Spinal lamina I projection neurons in the rat: Collateral innervation of parabrachial area and thalamus

Neuroscience

Printed in

Vol. 28, No. Britain

I, pp. 21-31,

0306-4522/8953.00+ 0.00

1989

Pergamon

Great

Press

plc

IBRO

SPINAL LAMINA I PROJECTION NEURONS IN THE RAT: COLLATERAL INNERVATION OF PARABRACHIAL AREA AND THALAMUS J. L. K. HYLDEN,* F. ANTON and R. L. NAHIN Neurobiology and Anesthesiology Branch, National Institute of Dental Research, National Institutes of Health, Bethesda, MD 20892, U.S.A. Abstract-A major ascending nociceptive pathway from spinal lamina I to the mesencephalon has previously been reported in the cat, rat and monkey. In the present paper, we have used single and double retrograde labeling techniques to describe this projection system and its collateralization to the thalamus in the rat. Injections of wheat germ agglutinin-horseradish peroxidase into the pontomesencephalic parabrachial area labeled cell bodies bilaterally in lamina I and deeper laminae of the spinal cord. Bilateral lesions of the dorsolateral funiculi at thoracic levels reduced labeling of lamina I neurons caudal to the lesions. Combined injections of fluorescent retrograde tracers into the lateral thalamus and parabrachial area resulted in double labeling of projection neurons in lamina I, lamina IV-VIII and the lateral spinal nucleus of the cervical and lumbar enlargements. Double-labeled neurons were especially abundant in lamina I. Thus, we have demonstrated a major lamina I projection through the dorsolateral funiculi to the parabrachial area with significant colateralization to the thalamus. Moreover, since more than 80% of retrogradely labeled lamina I spinothalamic tract cells had collaterals to the parabrachial area we have indirectly demonstrated the presence of a dorsolateral funicular pathway for lamina I spinothalamic neurons in the rat. More lamina 1 neurons were retrogradely labeled from midbrain injections as compared to thalamic injections. The significance of these findings rest on previous work in this and other laboratories and concerns the understanding of spinal nociceptive mechanisms. Lamina I projection neurons are primarily nociceptivespecific in their response properties and have been shown to project to both the midbrain and thalamus via the dorsolateral funiculus in a number of species. The role of this projection system in nociceptive transmission may lie in its ability to distribute precise information to multiple brain stem sites which in turn activate autonomic or affective responses or descending pain modulatory mechanisms.

Recent studies in our laboratory have been concerned with the characterization of the population of that respond exclusively or maximally to noxious lamina I neurons that projects to the mesencephalic stimuli.‘2~53Many lamina I neurons have been shown parabrachial areat (PBA).2’m23We observed that to project rostrally to innervate sites in the brain stem some nociceptive-specific lamina I projection neurons and thalamus. Individual, antidromically-identified ascend the spinal cord in the dorsolateral funiculi lamina I neurons projecting to the thalamus1~‘4~‘8*23~42~49 (DLF) and terminate in the midbrain PBA of the cat. and mesencephalon 22,23.32,33 have been characterized A few of these cells had demonstrable collateral with respect to their responses to stimulation of projections to the thalamus. the skin. Retrograde-labeling studies in the rat, The experiments described in the present report cat and monkey have demonstrated that significant were undertaken in order to characterize the lamina numbers of laminal I cells are labeled following I PBA projection system in the rat and to establish injection of tracer into the thalamussJ0~i7~20~4s~54 or a foundation for further investigations in this species. midbrain.2’,29,47,48,52,”Lamina I neurons are also retroIn addition, the collateral projection of some cells to gradely labeled from the parabrachial nucleus in the both the PBA and thalamus that we observed in the dorsolateral pons.“,37 cat23 have not been reported in the rat3’ We describe here the spinal projection to the PBA in the rat, the *To whom correspondence should be addressed. projection of lamina I fibers through the DLF and tThe midbrain parabrachial area was described as a region the collateral retrograde labeling of PBA/thalamic of the caudal midbrain surrounding the brachium projection neurons. conjunctivum that included portions of the parabrachial The marginal zone (lamina I) of the spinal and medullary dorsal horn is characterized by neurons

nuclei, nucleus cuneiformis and periaqueductal gray;21-23 it probably extends caudally into the dorsolateral pons as well.“.2*,37 Abbreviations: DAB, 3,3’-diaminobenzidine: DLF, dorsolateral funiculi; HRP, horseradish peroxidase; LSN, lateral spinal nucleus; PAG, periaqueductal gray; PBA, parabrachial area; TMB, tetramethylbenzidine; WGA, wheat germ agglutinin. NSC

2811-a

27

EXPERIMENTALPROCEDURES Projection neurons were labeled following unilateral injections of retrograde tracer into the DontomesenceDhalic PBA as follo& Adult male Sprague-Dawley’ rats (20@-275g) were anesthetized with Equithesin (chloral hydrate/pentobarbital) and placed in a stereotaxic appara-

28

J. L. K.

HYLDES rf al.

