Exp Brain Res (1987) 66:489-499 Experimental Brain Research 9 Springer-Verlag 1987

Enhancement of the responses of ascending tract cells in the cat spinal cord by acute inflammation of the knee joint

H.-G. Schaible, R.F. Schmidt, and W.D. Willis*

Physiologisches Institut der Universitfit, R6ntgenring 9, D-8700 Wfirzburg, Federal Republic of Germany

Summary. 1. Recordings were made from 16 ascend- ing tract cells in the spinal cords of anaesthetized, spinalized cats before and after an acute arthritis was produced by injection of kaolin and carrageenan into the knee joint. 2. The responses tested routinely were to passive flexion of the knee, an innocuous movement. In some cases, responses to other move- ments were also tested, and changes in background discharge rates were monitored. 3. Control record- ings for a period of 1 h or in 3 cases of 3 h indicated that the responses to flexion were reasonably statio- nary. 4. Four tract cells that initially showed little or no response to flexion of the knee joint developed large responses within 1 to 2 h after inflammation of the joint. 5. Another 9 cells were tested that had responses to flexion of the knee joint prior to inflammation. In 6 cases, inflammation produced enhanced static or transient responses. In 2 cases, the effect of flexion was initially inhibitory or variable, but after inflammation these cells showed large excitatory responses. In the other case, inflammation had no effect. Background discharges were increased by inflammation in 6 of these 9 cells. 6. The effect of inflammation of the knee joint was tested on 3 tract cells that had no clearly defined receptive field in the knee. In 1 case, a response developed to knee flexion after acute inflammation was produced. In the other 2 cases, there were initially responses to knee flexion, but these were unchanged by inflammation. 7. Two of the cells tested had bilateral receptive fields in or around the knee joints. Inflammation of one knee joint enhanced the responses to flexion of the same but not of the contralateral knee in one case but greatly increased the responses to flexion of both

* Permanent address: Marine Biomedical Institute, University of Texas Medical Branch, 200 University Blvd., Galveston, TX 77550, USA

Offprint requests to: R.F. Schmidt (address see above)

knees in the other case. 8. Injections of prostaglan- din (PGE2) caused an enhancement of the responses to knee flexion beyond that caused by inflammation in 5 of 7 cases. One cell whose responses to flexion of the knee were unaffected by inflammation showed inhibitory responses to prostaglandin injections into the inflamed knee joint. 9. The effects of inflamma- tion on the responses of ascending tract cells of the spinal cord appear to serve as a useful neural model of the events responsible for the development of arthritic pain.

Key words: Joint - Pain - Inflammation - Spinal cord - Ascending tracts - Cat

Introduction

In the preceding paper (Schaible et al. 1987), the properties of ascending tract cells of the cat spinal cord responding to mechanical stimulation of the knee joint were described. The neurones sampled appear to be a subset of those previously described that could be activated by electrical stimulation of fine afferent fibres of the posterior articular nerve (Schaible et al. 1986). Some of the characteristics of these neurones are the following: 1)they can be found most readily in spinal segments L5 and L6; 2) they are located in both the dorsal and ventral horns; 3) they can be activated by stimulation of knee joint receptors at both innocuous and noxious intensities; and 4) they receive a convergent input from sensory receptors in the skin~ in muscle or both, as well as from the knee joint.

We speculate that at least some of these neurones are involved in signalling joint pain or in triggering reactions that accompany joint pain (Schaible et al. 1986, 1987). Although the evidence is incomplete,

490

some observations consistent with this hypothesis are that the knee joint inputs appear to be produced by receptors having properties that are appropriate for mediating joint pain (cf., Schaible and Schmidt 1983a, b; Grigg et al. 1986), and the convergent inputs from skin and muscle receptors onto these cells could help account for the tenderness in tissues overlying arthritic joints (Schaible et al. 1987).

The present study examines the role of ascending tract cells in joint pain further by describing the changes in the responses of the neurones to joint movements during the development of acute inflam- mation. The peripheral neural effects of chemically induced inf lammation produced by injections of kaolin and carrageenan into the knee joint cavity have been well characterized by recordings from fine afferent fibres supplying the knee joint (Coggeshall et al. 1983; Schaible and Schmidt 1985). Here, the central effects of acute inf lammation of the knee joint are examined by recordings of the responses of ascending tract cells in the spinal cord before and after injections of kaolin and carrageenan into the knee joint. A prel iminary report of this work has been published elsewhere (Schmidt et al. 1986).

Material and methods

The experiments were done on 16 of the same cats as those reported in the companion paper (Schaible et al. 1987). The animals were first anaesthetized with ketamine (50 mg/kg by intramuscular injection). After insertion of a cannula into the cephalic vein alpha-chloralose was repeatedly injected intraven- ously in doses of 20 mg/kg. Anaesthetic level was judged based on pupil size and a continuous recording of the systemic arterial blood pressure, using a heparinized cannnia in a carotid artery. In the preceding paper are described the techniques used for mainte- nance of the animal, dissection, recording from ascending tract cells identified by antidromic activation, construction of single pass peristimulus time histograms showing the background and evoked activity of the neurones, and the histological reconstruction of recording sites.

