Finally, total lyrnph node irradiation using protocols originally designed for
Hodgkin's discase has been used in a few patients at Stanford (Kotzin, Strober,
Calin, Hoppe & Kaplan, 1980). The virtue ofthis approach is that a wide experience
than a dozcn patients who received 3600 or 4000 rad in divided dose s. Four are said
to have improved with the lower dose, and thrce on the higher dose are said to be in
remission after several months. This therapy could prove to be beneficial, although
it can hardl y be said to be simple.
All ofthe methods under discussion suffer from their failure to deal selectively with
the force s leading to rheumatoid intlammation . Many of the toxicities arise out of
this unselectivity. One can hope that as understanding ofthc way environmental and
genetic factors interact to lead to the immune and intlammatory events these
approaches interrupt, more specific and safer thcrapy for rheumatoid arthritis will be
devised. Until then, immune thereapy is likely to rcmain the dominant approach to
Butler, W. T. & Rossen, R. D. (1973). ElTects of corticosteroids on immunity in man. I.
Cascatio, D. A. & Scott, J. L. (1979). Acute leukemia following prolonged cytotoxic agent
Cooperating Clinics Committee of the American Rheumatism Association (1970). A
controlled trial of cyclophosphamide in rheumatoid arthritis. New Eng. 1. Med.. 283,
De Seze, M. & Kahn, M. F. (1974). Immunosuppressive drugs in rheumatoid arthritis: c1inical
results. Adv. Clin. Pharmac.. 6, 89-97.
Fan, P. T., Yu, D. T. Y., Clements, P. J., Fowlston, S., Eisman,J. & Bluestone, R. (1978). ElTect
of corticosteroids on the human immune response: comparison of one and three daily 19
intravenous pulsesofmeth ylprednisolone. J ./ab. clin. Med., 91,625-634
IM M UNOREACTI VE METHODS AN D RHEUMATOID A RTHRITIS 309
Garrett, R. & Paulus, H. (1 980). Co mplicatio ns of intravenou s meth ylpredn isolone (MP-IV)
pul se therapy. Arth. Rh eumat.. in press.
Gott lieb, N. L., Riskin , W., Vidal , A., Hyer, F. & McKinney, C. (1980). Methylprednisolone
(MP) pul se therap y in rheumatoi d a rthritis. C/in. Res., 28, 146A.
G ubne r, R., August, S. & Gi nsbe rg, V. (1951). Th erapeutic suppression of tissue reactivity. 11.
Effect ofaminopteri n in rheumatoid art hritis and psoriasis. Am. 1. med. Sei.. 221, 176-182.
Jiminez Diaz, C,. Lopez Garcia, E., Merchante, A. & Perian es, J. (1951). Treatment of
rheumatoid arthritis with nitrogen nu stard. J. Am. med. Ass.. 147, 1418-1419.
Kahn , M. F., Arlet, J. , Bloch- Michel, H., Ca roit, M., Chao uat, Y. & Reni er , J. C. (197 9).
Leucem ies aigues apres traitem ent par agents cytotoxiq ues en rheumatologie. No uv. Presse
Karsh , J., Wright, D. G. , Klippel , J. H., Decker, J. L., De isseroth, A. B. & Flye, M. W. (1979).
Lymphocyte depl etion by contino us flow cell centrifugation in rheumatoid arthritis. Arth.
Karsh , 1.. Klipp el, J. H.. Plotz, P. H., Wright , D. G ., Flye, W. & Decker , J. L. (1 980).
Lymphapheresis in rhe umatoid arthritis: a rando mized tria l. Arth. Rheumat.. in press.
Kot zin , B. L., Stro ber, S., Calin, A., Hoppe, R. & Kapl an , H. (1980). Th e treatment of
rheuma to id arthritis with total lym ph oid irradiation. Arth . Rheumat.. in press.
Liebl ing, M., Lieb , E., McLau ghlin, K., Bloc ka, K., Furst, D., Nyman, K. & Paulus, H. (1980).
A doubl e-bl ind eross-over trial of meth ylprednisolon e pulse therap y in rheumatoid
a rthritis (RA). Arth. Rheumat.. in press.
Multicen tre Study G ro up. (1 978). Levami sole in rheumatoid ar thritis. Lancet, 2, 1007-1012
Parillo, J. E. & Fau ci, A. S. (1 979). Mec hanisms ofglucocorticoid ac tion on immune processes.
Ann. Rev. Pharmac. Toxicol.. 19, 179-201.
Paulus, H. E., Machl eder. H. 1., Levine, S., Yu , D. T . Y. & MacD on ald , N. S. (1 977).
Lyrnp hocyte invo lvement in rhe umatoid arthritis. Studies du ring thoracic duct dr ainage.
Arth. Rh eumat.. 20, 1249-1262.
Plot z, P. H., Klipp ei, J. H., Decker , J. L.. G rauman . D., Wolff, D., Brown , B. C. & Rutt, G.
