| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Pharmacology, University of Virginia, Charlottesville, Virginia 22908
Submitted 6 June 2003 ; accepted in final form 18 August 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
60–70% on the SSP-SAP-injected side. SSP-SAP spared VLM neurons devoid of NK1R-ir, such as the catecholaminergic cells and the precerebellar glutamatergic neurons. In the pre-Bötzinger region of the VLM the toxin killed glutamatergic neurons while sparing glycinergic and GABAergic inhibitory neurons. In the CVLM region 26% of the inhibitory cells were destroyed. In conclusion, the baroactivated depressor neurons of the CVLM do not appear to express NK1Rs but their activity is probably modulated by a population of excitatory NK1R-ir cells located in the VLM. The results also suggest that a region located below the CVLM (subCVLM) may contain an unrelated population of GABAergic depressor neurons that are NK1R-ir but are either not barosensitive or do not express Fos during baroreceptor stimulation.
sympathetic tone; blood pressure; respiration; pre-Bötzinger complex; respiratory rhythm generation; substance P; saporin
SSP-SAP lesions of the VLM NK1R-expressing neurons also reduce the hypotensive response produced by microinjecting the excitatory amino acid DL-homocysteic acid (DLH) into the caudal part of the VLM (CVLM) (37). This depressor response is classically attributed to the direct or indirect depolarization of a group of GABAergic interneurons whose hallmark is to be vigorously activated by stimulation of arterial baroreceptors (1, 29). These neurons, henceforth called CVLM baroactivated depressor neurons, contribute to the sympathetic baroreflex by inhibiting the presympathetic neurons of the rostral VLM (RVLM) during baroreceptor stimulation (41–43). The loss of DLH-induced hypotension after SSP-SAP injection into the VLM (37) indicates that the neural network involved in regulating sympathetic tone is also affected by the loss of some of the VLM NK1R-expressing neurons.
The present study is designed to examine why selective lesions of NK1R-ir neurons with SSP-SAP attenuate the depressor response to DLH injection into the CVLM. To ascertain that the lesion attenuates DLH-induced hypotension by preventing sympathoinhibition, we first repeated the initial observation of Wang et al. (37) while measuring the splanchnic sympathetic outflow. Next, we tested whether unilateral lesions of VLM NK1R-ir cells with SSP-SAP attenuate the activation of CVLM depressor neurons by baroreceptor stimulation in alert rats. Activation of these cells was gauged by Fos expression (4, 19, 39). After that, we examined whether the baroactivated depressor neurons express NK1Rs and therefore should be destroyed by SSP-SAP. This not being the case, we finally examined the phenotype of the VLM neurons that are destroyed by the toxin. We found that SSP-SAP destroys mostly excitatory neurons in the pre-Bötzinger region, whereas in the CVLM area, a moderate number of GABAergic and/or glycinergic neurons are eliminated. The results suggest that a heterogeneous population of VLM NK1R-expressing neurons contributes to the sympathoinhibition, which results from stimulating the CVLM with excitatory amino acids.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Unilateral selective lesion of NK1R-ir neurons in VRG. Surgical anesthesia was induced with a mixture of ketamine (75 mg/kg), xylazine (5 mg/kg), and acepromazine (1 mg/kg im). If insufficient, this dose was occasionally supplemented with 10% of the initial amount. The saporin conjugate SSP-SAP or free SAP (both from Advanced Targeting Systems; San Diego, CA) was administered into the left ventrolateral medulla by pressure injection (model PLI-100, Medical Systems; Greenvale, NY) using glass micropipettes with a tip diameter of 25 µm as described previously (37). Briefly, a longitudinally oriented row of three injections was made into the left VLM. Each injection consisted of 0.313 ng of SSP-SAP in 50 nl of sterile physiological saline delivered in several 5- to 10-nl boluses over 2 min. The first injection was made 1.9 mm lateral to the calamus scriptorius and 2.4 mm below the dorsal surface of the brain stem with the electrode pointing 30° from the vertical toward the front of the animal. The other two injections were placed 0.5 and 1 mm rostral to that point. Control rats received three injections of unconjugated SAP (0.313 ng and 50 nl each) at the same locations. After surgery, the rats were treated with an antibiotic (125 mg/kg im ampicillin; Bristol-Myers-Squibb; Princeton, NJ) and an analgesic (0.6 mg/kg sc ketorolac) and were returned to standard housing conditions. Animals were allowed to survive 2–3 wk. They did not exhibit any noticeable behavioral abnormality and grew normally.