The additional technique to be described here is the induction of acute inflammation of the knee joint. This was accomplished by injecting first kaolin (4%, 0.3-0.5 ml) and then 15 min later carrageenan (2%, 0.3 ml) into the knee joint. Each of these substances injected individually can induce an acute inflammation, but the combined use of these two agents appears to cause a greater degree of, inflammation that follows a more consistent time course than does either agent alone (Winter et al. 1962; Brune et al. 1974; Schaible and Schmidt 1985). Flexion and extension movements were repeated for a period of 5 rain following the kaolin and the carrageenan injections in this series of experiments.

The time course of the inflammation caused by kaolin and carrageenan injections into the knee joint has been estimated in two ways. In behavioural experiments, it has been found that cats develop guarding of the joint within about 60 min of injection (unpublished observations). In recordings of the discharges of fine joint afferent fibres, changes in both spontaneous and evoked activity have been found as a consequence of joint inflammation (Coggeshall et al. 1983; Schaible and Schmidt 1985). For these

reasons, we anticipated that we would need to record from a given neurone for an initial control period and then for at least several hours to observe the development of any changes in the responses of central neurones to knee joint inputs during the course of acute inflammation produced by kaolin and carrageenan injections. The long observation time required was feasible in experiments involv- ing recordings from antidromically identified ascending tract cells, since we could check that antidromic activation was still possible at intervals throughout the recording period. The latency of the antidromic action potential, in addition to spike configuration and receptive field properties, all helped to confirm the identity of the neurone. It should be noted that receptive field properties alone were often insufficient for identification of a neurone, since acute inflammation could alter at least some of the receptive field properties of the cell.

In some of the experiments, the effects of injections of prostaglandin (PGE2) into the inflamed knee joint were tested and compared with the effects of control injections of Tyrode solution.

Results

Control experiments

The exper iments to be reported here involved recordings from 16 of the ascending spinal tract cells described in the preceding paper (Schaible et al. 1987) before and after inducing acute arthritis by injecting kaolin and carrageenan into the knee joint cavity. All but 3 of the neurones had an excitatory input from receptors in the knee joint, as demon- strated initially by mechanical stimulation of the joint and as confirmed at the end of the experiment by probing the dissected knee joint. Further confirma- tion of a knee joint receptive field came in many instances from responses during the injection of kaolin and carrageenan due presumably either to the transiently elevated pressure within the joint capsule or to chemical stimulation of joint receptors.

The standard test employed in these experiments was the response to flexion of the knee joint, starting from the midposit ion of the joint's range. This is an innocuous stimulus, and we were interested in the possible enhancement of responses to this stimulus following the development of acute inflammation, since knee joint receptors supplied by fine afferent fibres have been shown to demonstrate greatly enhanced responses to this form of stimulation fol- lowing joint inf lammation (Grigg et al. 1986; Schai- ble and Schmidt 1985). Such responses are in contrast to those of fine afferents in normal joints, many of which do not respond to innocuous joint movements (Schaible and Schmidt 1983b). In addition, we moni- tored background firing rates of the neurones, and we sometimes tested the responses of the neurones to other stimuli.

Since the effects of acute inf lammation of the knee joint were manifested over a prolonged time

491

Imp/response

200

100 ! .~ I .,,

L 4 ~ . ~ .v ~.r.. ~i . / \ 0 V \\ .~ b"

t-, "~ ix

'v~ / " , A

\ I I

I I

II I

i i i i I ~) 410 I 810 1'20 160 2(30 min

-100

-200

Fig. 1. Control experiments demonstrating that the responses of 3 different ascending tract cells to flexion of the knee joint did not increase over a 3 h recording period. The responses were mea- sured as the change from background discharge frequency during passive flexion of the knee for periodes of 30 s

course (latency of about i to 2 h), it was necessary to ensure that the responses of the ascending tract cells were stable before inducing inflammation. In most experiments, this was done by recording the activity of the cell under investigation for at least one hour (after initial characterization of the receptive field properties) before injecting the knee joint with kaolin and carrageenan. However, to be sure that the activity of the cells could be stable for the entire duration of most of the experiments, we followed the responses of three cells for 3 h before inducing inflammation. The graph in Fig. 1 shows the responses of the 3 cells to knee flexion (plotted as the difference between the activity during flexion and the background activity during the 30 s period prior to flexion). It can be seen that the responses of 2 of these cells were relatively stable over the 3 h period. In one case, there seemed to be a tendency for the response to become inhibitory over time, although there was a small excitatory effect initially and again just before the injection of kaolin. None of the cells showed any increase in background activity over the 3 h of the control period.