(1979). Bladd er complications in pat ient s receiving cyclophos phamide for systemic lupus
erythema tos us or rheumatoid arthritis. Ann. intern. Med., 91,221-223.
Polley, H. F. & Slocumb, C. H. (1976). Behind th e scenes with co rtiso ne and ACTH. Ma yo C/in.
Roth well. R. S., Gordo n, P. A., Davis, P., Dasgupta. M. K., Johny, K. V., Ru ssell , A. S. &
Percy, J. S. (1980). A controlled study of plasm a exc ha nge in the treatme nt of seve re
rheumatoid art hritis. Arth. Rheum at.. in press.
Smyth, C. J., Bartholomew, B. A., Mills, D. M., Steige rwald, J. c., Stro ng, S. J. & Recart, S.
(1 975). Cyclo phos pha mide therapy for rheumatoid art hritis. Arch. l nt. Me d., 135,789-793.
Townes, A. S., Sowa, J. M. & Shulman , L. E. (1 976). Co ntrolled tr ial of cyclophosph amide in
rheumatoid arthritis. Arth . Rh eumat.. 19, 563-573.
Ueo, T ., Tanaka, S., Tominaga, Y., Ogawa, H. & Sukurami , T. (1 979). The effect of thoracic
du ct drainage on Iymphocyte dynamics and cli nical sympto ms in pati ent s with rheumatoid
arthritis. Arth . Rh eumat.. 22, 140 5-1412.
Urowitz, M. B., Hunt er , T., Bookrnan , A. A. M., Go rdo n. D. A., Sm ythe, H. A. & Ogryzlo , M.
A. (1974). Aza thiopri ne in rheu matoid arthritis : a dou ble-blind study comparing full dose
to halfd ose. J. Rheum at.. 1, 274-281.
Wallace, D. L., Gold finger. D., G ratt i, R., Lowe, c., Fan, P., Blueston e, R. & K1inenberg, J. R.
(1979). Plasmapheresis and Iymphopl asmapheresis in the management of rheumatoid
arthritis. Arth. Rheumat.. 22, 703- 710.
Web eI, M. L., Ritts, R. E., Ta swell, H. F., Don ad io, J. V. & Woods, J. E. (1974). Cellular
irnmunity after intravenou s adminstratio n of meth ylpredni solone. J. lab. clin. Med., 83,
Departm ent ofNeurological Su rgery RI-20,
Seattle , Was hingto n 98195, USA
The publication by Melzack & Wall (1965) of a novel concept of the anatomy and
physiology of pain marked the advent of a renaissance in the study of this major
cause of human sufTering. The ph ysiological, anatornical , pharmacological and
psychological substrates ofpain have become th e subjects ofa wide range ofbasic and
c1 inic al research projects, A conceptual framework to integrate information on pa in
will be given. It is ofconcern that each scientist tends to see onl y those aspe cts ofthe
' Puzzle of Pain' which are explicable with in the con ceptual framework oftheir own
scientific discipline; yet, th e problems encountered by the c1inician suggest that the
biomedical model which forms the corne rstone of mod ern scientific endeavour does
not ofTer rat ional solutions to the readil y apparent issues. It has been stated by Engel
(1977) that: '. . . all med icine is in crisis and, further, that medicine's crisis deri ves
from the same basic fault as psychiatry's, narnely, adherence to a model of disease no
longe r adequate for the scientific task s and social responsibil ities of eith er medicine
or psychiatry. The importance ofhow ph ysicians conc eptualize disease deri ves from
how such concepts det ermine what are con sidered the proper boundaries of
profession al respon sibil ity and how they influence att itudcs toward and beh avior
It is essenti al to create a model for the phenomena of pain whic h will pro vide both
c1inician s and basic scientists with a conceptua l fram ework which will appropriately
int egrat e the observations made in several disciplines.
The traditional model of Western Med icine is deri ved from Descartes, who
fractured human existence into 'rnind' and 'body', For hirn, pa in was a reflex
response to a ph ysical stimulus; pred ictable and expli cable if th e stimulus was
known. For us, it is apparent that the ph ysical characteristics ofthe pain stimulus are
often indeterminate. Ju st what is the aetiology ofchronic low back pain - is there any
nox ious stimulus? Secondly, even when the stimulus is more apparent, as in acute
pain, the stimulus alone does not predi ct the indi vidual's response. If you do not
con sider the model one has for an illn ess important, con sider how you would treat a
patient using SI. Augustine's concept: 'All diseases ofChristians are due to demons.'
A more serviceable model can be bu ilt around four concepts: nociception, pain,
suffering and pain behaviour(Figure I). Nociception is defmed here as the detection of
potentially tissue -damaging thermal or mechanical energy by specialized nerve
endings connected to A t5 and C fibre afTerents. A wide variety of anatomical and
ph ysiological studies has conclusively shown that nociception is a specific
peripheral sensory system. It is not due to excessive or patterned activation of larger
afTerent fibres invol ved in light touch or proprioception.