tus. A hole was drilled through the skull in order to allow insertion of a Hamilton microliter syringe needle into the brain stem according to the atlas of Paxinos and Watsot? (0.1 mm caudal in interaural zero and 1.7mm lateral to midline at 3.2mm dorsal to zero). Following injection of horseradish peroxidase conjugated to wheat germ agglutinin (WGA-HRP, 0.2 ~1, 25% in saline; Sigma), the skull opening was filled with sterile bone wax and the incision was closed with sutures. In some animals, a laminectomy was performed to expose the lower thoracic spinal cord, where bilateral DLF lesions were made. For this purpose, the dura was retracted to expose approximately one spinal segment, a few drops of lidocaine were placed on the cord dorsum and bilaterai lesions were made between the dorsal root entry zone and the dentate ligament using iridectomy scissors. The opening was covered with sterile Gelfoam and the wound was sutuied in layers. At 48-72 h after WGA-HRP iniection. rats were deeolv . . anesthetized with sodium pentobar-bital and perfused transcardially with I .5% paraformadehyde/l.5% glutaraldehyde in 0.1 M phosphate buffer. The brain and spinal cord were removed and placed in cold fixative with 30% sucrose overnight and then transferred to cold phosphate-buffered sucrose for 24-48 h. Brain stems were cut into SO&m transverse sections on a freezing microtome. Every third section was reacted with 3,3’-diamino~nzidine (DAB), mounted and counte~tained with Neutral Red. DAB was used to demonstrate the injection sites as this is the chromogen that most accurately reflects the area from which retrogade transport occurs.‘7.M.5’Spinal cord segments containing DLF lesion sites were embedded in gelatin, cut into 100 pm transverse sections, mounted serially and counterstained with Neutral Red. Camera lucida drawings were made to reflect the maximum extent of tissue damage bilaterally. The cervical and lumbar enlargements were cut into 50pm transverse sections and reacted with tetramethylbenzidine (TMB) by the method of Mesulam.” Using brightfield microscopy. retrogradely labeled neurons were plotted on spinal cord diagrams, and the mean number of cells per SOprn section was determined in at least ten randomly selected sections per segment. The presence of projection neurons that innervated both the PBA and thalamus was detected using double retrograde transport of two different fluorescent -tracers. Relatively laree iniections of Fluoro-Gold (0.3 ~1, 3% in saline; Fl;oro&rome Inc.) were made into the lateral thalamus (6.0 mm anterior to interaural zero and 2.7 mm lateral at a depth of 3.5 mm dorsal to the zero plane38) followed by two injections of rhodamine-labeled latex microspheres (0.3 pi, aqueous suspension of spheres. 0.02-0.2 pm diameter; Tracer Technology) into the ipsilaterai PBA. Two injections of microspheres (separated rostro~audally by 0. i--O.3mm) were used in an effort to increase the effective area from which retrogade transport occurred as injections of this tracer resulted in small injection sites with minimal spread. Surgical procedures and PBA coordinates were as described for WGA-HRP injections. Rhodamine-labeled microspheres were chosen as the retrograde marker for PBA injections because they have been shown not to be taken up by axons of passage. 39,40Fluoro-Gold, on the other hand, may be taken up by axons and was therefore only used for the more rostrai (e.g. thalamic) injections where this property is less problematic. Fourteen days after the injection of fluorescent tracers, rats were perfu&d with cold 4% parafo~aIdehyde in 0.1 M DhOSDhate buffer. The brain and spinal cord were postfixed f over&ght and subsequently transferred to 30% sucrose in phosphate buffer. Coronal sections of the brain (50 pm, unstained) were examined with both brightfield and epifluorescence optics in order to reconstruct the injection sites. The cervical and lumbar enlargements were cut into 25 pm sagittal sections on a freezing microtome and mounted immediately from ice-cold phosphate-buffered saline. In

some rats, the caudal medulla was cut as well (?S~tm. transverse sections). Tissue sections were examined uncoverslipped with a Zeiss epifluorescence microscope at a magnification of 400 x The appearance of retrogradely-transported FluoroGold was characterized by fluorescent, gold-colored granules within the cytoplasm of cells when viewed with ultraviolet light (Ziess U. V. filter set G365). Rhodamine microspheres were viewed with green light (Zeiss Rhodamine filter set H546) and appeared as distinct fluorescent-red dots within retrogradely-labeled cells. Double-labeled neurons were counted in the contralateral cervical and lumbar enlargements. ~psilateral spinal cord sections were not examined rigorously because of the reiatively small number of ipsilateml~y-projecting spinothalamic neurons in the rat.

RESULTS

Injections of WGA-HRP were centered on the area surrounding the brachium conjunctivum at the pontomesencephalic junction and included varying portions of the parabrachial nuclei, ventrolateral ~~aqueductal gray (PAG) and nucleus cuneifo~is (Fig. 1). The distribution of retrogradely-labeled cells was examined in the cervical and lumbar enlargements of five rats following such injections into the PBA. The pattern of labeling was similar in all animals and is illustrated in Fig. 2A for one animal. Labeled cells were found bilaterally in both enlargements with 5741% of the cells being on the side contralateral to the injection site. A particularly high concentration of ceils was noted in lamina I. More than one-third of the labeled cells in the cervical and lumbar enlargements were located in lamina I; 6540% of lamina I cells were on the contralateral side. These cells were located along the mediolateral width of lamina I with a tendency to be clustered beneath the dorsal root entry zone (Fig. 3). In transverse section, labeled lamina I neurons usually appeared as small fusiform or multipolar neurons with dendrites largely confined to lamina I. Other labeled cells found in the spinal gray matter were concentrated in the lateral part of lamina V or scattered in laminae IV-VIII and X. These deeper cells usually appeared as medium-to-large multipolar neurons and showed only a slight contralateral predominance (.51-58%). In addition, all animals demonstrated retrogradely-labeled cells in the lateral spinal nucleus (LSN).“j These cells were located just beyond the lateral extension of the dorsal horn but were distinct from lateral lamina I cells in that they were surrounded by white matter and their orientation did not conform to the dorsal horn border. Cells in the LSN showed no contralateral predominance. Animals with bilateral DLF lesions at lower thoracic levels (n = 2) had the same distribution of labeled neurons in the cervical spinal cord as described for intact control animals, but the lumbar enlargement was almost devoid of lamina I neurons on both sides (Fig. 2B). Table 1 lists the mean number of cells per