The neurones whose control responses are plot- ted in Fig. 1 seemed to be representative of the population of neurones with knee joint receptive fields that were found in the dorsal horn. Two were classified as "sdj" cells, since they had receptive fields in the skin and in deep structures of the limb, as well as in the knee joint. The other neurone was a "dj" cell, having both a deep and a knee joint

receptive field. All three cells responded to move- ments of the knee joint within the normal working range, and all had focal receptive fields to probing the knee joint through the skin. When kaolin was injected, all 3 cells showed a response during the injection (although insertion of the needle by itself did not cause a response).

Effect of inflammation on cells lacking background activity and responses to normal movements of the knee joint

The most dramatic effects of inflammation of the knee joint were produced in experiments in which the ascending tract cells had no background activity and in which there were no or only weak responses to innocuous joint movements during the control period. These conditions held for 3 cells, all of which were located in laminae VII or VIII. The cells were classified as "dj" neurones, since they had a conver- gent input from muscle as well as from the knee joint. Responses to probing the knee joint showed that the receptors had moderate to high thresholds. Innocu- ous joint movements produced either no response or in one case only a few impulses. One of the cells had no receptive field in the ipsilateral knee joint but did have a receptive field in the contralateral knee joint. This neurone was activated by ipsilateral knee joint movements (presumably because of activation of deep receptors, probably in muscle), but not by contralateral movements. In all three cases, insertion of a needle into the knee joint caused a discharge, and injection of kaolin caused additional activity.

Figure 2 shows the responses of one of these neurones to knee joint movement before and after the injection of kaolin and carrageenan into the knee joint. Figure 2A depicts the receptive field distribu- tion of this "dj" neurone. In Fig. 2B, the histogram indicates that the cell[ had practically no response to innocuous movements of the knee joint, such as flexion, outward rotation, or inward rotation, although there were responses to forced outward and inward rotation. In Fig. 2C, the segments of histo- gram at the left show again that during the control period there was little or no response to an innocuous flexion movement of the knee. Following the injec- tion of kaolin and carrageenan, flexion movements began to evoke a response by 92 min after the injection, and this response increased dramatically until very large responses were observed at 289 min after the injection.

Figure 3 shows the enhancement of the responses of this neurone and of two others to knee flexion before and after acute inflammation of the knee joint

492

C Imp/s

150

A B Imp/s

...... ::f L knee/~ int ju 0 " '

Flex. OR/n.OR IR/n.IR

lOO

50

o control

Imp/response

7OO

60O

500

400

3OO

2o0

lO0

0

Kaolin

Flexion

Injection of 30s Kaolin and

Carrageenan L

1 ,1 L 68 92 122 126 134 161 219 287

minutes after injection of Kaolin

lilt/Ill

20 60 100 140 180 220 260 minutes after injection of Kaolin

289

A

B

60

30

0

60

30

0

Imp/s

"a..r Flex. Inflamm.

- - 30s Control

n_.j- "L.J- " l . J -

Imp/s

q r- q -u - q -u - ~

40 42 44 160 165 minutes after injection of Kaolin

Control

-t_r Flex. Inflamm,

-- 30s

" t - J - "L- / - - t - J - - t _ t -

67 72 -t--t- "t--r"

202 207 minutes after injection of Kaolin

Fig. 2A-C. Responses to flexion induced by acute inflammation of the knee joint. The receptive fields of a "dj" cell in the knee and muscle are drawn in A. The single pass peristimulus time histogram in B shows that the cell initially had little or no response to innocuous movements of the knee, such as flexion (Flex.), outward rotation (OR) or inward rotation (IR), although there were responses to noxious movements, such as noxious outward and inward rotation in.OR and n.IR). The time course of development of responses to flexion movements is shown in C. The arrow indicates the time of injection of kaolin and carrageenan into the capsule of the knee joint

Fig. 3. The time course over which responses to knee flexion developed following inflammation is plotted for 4 cells. Three of the cells had no initial background activity (including the one illustrated in Fig. 2, shown here by the filled circles), whereas the other did. The small arrows indicate the time of the injection of earrageenan in the experiments

Fig. 4A, B. Increases in responses to knee flexion induced by acute inflammation of the knee joint. The effects of inflammation are shown for two different cells. In A, inflammation increased both the phasic and the static components of the responses, whereas in B just the phasic components were enhanced. However, the background discharges of the cell illustrated in B were also increased

Imp/response

500

400

300

200

100

[1

Ik.

