Reciprocal parabrachial-cortical connections in the rat

Brain Research, 242 (1982) 33-40 Elsevier Biomedical Press

33

Reciprocal Parabrachial-Cortical Connections in the Rat CLIFFORD B. SAPER Department of Neurology, New York Hospital Cornell University Medical Center, 525 E. 68th Street, New York, N Y 10021 and Departments of Neurology and Neurological Surgery (Neurology) and Anatomy and Neurobiology, Washington University School of Medicine, 660 South Euclid, St. Louis, MO 63110 (U.S.A.)

(Accepted November 19th, 1981) Key words: parabrachial nucleus - - Kolliker-Fuse nucleus - - wheat germ agglutinin - - horseradish peroxidase

The projection from the parabrachial nucleus (laB) to the cerebral cortex in the rat was studied in detail using the autoradiographic method for tracing anterograde axonal transport and the wheat germ aggiutinin-horseradish peroxidase (WGA-HRP) method for both anterograde and retrograde tracing. PB innervates layers I, V and VI of a continuous sheet of cortex extending from the posterior insular cortex caudally, through the dorsal agranular and the granular anterior insular cortex and on rostrally into the lateral prefrontal cortex. Within the prefrontal area, PB fibers innervate primarily layer V of the ventrolateral cortex caudally, but more rostrally the innervated region includes progressively more dorsal portions of the prefrontal area, until by the frontal pole the entire lateral half of the hemisphere is innervated. This projection originates for the most part in a cluster of neurons in the caudal ventral part of the medial PB subdivision, although a few neurons in the adjacent parts of the PB, the Kolliker-Fuse nucleus and the subcoeruleus region also participate. After injection of WGA-HRP into the PB region, retrogradely labeled neurons were found in layer V of the same cortical areas which receive PB inputs. The importance of this monosynaptic reciprocal brainstem-cortical projection as a possible anatomical substrate for the regulation of cortical arousal is discussed. INTRODUCTION The parabrachial nucleus of the pons (PB) is the main relay area for ascending visceral afferent inputs from the nucleus of the solitary tract to the forebrain (refs. 9-11, 14). Although it had long been thought that visceral afferent inputs to the cortex must require an additional relay in the thalamus, we have recently reported that in the rat some PB fibers bypass the thalamus and innervate certain cortical areas directly 12-~4. These regions include the insular cortex, the lateral prefrontal cortex, the septo-olfactory area and possibly the adjacent infralimbic cortex. The origin of this projection within PB was inferred from autoradiographic studies to be within the medial PB subnucleus (PBm), but finer localization was not possible 14. I have therefore further explored this system of connections using the method of anterograde and retrograde transport of wheat

germ agglutinin-horseradish peroxidase conjugate ( W G A - H R P ) as stained by the tetramethylben~dine method. By placing injections in both the PB-innervated cortical areas as well as in PB itself, it has been possible to distinguish that the insular and lateral prefrontal cortex of the rat are reciprocally connected with PB. METHODS A previously reported 14 series o f 35 autoradiographic experiments, in which tritium-labeled amino acids were injected into PB and adjacent areas, was systematically reviewed. The detailed distribution of the PB--cortical projection was plotted (see Fig. 1) as a reference for placement of W G A - H R P injections. Injections consisting of 20-40 nl of 0.5-1~o W G A - H R P (Sigma) in 0.9 ~o saline were stereotaxically placed in the PB-innervated regions of cerebral cortex in 15 albino rats, and injections were placed

* Address for correspondence: Department of Neurology, Washington University Medical School, 660 South Euclid, St. Louis, MO 63110, U.S.A.

34

R21a ~

LC1a

d

d

e

e

Fig, 1. Diagram to illustrate the reciprocity of parabrachial-cortical connections. In experiment R21 a--e, triangles schematically represent the distribution of retrogradely labeled neurons seen following an injection of WGA-I-IRP into the parabrachial region. In experiment LCI a-e, black dots schematically illustrate the pattern of autoradiographic axonal labeling seen after injecting PB with tritiated amino acids. The shaded areas in R21f and L C l f show the extent of the injection site in each experiment. Note the close correspondence between the regions of insular and lateral prefrontal cortex which receive parabraehial projections (LC1) and the cortical areas which project back upon the parabrachial region (R21). For abbreviations see list at end of paper.