Physiological recordings in anesthetized rats. The experiments were done in urethane-anesthetized and ventilated rats as described previously (14, 37) except that to preserve intact baroreflexes, the rats were not vagotomized. Briefly, anesthesia was induced with 5% halothane in 100% oxygen and maintained during surgery with 1.5–1.7% halothane in 100% oxygen via a tracheal cannula (60 cycles/min; 1 ml/100 g). End-expiratory CO2 was kept between 4.5 and 5% and rectal temperature between 37.5° and 38.5°C. The femoral artery and the femoral vein were catheterized to record arterial blood pressure (BP) and to administer drugs, respectively. The splanchnic nerve was dissected through a retroperitoneal approach, and bipolar recordings were made from a segment of nerve located distal to the suprarenal ganglion (14). The nerve and the bipolar electrode were insulated with polyvinylsiloxane impression material (Carlisle Laboratories; Rockville Center, NY). A concentric bipolar stimulating electrode (Rhodes Medical Instruments; Woodland, CA; diameter 250 µm; tip separation 500 µm) was placed in the fascia surrounding the mandibular branch of the facial nerve on the right side of the rat. This electrode was used to locate the caudal pole of the facial motor nucleus by means of antidromic field potential recordings (monophasic square pulses, 100 µs, 0.5–2 mA, 1 Hz) (3). After completion of surgery, halothane was slowly replaced by urethane 1.0 g/kg iv with supplements of 0.1 g/kg each approximately every 90 min or as required. After 1-h equilibration, the muscle relaxant pancuronium was administered (1 mg/kg iv, with 0.3–0.5 mg/kg supplements as required) and electrophysiological recordings were initiated. The end-expiratory CO2 was maintained at 4% to 4.5%. Under paralysis, anesthesia was maintained at the level at which a strong pinch of the hindpaw produced negligible effects on phrenic nerve discharge rate and blood pressure (<10 mmHg). DLH (10 mM in sterile physiological saline) was mixed with 1% green fluorescent beads (Lumafluor; Naples, FL) to identify the injection sites. Pressure injections of 10 nl were made as described previously (37). The first injection was placed in the RVLM on the left side at the level of the caudal end of the facial motor nucleus. The pipette tip was guided to this site by first mapping the field potential elicited in the facial motor nucleus by stimulating the mandibular branch of the facial nerve (37). Subsequent injections were made by moving the location of the pipette tip in steps of exactly 200 µm in the caudal direction. A total of 12 injections separated by 5 min were made 0–2,200 µm behind the reference point (caudal end of facial motor nucleus) (bregma –11.7 to –13.9 mm). Then the pipette was moved symmetrically to the contralateral side (3.8 mm to the right) and the process of sequential injections was repeated.
The splanchnic nerve discharge (SND) was amplified, filtered (200–3,000 Hz), full-wave rectified, and integrated using a sample hold integrator with a 0.1-s reset time (15). Analog signals (BP, end-expiration CO2, raw SND, and integrated SND) were digitized using a Power 1401 interface and version 3 of the Spike2 software (both from Cambridge Electronics Design; Cambridge, UK). The noise level of the SND recordings was identified at the beginning of the experiment during injection of a bolus of L-phenylephrine hydrochloride (PE; 2 µg/kg; Sigma; St. Louis, MO) that saturated the baroreflex. At the end of the experiment the noise level was again determined by measuring the residual activity after administration of a high dose of the sympatholytic drug clonidine (0.1 mg/kg iv). The two noise determinations matched, and their mean value was subtracted from the recorded voltage to measure the amplitude of the SND. All physiological variables [arterial pressure (AP), end-expiration CO2, and integrated SND] were analyzed and measured with the Spike 2 software. The effect of DLH injection on BP and SND is expressed as a percent change from the preinjection baseline. Baseline values were measured during the 10 s preceding DLH injection. Peak effects were measured by averaging BP and SND during a period corresponding to three ventilation cycles (3 s) centered on the visually identified nadir or apex of the responses.
Production of hypertension in conscious rats. Surgery was performed while the rats were under anesthesia with ketamine, xylazine, and acepromazine (see above). Postsurgical treatment consisted of ampicillin and ketorolac as described above. Rats used for Fos histology were instrumented with a single catheter in the left jugular vein that was exteriorized at the back of the neck. The animals were then placed on a tether in a recording chamber located in a quiet environment at 75°F with free access to food and water and their normal 12-h light-dark cycle. Two days later, a 25-min intravenous infusion of a fixed amount of the vasoconstrictor PE was administered to the rats to elevate their blood pressure (0.6 µg/µl in sterile saline; 12 µl/min; 18–22 µg · kg–1 · min–1). Control rats received a saline infusion matched in volume and duration. These animals were euthanized 2 h after the end of the intravenous infusion of PE or saline with an overdose of urethane. They were then fixed with formaldehyde as described below.
The effect produced by infusion of PE or saline on BP and HR was examined in a separate group of 10 age-matched rats instrumented with a femoral artery catheter in addition to the venous catheter. These rats were not used for Fos histology because placement of a femoral catheter was found to produce an unwelcome background Fos expression in the CVLM and nucleus tractus solitarius (NTS). In 5 rats PE increased BP by an average of 46.4 ± 2.2 mmHg from a baseline level of 118.3 ± 2.0 mmHg (P < 0.05) and decreased HR by 118.6 ± 10.0 beats/min (from a resting level of 378 ± 5 beats/min). BP returned to control levels within 10 min after the end of the PE infusion. In the other five rats, saline infusion was without effect on BP [resting level: 126.2 ± 2.6 mmHg; mean level during saline infusion: 123.1 ± 2.7; not significant (NS)] or HR (resting level: 367.9 ± 3.3 beats/min; mean level during saline infusion: 360 ± 2.9; NS).