Kaolin

A

,a3

2'0 60 100 140 180 220 260 minutes after injection of Kaolin

Imp/response B 500-

200-

1 too

o t 20 60 100 140 180 220 260

Kaolin minutes after injection of Kaolin

Fig. 5A, B. Enhancement of the responses to knee by acute inflammation of the knee joint. The effects of acute inflammation are shown for 6 ceils. In 5 cases, the responses to flexions were increased, whereas in 1 case they were not. In A is plotted the time course of the changes, whereas in B are shown the mean responses before and after inflammation (+ 1 S.D.). The arrows in A indicate the time of the injection of carrageenan

(filled triangles, open and filled circles). None of the cells had an appreciable response before injection of kaolin and carrageenan into the knee joint. One of the cells developed a response by 55 min after the injection, and the other 2 cells (including the one illustrated in Fig. 2) had a response by 120 min. The response amplitudes may not have reached a plateau at the times of the last tests.

In all three cases, following inflammation, a response developed to innocuous extension of the knee where none had been observed during the control period. Similarly, in 2 cases inflammation induced a response to innocuous outward rotation of the knee.

493

Effect of inflammation on cell with background activity but without a response to innocuous joint movements

Another neurone in lamina VII had some initial background activity but little or no response to innocuous movements of the knee joint. This cell was classified as a "sdj" neurone. The receptive field demonstrated by probing the knee joint had a high threshold, and the cell could be excited by noxious joint movements. It was also excited by the injection of kaolin into the knee joint.

The development of a response to flexion move- ments of the knee after injection of kaolin and carrageenan is plotted in Fig. 3 (open squares). A clear response to flexion was seen by 57 min after injection.

Effect of inflammation on cells with background activity and responses to innocuous joint movements

The effects of inflammation were tested on 9 cells that initially had background activity and responses to innocuous movements of the knee joint. Three of these cells have already been described in relation to Fig. 1. Most of the 9 cells appeared to be located in or near lamina V but 2 were in lamina VIII. The classification of 7 of the cells was "sdj"; one was a "sj" cell; and the other was a "dj" cell. The thresholds for local mechanical stimulation of the knee joint varied from low (5 cells) to moderate (3 cells) to high (1 cell). All were excited by insertion of the injection needle into the knee joint and/or by the injection itself. All of the cells responded to innocu- ous movements of the knee; in some cases, there was both an excitatory and an inhibitory response.

In 5 cases, there was a distinct increase in the responses to flexion of the knee following the induc- tion of acute inflammation. This is illustrated in Fig. 4A for one cell. Prior to inflammation, each episode of knee flexion evoked a transiently increased discharge of the cell. Following injection of kaolin and carrageenan into the knee joint, the responses to flexion remained constant initially, but they were dearly increased by 160 min after the injection. It is noteworthy that the responses now included not only a transient discharge at the onset of flexion but also a static discharge during maintained flexion.

In a sixth case, the total number of discharges evoked by knee flexion was unchanged by inflamma- tion, but the sizes of the transient discharges pro- duced by flexion were increased. Furthermore, the

150-

100-

50-

Kaolin

Imp/s A

_r-t. Ext.

-t_r Flex. 30s

494

. t ' - L J - - L . / "1 . "l--r" "t--r" - L J " -L-r"

Control movements

Inflam

" l _ r - "1...I- ...r-L

180 185 188 minutes after injection of Kaolin

Imp/response

1400 I

1200

1000

800[

600

400

200

0

-200

-400 w

20 60 100 140 180 220 minutes after injection of Kaolin

B

Fig. 6. A Conversion of an inhibitory response to an excitatory response following acute inflammation. The neurone whose responses are shown was initially inhibited by knee flexion and excited by knee extension. Following injection of kaolin and carrageenan, the cell was excited by knee flexion, and the response to extension was enhanced. B The time course over which initially inhibitory or mixed responses were converted to excitatory responses to knee flexion is shown for 2 cells (including the 1 illustrated in A, indicated here by the filled circles). The arrows indicate the time of the injection of carrageenan

level of background activity was considerably enhanced. These changes are shown in Fig. 4B.

For another cell, there was a response to knee flexion that depended in part on mechanical stimula- tion of the cutaneous receptive field on the leg. The response to flexion was unchanged' following inflam- mation in this case.

The time course of the changes in responses to knee flexion produced by inflammation for 6 neurones are shown in Fig. 5A. Five of the cells showed increased responses within 120 min after injection. One cell showed little or no change in response (open squares). The graph in Fig. 5B gives a comparison of the control responses and the last responses after the induction of inflammation for the same 6 neurones. The vertical lines indicate one standard deviation. The responses were significantly

enhanced in 5 cases but were unchanged in the other case.

The other 2 neurones showed inhibition or a variable response (either excitation or inhibition on different trials) before inflammation. After inflam- mation, these two neurones had large excitatory responses to knee flexion.

Figure 6A illustrates a case in which the initial response to knee flexion was an inhibition of back- ground activity. Knee extension, on the other hand, produced excitation. After injection of kaolin and carrageenan into the knee joint, the responses to knee flexion changed from inhibitory to excitatory within 100 min. The responses to knee extension were enhanced by inflammation.