35 in PB in 3 further rats. Following 48 h survival the animals were reanesthetized and perfused through the heart with 0.97o saline containing 10 U/ml heparin followed by 1 700 paraformaldehyde-l.25 glutaraldehyde in 0.1 M phosphate buffer (pH 7.2-7.3). Brains were removed, stored 1-5 days at 4 °C in 3070 sucrose in 0.1 M phosphate buffer, and then frozen sectioned at 50 #m. Serial sections of the PB region and a 1-in-3 series of sections of the rest of the brain were reacted with tetramethylbenzidine for peroxidase by the method of de Olmos et al. 3 (using 0.1 ~ hydrogen peroxide) and mounted on glass slides. Sections were stained with 1/4 700thionin, and coverslipped with Histoclad mounting medium. RESULTS

Autoradiographic experiments The general pattern of the PB-cortical projection which was reported in an earlier communication 14. will be considered here in greater detail (see Fig. 1). As in that earlier report, a single case is illustrated (see experiment LC1, Fig. 1), but the pattern of labeling seen in this experiment was typical of the PB--cortical projection. The PB fibers enter the cerebral cortex in two distinct bundles, The medial fibers appear to be a rostral continuation of the central tegmental tract-medial forebrain bundle pathway (see ref. 14, Fig. 4). These axons turn medially at the septal level and follow the lateral margin of the nucleus of the diagonal band dorsally. They can be followed further rostrally to the level of the genu of the corpus callosum, into the septo-olfactory area. A few labeled fibers appear to reach as far as layer V of the infralimbic cortex. Fibers to the lateral wall of the hemisphere arise from the dorsal tegmental bundle-zona incerta pathway. These axons can be followed into the substantia innominata where they turn laterally, enter the external capsule and are distributed to the cortex of the lateral wall of the hemisphere. At the rostral septal level PB fibers are distributed primarily in layer I and the deep part of layer III and in layer V in the posterior insular cortex and in both agranular and granular parts of anterior insular cortex (terminology of Krettek and Price)6. More rostrally, the labeled fibers are found primarily in layer V of the

ventral part of lateral frontal cortex. As the frontal pole is approached, the labeled axons turn dorsally until, at the frontal pole, layer V of the entire dorsolateral half of the hemisphere is innervated.

WGA-HRP experiments (A) Insular cortex The injection site included portions of the insular cortex in 9 experiments. In each of these the pattern of labeled parabrachial--cortical connections was similar. Following injections involving the granular insular cortex and dorsal agranular insular cortex retrogradely labeled neurons are found primarily in a cluster of cells in the caudal medial part of the medial parabrachial nucleus (see experiment R31, Fig. 2). These neurons tend to be horizontally oriented and fusiform or multipolar in shape, occupying an area within the ventral part of the medial parabrachial nucleus and extending into the subjacent subcoeruleus region (see Fig. 3). More caudally the labeled cells tend to be found most medially, adjacent to the mesencephalic trigeminal nucleus. A few scattered neurons are found in the remainder of the medial parabrachial nucleus, the most dorsal part of the Kolliker-Fuse nucleus and in the most medial parts of the superior cerebellar peduncle and overlying lateral parabrachial nucleus. When injection sites include the ventral agranular insular cortex, retrogradely labeled neurons are also found along the dorsolateral edge of the superior cerebellar peduncle. Anterograde axonal labeling within PB is distributed in much the same pattern as retrograde neuronal labeling. Injections in the more dorsal parts of the insular cortex demonstrate anterograde ~projections primarily to the more ventral and media~ parts of PB, while injections which also include the ventral insular cortex produce more widespread anterograde labeling including the area dorsally adjacent to the superior cerebellar peduncle. (B) Lateral frontal pole cortex Three injections involved the lateral part of the frontal pole. In these cases (see experiments R13 and j' R41, Fig. 2) the pattern of retrograde labeling in PB is similar to that seen after insular cortex injections. Most labeled cells are seen in the medial part of

36

R13

R41

R31

R27

R37

Fig. 2. Diagram to show the areas in the pontine tegmentum from which cortical projections arise. For each experiment, the extent of the cortical injection site with W G A - H R P is illustrated in the figure on the left as a shaded area. The pattern of retrograde neuronal labeling is represented schematically on the right by triangles, For abbreviations see list at end of paper.