Histology. After completion of the DLH microinjections (anesthetized rats) or the infusion of PE or saline to awake rats, a lethal dose of urethane was administered. Immediately after cessation of breathing movements the rats were perfused transcardially with 250 ml of PBS (pH 7.4), followed by 500 ml of 4% phosphate-buffered (0.1 M; pH 7.35) paraformaldehyde (EM Sciences; Ft. Washington, PA). The brain stem was removed and postfixed overnight with the same fixative at 4°C. Series of coronal sections (30 µm) were cut through the medulla oblongata with the use of a vibrating microtome (model VT 1000S, Leica Instruments; Nussloch, Germany).
Immediately after the brain was sectioned, a preliminary check of the accuracy of the DLH injection sites was made by examining the location of the fluorescent microbeads using an ordered 1/6 series of sections 180 µm apart, which were mounted, air dried, and coverslipped with Krystalon (EM Industrial; Gibbstown, NJ). More precise identification of the injection sites was made later in material processed for the immunohistochemical detection of tyrosine hydroxylase (TH) and NK1R-ir. All other histological procedures were done using sections stored in cryoprotectant solution at –20°C for up to 2 mo (30% RNAse free sucrose, 30% ethylene glycol, and 1% polyvinylpyrolidone-40 in 100 mM sodium phosphate buffer, pH 7.4).
Standard immunofluorescence procedures were used to detect Fos-ir nuclei, TH, NK1Rs, or pairs of these markers (for details, see Ref. 38). For the simultaneous detection of NK1R and TH, sections were incubated for 72 h at 4°C with a rabbit polyclonal antibody against NK1R (Chemicon International; Temecula, CA; 1:1,000 dilution) and a mouse monoclonal antiserum against TH (Chemicon, 1:2,000 dilution), after which the sections were incubated for 45 min with species-specific secondary antibodies F(ab')2 fragment goat anti-rabbit IgG conjugated to Cy3, dilution 1:200 (Jackson ImmunoResearch Laboratories; West Grove, PA); goat antimouse IgG conjugated to Alexa 488; 1:200 dilution (Molecular Probes; Eugene, OR). In some cases, immunoreactivities to Fos and NK1R were detected simultaneously by incubating sections for 24 h at 4°C with a rabbit polyclonal antibody against Fos (sc-52, Santa Cruz; Santa Cruz, CA; 1:2,000 dilution) and a guinea pig antibody against the NK1R (NK1 gp, Chemicon, 1: 2,000 dilution) after which the sections were incubated for 45 min at room temperature with species-specific secondary antibodies conjugated with a fluorochrome (goat anti-guinea pig IgG conjuguated with Cy3; Jackson ImmunoResearch, 1:200 dilution and goat anti-rabbit IgG conjuguated with Alexa 488; Molecular Probes; 1:200 dilution). In some cases NK1R-ir was detected using the rabbit anti-NK1R antibody (Chemicon; 1:2,000 dilution; 48 h at 4°C) and a biotinylated donkey anti-rabbit IgG (Jackson; 1:400), followed by Vector ABC Elite kit and Vector VIP reagent as the reporter (Vector; Burlingame, CA).
The sections were mounted in sequential order onto gelatin-coated slides. All sections except those processed with the Vector VIP method were dehydrated and delipidated through graded alcohols and xylenes. Finally, coverslips were affixed with DPX mounting medium (Aldrich; Milwaukee, WI). Sections with Vector VIP reagent were dipped for 15 s in 95% and 100% ethanol, followed by xylenes, and coverslips were affixed with Vectamount (Vector). No staining was observed in the absence of the primary antibodies. The distribution of NK1R-ir and TH-ir neurons within the VLM conformed to prior descriptions (37).
The mRNAs encoding glutamic acid decarboxylase-67 (GAD-67), glycine transporter 2, or vesicular glutamate transporter 2 were detected by in situ hybridization using digoxigenin-labeled probes. Previously described methods were used without modification (31–33). The plasmid (pBluescript SK+) containing the full-length 3.2-kb GAD-67 cDNA was kindly supplied and previously characterized by A. J. Tobin of the University of California at Los Angeles (32). A pBluescript plasmid containing a 3.1-kb partial cDNA insert encoding glycine transporter 2 was kindly provided by Dr. N. Nelson of Tel Aviv University (for details, see Ref. 31).
Mapping and imaging. Each histochemical procedure was performed using one in six series of sections that were mounted on slides in sequential rostrocaudal order. The section that contained the caudal end of the facial motor nucleus was identified under darkfield illumination and was assigned the level bregma –11.6 mm after the nomenclature of Paxinos and Watson (23). Levels caudal to this reference section were determined by adding a distance corresponding to the interval between sections (180 µm) multiplied by the number of intervening sections. The rostrocaudal levels reported in the present paper are therefore directly comparable to those reported in all our prior publications. They correspond very closely to the bregma levels reported in the atlas of Paxinos and Watson (23).
Sections were examined with a Leitz microscope using brightfield or epifluorescence as appropriate. The section outlines, major landmarks and the location of the cells of interest were drawn or plotted using a Lucivid camera (MicroBrightfield; Colchester, VT) and a motor-driven microscope stage (Ludl Electronic Products; Hawthorne, NY) controlled by the Neurolucida software (MicroBrightfield) as described previously (32, 37). As a general rule, cell nuclei were clearly visible and only cell profiles that included a nucleus were counted. The only exception was the NK1R-ir neurons in which immunoreactivity lined the somatic and dendritic membranes. NK1R-ir profiles were counted as cells only when a soma of at least 10 µm in diameter contiguous with two or more primary dendrites was present. Cell counts were not corrected for section thickness because we sought to determine the percent changes in cell numbers caused by a lesion, not the absolute numbers of cells. The Neurolucida files were exported to the Canvas software drawing program (Deneba Software; Miami, FL) for final modifications and printing.