29

Lamina I projection neurons PEA 31

PEA 61

PEA

62

PBA 69

PBA

70

Fig. 1. Camera lucida drawings of sections through the brain stem of five rats indicating the locations of injections of 0.2 ~1 WGA-HRP (shaded areas). The maximum extent of tissue damage in two animals with bilateral DLF lesions (dashed lines) at the lower thoracic level is shown below the corresponding injection site. BC, brachium conjunctivum; IC, inferior colliculus; PAG, periaqueductal gray.

A.

Control

B.

DLF Lesions

Cervical

a

ipdbmral

contmlamml

Fig. 2. Distribution of retrogradely-labeled neurons in the spinal cord of one control rat (A, injection number PBA 70 from Fig. 1) and one rat that had bilateral DLF lesions (B, number PBA 61 from Fig. I). Each section shows the cells found in 5 randomly selected 50 pm sections from either the cervical or lumbar enlargement.

30

J. L. K.

HYLDEN P/ ol

Fig. 3. Photomicrograph of a transverse section (50nm) of the lumbar enlargement that contained retrogradely-labeled PBA projection neurons in lamina I. Labeled neurons were observed along the width of lamina I, but tended to cluster near the dorsal root entry zone. This figure shows a band of lamina 1 neurons extending from the dorsal root entry zone (upper right) around the lateral edge of the dorsal horn (lower left). The white matter of the dorsal columns (DC) can be seen in the upper right corner. Calibration

bar: 50 film.

Fig. 4. Camera Iucida drawings of sections through pairs of injection sites in four rats. Upper figures indicate the extent of injection of Fluoro-Cold into the lateral thalamus (shaded areas). Lower figures indicate the extent of injections of rh~amine-Ia~I~ latex microspheres into the PBA (solid areas). Since tissue sections containing fluorescent dyes were not dehydrated, cleared or counterstained, nuclear boundaries and other details were not readily apparent. BC, brachium conjunctivum; CP, cerebral peduncle; fx, fornix; IC, inferior colliculus; mt, mammilothalamic tract; SC, superior colliculus; V, ventricle; VB, ventrobasal thalamus.

31

Lamina I projection neurons Table I. Retrograde labeling of parabrachial area projection neurons in lamina I, deeper laminae and

lateral spinal nucleus: effect of bilateral dorsolateral funiculi lesions Mean number of cells/section*

Rat No. Control PBA 31 PBA 69 PBA 70 DLF lesion PBA 61 PBA 62

_

Lumbar

Cervical

Percentage of cellst located in lamina I

I

Deep

LSN

I

Deep

LSN

Cervical

Lumbar

9.1 5.4 6.8

10.5 9.2 10.2

5.5 5.1 3.7

9.5 12.1 10.7

13.0 13.1 14.4

4.5 6.4 6.0

41.3 + 3.6 37.1 * 3.1 41.3 f 2.6

42.2 + 4.3 46.4 + 2.4 43.8 & 3.9

4.8 5.3

11.5 9.4

2.8 2.4

0.3 1.5

10.9 9.4

0.7 2.2

29.2 + 3.7 36.0 & 2.2

2.1 f 1.41 13.0 f 4.51

*Mean cells/section determined from 10 randomly-selected 50 pm sections. tMean percentage of cells (excluding LSN cells) in lamina I f S.E.M. $The percentage of retrogradely-labeled lamina I cells per 50 pm section in the lumbar enlargement was significantly less than in the cervical enlargement (P < 0.001, Student’s r-test). These values were also sianificantlv less than the percentage of labeled lamina I cells in the lumbar enlargements of control ktimals (k < 0.001, ANOVA). 50pm section and the percentage of those cells (excluding LSN cells) that were located in lamina I in the cervical and lumbar enlargements of three control animals and two animals with bilateral DLF lesions. The injection sites and lesions for these five animals are illustrated in Fig. 1. The lesions in rat PBA62 spared the lateral-most portion of the DLF: the lesions in rat PBA61 were more complete. The data in Table 1 indicate that although in control animals the proportion of labeled cells in lamina I of the lumbar enlargement was the same or even greater than in the cervical enlargement, there was a significant decrease in the percentage of labeled cells located in lamina I caudal to the lesions. We can estimate the percentage decrease as 93% and 71% for rats PBA61 and 62; this is based on an expected value of 44% of retrogradely-labeled cells in lamina I as observed in the lumbar enlargement of control rats (see Table 1). The number of labeled cells in deeper laminae was not affected by DLF lesions. The number of labeled cells in the LSN was significantly decreased in the animal with the most extensive DLF lesions (P < 0.01, Student’s t-test). Injections of Fluoro-Gold into the thalamus were centered on the ventrobasal complex and included the bulk of the lateral thalamus (with occasional caudal spread to the level of the lateral geniculate). These injections appeared as a region of necrosis surrounded by a bright region in which both the neuropil and cells fluoresced.46 Injections of rhodamine microspheres into the PBA were very restricted and exhibited minimal spread into the surrounding tissue.25 The extent of four sets of combined thalamic/PBA injections are illustrated in Fig. 4; these four sets of injections gave the highest yield of double-labeled cells. In a few cases, the midbrain injection was located slightly more dorsal (in the inferior colliculus); there was little retrograde labeling in the spinal cord from these dorsal injection sites. The distribution of retrogradely-labeled cells in cervical and lumbar enlargements following injections of rhodamine microspheres into the PBA