The time course of the changes in the responses of this neurone to knee flexion are graphed in

A

C

60

40

20

20

10

Imp/s

I// left knc 1 inflame

B

300

200

100

0

D

120 -

80-

40-

Imp/s

r . l .

-L_r" ~ -L J - " l_ l -

Control

Flex. - 30s

right knee inf lamed I r.

r. I . I~

,,] b,

Control

0

I m p / r e s p ~

i 120 160 200 aoo Kaolin minutes after Kaolin

Imp/response 2

0 le f t leg

40 80 120 t60 200 Kaolin minutes a f ter Kaolin

495

Fig. 7A-D. Effects of acute inflamma- tion of 1 knee joint on the responses to flexion of both knees. For the cell whose responses are illustrated in A, B, inflam- mation of 1 knee joint increased the response to flexion of that knee but had little effect on the responses to flexion of the eontralateral knee. However, for the cell whose responses are shown in C, D, inflammation of the right knee induced response to flexion in this joint and enhanced the response in the left knee

Fig. 6B (filled circles). In addition, the changes in the responses of another neurone to knee flexion are shown to change from a variably excitatory or inhibitory action to excitation following inflamma- tion (open circles).

The effects of inflammation on the background discharges of these cells were variable. In 6 cases, the background discharges were increased following inflammation (e.g. Fig. 4B), whereas in the other 3 cases the background activity was not changed.

Effects of inflammation on cells lacking clearly defined receptive fields in the knee joint

One cell located near the border between laminae V and VI was classified as a "d" cell because of its response to squeezing the quadriceps muscle and to pressure on the tibia and on the foot. Probing the knee joint had no effect. Insertion of the injection needle also had no effect, but the injection of kaolin did evoke a few discharges. No receptive field was found in the dissected knee joint at the conclusion of the experiment. The neurone only responded to noxious movements of the knee prior to inflamma- tion, but after inflammation it developed clear responses to flexion movements of the knee, and squeezing the lateral aspects of the knee now evoked activity.

Another neurone that appeared to be located in lamina VI was classified as an "sd" cell on the basis of an excitatory input from the skin and a deep input

from muscles in the thigh and leg. No receptive field could be demonstrated in the knee joint by probing, either before inflammation or after dissection of the knee joint at the end of the experiment. Nor were there any responses to injection. The cell responded to innocuous movements of the knee. Inflammation did not change the responses to knee flexion.

Finally, a third neurone located in the ventral horn (depth of 3.63 mm) was classified as a "d" cell because of its responses to compression of the quadriceps muscle. There was no apparent receptive field in the knee joint as judged by negative results of probing (before inflammation and after dissection) and injection of the joint. The responses of the cell to knee flexion were unchanged after inflammation. The inhibitory effect of injections of prostaglandin into the knee joint on this cell will be described later.

Effects of inflammation on cells with bilateral receptive fields

The responses of one of the neurones described previously to flexion of each knee joint were tested before and after inflammation. The cell was classified as a "sdj" cell, and it was located in lamina VIII. There was a receptive field in the left knee joint but not in the right knee joint. Flexion of the left knee evoked small responses before inflammation, whereas flexion of t]he right knee had no excitatory action. After inflammation of the left knee, the responses to flexion of the left knee were greatly

496

A

2400

2000

1600

1200-

800 -

400 -

0

Tyrode PGE 2 0.3pg

mO\ %

PGEz 3pg

C

B

200

100-

300 9

200-

100"

O-

Imp/s

PGE 2 0.3 IJg

-L.a-

Imp/s ~_r Flex.

- 30s

PGE z 3pg

1 J

"1..3"

Fig. 8A-C. Action of prostaglandin (PGE2) on the responses to knee flexion after the development of acute arthritis. The graph in A shows that the responses to knee flexion (open circles) and the background activity of the cell (filled circles) were only slightly affected by injection of Tyrode solution into the knee joint. However, injections of 0.3 or 3 ~tg of PGE2 caused greatly increased responses. Examples of these changes are shown in the single pass poststimulus time histograms in B, C. Note that the injection of PGE2 itself caused discharges of the neurone

enhanced, starting within 120 min, but no response developed to flexion of the right knee. This neurone can be regarded as a control for the cells whose responses will now be described.

Two neurones responded to inputs from both sides of the body. One of these cells was located in lamina V, and the other was in lamina VII. It was of interest to examine the effects of inflammation of one knee, joint on these cells, one of which had a receptive field in the contralateral knee joint.

The cell in lamina V was classified as a "sdj" cell. It had a local receptive field in the left knee joint, but not in the right knee joint. Innocuous movements of the left knee produced responses, but noxious move- ments of the right knee were required before the cell discharged. As shown in Fig. 7A, B, inflammation produced an enhancement of the responses to flexion of the left knee (and also to rotations of the left knee; not illustrated), but there was little or no change in the responses to flexion (or rotation) of the right knee.