37 the pontine tegmentum in this experiment than after insular or lateral frontal cortical injections. Labeled cells are seen, as in the other experiments, in PBm, the underlying subcoeruleus region, along the dorsolateral edge of the superior cerebellar peduncle and in the dorsal part of the Kolliker-Fuse nucleus. However, additional labeled neurons are seen more widely distributed in PB, especially in a cluster along the dorsolateral edge of the superior cerebellar peduncle. Other labeled cells are found in the medial pontine tegmentum, in a distribution similar to that seen after infralimbic and prelimbic cortical injections which spare the septo-olfactory area (see experiment R27, Fig. 2). These cells are seen in the central gray matter surrounding the dorsal tegmen-

PBm and in the subjacent subcoeruleus region. A few neurons are also labeled in the most medial part of the lateral PB subnucleus (PB1) and along the dorsolateral edge of the superior cerebeUar peduncle and trailing off into the dorsal part of the KollikerFuse nucleus. Light, apparently anterograde, labeling is also seen in this same distribution.

( C) Infralimbic cortex and septo.olfactory area The infralimbic cortex was labeled by four injections, but in only one of these cases was there retrograde labeling seen in PB. In this experiment (R37 in Fig. 2) the septo-olfactory area and lateral septum were also included in the injection site, Retrogradely labeled neurons are much more widely distributed in

,h ,.:Tr o J

~' "tllTI I

q

Fig. 3. A: photomicrograph of WGA-HRP injection site in experiment R31. Scale -- 1 mm. B: photomicrograph of WGA-HRP

injection site in experiment R41. Scale -- 1 mm. C: photomicrograph of retrogradely labeled neurons in parabrachial region following injection shown in A. Arrows indicate clusters of labeled neurons. Scale -- 100/~m. D: photomicrograph of retrogradely labeled neurons in parabrachial region following injection shown in B. Arrows indicate clusters of labeled neurons Scale = 100 ~zm. For abbreviations see list at end of paper.

38 tal nucleus of Gudden and in the dorsal part of the raphe pontis nucleus, extending laterally ventral to the medial longitudinal fasciculus. A few labeled cells are found in a similar distribution on the contralateral side of the pons. Comparison of the results of experiment R27 with R37 indicates that the retrograde labeling of PB neurons in the latter case is due to the inclusion of the septo-olfactory area, rather than the infralimbic cortex, in the injection site. Anterograde axonal labeling in experiment R37 is seen in the same areas of the central gray matter and PB which contain retrogradely labeled neurons.

( D ) Parabrachial nucleus In 3 experiments the injection site included PB. Each injection was large and also included portions of the adjacent central gray matter, the mesencephalic, principal sensory, and motor trigeminal nuclei and the subcoeruleus region. Thus these experiments can only serve as an indication of sources of projections to the dorsolateral pontine tegmentum. Nevertheless, the correspondence between the distribution of retrogradely labeled cortical neurons in these experiments and the pattern of anterograde transport of labeled amino acids to cortex after PB injections is striking (see Fig. 1). In a typical experiment, R21, illustrated in Fig. 1, a continuous sheet of neurons in layer V is labeled, running from the posterior insular cortex caudally, through the dorsal agranular and granular insular areas rostrally and on forward through the lateral and dorsolateral prefrontal cortex to the frontal pole. Following injections in these cortical regions, as noted above, anterogradely transported label in the dorsolateral pons is seen only in PB. Therefore, the cortical neurons labeled by dorsolateral pontine injections appear to project to PB, rather than to adjacent regions included in the injection site. The infralimbic cortex also contains retrogradely labeled neurons after dorsolateral pontine injections. However, following injections involving infralimbic cortex but not the septo-olfactory area (Fig. 2, R27), labeling of neurons and fibers in the dorsolateral pons is found only in the central gray matter. This suggests that labeling of infralimbic neurons after dorsolateral pontine injections may be due to involvement of the central gray matter by the injection site.