Photographs of the fluorescent material labeled with Alexa 488 and Cy3 were made using a two-color Olympus BX50 WI confocal microscope equipped with a Krypton and Argon laser as described previously (37). The images were scanned through x40 or x60 objectives, acquired at a resolution of 1,024 x 1,024 pixels and stored in 24-bit TIFF format. TIFF files were imported into Adobe Photoshop (version 5.0.1; Adobe Systems, Mountain View, CA). Nonradioactive in situ hybridization (brightfield) or darkfield-illuminated sections were photographed with a 12-bit color CCD camera (Cool-Snap, Roper Scientific, Tuscon, AZ; resolution 1,392 x 1,042 pixels) (for details see Ref. 35).
Statistics. All data are presented as means ± SE. Groups were compared by one-way ANOVA, ANOVA with repeated measures, followed by a Tukey-Kramer post hoc test or paired t-test as appropriate; significance was defined as P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Microinjections of DLH were made on both sides of the VLM in 6 rats that had received SSP-SAP into the left VLM 2 wk before the physiological experiments. The DLH was targeted to the respiratory column as described before (37), and injections were spaced at 200 µm intervals from the caudal end of the facial motor nucleus (bregma –11.5 mm to bregma –13.9 mm). DLH produced the same effects on BP as in our prior experiments (37). On the intact side of the medulla oblongata, a small and short-lived increase in BP was observed after DLH injection at the most RVLM levels (Fig. 1B, trace not shown). Caudally, the effect of DLH gradually changed into a large and long-lasting drop in BP (Fig. 1, A and B). The DLH-induced pressure drop reached a maximum around bregma –13.3 mm and rapidly waned at more CVLM levels (Fig. 1B, level –13.9 mm). On the toxin-injected side, DLH decreased BP and SND very little when injections were targeted to the CVLM (bregma levels –12.7 to –13.9 mm; representative trace in Fig. 1A lesioned side; summary in Fig. 1B) but the pressor and sympathoexcitatory effects of DLH injections into the RVLM (levels –11.5 to –12.3 mm) were the same as on the contralateral side (Fig. 1B). The changes in SND caused by DLH were consistent in amplitude and kinetics with the changes in BP as illustrated in Fig. 1, A and C.
|
The medulla oblongata of the six rats was processed for identification of NK1R-ir and TH-ir cells. SSP-SAP produced a virtually complete loss of NK1R-expressing neurons in the entire ventrolateral quadrant of the medulla oblongata in all rats. Within the region of the VRG (rVRG), NK1R-ir neurons were essentially undetectable at the level of the pre-BötC (bregma level: –12.4 to –12.9 mm, Fig. 2, C and D) and caudal to that level down to at least bregma level –13.5 mm (bregma level –13.0 mm illustrated in Fig. 2, E and F). The intensely stained neurons that are typical of the pre-BötC area and immediately adjacent rVRG were entirely destroyed (Fig. 2, C and D). In addition, SSP-SAP also consistently eliminated the somewhat less intensely stained NK1R-ir cells that lie between the VRG and the ventral surface of the medulla (Fig. 2, C–F). A substantial collection of these medium intensity NK1R-ir cells lies in and below the CVLM region. This region, henceforth called the subCVLM, is identified by a dotted trace in Fig. 2F.
|
The lesions, although large, had consistent boundaries. The NK1R-ir neurons located within the compact portion of the nucleus ambiguus were only marginally affected (white arrowheads in Fig. 2, A–D). NK1R-ir cells located dorsal to this nucleus or within the trigeminal nucleus or in the midline medulla were spared (not shown). The NK1R-ir neurons located at the RVLM close to the ventral surface of the brain (retrotrapezoid region) were also spared (Fig. 2, A and B). Finally, the contralateral uninjected side of the medulla also appeared intact (see also below). The TH-ir neurons of the VLM, though intermingled with the NK1R-ir cells, appeared histologically intact at all levels (green cells in Fig. 2, A–F). To quantify the extent, selectivity and reproducibility of the lesions, we counted the number of NK1R- and TH-ir neurons located at various levels of the VLM in four of the six rats. Neurons were counted within the lateral quadrant of the medulla oblongata defined in Fig. 3A. While the reduction of the number of NK1R-ir neurons was massive and highly significant statistically (Fig. 3B), the number of TH-ir neurons was unchanged throughout the VLM (Fig. 3C).
|
CVLM neurons that express Fos after treatment with PE do not express detectable levels of immunoreactive NK1 receptors. The simplest explanation for the loss of DLH-induced sympathoinhibition after treatment of the VLM with SSP-SAP would be that the toxin physically destroys the GABAergic interneurons of the CVLM the activation of which is thought to mediate the hypotension elicited by baroreceptor activation (12, 16, 29, 30). These neurons (CVLM baroactivated depressor neurons) can be identified histologically because they express Fos in animals subjected to a short episode of PE-induced hypertension (4, 20, 29, 39). Thus to test whether the baroactivated depressor neurons express NK1Rs we infused PE intravenously to a group of four awake rats and we determined whether any of the Fos-ir neurons of the CVLM were NK1R-ir. Four rats infused with saline served as controls.