A different result was obtained for another cell with a bilateral receptive field. This cell in lamina VII was a "dj" cell. It had a deep receptive field in the quadriceps muscle of the left leg and a receptive field in the knee joint on the right side. The cell responded initially to flexion of the left knee but not of the right knee. After inflammation of the right knee, the flexion in the right knee induced discharges and the flexion responses in the left knee were enhanced as shown in Fig. 7C, D.

Effects of prostaglandin injections

After inflammation was induced, the effects of intraarticular injections of prostaglandin (PGE2) were tested on the responses of 8 cells to knee flexion. Injections of an equivalent volume of Tyrode solution served as a control. The Tyrode solution itself had an effect in at least 7 cases. However, in 5 cases, the PGE2 injections enhanced the responses to knee flexion more than did the Tyrode solution; in the other two cases, injections of Tyrode solution had about the same effect as did PGE2 injections.

An example of the action of injections of Tyrode solution and of PGE2 are shown in Fig. 8. The histogram segments in Fig. 8B, C show the effect of flexion after inflammation (left-most responses), fol- lowed by responses to injection of PGE2 and the enhanced responses to knee flexion after PGE2 injection. Two doses of PGE2 were used, 0.3 ~tg and 3.0 ~g. The graph in Fig. 8A shows the responses to flexion before injection of Tyrode solution and the enhanced responses to flexion after Tyrode solution and after 0.3 ~g and 3.0 ~g of PGE2 (open circles). The background discharge rate of the cell is also shown before and after these various treatments (filled circles).

One cell developed an inhibitory response to knee flexion after injections of PGE2. This neurone was mentioned earlier in reference to cells lacking receptive fields in the knee joint. Evidently, the cell had a latent inhibitory receptive field in the knee joint that was made overt when PGE2 was injected.

497

Discussion

The injection of kaolin and carrageenan into the knee joint produced changes in the responses of most of the neurones investigated to innocuous move- ments of the knee joint. For 4 cells, no responses to flexion of the knee were seen initially, but responses developed after injection. For 6 other cells, the initial responses to flexion were enhanced. In two instances, inhibitory or mixed responses to knee flexion became purely excitatory responses. Kaolin and carrageenan injections were effective for 12 of 13 cells that had a demonstrable mechanical receptive field in the knee joint. When no such receptive field could be demonstrated, injection of the joint pro- duced enhancement of the responses to knee flexion in only 1 of 3 cases.

We do not believe that these increases in responses to knee flexion were due to nonstationar- ity. None of the cells showed a reduction in responses over time, as might be expected if we were recording from a population of cells whose responses were changing with time. For example, if the preparations had been deteriorating, presumably at least some of the responses would have diminished. Furthermore, in 3 control experiments we followed the responses of neurones for 3 h before injecting the knee joint. In none of these cases did the response to knee flexions increase during the control recording period. Simi- larly, there was no trend for the control responses recorded during at least 1 h before the other injec- tions to increase.

It was particularly interesting that several of the neurones had no initial responses to flexion of the knee joint and yet they developed such responses after inflammation. This behaviour has a parallel in the response properties of joint receptors with fine afferent fibres. Many of the fine afferent fibres innervating the normal knee joint fail to respond to innocuous movements of the knee, but comparable fibres do respond following injection of kaolin and carrageenan into the joint; in fact, some fine joint afferent fibres do not even respond to noxious joint movements until the joint is inflamed (Coggeshall et al. 1983; Schaible and Schmidt 1985). Presumably, this change in the response properties of knee joint receptors underlies most of the changes observed in this study.

Various phenomena of sensitization by inflamma- tion have also been found for the responses of thalamic and cortical neurones to somatic stimuli in the polyarthritic rat (Gautron and Guilbaud 1982; Lamour et al. 1983; Kayser and Guilbaud 1984; Guilbaud 1985). In the same model superficially located spinal dorsal horn cells show sensitization to

stimulation of inflamed skin areas (Menetrey and Besson 1982).

In one of the cases in which there was a bilateral receptive field, injection of the left knee caused an enhancement of the responses to flexion of the left but not of the right knee. This observation is consistent with the idea that sensitization of knee joint receptors can account for most of the changes in responses. However, the other experiment involving a cell with a bilateral receptive field suggests that a central mechanism may be at work as well. In this case, injection of the right knee induced reactions to flexion in the left knee. Receptors could have been sensitized only in the right knee joint, and so the enhanced responses to flexion of the left knee must have been due to an increased excitability of the cell or of interneuronal circuits presynaptic to the cell (cf. Price et al. 1978; Kenshalo et al. 1982). Evidence for a central contribution to enhanced responses to somatic stimuli during acute inflammation or after injury has recently been reported for thalamic neurones (Benoist et al. 1985) and motoneurones (Woolf 1983).