DISCUSSION These experiments indicate that there is a continuous sheet of cerebral cortex, extending from the insular cortex rostrally along the lateral portion of the frontal cortex and encompassing the entire dorsolateral half of the frontal pole, which is reciprocally connected with the parabrachial nucleus in the rat. The parabrachial efferent projection appears to arise primarily in two cell clusters in caudal PBm, one ventromedially located and the other at the ventrolateral edge of the superior cerebellar peduncle. A few other neurons in PBm, the medial part of PBI and the dorsal part of the Kolliker-Fuse nucleus also contribute to this projection. The parabrachialcortical fibers represent a rostrolateral extension of the dorsal tegmental bundle-zona incerta ascending pathway. The axons turn laterally through the amygdala and substantia innominata and pass into the external capsule. These fibers follow the lateral edge of the external capsule rostrally and innervate layers I, V and possibly VI of the insular cortex and layer V of the lateral and the dorsolateral frontal cortex. Neurons in these areas of cerebral cortex project back upon the parabracbial region. Retrograde transport studies indicate that this projection takes origin exclusively from layer V neurons. The details of the descending pathway are not known, but, based on anterograde transport of WGA-HRP, it appears that the descending fibers innervate primarily the area of PB which projects to the cerebral cortex. Although neurons in the deep layers of the infralimbic cortex were retrogradely labeled after a dorsolateral pontine WGA-HRP injection, no connections between PB and the infralimbic cortex could be demonstrated; the infralimbic and prelimbic cortex instead appear to be reciprocally connected with the pontine central gray matter. The PB projection to the septo-olfactory area was confirmed by anterograde autoradiographic tracing after tritiated amino acid injections into either PBm or PB1. These results confirm and extend our previous observations on PB-cortical connections 14. We had earlier noted on the basis of the placement of multiple labeled amino acid injections, that this projection seemed to originate in PBm, and the present findings identify the specific neurons of origin. Pre-

39 vious autoradiographic studies suggested a sparse PB innervation of infralimbic cortex, but the present retrograde tracer experiments are unable to confirm this finding. It seems most likely that if any PB fibers reach this area their numbers are too few to allow effective retrograde uptake. The reciprocal projection from the infralimbic cortex to the dorsolateral pontine tegmentum appears to be limited to the central gray matter. Two other groups of investigators have also reported PB-cortical connections, using retrograde transport of H R P after cortical injections. The findings of Divac et al. 4 in the rat are difficult to interpret because only a single illustration of the location of the labeled neurons was included in the text. The illustrated area appears to be rostral to the PB-cortical projection neurons described here. The region may correspond, however, to the pedunculo-pontine tegmental nucleus, which appears in the present experiments also to project to the frontal cortex (unpublished observations). The retrogradely labeled neurons indicated by Llamas et al. s in the cat clearly included cells within PB. However, locus coeruleus neurons, which also project to the frontal cortex, are found scattered within PB in the cat 2, so that the cross-species correspondence of these findings with the present results is not clear. The significance of the PB-cortical reciprocal connections is not known. The insular cortex is also reciprocally connected with the ventroposteromedial parvocellular thalamic nucleus (VPMpc, also known as the ventromedial basal nucleus). VPMpc is itself in receipt of a major projection from PB and processes taste, as well as possibly other visceral afferent information. In this regard it is of interest that

the insular cortex is heavily interconnected with a number of other central autonomic structures. It has recently been proposed that the insular cortex serves as a cortical representation of the central autonomic system (see refs. 12 and 13). The dorsolateral prefrontal cortex, however, does not appear to be connected with other central autonomic areas, nor is it included within the cortex associated with the mediodorsal thalamic nucleus1, 6 although in the present experiments it does receive afferents from the basolateral amygdaloid nucleus (unpublished observations). While this region has been considered to be a motor or premotor area in the rat6,7,15A7, its heaviest thalamic connections in this series of experiments have been with the intralaminar nuclei: centromedial, paracentral and parafascicular. These regions also receive a dense PB input. Thirty years ago, Starzl et al. 16 found that, following single electrical shocks to the brainstem 'reticular activating system' of the cat, cortical potentials were most reliably and intensely evoked in the dorsolateral frontal pole cortex. They traced two ascending pathways to this area, via stimulation and lesion experiments: (i) through the intralaminar thalamic nuclei and (ii) through the hypothalamus. The observations of Llamas et al. 8 suggest that a system of connections similar to those demonstrated here in the rat, probably exists in the cat. If so, the present results may provide an anatomical substrate for Starzl et al.'s electrophysiological observations. The importance of these connections in cortical arousal or attentional mechanisms, however, remains to be explored.

ABBREVIATIONS USED IN FIGURES

ILC KF LS MCP MLF MNV MoV MTV NAc

AC AId,v AON BC BST C CG CPU DBB DTN EN GI

anterior commissure agranular insular cortex, dorsal and ventral subdivisions anterior olfactory nucleus brachium conjunctivum bed nucleus of the stria terminalis clanstrum central gray caudate-putamen nucleus of the diagonal band of Broca dorsal tegmental nucleus endopiriform nucleus granular insular cortex

OB

OLT PBI,m PIR PSV

infralimbic cortex Kolliker-Fuse nucleus lateral septal nucleus middle cerebellar peduncle medial longitudinal fasciculus mesencephalic trigeminal nucleus motor trigeminal nucleus mesencephalic trigeminal tract nucleus accumbens olfactory bulb olfactory tubercle parabrachial nucleus, lateral and medial subdivisions piriform cortex principal sensory trigeminal nucleus