In the intact animal, nociception usually leads to pain, which is defmed here as a
perceived noxious input to the nervous system. The linkage between nociception and
pain can be interrupted by surgical, pharmacological or psychological means;
emphasizing the point that nociception is a peripheral event ; pa in is a feature of the
spinal cord and brain. Not only can we have nociception in the absence of pain , but
there is c1early pain without nociception. Consider the long list of pain states
associated with injuries to the central or peripheral nervous system, often called
'central pain states' (Table 1). Every one ofthese illnesses is a chronic problem which
exists in the absence of any evidence for ongoing nociceptive act ivity. Indeed, in
Table 1 Pain syndromeswithout nociception.
phantom-limb pain not only is there no evidence for tissue damage where the pain is
perceived, but the entire painful body part is absent. A meaningful concept of pain
must include an explanation for those types ofpain unrelated to noxious stimulation.
The absence of such a concept leads to absurd surgical and pharmacological
therapies, based more upon social convention than scientific realities. For example, if
you have the pain of tic douloureux in your nose , no one will advocate surgical
excision of that part of your physiognomy. In contrast, myriads of patients with tic
douloureux experienced in their teeth have those cosmetically and functionally
important structures avulsed, even though there is not a shred of evidence that dental
extractions ever eliminate the pain of tic douloureux. A similar error is made by
prescribing narcotics for chronic usage to patients with tic douloureux: these drugs
are antinociceptive and do not alleviate central pain states unless soporific levels are
attained. The neural mechanisms of central pain states are not understood: it is
tempting to refer to them as 'errors in information processing' or 'short circuits'. It is
essential when evaluating studies of pain to discriminate pa in due to nociception,
from pa in due to something wrong within the nervous system. They are equally real ,
but may have widely disparate anatomical and physiological substrates. Experi-
mental pain stud ies, with a few recent exceptions, have focu sed upon nociception, but
the fmd ings are often extrapolated to the c1inical sett ing where th e pre sence of
nociception is dubious. Cognitive factors can be shown to modify the response to an
experimenter -eontrolled noxious stim ulus, but they may not act in a sim ilar fashion
in c1inical pain problems which are c1early not tightl y linked to noxious events. It is
not at all c1ear that c1assic laboratory sensory physiology studi es ha ve any relevance
Pain usually leads to suffering, which can be defmed as a negative affective
response engendered by pain and by such diverse phenomena as depression,
isolation, an xiet y or fear. People ma y suffer for a variet y ofreasons; Engli sh speakers
tend, however , to use the language of pain to describe all the phenomena ofsujJering.
A book entitled Worlds 0/Pain is, in fact, about the suffering of a poor, im migrant
population in the Chicago stockya rds at the turn of the century. A J ust Meas ure 0/
Pain described penitentiary life in the earl y industrial age. Neither volu me discussed
pain; both a re concerned with suffering. Consider this quotation: ' Art is pain. It must
demand for an anesthetic.' (Levin .Horizon Magazine). Finall y, we have all probably
used the phrase, 'So-and-so is a pain in th e neck' (or elsewhere). We do not mean our
neck hurts, nor that 'so-and-sc' resides in our neck . We do mean th at 'so-and-so' is
ca using us to suffer. Similar lin gui stic examples abound. The way language is utilized
clearly constrai ns th e way both patients and ph ysician s th ink. Our health care
deli ver y syste ms reinforce thi s error: patients who compl ain of pa in a re pr esumed to
ha ve a noxiou s aet iol ogy a nd the pot enti al role of affecti ve factors is ignored. Illich
(1976) sta ted eloquently: 'rn edica l civilizatio n, ho wever, tends to turn pain int o a
technica l matter and ther eb y deprives suffering ofits inh erent person al meaning.'
It is im portant to recogn ize that nociception, pain. and suffering are pri vate,
internal events that cannot be qu antifred or proven to exi st. SujJering usually leads to
pa in behaviour, which is the interaction between the individual and the surro unding
world. Pain behaviour is defmed as an y and all outputs of the individual which
reasonable observers would characterize as suggesting pain, such as , but not limited
to : po sture, facial expression , verbalizing, lying down , taking medicines, seeking
medical assistance or receiving compensation. Like an y other form of behaviour,
pa in beha viours are measurable. Indeed, it is the patient's pa in behaviour wh ich the
ph ysician eva luates in the establishment of diagnosis and treatment outcome. All
behaviours are influenced by the ir consequences, especially ifthey ha ve been present
over time. En vironmental contingencies will alter the likelihood ofthe elaboration of
a specific behaviour. Neuros urgeons and orthopaedists in the USA ha ve long c1aimed
that back surgery was more successful on pri vate patients than on those covered by
workrnan's com pe nsation. Such a recogn ition of beh avioural factors is in no wa y to
exclude psychodynamic explanations of human behaviour, for it is obvious that both
past affect ive experiences and present contingencies will influence behaviour. It is
crucial to acknowledge that events outside of the patient often playa major role in
the genesis ofpa in beha viour.