The CVLM of the PE-infused rats contained a tight cluster of Fos-ir nuclei (Fig. 4A), whereas the corresponding area of control rats had very few Fos-ir nuclei (Fig. 4B). NK1R-ir was abundant in the region of the Fos-ir neurons (red processes in Fig. 4, A and B, insets), but it was undetectable in the Fos-ir neurons (Fig. 4A, inset). The results were analyzed in quantitative fashion by counting the number of Fos-ir cells present in the left VLM from bregma –12.1 to –13.6 mm (9 sections). As shown in Fig. 5, administration of PE increased significantly the number of Fos-ir present in the VLM, relative to the control group. This increase was particularly dramatic caudal to bregma –12.6 mm. NK1R-ir was detected in no more than one Fos-ir neuron per PE-treated rat.
|
|
Unilateral lesions with SSP-SAP reduce the number of VLM neurons that express Fos after PE infusion in awake rats. The next experiment was designed to test whether lesion of the NK1R-ir neurons of the VLM interferes with the ability of the baroactivated depressor neurons to express Fos after PE-induced hypertension. SSP-SAP was injected into the left VLM of four rats and a group of four control rats received injections of unconjugated SAP (control toxin) in the same location. Both groups of rats received an infusion of PE while alert 2 to 3 wk after the injection of SSP-SAP or SAP, and their medulla oblongata was processed to reveal Fos-ir nuclei and NK1R-ir cells.
SSP-SAP again produced a near complete disappearance of NK1R-ir processes in the VLM, whereas unconjuguated SAP produced no detectable effect (Fig. 6, A–D). The lesion caused by SSP-SAP appeared strictly unilateral given that the uninjected side of the rats treated with SSP-SAP was undistinguishable from that of rats treated with SAP or with nothing at all (compare Fig. 6, B and D, and Fig. 2). The number of Fos-ir nuclei was considerably reduced in SSP-SAP-treated rats on the lesioned side (Fig. 6A) compared with the control side of the same rats (Fig. 6B) or compared with either side of the rats treated with the unconjugated SAP (Fig. 6, C and D). In contrast, the number of Fos-ir neurons present in the nucleus of the solitary tract was unaffected by treatment with SSP-SAP indicating that the toxin did not interfere with processing of baroreceptor information upstream of the CVLM. Figure 6E illustrates the appearance of the NTS dorsomedial to the tractus solitarius on the side ipsilateral to an injection of SSP-SAP into the VLM. Figure 6, F and G, shows that this region contained approximately the same number of Fos-ir nuclei in a rat that received an injection of unconjuguated SAP into the VLM (Fig. 6F) or in a rat that received no injection of any kind (Fig. 6G). Finally, Fig. 6H shows that this region of the NTS did not contain Fos-ir nuclei in a rat that received saline instead of PE. These results were quantified by calculating the total number of Fos-ir neurons present in 3 anatomically matched coronal sections of the CVLM in each rat (bregma –12.8 to –13.4 mm) and in three anatomically matched sections of the NTS (mid-area postrema level and two more caudal sections separated by 180 µm). Counts were made on each side of the midline. As shown in Fig. 7A, the number of Fos-ir neurons detected in the CVLM was significantly reduced on the side injected with SSP-SAP. The number of Fos-ir neurons present in the uninjected side was the same in rats treated with SSP-SAP or unconjuguated SAP, and the number of Fos-ir neurons was not significantly reduced on the SAP-injected side. Finally, as shown in Fig. 7B, the number of Fos-ir neurons present in the nucleus of the solitary tract was unaffected by unilateral injection of either SSP-SAP or SAP into the VLM. NK1R-ir in the NTS was also unaffected by SSP-SAP injections into VLM (result not illustrated).
|
|
Effect of SSP-SAP on GABAergic and glycinergic neurons in VLM. As shown before, injection of SSP-SAP into the VLM spared the TH-ir neurons, consistent with the fact that these cells lack NK1R-ir (6, 37). Clearly, several VLM glutamatergic cells are killed by SSP-SAP because virtually all the highly NK1R-ir cells of the VLM that reside in the pre-Bötzinger region contain the mRNA that encodes vesicular glutamate transporter 2 (9) and these cells were destroyed by the toxin (Fig. 2C). The next experiments were designed to determine whether SSP-SAP also destroys GABAergic and glycinergic neurons in the VLM and whether the loss of this type of cell is region specific. Two areas were selected for analysis namely the pre-Bötzinger region that probably plays a specialized role in respiration rhythm generation and the CVLM that is usually associated with blood pressure control, although this region overlaps considerably with the rVRG (34, 39).
These experiments were done in three additional rats that received unilateral injections of SSP-SAP into the VLM according to the standard protocol and, 2 wk later, received an infusion of PE while awake so that CVLM neurons could also be identified by the presence of Fos-ir. The expected loss of NK1R-ir neurons in the VLM was verified by processing a series of coronal sections for NK1R-ir (data not shown).