(Fig. 5) was essentially the same as that described above for WGA-HRP injections (compare Fig. 5 with Fig. 2). Retrograde labeling of spinothalamic tract neurons with Fluoro-Gold was similar to that described previously for injections of Fluoro-Gold36 or HRP” into the lateral thalamus of rats. It is important to note the possibility that Fluoro-Gold may label cells other than those terminating within the injection site via retrograde transport from axons of passage. The only known system of spinal neurons that project rostra1 to the thalamus is the recentlydescribed spinohypothalamic tract.’ However, the cells of origin of the spinohypothalamic tract clearly have a different distribution in the spinal cord when compared to the distribution of spinothalamic tract cells. Whereas spinohypothalamic neurons have a bilateral distribution in the spinal cord, spinothalamic neurons labeled by Fluoro-Gold are found predominantly contralateral to the injection site. Therefore, it is unlikely that there was significant uptake of Fluoro-Gold by spinohypothalamic axons in the present experiments. All rats that had reasonable retrograde labeling from both injection sites (i.e. more than 20 labeled lamina I cells per enlargement; n = 10 rats) demonstrated some degree of collateral labeling. The observed distribution of single and double labeled neurons is presented in Fig. 5-note that this figure is a diagrammatic representation of the distribution of single- and double-labeled cells in the lumbar enlargement. Double-labeled neurons were especially abundant in lamina I. In seven rats, the percentage of lamina I spinothalamic tract cells double-labeled from the PBA ranged from 50 to 98%. A total of 1077 lamina I spinothalamic projection neurons were identified in the contralateral cervical and lumbar enlargements and trigeminal nucleus caudalis; greater than 80% of these were double-labeled from the PBA (the overall mean for cervical and lumbar enlargements was 85%, see Table 2). The percentage of double-labeled thalamic projection neurons in lamina I of the trigeminal nucleus caudalis was lower than in

32

J.

L. K.

0

PBA

A

Tholamus

n

PBA/Tholamur

HYLDEN CI (I/.

Fig. 5. The locations of PBA projection neurons (open circles), thalamic projection neurons (open triangles) and neurons having collateral projections to both PBA and thalamus (filled squares) are indicated in schematic drawings of 3 representative parasagittal sections (25 pm) through the lateral spinal nucleus (A), lateral dorsal horn (B) and medial dorsal horn (C) of the lumbar enlargement of one rat. Shaded areas indicate gray matter. Drawings A, B and C represent the distribution of cells observed in L4 through L6. The data are compressed rostrocaudally in order to fit within the figure. The transformation of this cellular distribution to the transverse plane is represented in D. Arrows a, b, and c indicate the approximate mediolateral location of parasagittal sections A, B and C.

the spinal enlargements. This is probably due in part to the fact that the PBA projection at this level is primarily ipsilateral (63%), whereas the thalamic projection is contralateral (91%). Thus, the vast majority of lamina I spinothalamic tract neurons in the cervical and lumbar enlargements had collateral branches terminating in the PBA. The converse relationship was not true. The majority of lamina I PBA projection neurons (S-96%) were only singlelabeled (Table 3). In general, injections of retrograde tracer into the PBA labeled 2-10 times as many

lamina I neurons as thalamic injections. Figure 6 illustrates a cluster of retrogradely-labeled PBA projection neurons in lateral lamina I (Fig. 6B); one of the cells was double-labeled from the lateral thalamus (Fig. 6A). Some spinothalamic projection neurons located in deeper laminae (IV-VIII) were also double-labeled from the PBA, but the occurrence of such double-labeled cells was less common than in lamina I (Table 2). A total of 416 spinothalamic tract neurons were identified in the deeper laminae of the

Table 2. Percentage of double-labeled spinothalamic neurons in contralateral lumbar spinal cord and trigeminal nucleus caudalis Lamina I Cervical enlargement Lumbar enlargement Trigeminal n. caudalis (contralateral) Trigeminal n. caudalis (insilaterall

Laminae IV-VIII

cervical and LSN

84% (562/663, n = 6);

14% (22/158, n = 5)

45% (22/48, n = 5)

88% (234/264, n = 6)

35% (91/258, n = 5)

14% (44159, n = 5)

_

32% (45/141, n = 2) 44% (419. n = 2)

*Numbers in parentheses refer to the number of cells counted (double-labeled over total) in n rats.