The ta sk of the clinician is to determine which of the four factors, nociception,
pa in, sujJering and pain beha viour are playing significant roles in the gene sis of the
patient's problem, and then to direct therapies at the appropriate aetiological factors.
The obl igation of the basic scientist is to recognize th at experiments invol ving
portions of the neura l or psychological substra tes of pain do not enco mpass the
uni verse of phenomena rel ated to pain. It is for these reasons th at the differences
between acute, chro nic and experimental pain must be recognized. What they have in
common is very limited. Experimental and acute pain almost exclusivel y involve
nociception. C hron ic pa in usually has little evidence for any nociceptive activity and
is more often related to pa in in the ab sence of nociception or to affective and environ-
mental factors. This is reflected in the frequent meaninglessness of data deri ved from
experimental pain studies when applied to patients with chronic pa in. Indeed, close
scrutiny has led to the viewpoi nt that chronic pain and acute pain ha ve nothing in
com mon save for the four letter word, ' pain'. Therapies, whether pharmacological ,
ph ysical or surgical , which are highly successfu l in contro lling acute pain are not
only in effective in the patient with chro nic pain, but often add to th e patient's pain
behaviour. A significa nt number of chro nic pa in patients, who are surgical veterans ,
professional doctor-shoppers and ma ssive over-ut ilizers of drugs, have been seen.
They rea ch the pain clinic carryin g shoppi ng bags full of medicat ions; ph ysic ians
have become a way -sta tion in the pati ent-prescription relationship. The pain
behaviour of these patients is flor id. Doing nothing other th an helping them to cea se
their medications a nd increase activit y levels has, for many of these unfortunates,
el im inated their pain behaviour. Na rcotics and sedative-hy pnotic drugs (diazepam is
and fall ing blood levels of th ese drugs. Similarly, inactivity and immobilization,
which effectivel y palliate acute pain, only add to chronic pa in. The sign ificant factor
agai n appears to be the prominent role of nocicept ion in acute pain and its absence in
chronic pain and the importance ofaffective and behavioural factors in chronic pain .
lt is important to discriminate between chronic pain due to cancer and that due to
a benign disease process. The pain associ ated with uncontrolled cancer can be
conceiv ed as du e to continuous and often increasing tissue damage; hence, it is best
described as long-standing ac ute pain and sho uld be tre ated with therapies aimed at
reducing nocicept ion . Ben ign di sea ses causing chronic pain are rarel y associated with
nocicept ion and are rarely ameliorated by anti-nociceptive me asures.
Ad van ces in th e therapy of ac ute and chronic pa in are dependent upon th e
achievement of a more profound understanding of'the substra tes of nociception. pain,
suffe ring and pain behav iour. The papers in th is section address differ ent as pects of
the ' Puzz le of Pain' and have been selected to represent some of th e recent major
advances in the study of th e many facets of pain. An atomical , ph ysiological and
pharmacological studies have begun to illumi na te some o fthe bases for chronic pain,
th e placebo response and drug effects. No single methodological approac h is likel y to
expl ain all of th e known phenomena . lt is ob vious that th e br ain is th e organ of the
mind and th e source of all complex behaviours. A lthough we may not understand
th e electroc hemica l processes by which th e brain works, we do know which orga n to
A scheme for interpreting th e many phenomena of pain utilizing the concepts of
nociception, pain, suffering and pa in behaviour is de scribed. This conceptua l
approach offers a framework for the a na lysis ofboth clinical and rese arch studies on
Engel. G. (1 977).The need for a new medical model: A challenge for biomedicine. Science. 196,
IgnatiefT, M. (1 978). A Just Measure of Pain: The Penitentiary in the Industrial Revolution. New
IIIich, I. (1 976). Medical Nemesis. New York: Random House.
Melzack, R. & Wall, P. D. (1 965). Pain mechanisms: A new theory. Science, 150, 971-979.
Ruben, L. (1977). Worlds ofPain : Life in the Working Class. New York: Basic Books.
II. In stitut e 0/ Physiology, University 0/Heidelberg,
Im Ne uenheimer Feld 326, D-6900 Heidelberg, GFR
An appropriate discussion of nociceptive pathways must consider three levels of
processing: peripheral , spinal and supraspinal. A large amount ofneurophysiological
data has accumulated on peripheral and on spinal mechanisms, but thi s has certainly
not led to a complete understanding, our knowledge on supraspinal mechanisms of
The ' puzzle of pain' (Melzack , 1973) becomes even more difficult to deal with
when we concede that ditTerent kind s ofpain states have ditTerent pathophysiological
mechanisms. The sort of pain commonly studied by phy siologists is initiated by the
exc itation of nociceptors. Th e same is true for some pain states wh ich are of clinical
interest. Pain in inflammation for example, is due to the excitation of nociceptors
which have been sensitized by chemical agents. On the other hand, pain attacks in
trigeminal neuralgia, which are triggered by very light mechanica1 stimuli, are
pre sumabl y initiated by the excit ation of sensitive mechanoreceptors with thick
myelin ated Aß atTerents (Lindbiom & Verillo, 1979). The pathophysiological
mechan isms of thi s disease involve an abnormal excitation of central noc iceptive
pathways by non-nociceptive peripheral input. To mention a third case : pain of
patients with herniated disc usuall y will be projected into the periphery, but
ob viou sly has its origin not in the peripheral receptors, but rather in a direct action
on atTerent fibres in nerve roots.
