal of the N-terminal amino acid
Dipeptidyl aminopeptidases
Removal of N-terminal dipeptides
Neutral endopeptidase
Cleaves at the N-terminus of
hydrophobic residues
enzyme (ACE)
Removal of C-terminal dipeptides
Removal of the C-terminal amino acid
Removal of C-terminal amide
Leucine aminopeptidase EC 3.4.1 1.1 ; alanine aminopeptidase
EC (recognized by monoclonal antibody [MAb]
CD13); cystyl aminopeptidase EC 3.4.1 1.3; aminopeptidase B
EC; aminopeptidase A EC; aminopeptidase P
EC 3.4.14.-, e.g., dipeptidyl peptidase IV EC
(recognized by MAb CD26)
EC; synonyms: CALLA (common acute lymphoblastic
leukemia-associated antigen), enkephalinase (recognized by
MAb CD10)
ACE EC (also known as kininase 11) also has
endopeptidase activity; first described as an important
component of the renin-angiotensin-aldosteronesystem; in
addition to angiotensin I, substrates include substance P and
Carboxypeptidase A EC; carboxypeptidase B EC; carboxypeptidase N EC; carboxypeptidase E
EC (enkaphalin convertase)
Discovered in human platelets, also acts as a carboxypeptidase
tors and that its slow degradation results in SP accessing many sites remote from the sites of release.
The coexistence of two or more neuromodulators in the same axon challenges the old idea that any
given neuron utilizes only one neurotransmitter at its
various synapses. It seems unlikely that coexistence
reflects an ancestral vestigium of a function that became obsolete during phylogenesis or development. It
seems more likely that the neurons contain a primary
and fast neurotransmitter/receptor system (such as
Gltdionotropic non-NMDA receptors in the PAN/
projection neuron synapse) as well as other more
slowly acting neuromodulators (exemplified by the
Glu/NMDA and SP/NK, and others mentioned above,
which have been indicated in central hyperexcitability
and neuroplasticity). In this context it is of interest
that Glu may, in addition to synaptic receptors, have
another set of receptors at nonsynaptic sites, suggesting the possibility of interactions outside specialized
areas such as terminals (suited to release and/or reuptake of transmitters) and axons (suited to K+/Na+
exchange during electrical activity).
In addition to transmitter/modulator versus receptor mismatch, there is increasing evidence of the
ability of the microglial cells to take up and handle
transmitters/modulators. Nerve cells contain areas either in contact with glial cells/processes and bare areas
in addition to areas taken up by synaptic contacts. The
ionic and transmitter/modulator environment may affect the nerve cells in a significant way. For example,
it has been reported that a defective high-attnity glial
Glu uptake system and increased extracellular Glu
may lead to excitotoxicity.
As is clear from the above, neuropeptide degradation rather than reuptake by the axon terminals
contributes to the site- and time-specific neuropeptide
action or, alternatively, to the lack of it, i.e., to effects
exerted via volume transmission. Therefore, the enzymes responsible for neuropeptide degradation deserve special attention (Table 7).
In addition to amino- and carboxypeptidases
(including dipeptidyl peptidases), angiotensinconverting enzyme (ACE) and neutral endopeptidase
(NEP) are of particular interest. NEP was first described as a tumor-associated antigen, CALLA (the
common acute lymphoblastic leukemia-associated antigen). The complementary DNA sequence of CALLA
disclosed identity with NEP and with enkephalinase.
NEP has been localized to, e.g., early lymphoid progenitor cells, fibroblasts, granulocytes, and the dorsal
horn of the spinal gray matter, with distribution that
overlaps with that of enkephalin- and SP-rich regions
(52). In addition to peptidases, neuropeptides may be
degraded or marked for degradation by various reactive oxygen species in an oxyspecies- and timedependent manner. Neuropeptide-degrading enzymes
are at least temporarily diminished in the cerebrospinal fluid in arthritis (53), which may make it possible
for SP, CGRP, and other neuropeptides to spread,
persist, and cause some of the prolonged excitability
changes evoked by brief noxious stimuli (50).
Persistence of NKA and the usually short halflife of SP in the dorsal horn can be explained by their
differential sensitivities to degrading enzymes. However, degradation does not necessarily lead to the loss
of activity. In the spinal cord, N-terminal SP frag-
ments SP,,, SP,_,, and SP,, that are produced upon
local degradation desensitize to SP-induced behaviors
(which indicate pain and hyperalgesia), probably via
action on p-opioid receptors, since they displace p
agonists and because naloxone prevents desensitization. C-terminal fragments SP3-ll and SP5-ll induce
SP-like behaviors, but do not desensitize.
