After injecting retrograde tracer fiuoro-gold (FG) into the parabrachial region(PB), caudal ventrolateral medulla(CVLM) and the fourth segment of cervical spinal cord (C4), respectively, neurons in laminae I ～ Ⅱ of the medullary dorsalhorn projecting to the above mentioned brain areas were observed. PB received projections from bilateral laminae I and Ⅱ withan ipsilateral dominance; CVLM and C4 received projections from ipsilateral laminae I and Ⅱ. Neurons projecting to C4 werevery sparsely distributed in laminae I and Ⅱ of the medullary dorsal horn. The projecting neurons in outer part of lamina Ⅱwere more than those in inner part of lamina Ⅱ . Combined with immunofluorescence histochemistry for calbindin-D28k(CB) andparvalbumin(PV), it was demonstrated that a part of neurons projecting to PB or CVLM showed CB-like immunoreactivity, butnone of them exhibited PV-like immunoreactivity. There were only a few neurons in lamina Ⅱ projecting to C4 and they exhibitedneither CB- nor PV-like immunoreactivity. The present study provides further evidence for the existence of projecting neurons inlamina Ⅱ and suggests that immunostaining against CB and PV may distinguish two neuronal subpopulations in lamina Ⅱ .

A cerebellar afferent connection from the periaqueductal grey (PAG) has been demonstrated in the rat by means of retrograde transport of horseradish peroxidase (HRP) in the present study. The projection is bilateral, but the projection from the ipsilateral side is predominant (3:1). Its main origin is the ventromedial and ventrolateral regions of middle and caudal parts of PAG (98.8％), and the fibers reach different cerebellar cortical regions: culmen, declive, folium vermis, tuber vermis, pyramis vermis, uvula vermis, lobulus quadrangularis, crus Ⅰ, crus Ⅱ, and paraflocculus. Most labelled neurons are medium sized, but some small neurons also appear to project to cerebellum. Only a few large neurons are retrogradely labelled at the most caudal end of the caudal part. Functionally, both cerebellum and PAG are related to visceral activities. Consulting the present experiment, we discussed the significant role of the PAG-cerebellar projection.

In order to study the bifurcate projections of midbrain periaqueductal gray and nucleus raphe dorsalis to nucleus accumbens and nucleus raphe magnus, the fluorescence double-labeled method was used in the present study. Bisbenzimide (Bb) and propidium iodide (PI) were injected into nucleus raphe magnus and unilateral nucleus accumbens stereotaxically according to the time period necessary for their axonal transport. The percentages of double-labeled neurons were 21％; PI single labeled neurons were 32％; Bb single labeled neurons were 47％. Most of the labeled neurons were located in the middle and caudal parts of periaqueductal gray and the nucleus raphe dorsalis, and most were medium sized and fusiform and triangular in shape.

The distributions of 5-HT-like and Met-ENK-like immunoreactive (5-HT-LI and Met-ENK-LI) structures in the nucleus accumbens (Acb) of rats were studied by immunohistochemical technique in the present study. Under light microscope, 5-HT-LI fibers and terminals could be seen in each subnucleus at different planes of Acb, but the 5-HT-LI fibers and terminals in the medial and ventral subnuclei were more than the dorsal and lateral subnuclei, the amount of 5-HT-LI fibers and terminals in the caudal segment were more than the rostral segment. According to the diameter, pathway, and number of varicosity, 5-HT-LI fibers could be divided into 3 types: (A) thick fiber (0.35—0.40?m); (B) medium fiber(0.20—0.30?m); (C) thin fiber (about 0.10?m). These 3 types of 5-HT-LI fibers were remarkable in the medial and ventral subnuclei of Acb. 5-HT-LI neuronal bodies did not observed in the Acb. A few scattered Met-ENK-LI neuronal bodies were seen in the ventral subncleus and ventral part of the medial subnucleus. Met-ENK-LI fibers and terminals distributed in all subnuclei and predominant in the medial and ventral subnuclei. The distributions of Met-ENK-LI structures were no differences between the rostral and caudal segments. All of the Met-ENK-LI fibers were thin and irregular and villi-like in shape. There were only a few varicosities on the MetENK-LI fibers. Part of Met-ENK-LI fibers looked like discontinued varicosities. Under electron microscope, 5-HT-LI axonal boutons formed symmetric and asymmetric synapses with non-5-HT-LI dendrites. Met-ENK-LI dendrites formed symmetric and asymmetric axo-dendritc synapses with non-Met-ENK-LI axonal boutons. These synapses were mainly observed in the medial and ventral subnuclei of Acb. The identity of 5-HT-LI and Met-ENK-LI structures, especially in the medial and ventral subnuclei, supported the physiological studies that 5-HT-LI ascending efferent fibers activated the Met-ENK-LI neurons and then the latter sent descending efferent fibers to lower brainstem structures to take part in antinociceptive functions.