In order to syste mise thc discussion, mainly phy siological mechanisms of
nociccption will be discussed, that is, peripheral and central nervous events induced
by phy sical or chemical noxious influences. At the end of this paper the relationship
ofnociception to pain will be considcred briefly.
Peripheral mechanisms ofnociception
Although very few scientists would now den y the existence ofnociceptors, their study
still involves problems. Morcover, the significance of nociceptors for the initiation
and even more for thc subsistence of pain is qu ite controversial (Wall, 1979).
Nociceptors a re found among the atTerent nerve fibres with slow conduction velocity.
Elcctrical stimulation of cutaneous nerves becomes painful when the stimulus
surpasses the threshold of thin myelinated Ab fibres, Pain becomes stronger when
unmyelinated Cvfibres are recruited. Experiments in which the myelinated fibres
picture of the relation between fibre c1asses and pain. First, the c1ear relationship
between electrical stimulus strength and the diameter of the fibres wh ich are
recruited holds only with direct stimulation of exposed nerve trunks, but not with
transcutaneous stimulation of nerve endings. This fact has often been neglected by
researchers. Furthermore, Willer, Boureau & Albe-Fessard (1978) have shown that in
human subjects even stimuli which elicited only an Aß-volley in the compound
action potential ofthe sural nerve are judged painful and induce nocifensive reflexes
when high frequency repetitive stimulation is used instead of single shocks. These
experiments indicate c1early that temporal summation is involved in nociceptive
processing. They might further suggest that there are some nociceptive elements also
Nociceptors have been studied for many ycars by single fibre recording in animal
experiments (Zotterman, 1939; Iggo, 1959; Bessou & Pearl, 1969; Burgess & Perl,
fibres of conscious human subjects (Hagbarth & Vallbo, 1967). Figure I shows the
experimental arrangement. It can be seen that thick myelinated, thin myelinated, and
unmyelinated ftbres can be recorded with this technique.
Figure1 Percutaneous reeording (ree) of single units from the superficial radial nerve and
electrieal stimulation (stirn) ofthe fibres in their reeeptive fields, Speeimen reeords on the Jeft
show reeordings from a thiek myelinated, a thin myelinated, and an unmyelinated nerve fibre
resp~etively (Unpublished observations from experiments of the author, together with
Adnaensen, H., Gybels,J. & van Hees,J.)
NOCICEPTIVE PATHWAYS AND TRANSMITTERS 319
Studies with this technique of 'rnicroelectroneurography' have shown that in man
virtually all Cvfibre nociceptors belong to a fairly uniform group of ' polymodal
nociceptors' in that they can be excited by thermal, mechanical and chemical
noxious stimulation. Besides this receptor class, specific warm receptors have been
described among the Cvfibre afferents (Konietzny & Hensel, 1979). Sensitive
mechanoreceptors with Csfibres which are found in the cat's hairy skin are
apparently lacking in human skin (van Hees, 1979). In the A O-group there are also
polymodal nociceptors, but in addition high threshold mechanoceptive units, which
cannot be excited by heat (Adriaensen, Gybels, Handwerker & van Hees, 1980). As a
whole, the receptor spectrum in human skin is almost identical to that found in the
monkey, and also highly congruent to that found in the cat.
The main benefit from experiments conducted with conscious human subjects is
the possibility to directly compare fibre discharges and reports ofthe subjects on their
sensations. Thus far , it has been shown that a degree ofsummation ofimpulses from
nociceptors is usually needed to induce pain (Adriaensen et al., 1980). On the other
hand, evidence has been found that the discharge patterns of single C-fIbre
nociceptors discriminate different levels of noxious stimulation as weil as ratings of
the sensation by the subject (Gybels, Handwerker & van Hees, 1979).
A furt her point regarding the nociceptor concept entails two aspects: a) nociceptors
are receptors excited exclusively or mainly by noxious stimuli, that is, by stimuli
which can induce phlogistic reactions; and b) activation of nociceptors causes - due
to their central connections - pain sensations in man and nocifensive reflexes in man
and animals. Many researchers, especially when dealing only with peripheral
nervous events, classify nerve endings as nociceptive only by the first of these two
criteria. This might be misleading, since it is impossible to test the responses of a
receptor to all physical and chemical stimuli possibly occurring in its physiological
environment. This argument is even stronger for receptors in deeper tissues, for
example, in muscles and viscera. Perhaps receptors for metabolic changes are
sometimes mistaken as nociceptors.
Modulator substances in the peripheral nociceptive system
When considering the pharmacology ofthe peripheral nociceptive system, one must
a) Substances produced in inflamed tissue which induce or at least modulate the
b) Substances, which are secreted from nociceptive nerve endings themselves in the
chemical stimulation, possibly resulting from the generation of chemical mediators.