One series of coronal brain sections was reacted for identification of neurons containing GAD-67 mRNA, Fos-ir, and TH-ir. The three brain sections that contained the highest number of Fos-ir neurons in the VLM were selected as representative of the CVLM region. A 750 µm x 750 µm box was drawn around the cluster of Fos-ir neurons on the intact side that defines the CVLM (Fig. 8A). An identical box was placed in a symmetrical location on the lesioned side, and neurons of interest were counted within these boxes. As shown in Table 1, the toxin reduced massively (63%) the number of Fos-ir cells that contained GAD-67 mRNA, whereas the number of TH-ir neurons located within the same region was unaffected. Interestingly, the number of TH neurons that expressed Fos was also unchanged by the toxin. Thus the toxin reduced the activation of the GABAergic baroactivated depressor neurons by hypertension but it did not alter the activation of the catecholaminergic neurons by the same stimulus. The toxin also reduced the total number of GAD-67-containing neurons within the same region by 26% (Table 1). The loss of GABAergic neurons in the region of the CVLM is illustrated in Fig. 9, A1–A2. Note that the neuronal loss was most prominent in the region defined previously as the subCVLM.
|
|
|
To determine whether the loss of GABAergic neurons was specific to the CVLM or uniform throughout the VLM we also counted the number of GAD mRNA expressing neurons within the pre-Bötzinger region. This region was defined by a 750 µm x 750 µm box placed just below the compact nucleus ambiguus with its medial edge aligned with the lateral edge of the inferior olive as indicated in Fig. 8B. The box was centered on the region of the VLM that contains the intensely NK1R-ir neurons (Fig. 2D). Another series of sections was processed for identification of Gly-T2 mRNA and Fos. This material was used to count the number of glycinergic neurons present in the CVLM and the pre-Bötzinger region also as defined in Fig. 8, A and B. As shown in Fig. 8C, SSP-SAP affected GABAergic and glycinergic neurons differentially in the two VLM regions examined. Whereas 25% of GABAergic and glycinergic neurons were killed in the CVLM region (glycinergic cell loss illustrated in Fig. 9B1–2), the cell loss within the preBötzinger region was <10% and did not reach statistical significance (Fig. 8C). The loss of glycinergic neurons, like that of the GABAergic ones was more pronounced near the ventral surface at the lower edge of the region defined as the subCVLM (Fig. 9, B1 and B2). Finally, as shown previously (4), CVLM glycinergic neurons were not Fos-ir (result not illustrated).
To test further the specificity of the lesion in the CVLM region, a series of sections was processed for detection of VGLUT1 mRNA, a marker of the lateral reticular nucleus and other structures that contribute mossy fiber inputs to the cerebellum (11). As shown in Fig. 9, C1 and C2, the lateral reticular nucleus was completely spared by SSP-SAP despite its immediate proximity to the CVLM. The integrity of the LRN was so obvious that cell counts were not thought to be necessary.
The destruction of numerous GABAergic and glycinergic neurons in the CVLM region suggests that this region must contain NK1R-expressing neurons. This conclusion is in agreement with our previous observations (38) reiterated in the present study (Fig. 2F) but contrary to the notion that the NK1R-ir neurons of the VLM are confined to the region of the pre-BötC (7). To settle this issue, we examined NK1R-ir specifically in the CVLM region using an amplified peroxidase reaction with Vector VIP as substrate. This method revealed very clearly that the CVLM region contains numerous medium intensity NK1R-expressing neurons (Fig. 9, D1 and D2). The region photographed in this figure is just dorsal to the LRN in the coronal plane shown in Fig. 9, C1 and C2. This improved method still failed to detect NK1R-ir in the LRN consistent with the fact this nucleus was spared by SSP-SAP.
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
"Depressor" neurons of CVLM: are there more than one type? Microinjection of excitatory amino acids into the CVLM was shown in 1983 to produce large reductions of BP (depressor effect) in anesthetized rats (41). The depressor effect of glutamate was tentatively attributed to the activation of GABAergic interneurons that are part of the sympathetic baroreflex circuitry (41–43). Sapru's original view was that these GABAergic neurons are normally activated by baroreceptor stimulation and trigger the sympathetic baroreflex by inhibiting presympathetic excitatory cells located in the RVLM. Subsequent research has demonstrated that the CVLM does indeed contain baroactivated GABAergic neurons with the requisite axonal projections (12, 13, 16, 18, 29, 30, 36, 39). However, the existence of these neurons does not prove that the sympathoinhibition caused by injection of excitatory amino acid in CVLM is due exclusively to their activation to the exclusion of some other neighboring cell group. In fact, the existence of baroinsensitive depressor neurons in the VLM has been proposed before though their location was considered to be caudal to the barosensitive ones rather than below them (5).
Fos expression by CVLM neurons after PE-induced hypertension requires the integrity of the buffer nerves and is therefore attributed to the activation of arterial baroreceptors (4, 27). However, some contribution of cardiopulmonary receptors to the activation of CVLM neurons by infusion of PE is not ruled out by the available evidence. Within the CVLM region a large majority of these Fos-ir neurons are GABAergic, the remainder having a glutamatergic or mixed catecholaminergic and glutamatergic phenotype (4, 20, 39, present results). The depressor region of the VLM (from bregma –12.5mm to –13.9 mm in the present work) roughly coincides with the anatomical distribution of the VLM GABAergic neurons that express Fos in animals subjected to infusion of PE (4, 39). Therefore, the Fos-ir neurons present in the CVLM of rats subjected to PE-induced hypertension are thought to be the same population of cells as the CVLM GABAergic baroactivated neurons identified electrophysiologically (12, 30).