33

Lanka I projection neurons Table 3. Percentage of double-labeled parabrachial area projection neurons in contralateral cervical and lumbar spinal cord

Cervical enlargement Lumbar enlargement

Lamina I

Laminae IV-VIII

LSN

34% (562/1647, n = 6)*

35% (22162, n = 5)

30% (22170, n = 5)

42% (91/214, n = 5)

37% (44/l 18, n = 5)

31% (2341734, n = 6)

*Numbers in parentheses refer to the number of cells counted (doubIe-ladle

contralateral cervical and lumbar enlargements; 27% of these were double-labeled from the PBA. Spino thalamic neurons located in the LSN were observed less frequently than cells in laminae I and IV-VIII, but a high proportion of LSN cells were doublelabeled. A total of 107 LSN spinothalamic tract cells were observed in five rats (Table 2); 62% of these were double-ladle from the PBA. DWXJSSION

In the first part of the present study, relatively large injections of WGA-HRP were made into the PBA of the rat in order to label as completely as possible the population of projection neurons. Swett et aL4’ have

over total) in n rats.

analysed the labeling of lamina I spinomesencephalic cells following small injections of WGA-HRP in the rat. They concluded that even small injection sites in the caudal, medial midbrain, that included portions of lateral PAG and nucleus cuneiformis, labeled large numbers of lamina I neurons. This same region of the caudal midbrain (n. cuneiformis and PAG, or the area referred to as the midbrain PBA) is an area heavily labeled by HRP or WGA-HRP anterogradely transported from cat spinal cord5s” and is a site from which lamina I spinomesencephalic neurons can be antidromically activated in the cat22,23and rat.33 In addition, significant labeling of lamina I neurons is seen after injections of WGA-HRP immediately caudal to the midbrain PBA, into the dor-

Fig. 6. Photomicrographs of a parasagittal section through lateral lamina I of the cervical enlargement as observed under U.V. light (A) or green g&t (8). The Fluoro-Gold labeled spinothalamic cell seen in A was Iaheled with rhodamine microspheres transported from the PBA (arrows). Note that, in cells labeled with rhodamine microspheres, indi~d~l spheres can he identified in the cytoplasm whereas Fluoro-Gold appears to be more granular. Many PBA projection neurons were only single-Ia~Ied. x 500.

34

J. L. K.

HYLDEN ci al.

solateral pons, in the rat” and cat;“7 these pontine projection neurons have physiological characteristics similar to the mesencephalic projection neurons.28 Thus, it appears that lamina I projection neurons innervate a region of the brain stem that extends from &he dorsolateral pons into the caudal midbrain. In the present study, the PBA injections encompassed this region of the brain stem at the level of the pontomesencephalic junction. The distribution of retrogradely-labeled PBA projection neurons we observed has features in common with the reported dist~butions of spinomesencephalic neurons in the rat,29,3s,47cat2’.52 and monkey4R.54and of spinal neurons projecting to the dorsolateral pons of the rat” and cat.” The most striking characteristic of this distribution is the abundance of lamina I neurons; in all of the above studies a prominent, bilateral lamina I projection was reported (contralateral predominance). In the present report, more than one-third of the retrogradelylabeled cells in lumbar and cervical enlargements were located in lamina I. A second common group of cells, those found in lateral lamina V, has been reported for the spinomesen~phalic projection in the rat,i5 cat2’.s’ and monkey.4X,54 Spinomesencephalic cells in the LSN have been observed previously in the rat;‘5.47 as in the present study, LSN cells had a bilateral distribution with no contralateral predominance. In the rats with bilateral DLF lesions, the number of retrogradely-labeled lamina I neurons in the lumbar enlargement was greatly reduced compared to the cervical enlargement. This finding was true even in the case with relatively small, dorsally-placed lesions. We can conclude that the majority of lamina I PBA projection neurons in the rat ascend through the DLF, as we have demonstrated previously in the cat2’ and as has also been demonstrated by electrophysiological techniques for some lamina I neurons in the rat.32.33The labeling of cells in the LSN was also decreased by DLF lesions; thus, these cells appear to project their axons through the DLF as well. The labeling of PBA projection neurons located deeper in the gray matter was not affected by DLF lesions; therefore, these deeper neurons must project their axons rostrally through other ascending pathways. Our findings of collateral labeling of PBA/thalamic projection neurons in the rat were predicted from previous work on lamina I PBA projection neurons in the cat-in which we were able to antidromically activate some cells from both the midbrain and thalamus.22.23 The present study adds to previous reports of the collateralization of spinothalamic neurons by demonstrating that a significant number of spinothalamic tract cells send collaterals to the PBA in the rat. Indeed, the majority of lamina I neurons that projected to the lateral thalamus also projected to the PBA. These lamina I PBA/thalamic projection neurons appear to be from a different population than reported thalamic/medullary projection neu-