The prolonged time course characteristic of the development of acute arthritis of the knee joint and of the enhancement of the responses of spinal cord ascending tract cells to innocuous flexion movements following injections of kaolin and carrageenan into the knee joint parallels that described for the development of edema in the rat's foot after car- rageenan injection (Crunkhorn and Meacock 1971; Di Rosa et al. 1971; Garcia Leme et al. 1973) and of signs of joint inflammation after injection of urate crystals (Faires and McCarty 1962; Rosenthale et al. 1966; Okuda et al. 1984). We suggest that a similar process occurs in arthritis due to pathological proces- ses and that increased responsiveness of ascending tract cells in the spinal cord can help account for the hyperalgesia and other responses associated with arthritic pain.

Sensitization of ]knee joint receptors following injection of kaolin and carrageenan into the joint capsule is presumed to be a result of the release of active substances into the joint cavity as part of the process of inflammation (Schaible and Schmidt 1985). The mechanism might be similar to that of the inflammatory response to carrageenan injected into the rat foot (Winter et al. 1962; Crunkhorn and Meacock 197!) or urate crystals injected into a joint cavity (Faires and McCarty 1962; Rosenthale et al. 1966; Brune et al. 1974; Schumacher et al. 1974; Okuda et al. 1984). It would appear that the inflam- matory responses to these manipulations result from the release of chemical mediators, including biogenic amines, such as histamine and serotonin, kinins, and

498

prostaglandins (Willis 1969; Van Arman et al. 1970; Di Rosa et al. 1971; Garcia Leme et al. 1973; Lewis et al. 1975; Higgs and Salmon 1979; Holsapple et al. 1980).

Consistent with a role of prostaglandins are the observations in the present study of the effects of injections of prostaglandin (PGE2) into the inflamed knee joint. PGE2 injections caused a further enhancement of the responses of the cells to knee flexion. A reasonable working hypothesis is that the release of endogenous prostaglandins by the inflam- matory process contributes to the sensitization of joint receptors to mechanical stimuli. One experi- mental approach to test this hypothesis would be to determine the effects of prostaglandin synthesis inhibitors on the ability of injections of kaolin and carrageenan to enhance the responses of spinal cord neurones to innocuous joint movements. Further work is also needed to assess the role of kinins and of biogenic amines in the development of the acute arthritis induced by injections of kaolin and car- rageenan into the knee joint. Judging from the time course of the neuronal changes in the spinal cord after the injection of kaolin and carrageenan a significant role of the mediators of the first inflam- matory phase (histamin and serotonin) seems to be unlikely. But it has to be kept in mind that central neurones do not necessarily reflect the activity in single afferent units. In fact, the time course of sensitization in single afferent units varies consider- ably (unpublished observations).

We believe that these experiments provide evi- dence useful for furthering our understanding of the mechanisms of arthritic pain. It remains for future experiments to determine if the pathways involved in transmitting sensory information to the thalamus and cerebral cortex operate in a different manner from those responsible for triggering suprasegmental reac- tions to painful stimuli. For example, we presume that ascending tract neurones that respond to bilat- eral inputs from the knee are more likely to be involved in nociceptive reactions than in signalling joint pain, since arthritic pain can be localized to a particular joint, at least in humans.

Acknowledgements. The authors thank Annelie Pfeffer for her expert technical assistance, Christiane Jansen for help with the histology, and Margrit D. Derrick for typing the manuscript. The work was supported by the Deutsche Forschungsgemeinschaft. Dr. Willis was the recipient of a fellowship from the Florence and Marie Hall Endowment for Programs of Excellence in Education in the Medical Sciences and of an Alexander von Humboldt Senior U.S. Scientist Award.

References

Benoist JM, Kayser V, Gautron M, Gullbaud G (1985) Changes in responses of ventrobasal thalamic neurons during car- rageenan-induced inflammation in the rat. In: Fields HL, Dubner R, Cervero F (eds) Advances in pain research and therapy, Vol 9. Raven Press, New York, pp 295-303

Brune K, Walz D, Bucher K (1974) The avian microcrystal arthritis. I. Simultaneous recording of nocieeption and tem- perature effect in the inflamed joint. Agents Actions 4:21-33

Coggeshall RE, Hong KAP, Langford LA, Schaible H-G, Schmidt RF (1983) Discharge characteristics of fine medial articular afferents at rest and during passive movements of inflamed knee joints. Brain Res 272:185-188

Cnmkhorn P, Meacock SCR (1972) Mediators of the inflamma- tion induced m the rat paw by carrageenin. Br J Pharmacol 42:392-402

Di Rosa M, Giroud JP, Willoughby DA (1971) Studies of the mediators of the acute inflammatory response induced in rats in different sites by carrageenan and turpentine. J Pathol 104: 15-29

Faires JS, McCarty DJ (1962) Acute arthritis in man and dog after intrasynovial injection of sodium urate crystals. Lancet 2: 682-684