Among them, prostacyclin (PGh) and prostaglandins E (PGE) seem to play a major
role . Other mediators, such as plasmakinins (bradykinin), 5-hydroxytryptamine
(5HT) and histamine, have synergistic effects (Ferreira & Nakamura, 1979;
Handwerker, 1980). In neurophysiological experiments on single peripheral
nociceptive units, it has been shown that nociceptors are directly excited,
for example, by bradykinin and 5HT, and that their response to thermal
stimulation is increased by all of the above mentioned substances and by Car' ions
A cyclic AMP and Ca:" depcndcnt mechanism seerns to be involved in PGE
hyperalgesia, since adenylate cyclasc activators such as catecholamines may increase
hyperalgesia (Ferreira & Nakamura, 1979), a mechanism which might be involved in
some c1inicalsyndromes in which sympathetic disorders are accompanied by pain .
It is a widely accepted hypothesis that peripherally acting 'aspirin-like' analgesics
exert their etTect by inhibiting cyclo-oxygenase, and thus preventing the release of
PGh or PGE from inflamed tissue . These drugs do not seem to block directly the
hyperalgesie action of prostaglandins, or the direct action of noxious stimuli on
Recently a peripheral action of morphine has been demonstrated, which appears
to take place somewhat later in the sequence of biochemical events inducing
hyperalgesia: at a stage that precedes the formation of cyclic AMP. Since naloxone is
a synergist rather than an antagonist in this action of morphine, a new type of
morphine receptor (N receptor) has been assumed at the nociceptive nerve endings
Peripheral endings of slowly conducting atTerent nerve fibres release a vaso-active
substance. Thi s has been proven by the fact that stimulation ofthe atTerentC-fIbres in
the peripheral stump of cut dorsal roots induces vasodilatation in the respective
dermatome and myotome (Hinsey & Gasser, 1930; Hilton & MarshalI, 1980). It is
unclear whether one substance or rather a group of substances is involved in this
phenomenon. The participation of a small peptide, of substance P, has been
suggested, since recent histochemical studies have shown that about 20% of
the cells in the dorsal root ganglia and many nerve endings in the periphery contain
substance P. These cells have small somas and their axons are smalI , probably
unmyelinated or thinly myelinated. Since substance P containing nerve fibres have
also been found among the tooth pulp atTerents, which are considered to mediate
only nocicept ive information, it is rather likely that nociceptors are among the
substance P containing neurones (Olgart, Gazelins, Brodin & Nilsson , 1977; Olgart,
Hökfelt, Nilsson & Pernow, 1977).
Synthetic substance P has been shown to induce vasodilatation. Since, however ,
the rate of secretion from peripheral nerve endings is unclear, it is also unclear
whether substance P is really the crucial agent in the phenomenon of vasodilatation
induced by antidromic activation of atTerents (Nicoll , Schenker & Leeman, 1980;
It is further unclear whether all nociceptors contain this peptide, and whether
nociceptors are the most important group ofthose atTerents which induce the atTerent
Unfortunately, no antagonist to substance P is yet available. It has been shown ,
however, that neonatal pretreatment of rats with capsaicin (50 mg kg-I) induces
extensive axonal degeneration in neurones containing substance P. After this
treatment rats are insensitive to many pain-producing chemical stimuli and also, to
some extent, to noxious mechanical stimulation (Hayes & Tyers, 1980). There are,
however, conflicting reports on the influence ofneonatally adm inistered capsaicin on
heat induced nocifensive reactions: while some authors claim a slight increase in
reaction times in the 'hot plate' and 'tail flick' tests, others deny an etTect on heat
induced nocifensive reflexes (Holzer, Jurna, Gamse & Lembeck 1979; Hayes &
These results may suggest that at least some of the nociceptors do not contain
substance P. On the other hand , there are many substance P containing nerve fibres
in the vagus nerve , which presumably are not all nociceptive, some of which being
perhaps connected with baroreceptors and /or chemoreceptors (Cuello & McQueen,
1980) and with sensitive lung receptors. This assumption is further suggested by the
fact that these types of atTerents are excited by acute capsaicin application (Virius &
NOCICEPTIVE PATHWAYS AND TRANSMITTERS 321
In summary, it is still unclear whether all nociceptors contain substance P, but it is
very likely that so me non-nociceptive primary afferents contain this peptide. Even in
those nociceptors which contain substance P and can release it from their central
endings in the substantia gelatinosa, its role in peripheral nociceptor mechanisms is
still unclear. Close arterial application of substance P has shown that it is not very
effective in exciting nociceptors (Kumazawa & Mizumura, 1979). Nevertheless it
might act as as a modulator of peripheral nociception. This is consistent with the
close association ofsubstance P containing afferent terminals with blood vessels and
with their ultrastructural appearance resembling that of terminals with known
endocrine functions. On the other hand, substance P might be a classical transmitter
in the synapses between primary nociceptive afferents and secondary neurones in the
dorsal horn (Nicoll et al. , 1980).