rons, since such neurons have not been identified in lamina I.26.4’In addition, we noted that the relative magnitude of lamina I labeling from the PBA was greater than the labeling from the lateral thalamus (about 30% of the number of lamina I PBA neurons). The present investigation demonstrates the presence of a direct projection system from lamina I, through the DLF, to the PBA with major collateralization to the thalamus in the tat. We know that neurons in this lamina I projection pathway are largely nociceptive-specific22.23 and likely to be involved in the transmission of nociceptive information. Two major features of the pathway, the DLF projection and the extent of spinothalamic collateralization, are of considerable interest in the understanding of spinal nociceptive mechanisms and will be discussed below. Our observations of a DLF projection for lamina I neurons in the rat is in agreement with such observations for spinothalamic lamina I neurons in the cat24 and monkey’ and for spinomesencephalic2’ and indeed all lamina I projection neurons in the cat.’ The earliest description of a DLF projection was made by Zemlan et al.,‘” in the rat. These authors used spinal cord white matter applications of HRP to demonstrate labeling in more caudal segments. In the present study, the DLF projection was demonstrated for lamina I neurons with axon terminals in the PBA, but, due to the high degree of collateralization, can also be inferred for the majority of lamina I spinothalamic neurons in the rat. This conclusion is likely because, on average, more than 80% of lamina I spinothalamic neurons had terminations in the PBA and lamina I PBA projection neurons send their axons almost exclusively through the DLF (a 93% decrease in labeling was noted caudal to complete DLF lesions). Thus, a dorsolateral spinothalamic pathway has now been demonstrated directly in the monkey and cat and indirectly in the rat. The high proportion of double-labeled lamina I spinothalamic cells demonstrated in the present study with pontomesencephalic PBA injections has also been seen with more discrete injections into the PAG. Pechura”@ showed a significant collateralization of both lateral and medial spinothalamic tract cells to the PAG in the rat. This was especially striking for lamina I cells projecting to the lateral thalamus; approximately one-third of such cells were doublelabeled from the PAG. Since the pontomesencephalic terminal field of lamina I PBA projection neurons extends into the PAG, double-labeled cells described by Liu,” Pechura”‘~” and Carlton et al.’ can be considered to be a subset of the larger population of PBA/thalamic projection neurons. The fact that ~cMahon and Wal13”failed to observe any lamina I mesencephalic projection neurons with thalamic collaterals in their antidromic mapping experiments may be explained by the observation that a smaller percentage of the larger population of PBA projection neurons send collaterals to the thalamus or

Lamina I projection neurons may have been due to the technical difficulty inherent in the antidromic activation of small caliber axons. The presence of thalamic/mesencephalic projection neurons has been demonstrated using double antidromic activation in cats22.23and monkeys.‘9,43 A recent preliminary report has demonstrated the presence of spinothalamic tract neurons in laminae I and V of the lumbar enlargements of monkeys with collaterals to the PAG.% Thus, as with the DLF pathway for lamina I neurons, the collateral projection of lamina I neurons to the PBA and thalamus has also been demonstrated in rat, cat and monkey. It is apparent from the present study, as well as the other studies discussed above, that the lamina I projection system is involved in distributing information to various brain stem sites. The information these neurons carry is of a very specific type. The majority of lamina I projection neurons have slowly-conducting axons and respond exclusively to noxious stimuli applied to well-defined receptive fields.‘3~23~27~32.34 The slow conduction velocities of these neurons, together with a lack of a direct involvement of DLF pathways in the detection of noxious stimuli,50 make it unlikely that nociceptivespecific lamina I projection neurons play a major role in the sensory-discriminative aspects of pain. This idea is supported by recent evidence indicating that nociceptive-specific neurons are not involved in the encoding process by which monkeys perceive the intensity of noxious heat stimuli.3’

35

The relative abundance of nociceptive-specific lamina I PBA projection neurons leads us to believe that this projection system is important for some very basic component of the pain response. We have previously postulated that lamina I PBA projection neurons are likely to be involved in the activation of autonomic reflexes or pain modulatory mechanisms.2’.23 Thus, activation of lamina I neurons by noxious stimuli may: (1) activate somatovisceral reflexes via input to the PBA of the dorsolateral pons/caudal mesencephalon,45 (2) trigger affective responses to acute pain via connections from the PBA to the hypothalamus6,‘5 and amygdala,45 (3) activate descending modulatory systems at the level of nucleus cuneiformis/PAG through connections to the nucleus raphe magnus. Lamina I neurons with projections to the thalamus appear to belong to a subpopulation of lamina I PBA projection neurons and as such share the properties of the larger group of cells, i.e. nociceptive-specific, small receptive fields, slowly-conducting axons that project via the DLF. As with lamina I PBA projection neurons, it is also unlikely that lamina I spinothalamic neurons are responsible for encoding the sensory-discriminative aspects of pain. Instead, these cells may play a role in stimulus localization or may provide more discriminative information at suprathreshold intensities. Acknowledgemenfs-The authors wish to thank Drs Cl. J. Bennett, E. H. Chudler and R. Dubner for critical comments.

REFERENCES

Albe-Fessard D., Levante A. and Lamour Y. (1974) Origin of spinothalamic and spinoreticular pathways in cats and monkeys. In Aduances in Neurology (ed. Bonica J. J.) Vol. 4, pp. 157-166, Raven Press, New York. Apkarian A. V., Stevens R. T. and Hodge C. J. (1985) Funicular location of ascending axons of lamina I cells in the cat spinal cord. Brain Rex 334, 160-164. Apkarian A. V., Stevens R. T. and Hodge C. J. (1987) The primate dorsolateral spinothalamic pathway. Sot. Neurosci. Abst. 13, 580.

Basbaum A. I and Fields H. L. (1978) Endogenous pain control mechanisms: review and hypothesis. Ann. Neural. 4, 451462.