Although the role of neuropeptide degradation
in the dorsal horn seems particularly interesting, it is
also altered in the periphery, i.e., in the arthritic joint.
In the peripheral tissues, neuropeptide molecules may
have to diffuse up to 1,000 pm to reach the postjunctional target cell receptor, whereas the distance from
the presynaptic axon terminus to the postsynaptic
membrane is only 15-25 nm. ACE and, in particular,
NEP are increased in the synovial tissue and fluid of
RA patients (5435). This may lead to accelerated
neuropeptide degradation at the site of inflammation.
Many of the clinical conditions that point to a
possible involvement of neural mechanisms in arthritis
nonselectively involve different types of neurons (peripheral nerve lesions) or neurons that may not be so
important in arthritis (upper or lower motor neurons).
A more detailed review of specific nerve fibers and
transmitters/modulatorsand their alterations in arthritis suggests that peripheral and spinal neural mechanisms may play a role in arthritis, particularly with
regard to inflammation and pain.
We are most grateful to our international hosts
Professor Paul Davis and Professor Anthony S. Russell
(Edmonton), Dr. Jon D. Levine (San Francisco), Dr. Robert
J. Winchester (New York), and Professor Manfred Zimmermann (Heidelberg) for their support during this work and
during our sabbatical years.
Levine JD, Collier DH, Basbaum AI, Moskowitz MA, Helms
CA: Hypothesis: the nervous system may contribute to the
pathophysiology of rheumatoid arthritis. J Rheumatol 12:411, 1985
Kidd BL, Mapp PI, Gibson SJ, Polak JM, O’Higgins F, Buckland Wright JC, Blake DR: A neurogenic mechanism for symmetrical arthritis. Lancet 2:1128-1130, 1989
Matucci-Cerinic M, Partsch G: The contribution of the peripheral nervous system and the neuropeptide network to the
development of synovial inflammation (editorial). Clin Exp
Rheumatol 10:211-215, 1992
Von Euler US, Gaddum JH: An unidentified depressor substance in certain tissue extracts. J Physiol 72:74-87, 1931
98 1
5. Konttinen YT, Shiozawa S, Hukkanen M, Santavirta S: Neural
influences in musculoskeletal diseases, Rheumatology, State of
the Art. Edited by G B6lint. Apsterdam, Elsevier Science, 1992
6. Melzack R, Wall PD: Pain mechanisms: a new theory. Science
150:971-979, 1%5
7. Konttinen YT, Sorsa T, Santavirta S, Russell A: Via dolorosa,
from the first to the second station. J Rheumatol 21:783-787,
8. Konttinen YT, Koski H, Santavirta S, Hukkanen M, Soinila S:
Nociception, proprioception and neurotransmitters in the cervical spine, Disorders of the Cervical Spine. Edited by JH
Bland. Philadelphia, WB Saunders Company, 1994
9. Levine JD, Fields HL, Basbaum AI: Peptides and the primary
afferent nociceptor. J Neurosci 13:2273-2286, 1993
10. Devillier P, Weill B, Renoux M, Menkes C, Pradelles P:
Elevated levels of tachykinin-like immunoreactivity in joint
fluids from patients with rheumatic inflammatory diseases (letter). N Engl J Med 314:1323, 1986
11. Konttinen YT, Rees R, Hukkanen M, Gronblad M, Tolvanen E,