40 SOA T TTd,v

septo-olfactory area trapezoid nucleus taenia tecta, dorsal and ventral subdivisions

ACKNOWLEDGEMENTS

I II IV V-VI

lamina 1 of cortex laminae II, lII and IV of cortex laminae V and VI of cortex

use o f material p r o d u c e d in c o l l a b o r a t i v e efforts and

work was supported in part by USPHS Grants NS1275 and NS3346, and by American Heart Association, Missouri Affiliate Grants-in-Aid to the author and to Dr. Loewy. The author is recipient of

b o t h he an d Dr. F r e d P l u m f o r e n c o u r a g e m e n t and

T e a c h e r - I n v e s t i g a t o r D e v e l o p m e n t A w a r d NS0631-

helpful discussions o f the issues in this paper. This

01 from N I N C D S .

REFERENCES

demonstrated by retrograde transport of horseradish peroxidase, Brain Research, 89 (1975) 331-336. 9 Loewy, A. D. and Burton, H., Nuclei of the solitary tract: efferent projections to the lower brain stem and spinal cord of the cat, 3. comp. Neurol., 181 (1978) 421-450. 10 Norgren, R., Taste pathways to hypothalamus and amygdala, J. comp. Neurol., 166 (1976) 17-30. 11 Ricardo, J. A. and Koh, E. T., Anatomical evidence of direct projections from the nucleus of the solitary tract to the hypothalamus, amygdala, and other forebrain structures in the cat, Brain Research, 153 (1978) 1-28. 12 Saper, C. B., Convergence of autonomic and limbic connections in insular cortex of the rat, in preparation. 13 Saper, C. B. and Loewy, A. D., Connections of visceral afferent cortex in the rat, Neurosci. Abstr., 6 (1980) 116. 14 Saper, C. B. and Loewy, A. D., Efferent connections of the parabrachial nucleus in the rat, Brain Research, 197 (1980) 291-317. 15 Settlage, P. H., Bingham, W. G., Suckle, H. M., Borge, A. F. and Woolsey, C. N., The pattern of localization in the motor cortex of the rat, Fed. Proc., 8 (1949) 144. 16 Starzl, T. E., Taylor, C. W. and Magoun, H. W., Ascending conduction in the reticular activating system, with special reference to the diencephalon, J. Neurophysiol., 14 (1951) 461-477. 17 Woolsey, C. N., Settlage, P. H., Meyer, D. R., Soncer, W., Pinto Humay, T. and Travis, A. M., Patterns of localization in precentral and 'supplementary' motor areas and their relationship to the concept of a premotor area, Res. Publ. Assoc. Res. nerv. ment. Dis., 30 (1950) 238-264.

The a u t h o r thanks Dr. A r t h u r D. L o e w y for the

1 Beckstead, R. M., An autoradiographic examination of corticocortical and subcortical projections of the mediodorsal-projection (prefrontal) cortex in the cat, J. comp. Neurol., 184 (1979) 43-62. 2 Chu, N.-S. and Bloom, F. E., The catecholamine-conraining neurons in the cat dorsolateral pontine tegmenturn: distribution of the cell bodies and some axonal projections, Brain Research, 66 (1974) 1-21. 3 de Olmos, J., Hardy, I-L and Helmet, L., The afferent connections of the main and accessory olfactory bulb formations in the rat: an experimental HRP study, J. comp. Neurol., 181 (1978) 213-244. 4 Divac, I., Kosmal, A., Bjorkland, A. and Lindvall, O., Subcortical projections to the prefrontal cortex in the rat as revealed by the horseradish peroxidase technique, Neuroscience, 3 (1978) 785-796. 5 Hall, R. D. and Lindholm, E. P., Organization of motor and somatosensory cortex in the albino rat, Brain Research, 66 (1974) 23-38. 6 Krettek, J. E. and Price, J. L., The cortical projections of the mediodorsal nucleus and adjacent thalamic nuclei in the rat, J. comp. Neurol., 171 (1977) 157-192. 7 Krieg, W. J. S., Connections of the cerebral cortex. 1. The albino rat. A. Topography of the cortical areas, J. eomp. Neurol., 84 (1946) 275-277. 8 Llamas, A., Reinoso-Suarez, F. and Martinez-Moreno, E., Projections to the gyrus proteus from the brainstem tegmentum (locus coeruleus, raphe nuclei) in the cat,