Bjorkeland M. and Boivie J. (1984) The termination of spinomesencephalic fibers in cat: an experimental anatomical _’ study. Anat. Embryol. 170, 265-277. 6. Brodal A. (1958) The Reticular Formation of the Brain Stem. Anatomical Aspects and Functional Correlations. Thomas, Sminefield. 7. BurstLin R., Cliffer K. D. and Giesler G. J. Jr. (1987) Direct somatosensory projections from the spinal cord to the hypothalamus and telencephalon. J. Neurosci. 7. 41594164. 8 Burton H. and Craig A. D: Jr. (1979) Distribution of trigeminothalamic projection cells in cat and monkey. Brain Res. 161, 515-521. 9. Carlton S. M., Harmann P. and Willis W. D. (1987) Collaterals of spinothalamic tract (STT) cells to the PAG: a double-labeling study in the rat. Pain Suppl. 4, S400. 10. Carstens E. and Trevino D. L. (1978) Laminar origins of spinothalamic projections in the cat as determined by the retrograde transport of horseradish txroxidase. J. coma. Neural. 182. 151-166. II. Cechetto D. F., Standaert D. G. and Saper C. B. (1985) Spinal and trigeminal dorsal horn projections to the parabrachial nucleus in the rat. J. camp. Neural. 240, 153-160. 12. Christensen B. N. and Per1 E. R. (1970) Spinal neurons specifically excited by noxious or thermal stimuli: marginal zone of the dorsal horn. J. NeurophysioL 33, 293-307. 13. Craig A. D. and Knifliki K.-D. (1985) Spinothalamic lumbosacral lamina I cells responsive to skin and muscle stimulation in the cat. J. Physiol., Land. 365, 197-221. 14. Dilly P. N., Wall P. D. and Webster K. E. (1968) Cells of origin of the spinothalamic tract in the cat and rat. Expl Neural. 21, 550-562.

15. Edwards S. B. and de Olmos J. S. (1976) Autoradiographic studies of the projections of the midbrain reticular formation: ascending projections of nucleus cuneiformis. J. camp. Neural. 165, 417432.

36

J. L. K. HYLVFNrr ul.

16. Giesler G. J. Jr. and Elde R. P. (1985) Immunocytochemical studies of the peptidergic content of fibers and termmals within the lateral spinal and lateral cervical nuclei. J. Neurosci. 5, 1833-1841. 17. Giesler G. J. Jr., Menetrey D. and Basbaum A. 1. (1979) Differential origins of spinothalamic tract projections to medial and lateral thalamus. J. camp. Neural. 184, 107-125. 18. Giesler G. J. Jr., Menetrey D., Guilbaud G. and Besson J.-M. (1976) Lumbar cord neurons at the origin of the spinothalamic tract in the rat. Brain Rex 118, 320-324. 19. Girardot M.-N., Brennan T. J., Martindale M. E. and Foreman R. D. (1987) Effects of stimulating the subcoeruleus-parabrachial region on the non-noxious and noxious responses of T,_, spinothalamic tract neurons in the primate. Brain Res. 409, 19-30. 20. Hockfield S. and Gobel S. (1978) Neurons in and near nucleus caudalis with long ascending projection axons demonstrated by retrograde labeling with horseradish peroxidase. Bruin Res. 139, 3333339. 21. Hylden J. L. K., Hayashi H. and Bennett G. J. (1986) Lamina I spinomesencephalic neurons in the cat ascend via the dorsolateral funiculi. Somatosensory Res. 4, 3 l-41. 22. Hylden J. L. K., Hayashi H., Bennett G. J. and Dubner R. (1985) Spinal lamina I neurons projecting to the parabrachial area of the cat midbrain. Bruin Res. 33, 195-198. 23. Hylden J. L. K., Hayashi H., Dubner R. and Bennett G. J. (1986) Physiology and morphology of the lamina I spinomesencephalic projection. J. camp. Neural. 247, 505-515. 24. Jones M. W., Hodge C. J. Jr., Apkarian A. V. and Stevens R. T. (1985) A dorsolateral spinothalamic pathway in cat. Bruin Res. 335, 188-193. 25. Katz L. C., Burkhalter A. and Dreyer W. J. (1984) Fluorescent latex microspheres as a retrograde neuronal marker for in uiuo and in vitro studies of visual cortex. Nature 310, 498-500. 26. Kevetter G. A. and Willis W. D. (1983) Collaterals of spinothalami~ cells in the rat. J. camp. Neural. 215, 453-464. 27. Kumazawa T., Per1 E. R., Burgess P. R. and Whitehom D. (1975) Ascending projections from marginal zone (lamina I) neurons of the spinal dorsal horn. J. camp. Neurol. 162, 1-12. 28. Light A. R., Casale E. J. and Menttrey D. (1985) Spinal cord laminae I and II neurons which project to the contralateral pontine reticular formation in the cat. Sot. Neurosci. Abst. 11, 411. 29. Liu R. P. C. (1983) Laminar origins of spinal projection neurons to the periaqueductal gray of the rat. Brain Res. 264, 118-122. 30. Liu R. P. C. (1986) Spinal neuronal cohaterals to the intralaminar thalamic nuclei and ~~aqueductal gray. Bruin Res. 365, 145-150. 31. Maixner W., Dubner R., Bushnell M. C., Kenshalo D. R. Jr. and Oliveras J.-L. (1986) Wide-dynamic range dorsal horn neurons participate in the encoding process by which monkeys perceive the intensity of noxious heat stimuli. Brain Res. 374, 385-388.