Processing of nociception in the spinal cord
When regarding the first station of central processing of nociception - the dorsal horn
difficulties ofinterpretation increase. Dorsal horn neurones have two main functions
in the context of nociception: a) to mediate nocifensive segmental autonomic and
motor reflexes, and b) to transfer nociceptive information to the bra in. There are at
least two types of neurone which are candidates for these functions: units in Rexed's
lamina I, the Waldeyer cells, which receive relatively selective nociceptive input
from Ac5 and Cvfibres (Christensen & Perl, 1970; Kumazawa & Perl , 1978) and
another class of neurones, found mainly in lamina V, which receive input from
nociceptors and from sensitive mechanoreceptors with thick myelinated Aß
afferents as weil (convergent neurones).
One possibility to study the significance of any group of dorsal horn neurones for
nociception is to look for their projections into the anterolateral spinothalamic tract,
which is assumed to be the major oligosynaptic pathway carrying nociceptive
information to higher brain levels . Electrophysiological studies have revealed that in
the monkey the spinothalamic tract carries information on wide ranging cutaneous
stimuli including light touch, hair movement, pinch, pin prick and temperature
changes which are noxious or non-noxious (Albe-Fessard, Levante & Lamour, 1974;
Willis, Trevino, Coulter & Maunz, 1974). Morphological studies using tracers
transported with axoplasmatic flow have shown that cell bodies are found in lamina I
as weil as in deeper laminae. Convergent neurones in deeper laminae, however, seem
to contribute the major part ofthe axons of this tract (Trevino & Carstens, 1975).
Another argument speaks in favour of the convergent neurones: they have been
shown to encode the intensity ofnoxious stimulation, and their activity is modulated
by segmental and supraspinal control mechanisms which are known to control pain
reactions (Handwerker, Iggo & Zimmermann, 1975; Carstens, Yokota &
On the other hand, it is very unlikely that nociceptive information is processed in
the CNS by just one functionally uniform class of neurones and by just one single
ascending pathway. Karplus & Kreidl published, as early as 1925, results from
experiments in which the spinal cord of cats was hemisected on both sides, but at
different levels . All long ascending tracts were thus severed . Nevertheless, these
authors found that nocifensive reactions could be elicited from caudal body regions
as long as the distance between the hemisections was at least 4-5 spinal segments.
These results strongly suggest a role for chains of neurones with short axons in the
processing of nociception (Noordenbos, 1959). The existence of parallel channels
processing nociception is also suggested by the fact that neurosurgical lesions of the
anterolateral quadrant (chordotomy) suppress pain only for a limited period. Why
the chordotomy patients are at first pain free and then have aremission of pain after
several weeks or months is still an unsolved question.
Nociceptive transmitters and modulators in the spinal cord
Substance P is highly concentrated in the substantia gelatinosa of the spinal cord.
Since most ofthe terminals from slowly conducting primary afferents end there, one
may draw the conclusion that it is concentrated in these terminals. In fact, it has been
shown that substance P-like immunoreactivity in the dorsal horn dropped markedly
after the dorsal roots were cut (Takahashi & Otsuka, 1975). Iontophoretically applied
substance P has been reported to excite neurones that respond to noxious
cutaneous stimuli in the dorsal horn (Henry, 1976; Zieglgänsberger & Tulloch, 1979)
and in the nucleus caudalis of the trigeminus (Henry, Sessle, Lucier & Hu , 1980).
Thus substance P is a good candidate as the (or at least one) transmitter at the first
synapse of nociceptive pathways. On thc other hand, this peptide has a very slow
onset of action when applied iontophoretically leading some authors to conclude that
it might be a modulator rather than a true transmitter. This slow onset ofaction upon
iontophorctic application, however, might be a consequence of the low transport
number ofthis substance rather than a true pharmacologicaI characteristic.
The occurrence of substance P in the dorsal horn and in nucleus caudalis of the
trigeminal paralleis that of enkephalins, which are believed to be endogenous ligands
of the opiate receptor. In biochemical studies of slices of rat trigem inal nucleus
caudalis and of rat substantia nigra, Jessel & Iversen (1977) have shown that
morphine produces a dose-dependent, stereospecific, naloxone antagonizable
reduction of substance P release in the trigeminal nucleus, but not in the substantia
nigra which does not receive primary afferent terminals. These authors concluded
that enkephalinergic segmental interneurones might inhibit transmission through
substance P synapses between primary and secondary pain processing neurones,
by a presynaptic mechanism. This modern neuropharmacological hypothesis
resembles the 'closing mechanisrn' in the 'gate control' theory of Melzack & Wall
Interestingly terminals containing substance P-like immunoreactivity have been
found also in 5HT-containing fibres descending from the brain stern raphe
nuclei which are thought to inhibit transmission of nociception in the substantia
gelatinosa (Hökfelt, Ljungdahl, Steinbusch, Verhofstad, Nilsson, Brodin, Pernow &
Goldstein. 1978). The functional significance ofthis fmding is still unclear.