In summary, the CVLM contains a population of GABAergic interneurons that are activated by baroreceptors and contribute to the sympathetic baroreflex. These cells express Fos in rats subjected to PE-induced hypertension. The hypotension caused by injecting DLH into the CVLM of anesthetized rats is due to sympathoinhibition. The sympathoinhibition may be caused by at least three separate mechanisms: direct activation of the GABAergic baroactivated depressor neurons, indirect activation of these cells via local excitatory interneurons that are antecedent to them or activation of depressor neurons that are not barosensitive and the distribution of which overlaps with the GABAergic baroactivated depressor neurons.
Selectivity and extent of lesions caused by SSP-SAP. SSP-SAP is a conjugate of the ribosome inactivating toxin SAP and an analog of substance P. This conjugate binds with high selectivity and affinity to the NK1R and promotes internalization of the NK1R with the SAP conjuguate (40). The selectivity of SSP-SAP is attributed to the much lower probability of uptake of SAP by cells that are devoid of NK1Rs. In the present study, nonselective uptake was assessed by injection of unconjugated SAP in the same molar amount as the conjugate. As shown previously, injection of a matched dose of unconjugated SAP into the VLM did not create a histologically detectable lesion nor produce a demonstrable change in the physiological properties of the underlying network (6, 37). Thus the toxicity of SSP-SAP clearly required binding of the toxin to a receptor with high affinity for SSP. Given the selectivity of SSP for the NK1R over other neurokinin receptors, it is reasonable to assume that the cells that express the NK1R are preferentially if not exclusively targeted for destruction. However, the possibility that the toxin might also destroy some of the neurons that express the NK3 receptor cannot be eliminated entirely.
The selectivity of the lesions created by SSP-SAP was clearly very high given that both the TH-ir neurons of the VLM and the VGLUT1-mRNA expressing neurons of the lateral reticular nucleus were spared. The resistance of these two neuronal types is consistent with the fact that they lacked a detectable level of NK1R-ir. Additional evidence of specificity includes the fact that inhibitory neurons (GABAergic or glycinergic) were only partially destroyed in the region of the VLM where all identifiable NK1R-ir neurons were killed. Furthermore the destruction of inhibitory neurons was region specific. More precisely, the toxin destroyed about a quarter of the GABAergic and glycinergic neurons located in the CVLM, whereas the number of these neurons was not reduced significantly in the pre-Bötzinger region. In both regions, however, NK1R-ir neurons were eliminated by the toxin. The resistance of the GABAergic and glycinergic neurons of the preBötzinger region to SSP-SAP confirms and extends our previous observation that very few GABAergic or glycinergic neurons express detectable levels of NK1R-ir in this region (37).
As in our previous experiments, SSP-SAP produced a virtually complete loss of NK1R-expressing neurons in the entire ventrolateral quadrant of the medulla oblongata (37). The toxin-induced lesion was clearly not limited to the intensely NK1R-ir neurons of the ventral respiratory column that are characteristic of the pre-BötC region and rostral rVRG (7, 26, 38). SSP-SAP also consistently eliminated the somewhat less intensely stained NK1R-ir cells that lie between these VRG segments and the ventral surface of the medulla as well as the many NK1R-ir cells of the CVLM region. The lesions, though large, had consistent boundaries. The NK1R-ir neurons located within the compact portion of the nucleus ambiguus were only marginally affected, and cells located dorsal to this nucleus, within the trigeminal nucleus or in the midline medulla, were spared. Finally, NK1R-ir neurons located at the rostral end of the VLM (retrotrapezoid region) were also spared and the contralateral uninjected side of the medulla also appeared intact.
Are all baroactivated depressor neurons of the CVLM spared by SSP-SAP treatment? The interpretation of the present physiological experiments depends on whether SSP-SAP destroys the CVLM GABAergic baroactivated depressor neurons or simply renders them less sensitive to excitation by DLH or by baroreceptor activation because a local network of NK1R-expressing cells is destroyed by the toxin.
The evidence discussed in the previous section indicates that SSP-SAP spared two separate classes of VLM neurons that did not exhibit detectable NK1R-ir under the present conditions. These neuronal types are either in immediate proximity of the baroactivated depressor neurons (lateral reticular nucleus) or intermingled with them (TH-ir cells). Because the baroactivated depressor neurons identified by Fos immunoreactivity in PE-treated rats also lacked detectable NK1R-ir, we conclude that they too must have been spared. However, definitive proof of this assertion would require the availability of a diagnostic anatomical marker for these cells. Because no such marker is available, survival of every single baroactivated depressor neuron after SSP-SAP treatment is not definitively established.