32. McMahon S. B. and Wall P. D. (1983) A system of rat spinal cord lamina I cells projecting through the contralateral dorsolateral funiculus. .J. camp. Neurot. 214, 217-223. mapping of brainstem projections of spinal cord famina 33. McMahon S. B. and Wall P. D. (1985) El~trophysiolo~cai I cells in the rat. Brain Res. 333, 19-26. 34. Menetrey D., Chaouch A. and Besson J.-M. (1980) Location and properties of dorsal horn neurons at origin of spinoreticular tract in lumbar enlargement of the rat. J. Neurophysiof. 44, 862-877. 35. Menetrey D., Chaouch A., Binder D. and Besson J.-M. (1982) The origin of the spinomesencephalic tract in the rat: an anatomical study using the retrograde transport of horseradish peroxidase. J. camp. Neurol. 206, 193-207. 36. Nahin R. L. (1988) Immunocytochemical identification of long ascending peptidergic lumbar spinal neurons terminating in either the medial -or lateral thalamus in the rat. Brain Res. (in press). 37. Panneton W. M. and Burton H. (19851Projections from the naratri~eminal nucleus and the medullary and spinal dorsal horns to the peribrachial area in the’cat. -~e~ros~ie~~~ 15,‘779-797. 38. Paxinos G. and Watson C. (1982) The Raf Brain in Stereotaxic Coordinates. Academic Press, New York. 39. Pechura C. M. (1987) Laterally- versus medially-projecting spinothalamic neurons and their axon collaterals to the periaqueductal gray and medullary reticular formation in the rat, Ph.D. dissertation, Uniformed Services University of the Health Sciences. 40. Pechura C. M. (1987) Lateral versus medial spinothalamic neurons and their axon collaterals to the periaqueductal gray and medullary reticular formation in the rat. See. Neurosci. Absf. 13, 113. 41. Pechura C. M. and Liu R. P. C. (1986) Spinal neurons which project to the ~riaqueductal gray and the medullary reticular formation via axon collaterals: a double-label fluorescence study in the rat. Bruin Res. 374, 357-361. 42. Price D. D., Dubner R. and Hu J. W. (1976) Trigeminothalamic neurons in nucleus caudalis responsive to tactile, thermal, and nociceptive stimulation of monkey’s face. J. Neurophysiol. 39, 936953. of input to 43. Price D. D., Hayes R. L., Ruda M. A. and Dubner R. (1978) Spatial _ _.and temporal . . 1_transformations _^A ^~_ spinothalamic tract neurons and their relation to somatic sensation. J. NeurophysJol. 41, YM-~4/. 44. Ruggiero D. A., Ross C. A., Kumada M. and Reis D. J. (1982) Re-evaluation of projections from the mesencepbalic trigeminal nucleus to the medulla and spinal cord: new projections. A combined retrograde and anterograde horseradish peroxidase study. f. camp. Neurof. 206, 278292. 45. Saper C. B. and Lowey A. D. (1980) Efferent connections of the parabrachial nucleus in the rat. Brain Res. 19, 291-46. Schmued L. C. and Fallon J. H. (1986) Fluoro-Gold: a new fluorescent retrograde axonal tracer with numerous unique properties, Brain Rex 377, 147-154. 47. Swett J. E., McMahon S. B. and Wall P. D. (1985) Long ascending projection to the midbrain from cells of lamina I and nucleus of the dorsolateral funiculus. J. camp. Neurof. 238, 401416. 48. Trevino D. L. (1976) The origin and projections of a spinal nociceptive and the~or~ptive pathway. In Senf0r.r Funcfions qf the Skin of Primates with Special Reference to Man fed. Zotterman Y.), pp. 367--376. Pergamon Press, Oxford. 49. Trevino D. L., Coulter J. D. and Willis W. D. (1973) Location of cells of origin of spinothalamic tract in lumbar enlargement of the monkey. J. Neurophysfol. 36, 750-761. 50. Vierck C. J. Jr., Hamilton D. M. and Thornby J. I. (1971) Pain reactivity of monkeys after lesions to the dorsal and lateral columns of the spinal cord. Expl Bruin Res. 13, 14&158.

Lamina I projection neurons

31

51. Warr W. B., DeOlmos J. S. and Heimer L. (1981) Horseradish peroxidase: the basic procedure. In Neuronal Tract-Tracing Methods (eds Heimer L. and Robards M. J.), pp. 252-256. Plenum, New York. 52. Wiberg M. and Blomqvist A. (1984) The spinomesencephalic tract in the cat: its cells of origin and termination pattern as demonstrated by the intra-axonal transport method. Brain Res. 291, l-18. 53. Willis W. D. and Coggeshall R. E. (1978) Sensory Mechanfi~ of rhe Spinal Cord. Plenum, New York. 54. Willis W. D., Kenshalo D. R. Jr. and Leonard R. B. (1979) The cells of origin of the primate spinothalamic tract. J. camp. Neurol. 188, 543-574. 55. Zcmlan F. P., Leonard C. M., Kow L. M. and Pfaff D. W. (1978) Ascending tracts of the lateral columns of the rat spinal cord: a study using the silver impregnation and horseradish peroxidase techniques. Expi Neural. 62, 298-334. 56. Zhang D., Carlton S. M., Sorkin L. S., Harmann P. A. and Willis W. D. (1987) Collaterals of spinothalamic tract (STT) neurons to the PAG: a double labeling study in the monkey. Sot. Neurosci. Abst. 13, 113. (Accepted 18 May 1988)