Gibson SJ, Polak JM, Brewerton DA: Nerves in inflammatory
synovium: immunohistochemical observations on the adjuvant
arthritis rat model. J Rheumatol 17:15861591, 1990
12. Weihe E, Nohr D, Millan MJ, Stein C, Muller S, Gramsch C,
Herz A: Peptide neuroanatomy of adjuvant-induced arthritic
inflammation in rat. Agents Actions 25255-259, 1988
13. Basbaum AI, Levine JD: The contribution of the nervous
system to inflammation and inflammatory disease. Can J Physiol
Pharmacol69:647451, 1991
14. Colpaert FC, Donnerer J, Lembeck F: Effects of capsaicin on
inflammation and on the substance P content of nervous tissues
in rats with adjuvant arthritis. Life Sci 32:1827-1834, 1983
15. Deal CL, Schnitzer TJ, Lipstein E, Seibold JR, Stevens RM,
Levy MD, Albert D, Renold F: Treatment of arthritis with
topical capsaicin: a double-blind trial. Clin Ther 13:383-395,
16. Lembeck F, Donnerer J, Colpaert FC: Increase in substance P
in primpry afferent nerves during chronic pain. Neuropeptides
1:175-180, 1981
17. Donaldson LF, Harmar M , McQueen DS, Seckl JR: Increased
expression of preprotachykinin, calcitonin gene-related peptide,
but not vasoactive intestinal peptide messenger RNA in dorsal
root ganglia during the development of adjuvant monoarthritis
in the rat. Brain Res Mol Brain Res 16:143-149, 1992
18. Matucci-Cerinic M, Pignone A, Lotti T, Partsch G, Livi R,
Cagnoni M: Gold salts and somatostatin: a new combined
analgesic treatment for psoriatic arthritis. Drugs Exp Clin Res
185361, 1992
19. Levine JD, Clark R, Devor M, Helms C, Moskowitz MA,
Basbaum AI: Intraneuronal substance P contributes to the
severity of experimental arthritis. Science 226547-549, 1984
20. Lam FY, Ferrell WR:Inhibition of carrageenan induced inflammation in the rat knee joint by substance P antagonist. Ann
Rheum Dis 48:928-932, 1989
21. Levine JD, Fye K, Heller P, Basbaum AI, Whiting-O’KeefeQ:
Clinical response to regional intravenous guanethidine in patients with rheumatoid arthritis. J Rheumatol 13:1040-1043,
22. Sternberg EM, Chrousos GP, Wilder RL, Gold PW: The stress
response and the regulation of inflammatorydisease. Ann Intern
Med 117:854-866, 1992
23. Cunha FQ, Poole S, Lorenzetti BB, Ferreira SH: The pivotal
role of tumour necrosis factor alpha in the development of
inflammatory hyperalgesia. Br J Pharmacol 107:660-664, 1992
24. Men6trey D, Besson JM: Electrophysiologicalcharacteristics of
dorsal horn cells in rats with cutaneous inflammation resulting
from chronic arthritis. Pain 13:343-364, 1982
25. Sano H, Hla T, Maier JA, Crofford W,Case JP, Maciag T,
Wilder RL: In vivo cyclooxygenase expression in synovial
tissues of patients with rheumatoid arthritis and osteoarthritis
and rats with adjuvant and streptococcal cell wall arthritis.
J Clin Invest 89:97-108, 1992
26. Donnerer J, Schuligoi R, Stein C: Increased content and transport of substance P and calcitonin gene-related peptide in
sensory nerves innervating inflamed tissue: evidence for a
regulatory function of nerve growth factor in vivo. Neuroscience 49:693-698, 1992