Besides serotonergic descending pathways, noradrenergic projections from the
brainstem seem to inhibit processing ofnociception in the dorsal horn (Yaksh, 1979).
Supraspinal mechanisms ofnociception
Nociception in the spinal cord is characterized by parallel processing. At supraspinal
levels this feature is even more pronounced. No single 'pain centre' has been found in
the brain. lnstead, many brain nuclei seem to deal with nociceptive information.
Despite many atternpts, it has not yet been possible to identify an unequivocal
corticaI projection of nociceptive afferents except for the projections of tooth pulps
(Biedenbach, van Hassel & Brown , 1979). Several thalamic nuclei have bcen
implicated in nociceptive processing, but the functional significance of single unit
responses recorded in the thalamus is controversial. In addition, considerable species
differences occur in projections ofnociceptive elements ofthe spinothalamic tract to
different medial and lateral thalamic nuclei (Carstens & Trevino, 1978; Giesler,
Menetrey & Basbaum, 1979). Most subcortical structures which are involved in the
NOCICEPTIVE PATHWAYS AND TRANSMITTERS 323
processing of nociception are involved also in a multitude of other functions, for
example, the nuclei ofthe reticular formation, the monoaminergic nuclei ofthe brain
stern, the periaqueductal grey, limbic structures and the hypothalamus. It is one of
the fmal goals of pain research to fmd out how pain behaviour emerges from the
interaction ofditTerent areas ofthe brain.
In recent years, research has concentrated on pain control systems in the brain
possible role in the processing ofnociception.
Figure 2 summarizes a number of experiments which have becn conducted in the
rat, cat and monkey. Morphine microinjection or electrical stimulation in the ventral
periaqueductal grey matter (PAG) depresses the nociceptive responses of dorsal horn
neurones and supresses pain behaviour (Basbaum & Fields, 1978; Fields & Basbaum,
1978). Electrical stimulation of the medullary nucleus raphe magnus has a similar
etTect (Guilbaud, Oliveras, Giesler & Besson, 1977). It is controversal, whether
morphine mimics the etTect of electrical stimulation at this point too (LeBars,
Dickenson & Besson, 1980). It has been inferred from these fmdings that a descending
projection from the PAG, which might include endorphins activates neurones in the
nucleus raphe magnus directly or indirectly.
Figure2 Schematic summary of experiments on brain stern mechanisms of pain control
includingrat, cat,and monkey. (For furtherexplanation seetext).
These neurones, and those of the adjacent nucleus reticularis magnocellularis, are
in turn the origin of a descending pathway which is partly serotonergic and which
inhibits nociception at the spinal level (Fields & Basbaum, 1978).
The question arises as to the extent to which these midbrain and hindbrain control
centres are included in feedback loops, in that nociceptive and /or non-nociceptive
Raphe-spinal neurones in cat and rat have somatosensory receptive fields which
1980; Guilbaud, Peschanski, Gautron & Binder, 1980). In most cases this input is
recent studies that many neurones in the PAG, and in the enclosed nucleus raphe
dorsalis (DR) also receive somatosensory input in rats anaesthetised with chloralose.
In addition, it has been shown that most ofthe PAG-units are excited, whereas most
ofthe DR-units are inhibited by somatosensory input (Sanders, Klein, Mayer, Heym
& Handwerker, 1980). The functional implication of these somatosensory inputs to
the two best known areas involved in pain control, in the midbrain and medulla, is
still unclear, and is therefore labelIed by question marks in Figure 2. Are negative
feedback loops predominant, whereby non-noxious mechanoceptive input inhibits
nociception, as has been shown at spinal segmental level (Handwerker et al., 1975)?
Or is there a feedback inhibition by which nociceptive input limits itself? Or in
contrast, is positive feedback in the nociceptive system predominant? Several
hypotheses on the biological functions ofpain control systems have been inaugurated
(Basbaum & Fields, 1978; Besson, Chitour, Dickenson & LeBars, 1980), but our
knowledge is still too sparse. Thus it will remainone ofthe major challenges for pain
research in the next decade to fmd out what the major biological functions of the
brain stern pain control centres are .
Since this paper is concerned with nociceptive pathways, the variability of c1inical
pa in states was beyond its scope. As mentioned in the introduction, pain might
originate somewhere in the peripheral or in central nociceptive pathways. The
pathophysiological mechanism of a pa in state has to be considered for optimal
therapy. For example, not all pain states involve activation of nociceptors, and
therefore some types ofdrugs, such as the aspirin-like analgesics which act mainly on
the processes inducing sensitization ofnociceptors, might be less efTectivein them.
Pain is always a complex psychophysiological phenomenon and nociceptive
processing, as described in this paper, represents only one aspect.
The author is grateful to Dr E. Carstens for valuable suggestions and Mrs A. Müller
for carefully typing this manuscript. Experimental work of the author, described in
Comments
Post a Comment
اكتب تعليق حول الموضوع