Garcia Leme J, Hamamura L, Leite MP, Rocha e Silva M (1973) Pharmacological analysis of the acute inflammatory process induced in the rat's paw by local injection of carrageenin and by heating. Br J Pharmacol 48:88-96

Gautron M, Guilbaud G (1982) Somatic responses of ventrobasal thalamic neurons in polyarthritic rats. Brain Res 237:459-471

Grigg P, Schaible H-G, Schmidt RF (1986) Mechanical sensitivity of group III and IV afferents from posterior articular nerve in normal and inflamed cat knee. J Neurophysiol 55:635-643

Gullbaud G (1985) Thalamic nociceptive systems. Philos Trans R Soc London B 308:339-345

Higgs GA, Salmon JA (1979) Cyclo-oxygenase products in car- rageeuln-induced inflammation. Prostaglandins 17:737-746

Holsapple MP, Schnur M, Yim GKW (1980) Pharmacological modulation of edema mediated by prostaglandin, serotonin and histamine. Agents Actions 10:368-373

Kayser V, Guilbaud G (1984) Further evidence for changes in the responsiveness of somatosensory neurons in arthritic rats: a study in the posterior intralaminar region of the thalamus. Brain Res 323:144-147

Kenshalo DR Jr, Leonard RB, Chung JM, Willis WD (1982) Facilitation of the responses of primate spinothalamic cells to cold and to tactile stimuli by noxious heating of the skin. Pain 12:141-152

Lamour Y, Guilbaud G, Willer JC (1983) Altered properties and laminar distribution of neuronal responses to peripheral stimulation in the Sm I. Cortex of the arthritic rat. Brain Res 273:183-187

Lewis AJ, Nelson DJ, Sugrue MF (1975) On the ability of prostaglandin E1 and arachidonic acid to modulate experi- mentally induced oedema in the rat paw. Br J Pharmacol 55: 51-56

Menetrey D, Besson JM (1982) Electrophysiological characteris- tics of dorsal horn cells in rats with cutaneous inflammation resulting from chronic arthritis. Pain 13:343-364

Okuda K, Nakahama H, Miyakawa H, Shima K (1984) Arthritis induced in cat by sodium urate: a possible animal model for tonic pain. Pain 18:287-297

Price DD, Hayes RL, Ruda MA, Dubner R (1978) Spatial and temporal transformations of input to spinothalamic tract neurons and their relation to somatic sensations. J Neurophy- siol 41:933-947

499

Rosenthale ME, Kassarich J, Schneider F (1966) Effect of anti- inflammatory agents on acute experimental synovitis in dogs. Proc Soc Exp Biol Med 122:693-696

Schaible H~G, Schmidt RF (1983a) Activation of groups III and IV sensory units in medial articular nerve by local mechanical stimulation of knee joint. J Neurophysiol 49:35-44

Schaible H-G, Schmidt RF (1983b) Responses of fine medial articular nerve afferents to passive movements of knee joint. J Neurophysiol 49:1118-1126

Schaible H-G, Schmidt RF (1985) Effects of an experimental arthritis on the sensory properties of fine articular afferent units. J Neurophysiol 54:1109-1122

Schaible H-G, Schmidt RF, Willis WD (1986) Responses of spinal cord neurones to stimulation of articular afferent fibres in the cat. J Physiol 372:575-593

Schaible H-G, Schmidt RF, Willis WD (1987) Convergent inputs from articular, cutaneous and muscle receptors onto ascend- ing tract cells in the cat spinal cord. Exp Brain Res 66: 479-488

Schmidt RF, Schalble H-G, Willis WD (1986) Enhancement of activity of ascending spinal cord tract cells by acute inflamma- tion of the knee joint in the cat. Proc Int U Physiol Sc XVI: 207

Schumacher HR, Phelps P, Agudeto CA (1974) Urate crystal induced inflammation in dog joints: sequence of synovial changes. J Rheumatol 1:102-113

Van Arman CG, Carlson RP, Risley EA, Thomas RH, Nuss GW (1970) Inhibitory effects on indomethacin, aspirin and certain other drugs on inflammations induced in rat and dog by carrageenan, sodium urate and ellagic acid. J Pharmacol Exp Ther 175:459-468

Willis AL (1969) Release of histamine, kinin and prostaglandins during carrageenin-induced inflammation in the rat. In: Montegazza P, Horton EW (eds) Prostaglandins, peptides and amines. Academic Press, London, pp 31-38

Winter CA, Risley EA, Nuss GW (1962) Carrageenin-induced edema in hind paw of the rat as an assay for antiinflammatory drugs. Proc Soc Exp Biol Med 111:544-547

Woolf CJ (1983) Evidence for a central component of post-injury pain hypersensitivity. Nature 306:686-688

Received June 30, 1986 / Accepted October 27, 1986