27. Sluka KA, Dougherty PM, Sorkin LS, Willis WD, Westlund
KN: Neural changes in acute arthritis in monkeys. 111. Changes
in substance P, calcitonin gene-related peptide and glutamate in
the dorsal horn of the spinal cord. Brain Res Brain Res Rev
17:29-38, 1992
28. Oku R, Satoh M, Takagi H: Release of substance P from the
spinal dorsal horn is enhanced in polyarthritic rats. Neurosci
Lett 74:315-319, 1987
29. Almay BG, Johansson F, Von Knomng L, Le Greves P,
Terenius L: Substance P in CSF of patients with chronic pain
syndromes. Pain 333-9, 1988
30. Sluka KA, Westlund KN: An experimental arthritis in rats:
dorsal horn aspartate and glutamate increases. Neurosci Lett
31. Dougherty PM, Sluka KA, Sorkin LS, Westlund KN, Willis
WD: Neural changes in acute arthritis in monkeys. I. Parallel
enhancement of responses of spinothalamic tract neurons to
mechanical stimulation and excitatory amino acids. Brain Res
Brain Res Rev 17:l-13, 1992
32. Besson JM, Chaouch A Peripheral and spinal mechanisms of
nociception. Physiol Rev 67:67-186, 1987
33. Wall PD, Sweet WH: Temporary abolition of pain in man.
Science 155:10%109, 1967
34. Le Bars D, Dickenson AH, Besson JM: Diffuse noxious inhibitory controls (DNIC). I. Effects on dorsal horn convergent
neurones in the rat. Pain 6:283-304, 1979
35. Chen L, Huang LY: Protein kinase C reduces M 8 + block of
NMDA-receptor channels as a mechanism of modulation. Nature 356521423, 1992
36. Coderre TJ, Katz J, Vaccarino AL, Melzack R: Contribution of
central neuroplasticity to pathological pain: review of clinical
and experimental evidence. Pain 52:259-285, 1993
37. Malmberg AB, Yaksh TL: Hyperalgesia mediated by spinal
glutamate or substance P receptor blocked by spinal cyclooxygenase inhibition. Science 257:1276-1279, 1992
38. Bannwarth B, Demotes-Mainard F, Scheverbeke T, Dehais J:
Where are peripheral analgesics acting? Ann Rheum Dis 5214,
39. Dubner R, Ruda MA: Activity-dependent neuronal plasticity
following tissue injury and idammation. Trends Neurosci 15:
96-103, 1992
40. Schaible H-G, Neugebauer V, Cervero F, Schmidt RF: Changes
in tonic descending inhibition of spinal neurons with articular
input during the development of acute arthritis in the cat.
J Neurophysiol66:1021-1032, 1991
41. Weil-Fugazza J, Godefroy F, Manceau V, Besson JM: Increased norepinephrine and uric acid levels in the spinal cord of
arthritic rats. Brain Res 374:190-194, 1986
42. Godefroy F, Weil-Fugazza J, Besson JM: Complex temporal
changes in 5-hydroxytryptamine synthesis in the central nervous system induced by experimental polyarthritis in the rat.
Pain 28:223-238, 1987
43. Russell IJ, Vaeroy H, Javors M, Nyberg F: Cerebrospinal fluid
biogenic amine metabolites in fibromyalgidfibrositis syndrome
and rheumatoid arthritis. Arthritis Rheum 35550-556, 1992
44. Stein C, Cornisel K, Haimerl E, Yassouridis A, Lehrberger K,
Herz A, Peter K: Analgesic effect of intraarticular morphine
after arthroscopk knee surgery. N Engl J Med 325:1123-1 126,
45. Ferreira SH, Nakamura M: I. Prostaglandin hyperalgesia, a
cAMP/Caz+ dependent process. Prostaglandins 18:179-190,
46. Suzuki N, Yoshino S, Nakamura H:A study of opioid peptides
in synovial fluid and synovial tissue in patients with rheumatoid
arthritis. Arerugi 41:615420, 1992
47. Przewlocki R, Hassan AH, Lason W,Epplen C, Herz A, Stein
C: Gene expression and localization of opioid peptides in
immune cells of inflamed tissue: functional role in antinociception. Neuroscience 48:491-500, 1992
48. Crofford LJ, Sano H, Karalis K, Webster EL, Goldmuntz EA,
Chrousos GP, Wilder RL: Local secretion of corticotropinreleasing hormone in the joints of Lewis rats with inflammatory
arthritis. J Clin Invest 90:2555-2564, 1992
49. Agnati LF, Fuxe K, Zoli M, Ozini I, Toffano G, Ferraguti F: A
correlation analysis of the regional distribution of central enkephalin and beta-endorphin immunoreactive terminals and of
opiate receptors in adult and old male rats: evidence for the
existence of two main types of communication in the central
nervous system, the volume transmission and the wiring transmission. Acta Physiol Scand 128:201-207, 1986
50. Hope PJ, Jarrott B, Schaible H-G, Clarke RW, Duggan AW:
Release and spread of immunoreactive neurokinin A in the cat
spinal cord in a model of acute arthritis. Brain Res 533:292-299,
51. Roques BP, Fournie-Zaluski MC, Soroca E, Lecomte JM,
Malfroy B, Llorens C, Schwartz JC: The enkephalinase inhibitor thiorphan shows antinociceptive activity in mice. Nature
288:28&288, 1980
52. van Veldhoven P, Carton H: Enkephalinase A activity in
different regions of brain and spinal cord of normal and chronic
arthritic rats. FEBS Lett 138:7&78, 1982
53. Persson S, Post C, Holmdahl R, Nyberg F: Decreased
neuropeptide-converting enzyme activities in cerebrospinal
fluid during acute but not chronic phases of collagen induced
arthritis in rats. Brain Res 581:273-282, 1992
54. Lowe JR, Dixon JS, Guthrie JA, McWhinney P: Serum and
synovial fluid levels of angiotensin converting enzyme in polyarthritis. Ann Rheum Dis 49921-924, 1986
55. Sreedharan SP, Goetzl EJ, Malfroy B: Elevated synovial tissue
concentration of the common acute lymphoblastic leukaemia
antigen (CALLA)-associated neutral endopeptidase (
in human chronic arthritis. Immunology 71:142-144, 1990
Без категории
Размер файла
1 889 Кб
particular, pain, spina, treatment, periphery, inflammation, mechanism, neural, arthritis, references
Пожаловаться на содержимое документа