5-HT-, SP- and L-ENK-like (5-HT-LI, SP-LI and L-ENK-LI) immunoreactive ultrastructures of the ventrolateral subdivision (VLS) of midbrain periaqueductal gray (PAG) in the rat were observed by immunoelectron microscopical technique in the present study. 5-HT-LI neuronal cell bodies were frequently seen in the VLS of the PAG. 5-HT-LI dendrites were found to form axo-dendritic synapses with non-immunoreactive axon terminals, and the major form of this kind of synapse was asymmetric. A few 5-HT-LI axon terminals formed axo-axonic synapses with non-5-HT-LI axon boutons. 5-HT-LI and non-5-HT-LI axon boutons formed axo-dendritic and axo-somatic synapses with 5-HT-LI dendrites, non-5-HT dendrites and 5-HT-LI neuronal cell bodies, respectively. SP-LI fusiform neuronal cell bodies were only a few and formed axo-somatic synapses with non-SP-LI synaptic boutons which contained pleomorphic vesicles. SP-LI dendrites formed axo-dendritic synapses with non-SP-LI axon terminals, this kind of synapse was the main form of synapses formed by SP-LI structures. SP-LI axon boutons formed axo-somatic and axo-dendritic synapses with non-SP-LI neuronal cell bodies and SP-LI dendrites. L-ENK-LI neuronal cell bodies were also limited. The most common form of synapses of L-ENK-LI structures was L-ENK-LI dendrites formed axo-dendritic synapses with non-L-ENK axon terminals. Non-L-ENK-LI axon terminals constituted axo-somatic synapses with L-ENK-LI neuronal cell bodies. A few L-ENK-LI axon terminals formedaxo-dendritic synapses with L-ENK-LI dendrites. The majority of the synapses mentioned above contained spherical clear vesicles, but some were mixed with a few flat, oval and granular vesicles. The immunoreacrive products were located on the surface of vesicles or on the surface of membranous organelles in the cytoplasm.

The ultrastructure of the rat nucleus raphe magnus (NRM) was observed under transmission electron microscope in the present study. Most of the neurons of the NRM were medium and small sized fusiform and triangular in shape, they had spherical or ellipsoid nuclei and rather high nucleo-cytoplasmic ratio. Nuclear rodlets which composed of parallel filaments could be seen in some fusiform NRM neurons. There were numerous organelles in the cytoplasm. Axonal terminals apposed to most neuronal bodies and formed axo-somatic synapses. The predominant type of these synapses was symmetric. Sometimes, rod-like or spine-like cytoplasmic protrusions could be seen on the neuronal bodies, they often made axo-somatic synapses with axonal terminals. The neuropil of the NRM was quite complex. It was formed by transverse sections of myelinated fibers, unmyelinated fibers, synapses, and neuroglia. The axo-dendritic synapses were the major synaptic type in the neuropil. The predominant type of these axo-dendritic synapses was also symmetric and asymmetric axo-dendritic synapses. Some axonal terminals were arranged parallel with dendrites and formed symmetric synapses. Beneath subsynaptic membrane of some postsynaptic bags, there were some electrical dense spherules or bands which formed subsynaptic dense bodies. There were. no typical axo-axonic synapses in the NRM, but the parallelly arranged axonal terminals were often seen. Most of the presynaptic bags contained clear spherical vesicles or mixed with flattened and granular vesicles. Some postsynaptic bags were filled with flattened vesicles, they were also often mixed with spherical and granular vesicles.

The submicroscopic structure of nucleus accumbens (Acb) was observed under transmission electron micrscope in the present study. Most of the Acb neurons were spherical, with diameter ranging from 10 ?m to 15 ?m. The nucleus were ellipsoid with, sometimes, nucleolus and deep invaginations of nuclear membrane. The nucleocytoplasmic ratio was always rather high. Axo-somatic synapses were formed on the surface of most neuronal perikarya, predominantly of the symmetric type. The percentage of axo-somatic synapses in total number of Acb synapses was 2%. Generally, the outline of the dendrites was regular. Most of the thick dendrites made symmetric axo-dendritic synapses while the thin dendrites usually made asymmetric axo-dendritic synapses. Some of the irregular dendrites had spines. The spines were usually long and thin, with enlarged end and markedly thin neck, and formed asymmetric synapses with axonal terminals. The axo-dendritic synapes composed 97% of the total synapses within Acb. Small sized axons were frequently observed. The majority of the axon terminal segments contained clear round vesicles and made synapses en passant with the dendrites. Synapses en passant were also observed on the collaterals of axon. Two or more axonal terminals were often closely contacted, but no typical synaptic formations were found. Only a few axo-axonic synapses could be observed. Axoaxonic-dendritic synaptic complexes were also seen. The percentage of axo-axonic synapse was 0.5%. Most of the synaptic boutons mentioned above were mainly filled by spherical vesicles, some of them were filled with pleomorphic vesicles.