Given the assumption that the entire population of the baroactivated depressor neurons must have survived SSP-SAP, why would the number of cells that express Fos after PE-induced hypertension be reduced and why would the hypotension resulting from direct activation of CVLM neurons by DLH injection be also reduced? The reduction in the number of Fos-ir cells could signify that the baroactivated depressor neurons are less easily activated by baroreceptor inputs after the surrounding network of NK1R-expressing cells is destroyed. Similarly, the loss of a globally excitatory drive from the surrounding network of NK1R-expressing cells may cause the baroactivated depressor neurons to be less easily depolarized by DLH. This hypothesis is consistent with the fact that a large proportion of the NK1R-expressing neurons of the RVLM are glutamatergic (9). However, there are alternate explanations for the loss of the DLH-induced sympathoinhibition. As mentioned above, it has not been proven that the sympathoinhibition caused by DLH injection into the CVLM is exclusively due to the depolarization of the GABAergic baroactivated depressor neurons. The hypotension and sympathoinhibition could also be due to the activation of GABAergic depressor neurons that reside in or very close to the CVLM and are unrelated to the cells that express Fos after stimulation of arterial baroreceptors. Indeed it is clear from the present data that the population of CVLM GABAergic neurons is heterogeneous and does include NK1R-ir expressing cells that are destroyed by the toxin. The disappearance of some of the GABAergic neurons located in the subCVLM region is consistent with the presence of NK1R-expressing neurons in this area (Fig. 9, D1–D2). It is also consistent with our prior observation that a small fraction of VLM neurons located predominantly below the VRG contain GAD-67 mRNA (38). In summary, the NK1R-expressing GABAergic neurons of the subCVLM may be among the baroinsensitive depressor neurons that have been postulated previously (5).
Functional significance. The region of the VLM that extends from the CVLM to the facial motor nucleus is essential for blood pressure stabilization via the baroreflex, but this region also plays a major role in integrating the respiratory and circulatory components of blood gas homeostasis (1, 2). The theory that both respiration and the sympathetic outflow are controlled by a shared set of VLM "cardiorespiratory" interneurons has been proposed previously to account for the centrally generated respiratory oscillations exhibited by sympathetic vasomotor and cardiovagal neurons (8, 28). The current view is that the excitatory presympathetic neurons of the RVLM are the key convergence point between the medullary respiratory network and the sympathetic vasomotor efferent circuitry (8, 21). This interpretation remains tentative since the respiratory inputs to other identified components of the central network that govern sympathetic tone, including the baroactivated depressor neurons of the CVLM, have not yet been examined or reported. The baroactivated depressor neurons are intermingled with the inspiratory-augmenting bulbospinal neurons of the rVRG (34, 39) and are anatomically quite close to the pre-BötC, a region of the VLM that is critical for respiratory generation and or regulation (6, 7). This anatomic proximity suggests that the activity of the baroactivated depressor neurons could be regulated by VLM interneurons that are also part of the respiratory network. One interpretation of the present results is that after SSP-SAP, the excitability of the baroactivated depressor neurons of the CVLM decreases because these cells lose essential excitatory drives from some excitatory component of the respiratory network. This interpretation, depicted in Fig. 10, is consistent with the fact that the respiratory network contains numerous NK1R-ir glutamatergic neurons (9, 10) and is profoundly disrupted by lesions of the VLM with SSP-SAP (6, 37). A decrease in sympathetic drive to resistance vessels leading to increased peripheral blood flow may be called for to complement an increased respiratory drive under specific physiological circumstances. The present results suggest that some of the NK1R-expressing neurons of the VLM may link the respiratory network and the sympathetic system via the baroactivated depressor neurons of the CVLM. However, it is also possible that the decreased responsiveness of the CVLM baroactivated depressor neurons derives from the loss of excitatory and NK1R-ir neurons that are not involved in respiratory pattern or rhythm generation. Inhibition of CVLM baroactivated neurons accompanies and may contribute to the pressor responses caused by stimulation of somatic afferents (17, 30). Conversely, these neurons seem to mediate some of the hypotension produced by stimulation of the hypothalamus (44) or the greater splanchnic nerve (24, 25). These inputs may well be polysynaptic and could conceivably be relayed by VLM neurons that are NK1R-ir and eliminated by SSP-SAP.
|
The loss of the sympathoinhibitory response after SSP-SAP treatment could also result partly from the destruction of NK1R-expressing GABAergic or GABAergic/glycinergic neurons located in the CVLM/subCVLM region (Fig. 10). These cells may be the baroinsensitive depressor neurons whose existence was postulated by Cravo et al. (5). These neurons are thought not to be under baroreceptor control and therefore would not express Fos after infusion of PE, which is consistent with the results of the present Fos study.
In conclusion, we speculate that the loss of sympathoinhibitory response to DLH microinjection after lesions of the VLM with SSP-SAP may be due to the destruction of at least two types of NK1R-expressing neurons (Fig. 10). One type may be glutamatergic and is assumed to be located in the region of the pre-BötC. These cells may be cardiorespiratory interneurons that facilitate the activation of the CVLM GABAergic neurons that are activated by baroreceptor stimulation. The other NK1R-ir neurons may be baroinsensitive depressor neurons located in the subCVLM.
|
DISCLOSURES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. A. Mandel and A. M. Schreihofer Glutamatergic inputs to the CVLM independent of the NTS promote tonic inhibition of sympathetic vasomotor tone in rats Am J Physiol Heart Circ Physiol, October 1, 2008; 295(4): H1772 - H1779. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. V. Dzurik, A. Diedrich, B. Black, S. Y. Paranjape, S. R. Raj, D. W. Byrne, and D. Robertson Endogenous substance P modulates human cardiovascular regulation at rest and during orthostatic load J Appl Physiol, June 1, 2007; 102(6): 2092 - 2097. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