The transmission electron microscopic technique was used in the present study to observe the normal ultrastructure of the ventrolateral subdivision (VLS) of the rat midbrain periaqueductal gray (PAG). Most of the neurons in the VLS of PAG had medium-sized fusiform neuronal cell body, ellipsoid nucleus and irregular nuclear membrane with deep invaginations or infoldings. The nucleolus was electron dense, large and spheroidal. There were many kinds of organelles in the cytoplasm and the Golgi apparatus in some cells were circular or semicircular. The percentage of axo-somatic synapse in total synapses in VLS was 6% and most of them was of symmetric type (82%). Axo-dendritic synapses counted 94%, of these 66% were symmetric type, 34% were asymmetric. Some special types of axo-dendritic synapse were observed. e. g., axo-spine synapse, parallel synapse, and central dendritic synaptic glomerulus. The synaptic vesicles in the presynaptic bags of the synapses mentioned above were mainly clear spherical in type, and most of the clear spherical vesicle-filled bags contained a few flattened vesicles and granular vesicles. Only a few bags were mainly filled with flattened vesicles.

Immunoelectron microscopic technique was used in the present study to observe the serotonin (5-HT)-, substance P (SP)-, and leucine-enkephalin (L-Enk)-like immunoreactive ultrastructures in the nucleus raphe magnus (NRM) of the rat. 5HT-like immunoreactive (5-HT-LI) axonal terminals were found to form axosomatic, axo-dendritic, and axo-axonic synapses with non-5-HT-LI neuronal cell bodies, 5-HT-LI and non-5-HT-LI dendrites, and non-5-HT-LI axonal terminals respectively. Non-5-HT-LI axonal terminals formed axo-somatic and axo-dendritic synapses with 5-HT-LI neuronal cell bodies and dendrites. SP-LI (or L-Enk-LI) axonal terminals formed axo-somatic, axo-dendritic synapses with SP-LI (or L-Enk-LI) and non-SP-LI (or non-L-Enk-LI) neuronal cell bodies and dendrites, respectively. L-Enk-LI axonal terminals constituting axo-axonic synapses with L-Enk-LI axonal boutons were observed less frequently. The most common synaptic type made by 3 kinds immunoreactive profiles mentioned above was axo-dendritic synapses made by non-immunoreactive axonal terminals with immunoreactive dendrites. The majority of the immunoreactive axonal boutons were mainly filled by clear spherical vesicles, but sometimes were mixed with small number of flat and granular vesicles. The immunoreactive products were irregular electron-dense substances, and were located on both inner and outer surfaces of the vesicles, or on the surface of the membranous cell organelles in the cytoplasm, etc.

Objective To investigate the spatio-temporal expression of P2X3 receptor ( P2X3R) in rats with diabetic mechanical allodynia ( DMA ) .Methods DMA model in rats was induced by intraperitoneal injection of streptozocin ( STZ) .The von Frey filaments were applied to identify the changes of the paw withdrawal threshold ( PWT) in DMA rats.Immunofluorescence was employed to detect the spatio-temporal expression of P2X3R in the spinal dorsal horn (SDH), dorsal root ganglion (DRG) and hind paw skin on different time points after intraperitoneal injection of STZ , respectively.The protein expression of P2X3R in SDH and DRG was further semi-quantitatively analyzed by Western blotting.Results Compared with control group , intraperitoneal injection STZ induced significant mechanical allodynia indicated by the reduced PWT from 7 days, and which reached the peak on 14d and maintained to 28days (P<0.05). The expression of P2X3R in DRG neurons was significantly increased on 14 days and 21 days (P<0.05), while that in SDH and skin was markedly increased on 21 days and 28 days, compared with control group (P<0.05).Conclusion With the progress of DMA, the expression of P2X3R was significantly increased in the SDH, DRG and skin, which was almost parallel with the mechanical allodynia , but the changes in SDH and skin were 1 week later than that in DRG .These results suggest that P2X3R may play a key role in the maintenance of the DMA .